2012-05-01 [18:45:26.0000] "utf-eight" [18:50:54.0000] Hixie: I have finished that blog topic I had mentioned writing. What do I do now? Post it here? [20:28:43.0000] cbright6062: ping annevk, he's more up on the blog than i am [20:48:13.0000] cbright6062: I can give you an account [20:48:54.0000] cbright6062: could you email your post to annevk⊙oc ? [20:49:09.0000] cbright6062: if there's nothing weird I'll set you up [20:58:55.0000] annevk: how was the Notification meeting? [21:00:06.0000] went well I think [21:00:10.0000] posted a summary on the list [21:03:28.0000] k [21:16:29.0000] annevk: btw, have you looked at the Web Performance specs lately? plh tells me they are transitioning a bunch of them [21:16:40.0000] I remember you had commented on some of them before [21:17:01.0000] might be worth going back and checking to see whether they actually responded to your comments [21:37:44.0000] somdbody's doing spring cleaning! [21:38:44.0000] MikeSmith: is any F2F action happening yet? [21:38:57.0000] nope [21:39:08.0000] not for another 12 hours or so [21:39:15.0000] ah good [22:02:59.0000] annevk: you want a ride tomorrow? [22:03:10.0000] MikeSmith: I guess I should, but that group kind of gives the same vibe as geolocation [22:03:31.0000] yeah [22:07:11.0000] MikeSmith: it looks walkable... I kind of want to try walking [22:07:18.0000] ok [22:07:38.0000] my an plh going there pretty early anyway [22:07:44.0000] 8am [22:07:56.0000] ow [22:08:18.0000] might have to do the same in case the walking fails :p [22:09:18.0000] in other news, just finished Half-Life 2 [22:09:22.0000] cool game [22:11:01.0000] does Valve make any web-based games/ [22:12:01.0000] haven't seen any [22:12:38.0000] can't believe I never picked up on this before, still have Episode 1 & 2, and Portal 2 [22:13:26.0000] and maybe then they'll release whatever /r/gaming/ gets worked up about every other day [23:02:27.0000] (btw turns out i actually won't be around tomorrow, should be back wednesday) [23:03:35.0000] Enjoy May Day! [00:33:29.0000] /me is in Sunnyvale [00:33:55.0000] /me wonders where annevk / mikesmith are staying [01:05:50.0000] hello timeless! [01:19:22.0000] hi othermaciej [01:19:53.0000] /me discovers which channel/network this is and decides it's time to sleep [01:20:02.0000] /me has to wake up today and catch a cab or something [01:20:22.0000] see people in <8 hours [01:21:10.0000] later [02:58:35.0000] So in the Google Groups discussion I was just reading, 1) back/forward work in some crazy way I don't understand (they didn't act as inverses), 2) the place I'm scrolled to in the page isn't preserved on back/forward, 3) when I got to the end of the page it refused to let me scroll down and I have no idea why because there's no visible scrollbar. [02:58:37.0000] Yay for HTML5! [02:59:05.0000] Finally allowing us to create hard-to-use pages that ignore all established UI conventions but are fully standards-compliant! [02:59:34.0000] I guess it would have been even worse without all the new APIs . . . [03:08:11.0000] At least back and forward *sort of* work. [05:10:37.0000] timeless: "quality" inn [06:26:22.0000] there might actually be an observable difference too between just tokenizing and using an insertion mode [06:26:33.0000] consider e.g. "test" as input [06:48:48.0000] More extreme given foreign content and breaking out of it. [06:50:16.0000] example? [07:02:37.0000] annevk: No, too close to exam time :P [07:27:47.0000] Loading a page from cache, if it wasa originally sent with Set-Cookie, will that cookie be set again on cache load? [07:46:56.0000] gsnedders: exams [09:19:04.0000] good morning, Whatwg! [09:19:27.0000] where's Ms2ger? [09:23:10.0000] timeless, staying at best western, 2300 el camino [09:26:58.0000] AryehGregor: yt? [10:06:34.0000] annevk: You around? [10:06:54.0000] yeah [10:13:09.0000] annevk: Watch for an email from me. [10:13:16.0000] is https://twitter.com/#!/tabatkins/status/197373135153147906 an IETF troll [10:16:11.0000] annevk: Sent. [10:18:20.0000] thanks [10:18:30.0000] MikeSmith: Well, that too. [10:18:53.0000] cbright6062: okay looks fine [10:18:59.0000] let me get you an account [10:19:21.0000] annevk: Thanks. [10:20:40.0000] password will be emailed to you [10:20:44.0000] okay [10:20:53.0000] let me know if something does not work out [10:23:49.0000] dinosaurs rock [10:28:19.0000] !summon ms2ger [10:37:18.0000] dglazkov: you should look at your normative requirements in the Shadow DOM spec [10:37:38.0000] dglazkov: e.g. the bit about the constructor has no normative requirements [10:37:51.0000] dglazkov: while I guess you want implementations to do something with the constructor [10:38:07.0000] annevk: can you file a bug? I don't want to forget this. [10:39:54.0000] dglazkov: done [10:40:24.0000] annevk++ [11:17:35.0000] and posted. [11:18:30.0000] cbright6062: can you adjust the heading levels maybe? [11:18:47.0000] cbright6062: blog post itself is h2, anything inside ought to be h3 or lower [11:19:43.0000] okay [11:20:23.0000] fixed. [11:21:48.0000] annevk: I was just following the standard I've learned for heading levels. Fixed though, since you seem to use a different scheme. [11:42:29.0000] cbright6062: are you still getting and DB errors from wiki.whatwg.org? [11:46:14.0000] MikeSmith: No [11:46:23.0000] ok [11:57:53.0000] rafaelw_: btw, http://krijnhoetmer.nl/irc-logs/whatwg/20120501#l-448 [12:02:32.0000] rniwa, in general I'm not around during American working hours -- I'm now in Israel, GMT+0200. [12:02:41.0000] AryehGregor: ah, ok [12:02:43.0000] Right now I just logged on for a few minutes. [12:02:47.0000] AryehGregor: when can i talk to you? [12:02:55.0000] E-mail or bug report posting or whatever is best. [12:02:58.0000] AryehGregor: okay. [12:03:09.0000] AryehGregor: will do that. [12:03:46.0000] When I'm working it's generally between 8:30 and 6:00 my time, which is 10:30 PM to 8:00 AM Pacific time. [12:04:12.0000] And usually now more like 8:30 to 2:00, so even the East Coast I can't talk to much. [14:01:42.0000] annevk: does dom4 spec try to load something from whatwg? [14:01:51.0000] the latter seems to be very slow atm [14:02:57.0000] odd [14:03:09.0000] several people have reported several weird failure modes on that server today [14:03:12.0000] and i can't reproduce any of them [14:06:36.0000] smaug____: yes [14:06:45.0000] smaug____: stylesheet and script [14:06:52.0000] smaug____: maybe just a script these days [14:07:00.0000] that is silly ;) [14:07:07.0000] smaug____: it's at the end of the document [14:07:20.0000] given that whatwg.org isn't very reliable server [14:07:20.0000] and helps with cross-references [14:07:26.0000] w3.org isn't either [14:07:31.0000] true [14:07:40.0000] but the doc is in w3.org [14:07:56.0000] if you volunteer maintaining that file you could host and maintain it on w3.org and I'll link it, no problem [14:48:51.0000] By the way, just making a random observation: I find it amusing that the blog hasn't been formatted with the newer HTML coding. [14:50:57.0000] never taken the time to update the markup [14:51:16.0000] ah, makes sense. [14:51:25.0000] well we have one or two times, back then support for new elements was spotty [14:51:32.0000] ah [14:51:46.0000] the new markup is a lot more universally supported now. [14:52:02.0000] it's actually why I had originially formatted my article the way I did. [14:52:13.0000] Lol [14:52:26.0000] yeah, I guess if there's a good reason to update the markup again that aspect would be updated as well [14:53:20.0000] /me may have somewhat broken html5lib tip by making it assert where we previously did bogus stuff [14:54:34.0000] annevk: well, if you ever need a hand with it, I'd be willing to help. (updating code is actually one of my favorite things to do.) [14:55:58.0000] http://code.google.com/p/html5lib/issues/detail?id=202 [14:55:58.0000] if you're interested in doing that, sounds fine to me [14:56:16.0000] I'd have to figure out how to give you ssh access I guess [14:57:06.0000] annevk: I'm fully interested and fully willing to help. [14:57:09.0000] jgraham: see above html5lib bug [14:57:20.0000] cbright6062: so I can give you access via the WordPress interface right away [14:57:22.0000] jgraham: (202) [14:57:43.0000] cbright6062: ssh access is something Lachy can maybe arrange [14:58:30.0000] annevk: sounds fine to me. Just send me through the proper channels. As for the Wordpress side of things, I'm ready as soon as you are. [14:59:11.0000] seems whatwg.org is a little slow, but you'll be an Admin soonish which will allow you to update the pages [14:59:24.0000] not quite an ideal interface to edit markup, but it works [14:59:48.0000] annevk: I actually do it that way on a daily basis. Lol [15:00:35.0000] well hey, if it works for you :) [15:03:09.0000] cbright6062: you should be set now, have fun and don't kill it :) [15:03:35.0000] annevk: I noticed. And trust me, I won't kill it. [15:13:55.0000] anyone happen to remember a post with ian's argument against using URLs as identifiers for things that aren't actually resources (eg. XML namespaces)? [15:14:04.0000] would like to be lazy and point to it rather than make it from scratch, heh [15:17:12.0000] found one (webintents) [15:18:21.0000] You know what would make removing dependencies on implicit unicode/str conversion easier in html5lib? Not have the stdlib rely upon them. [15:42:11.0000] each year this push stuff [15:42:13.0000] meh [15:51:00.0000] dglazkov: what is the language in "Matching Insertion Points" trying to accomplish? [15:51:12.0000] dglazkov: subsetting Selectors in that way seems like a bad idea [15:51:26.0000] (Shadow DOM spec) [15:52:36.0000] dglazkov: "host element" -> "shadow host" 2012-05-02 [17:32:02.0000] annevk: Anything you specifically you would like to see? (such as updating the heading levels to the new standard scheme, etc). I'm just making sure to stay within reason, etc. [18:54:54.0000] hello! if I believe I've found some overlap in a working draft of CSSOM (compared to DOM2 core) with a notice about how the IDL is re-defining things, how would I go about helping the editor fix this? email them directly? scour IRC channels or lists? http://www.w3.org/TR/cssom-view/#dom-mouseevent-clientx seems to overlap with http://www.w3.org/TR/DOM-Level-2-Events/events.html#Events-MouseEvent-clientX (which is already recommended), for more conte [18:55:18.0000] Email www-style⊙wo about it. [18:55:34.0000] TabAtkins_: okie dokie [18:55:51.0000] That said, DOM Level 2 Events isn't used, iirc. [18:56:17.0000] TabAtkins_: it isn't used in what sense? like in the ES4 sense where it was dropped? [18:57:21.0000] It's superseded by better specs. [18:57:27.0000] TabAtkins_: ah, which specs? [18:58:04.0000] /me is inb4 DOM Level 3 events [18:58:29.0000] itym dom4 [18:58:42.0000] (fine, dom3 for the events themselves; but for the event model, dom4) [18:58:50.0000] they have the same conflicts [18:59:10.0000] All right then. [18:59:21.0000] client{X,Y} and screen{X,Y} are supposed to be on MouseEvent interface [18:59:21.0000] I wonder if we have an active editor for CSSOM-View... [18:59:23.0000] I dont' think so. [18:59:48.0000] yeah, was gonna email Anne, but she's may not be as active now? [18:59:53.0000] Anne's a guy. [18:59:59.0000] my bad [19:00:04.0000] And has left the CSSWG, so he's not editting that spec any longer. [19:00:09.0000] yeah [19:00:11.0000] So, yeah, that's an editorless spec. [19:00:49.0000] I didn't know that Anne left the CSSWG [19:01:18.0000] Yeah, he dropped out a few months ago. [19:01:25.0000] TabAtkins_: I don't supposed there'd be a point in mentioning this in my email to www-style and asking for volunteers? [19:01:30.0000] suppose* [19:01:41.0000] danbeam: Go ahead and mention it. Make sure to put [cssom-view] in the subject line. [19:01:47.0000] TabAtkins_: ok [19:01:52.0000] Just so that future editors of that spec will be able to find it. [19:01:56.0000] ya [19:02:22.0000] /me wonders what Glenn Adams has been doing lately, since he hasn't seen much activity in the CSSOM spec recently... [19:05:06.0000] TabAtkins_: so if CSSOM spec declares it added these to the MouseEvent interface, but they were in DOM Level 2 Events (which seems to be circa 2000, 11 years previous), should that part simply be removed from the CSSOM spec, essentially (given than an editor is found)? [19:05:27.0000] s/spec/working draft/ for CSSOM (if that matters?) [19:05:34.0000] Yeah, unless there's a good reason that CSSOM View should update the definitions, the DOM specs shoudl be definitive. [19:05:55.0000] Given that D3E is more recent than View, too. [19:06:00.0000] TabAtkins_: it seems to be the exact same thing [19:06:09.0000] Then yes. [19:06:11.0000] TabAtkins_: so nothing to update [19:22:10.0000] /me wonders why there's only one person from Google that is active on www-style [19:22:32.0000] kennyluck: I'm not particularly active, but I'm about to prove you wrong just the tiniest bit, :P [19:22:48.0000] kennyluck: who were you referring to, Tab? [19:23:05.0000] The editor that's listed currently. [19:25:24.0000] kennyluck: editor of ... what? [19:25:59.0000] danbeam, of CSSOM and CSSOM-view. [19:26:24.0000] kennyluck: ah [19:26:34.0000] oh, sorry I misunderstood you. [19:26:36.0000] Ye [19:26:45.0000] Yeah, I was referring to Tab of course. [19:27:01.0000] so you're talking about shans@? [19:27:08.0000] @google.com* [19:27:17.0000] /me thought that you were asking who Glenn Adams is. [19:28:15.0000] kennyluck: I know exactly who Glenn Adams is -- the smartest person I've [n]ever met with an @cox.com email! [19:28:34.0000] :P [19:30:15.0000] TabAtkins_: the editor's draft of CSSOM View seems to have new editors, should I be CC'ing/directing my question to shans⊙gc / glenn.adams⊙cc as well? [19:30:45.0000] *shrug*. He only reminds me of a long thread about CORS and @font that I didn't look into. [19:31:41.0000] kennyluck: haha, Access-Control-Allow-Origin: * everywhere! what's the worst that could happen?! [21:09:09.0000] cbright6062: whatever you think is appropriate is fine [21:10:08.0000] cbright6062: anything goes as long as it's still readable in recently released browsers [21:10:24.0000] annevk: Okay. And I will makeshore it is. [21:10:24.0000] cbright6062: at least as far as I'm concerned, others in this channel might have different views [21:10:45.0000] annevk: I mainly asked you since you have the main focus on the blog, from what I've seen :) [21:11:03.0000] I think in the end the person doing the work should decide [21:12:18.0000] lol [21:12:29.0000] well, I'm doing the work, but I don't want to upset anyone on here :P [21:13:37.0000] don't worry about it [21:13:45.0000] ask forgiveness, not permission [21:14:36.0000] danbeam: DOM Level whatever does not define hit testing and does not define the coordinates are given in CSS pixels [21:14:52.0000] danbeam: having said that, TabAtkins_ is right, I'm no longer an editor or participant in the CSS WG [01:26:12.0000] Poe's Law in action: http://stuffandnonsense.co.uk/blog/about/there_i_said_it [01:28:43.0000] hsivonen: I don't think there's any suggestion that's a parody? [01:30:15.0000] At least, if that's the intention — and I agree the content could be mistaken for parody — he fails to carry it through into the comments [01:31:35.0000] Sadly I think there is just a school of thought that browsers are canvases for the artistic vision of web designers rather than tools that are designed to enable their users to access information [02:06:29.0000] jgraham: Also, to pedant, transforms are Core Animation. [02:15:10.0000] gsnedders: Fair point [02:16:49.0000] But regardless, an API that exists only on OS X and their Windows-Safari port thing, seemingly. [02:17:27.0000] Yes, my underlying point stands [02:17:59.0000] (I wonder if Apple will ever release their Obj-C runtime for Windows?) [02:18:59.0000] (I can't see it happening, but it raises interesting questions about Xcode being ported for the sake of iOS) [02:22:58.0000] /me is puzzled by the last few lines of conversation [02:24:34.0000] othermaciej: The context is that -webkit- CSS properties, in particular transforms, often have different implementations of the gfx part [02:25:09.0000] different from what? [02:25:18.0000] In different consumers of webkit [02:25:26.0000] e.g. in Chrome vs Safari [02:25:42.0000] oh, yeah, different ports have different graphics back ends, not just for transforms for that matter [02:25:50.0000] (sorry that was very badly worded) [02:26:08.0000] in fact I think transforms don't even use CA if they are 2D and not animated [02:26:16.0000] othermaciej: The context is in Andy's blog post comments [02:26:38.0000] (3D transforms and transforms with animation/transition applied do force a CA layer, I think other transforms are just done w/ CoreGraphics) [02:27:01.0000] On the post hsivon linked to ther was lots of wailing about how Opera implementing some properties with the -webkit- prefix was evil because they might not have the same quality as "the" webkit implementation [02:27:17.0000] Although there was never actually a conccrete example of this being a problem [02:27:22.0000] And may differ subtly from it. [02:27:42.0000] (which there is some basis for, given gradient syntax fun) [02:27:47.0000] Or indeed any insight into why this isn't a problem for things without prefixes [02:28:19.0000] I think people are starting with the conclusion that vendor A supporting vendor B's prefix is bad, and from there looking for arguments that may justify it [02:28:47.0000] I'm not a huge fan of the situation but I can't blame anyone for doing whatever it takes to make their browser compatible with content [02:29:59.0000] /me would quite like Apple to drop prefixed versions of properties, as everyone has done (albeit slowly) [02:30:09.0000] Yeah, that would match my impression and explain why so many of the arguments seem so weak [02:30:42.0000] I also don't like the situation fwiw, but I think the problem is with the prefix system [02:30:55.0000] gsnedders: while that might be a good idea in general, for the most problematic properties, I am not sure that strategy is sane [02:31:22.0000] Yeah, and something really should've happened when MS were going to implement… -webkit-text-size-adjust, was it, several years back. [02:31:28.0000] because the specs are not advanced enough for the CSS WG to grant their blessing on shipping the unprefixed version at all, and the prefixed version is so widely used that not having it is major compatibility fail [02:31:32.0000] I don't think it's reasonable, for the same reason I don't think it's reasonable for everyone else not to just implement those properties [02:31:36.0000] othermaciej: For border-radius it's a change that can happen. [02:31:48.0000] othermaciej: For gradients, yeah, it's less clear. [02:31:58.0000] I think the CSS WG should confess that it failed and publish a CSS aliases spec [02:32:08.0000] othermaciej: I believe you're now the only ones to support prefixed border-radius, FWIW [02:32:13.0000] That makes support for the prefixed properties part of the standard [02:32:20.0000] gsnedders: for border-radius it sounds like it would break a bunch of sites (assuming Opera's reason for implementing -webkit-border-radius is sound) [02:32:44.0000] jgraham: In a wildcard sense, as glazou was proposing, or just specific ones? [02:33:09.0000] gsnedders: I have no idea what glazou was proposing [02:33:09.0000] othermaciej: That's a cosmetic issue, unlike the others. And it breaks them no more than Gecko dropping -moz- in 13. [02:33:12.0000] not sure why Safari should take a compat hit in a case where Opera won't [02:33:17.0000] But I would suggest specific ones [02:33:25.0000] jgraham: -*-foo is mapped to foo [02:33:55.0000] gsnedders: That makes no sense to me? What am I missing? [02:34:21.0000] I mean, it seems to have all the properties of not having a prefix system whilst still having a porefix system [02:34:26.0000] *prefix [02:34:35.0000] othermaciej: I'd like to minimize the list of required synonyms, so we both would take the hit. I'm not gonna propsoe that for things like gradients where things actively become unusable. [02:34:38.0000] I think prefix may be a failed experiment, at least as used today, but I am not sure how else to make it safe for browsers to experiment with extensions [02:35:11.0000] othermaciej: The only sane thing I have heard is for extensions to be behind runtime flags [02:35:14.0000] othermaciej: But given Moz has managed to drop prefix for border-radius it suggests it can be done. [02:35:18.0000] Or not in stable versions [02:35:42.0000] (Ourselves and MS never supported it with prefix) [02:35:51.0000] Of course that isn't going to work well if "experiment" really means "add new features before anyone else and encourage authors to use them" [02:36:20.0000] FWIW, border-radius would never have been aliased without the far more serious non-cosmetic issues. [02:36:37.0000] I don't think it would fly to gate availability of features to content authors on the CSS WG's time to get to CR [02:36:55.0000] I think the CSS WG is part of the problem [02:37:06.0000] Ideally I'd like to only bake -webkit-linear-gradient (or whatever the name is) into the platform. [02:37:12.0000] Because it's the only one sites *rely* on. [02:37:30.0000] So a solution that forces change on them seems like a positive thing [02:37:35.0000] (Possibly old flexbox, but hopefully what Google uses that on moves away from it) [02:37:39.0000] I wonder also how early implementation should work for APIs, where the prefix approach is even worse [02:38:07.0000] Because, as you say, people would push like crazy for a spec model that allowed fast iteration if their ability to ship was gated on fast itereation [02:38:08.0000] annevk was suggesting second impl impls without prefix [02:38:13.0000] (HTML attributes seem only about as bad as CSS properties when prefixed and no one holds those to a hard CR line) [02:38:41.0000] othermaciej: Given the CSS WG agreed to drop things earlier than CR hopefully stuff is better in future. [02:39:03.0000] gsnedders: by "drop things" you mean drop prefixes earlier than CR? [02:39:16.0000] othermaciej: yeah [02:39:20.0000] I thought the CSS WG explicitly refused to do that even in clear-cut cases [02:39:23.0000] maybe I am not up to speed [02:39:43.0000] I remember this in some f2f this be resolved. [02:41:08.0000] I gotta ask hober since he goes to those things [02:41:32.0000] I should probably just leave the WG given I never do anything anyway [02:41:50.0000] anyway, CSS WG is touchy about their prefixes but I wonder if we can come up with sane cross-browser best practices for prefixing of APIs and markup attributes (I presume everyone would agree that prefixed elements are a bad idea) [02:41:52.0000] Like, I every few months read the mailing list. Only otherwise do when people tell me to. [02:43:19.0000] (In an unrelated point, Safari uses NSURL for all HTTPS stuff, so doesn't use NSS, right?) [02:43:48.0000] I don't know what NSS is [02:44:05.0000] we use CFNetwork for networking, via NSURL APIs on some OS versions [02:44:25.0000] under the covers somewhere in there, I think OpenSSL is used for SSL stuff [02:44:38.0000] SSL impl, used by Moz/Chrome [02:45:06.0000] othermaciej: Shows how much time I've spent dealing with OS X that I don't even know that NSURL isn't current any more :) [02:45:24.0000] NSURL is current, CFNetwork is just the thing underneath it [02:45:40.0000] "some OS versions"? [02:46:16.0000] (seeming NSURL predates Safari) [02:48:12.0000] I think NSURLConnection was introduced along with Safari, actually. Looking at the docs: "Available in Mac OS X v10.2 with Safari 1.0 installed. Available in Mac OS X v10.2.7 and later." [02:48:54.0000] on some platforms we just go to CFNetwork directly [02:49:14.0000] Ah. NSURL itself goes back to 10.0 [02:49:32.0000] (don't believe it goes back to NS, though) [02:49:35.0000] and yes, NSURLConnection was in fact introduced with Safari, it was first implemented by the Safari team [02:49:37.0000] yeah, NSURLHandle existed before that, and is deprecated since 10.4 [02:49:48.0000] there was an older NSURLHandle which did not cut it for browser use [03:07:25.0000] /me guesses he's in for fun working out what has changed in the past three OS X releases when he gets a MBA once they start shipping with Ivy Bridge CPUs [03:12:25.0000] (Waiting mainly for the sake of GPU performance, which would probably be good enough to ditch my pretty old MBP) [03:32:45.0000] Hello, does anybody know if there is a limit to HTML offline storage? I'm trying to find out if videos can be stored in it? Does anybody know that? [03:51:10.0000] jgraham: It's quite possible that it isn't a parody, which is troubling. [03:57:15.0000] See, I've seen plenty of people call it brilliant parody, but I fail to see any evidence of this. [03:57:49.0000] othermaciej: I think the best practice for prefixing is not to [03:59:55.0000] I'm rather surprised that Opera isn't aliasing Animations. [04:00:15.0000] Once the aliasing code is there, why bother limiting what's aliased? [04:00:50.0000] hsivonen: The limitation is probably to avoid too much of a backlash. [04:00:55.0000] /me has no idea [04:01:13.0000] What I think is sad about Opera's announcement is that some things that will have -webkit-aliases weren't announced to have unprefixed aliases, too [04:01:42.0000] well, the other sad thing was blaming Web authors [04:02:03.0000] blaming them isn't that productive and even if true, doesn't help [04:02:19.0000] blaming the CSS WG might actually end up having helpful effects [04:02:29.0000] and would be more correct direction of blame anyway [04:03:15.0000] Except the discussions a couple of months back basically seemed resigned to the fact that non-WebKit browsers will have to support them, so all you can really do is blame prior WG decisions which have since been changed. [04:03:50.0000] I'm not on the CSS WG, so I don't understand why layout devs act like CSS WG has the power to decide about unprefixing [04:03:51.0000] The fact that plenty of sites we are just told, "we only support WebKit", is an issue purely down to developers, though. [04:04:21.0000] seems like Opera aliasing but not also unprefixing is part of acting like it's in the WG's power to decide [04:04:30.0000] gsnedders: ok [04:04:44.0000] hsivonen: It's not official WG policy, pretty much, it's just de-facto agreements amongst browser vendors who happen to be in the WG, pretty much. [04:04:48.0000] (wrt when to unprefix) [04:05:34.0000] why do people who can land code keep honoring those de facto agreements? [04:06:14.0000] Presumably fear of being called out on not supporting standards properly [04:06:18.0000] hsivonen: The biggest issue is people doing background-image: -webkit-gradient(bleh); and refusing to put background-color: black; or whatever, which is down to more than just prefix policy. [04:06:43.0000] gsnedders: interesting [04:06:47.0000] Or a belief that the prefixing is a good idea, in spite of the evidence that it is harmful in these cases [04:07:04.0000] hsivonen: Leading to white-on-white text, which is the real harm of this situation. [04:07:22.0000] hsivonen: Everything else is merely cosmetic, and wouldn't have lead us to this. [04:07:32.0000] Well white-on-white is an obvious "you can't use this site" bug [04:08:04.0000] Other things are "your browser sucks more than $webkitBrowser even though it doesn't actually" issues [04:08:15.0000] Which are also harmful [04:08:36.0000] But not as harmful as to motivate anyone to risk the PR backlash we're currently getting. [04:08:55.0000] I'm not sure I agree [04:10:43.0000] Would have been nice to get Mozilla to make the change at the same time though [04:10:57.0000] Or announce that they will [04:10:58.0000] Well, I don't think anyone has seriously considered doing it based upon cosmetic reasons [04:11:20.0000] jgraham: We rather had our hand forced, though, seeming someone posted what was W3C member-confidenial. [04:11:52.0000] gsnedders: I thought there have been cosmetic reasons under discussion [04:12:10.0000] gsnedders: I have certainly advocated aliasing to get cosmetic benefits [04:13:09.0000] IIRC IE Mobile's planned support for -webkit-text-size-adjust was for usability issues [04:13:48.0000] In the long term, cosmetic issues are just as important as "this site is broken" issues [04:14:00.0000] Maybe moreso, because they are harder to evangelise [04:14:14.0000] they have mindshare consequences [04:14:23.0000] I think the hope was to come up with some cross-browser agreement about prefixes for them. [04:14:29.0000] And UX consequences. [04:14:30.0000] And for the current set of prefixes to go away. [05:24:48.0000] what is an "XML Literal" and what purpose does it serve? [05:26:15.0000] /me finds http://www.jenitennison.com/blog/node/103 [05:49:38.0000] MikeSmith: Its purpose is to confuse people by effectively being just a string that happens to contain markup, but technically being a string which is the UTF-8 decoding of an exclusive Canonical XML document fragment [05:49:51.0000] ah [05:50:05.0000] so it's the canonical part [05:50:05.0000] Easiest to just think of it as a markup string [05:50:08.0000] OK [05:50:35.0000] but with some magic in the RDF parsers so that you can represent that string as actual markup [05:50:41.0000] I see [05:50:55.0000] I find that all sentences with "XML" in become transparent in meaning if you s/XML/pain/g [05:51:05.0000] heh [05:51:52.0000] so there context of this is, somebody is asking me if there's a "formal notion and/or algorithm that says whether two HTML5 snippets are identical or not" [05:51:59.0000] but snippets I guess they mean nodes [05:52:14.0000] (i.e. the magic lets you write

2

which turns into a string with value "2", I think) [05:52:45.0000] Philip`: yeah, that looks like the example that Jeni uses in her blog entry [05:52:54.0000] (per http://www.w3.org/TR/rdfa-syntax/#s_xml_literals except the spec forgets the xmlns, I think) [05:53:02.0000] OK [05:53:34.0000] I assume the point is to prevent you having to write escaped markup inside markup, since that's ugly [05:53:41.0000] I see [05:54:26.0000] so then I guess once you'd done this XMLLiteral thing, there are some cases where you want to be able to compare two XMLLiteral things to determine if they are identical with each other [05:54:44.0000] not clear to me what those use cases are [05:56:52.0000] Yeah, they're just strings, so you can do arbitrary stringy things to them, but they happen to have the C14N guarantee that string equality is implied by XPath data model equality (which is quite like DOM equality) [05:57:42.0000] OK [05:58:52.0000] which (I think) in theory means you could extract RDFa from some XHTML document 'A', and then you could fiddle with 'A' to produce 'B' (change all the namespaces, change irrelevant whitespace, etc, e.g. by passing through an XML parser then serialiser) and extract RDFa from 'B', and you'd get exactly the same output [06:04:47.0000] th [06:45:04.0000] about webvtt line wrapping, where should the line break go if there are two positions that would produce the same delta? [06:58:39.0000] zcorpan: up to the implementation, i think, the balancing definition isn't precise (and should be loosened a bit more, to explicitly allow approximations) [07:01:23.0000] zewt: that's annoying when writing tests :-) [07:02:02.0000] tests how? it depends on the implementation's and platform's font rendering anyway [07:02:27.0000] reftests [07:02:52.0000] fwiw i want it to allow approximations so later the balanced wrapping can be made into a generic css white-space mode (where requiring it to be optimal isn't practical) [07:03:19.0000] yeah i'm fine with allowing that, actually [07:03:28.0000] it's still annoying [07:03:57.0000] you can test extreme cases, which should always give results within certain limits [07:04:43.0000] eg. if "a b c d e f g" wraps to "a b c d e f" and "g" in normal wrapping, it should always wrap closer to "a b c d", "e f g" in balanced [07:05:32.0000] that is, in that case the balanced output should always be narrower than the regular output, even if it's an approximation [07:06:14.0000] though since the balancing behavior is currently a special case rule instead of a CSS rule, I guess you can't turn it off in order to easily compare [07:07:27.0000] hmm, also the spec doesn't say where to wrap words that are too long to fit [07:07:41.0000] i guess you could check progressively longer cues ("foo", "foo foo", "foo foo foo", ...), and check that the width increases, and then decreases [07:08:07.0000] that is, the width of the box will increase as long as it fits on one line, then it'll abruptly drop to about half the width as it inserts a break around the middle, and then start increasing again; that doesn't happen with regular wrapping [07:08:46.0000] not a thorough test by any measure, but should at least tell whether it's happening at all [07:09:52.0000] anyhow off to work, later [07:10:02.0000] see ya [07:27:41.0000] XML Core WG no like BOM [07:37:03.0000] ok please tell me the editor's draft which seems to give an error right now doesn't do this silly renaming http://www.w3.org/TR/css3-text/#overflow-wrap0 [08:28:34.0000] zcorpan: no, fantasai hasn't got the memo that silly renaming is a bad idea [08:44:13.0000] /me sees "MIME RFCs" on WHATWG, ignores [09:15:56.0000] Ms2ger: what do you mean? [09:16:11.0000] Hmm? [09:16:39.0000] about MIME RFCs [09:16:51.0000] also, do you know any constants for UNINITIALIZED but with a better name? [09:18:33.0000] "multipart/form-data filename encoding: unicode and special characters" [09:18:34.0000] No [09:20:44.0000] oh yeah [09:21:00.0000] I've been saying for ages for we should just define the details of that RFC in HTML directly [09:21:13.0000] rather than this rather obscure reference [09:21:40.0000] Hixie had some reason not to I think, but I'm not sure it's still valid... [09:22:00.0000] annevk, do you know who does what for the eventPhase thing, btw? [09:26:14.0000] Chrome/Safari/IE do 0 [09:26:18.0000] Gecko does 2 [09:26:20.0000] Opera does 1 [09:26:23.0000] per Travis [09:26:26.0000] I have not verified [09:29:12.0000] added a comment [09:31:06.0000] Oh, you went with Gecko? How strange ;) [09:41:13.0000] hsivonen: Can I take a few minutes of your time and chat about DocumentFragment.innerHTML? [09:43:23.0000] Ms2ger: it seemed somewhat sensible given the options [09:43:31.0000] Ms2ger: I didn't want to mint a new constant [09:43:37.0000] Ms2ger: Travis is more open to that [09:43:45.0000] I defer to smaug [09:46:39.0000] lets see if I can fix it before Travis [09:46:48.0000] prolly not, going to discuss CORS again [09:46:51.0000] soooo boring [09:47:19.0000] annevk: did we skip the DOM3/4 stuff? [09:47:42.0000] Also, what happened to From-Origin? [09:56:51.0000] Ms2ger: nobody expressed interest in implementing it [10:00:00.0000] weinig: no [10:00:07.0000] Ms2ger: nobody implemented it [10:00:15.0000] Thanks, annevk, MikeSmith :) [10:01:54.0000] mention of resistance from Mark and Tyler [10:02:09.0000] as if it's new news [10:03:49.0000] yeah, let's document the rationale for the platform in every single spec [10:04:02.0000] \o/ [10:09:16.0000] annevk: i don't recall if there was any specific reason other than not pissing off people we don't need to piss off more, and avoiding extra work. [10:14:34.0000] rafaelw_: I think hsivonen hates you. [10:14:50.0000] rafaelw_: did you accidentally send him spam at some point? [10:14:52.0000] :) [10:15:06.0000] Accidentally? :) [10:15:10.0000] /me waves at dglazkov [10:15:25.0000] aw crap, forgot to good morning everyone [10:15:30.0000] good morning, Whatwg! [10:15:31.0000] :) [10:15:43.0000] Ms2ger: are you at the F2F? [10:15:50.0000] Maybe? :) [10:16:34.0000] okay peeps, let's smoke Ms2ger out [10:24:58.0000] rafaelw_: hsivonen typically is around during office hours in whatever weird timezone Finland's in [10:25:45.0000] EEST perhaps [10:26:47.0000] jgraham, yours +1 :) [10:27:34.0000] Oh look, the little graph at http://gavinsharp.com/irc/whatwg.html is very informative in this regard [10:28:26.0000] Ms2ger: I know what the GMT offset is (except when I don't), I just didn't know what it was called [10:33:24.0000] hmm, I guess I lost the customizatino I had that included the time zone being used in those stats [10:34:12.0000] jgraham: Thanks. =-) [11:09:26.0000] Hixie: hey, when do you have time to address the WebSocket issues? [11:27:27.0000] Can sb. tell me what is supposed to happen to the private key generated by the keygen element? How can I use it? [11:43:34.0000] annevk, can you file http://dvcs.w3.org/hg/domcore/rev/98a9587a515c on Gecko? [11:44:48.0000] sure [11:44:56.0000] Ta [11:48:25.0000] https://bugzilla.mozilla.org/show_bug.cgi?id=751286 [11:48:46.0000] guess I should file one on WebKit too [11:51:29.0000] Hmm, eventPhase can return 0 in Gecko too [11:52:02.0000] Gecko -> weird [11:52:06.0000] Truth [11:53:17.0000] weinig++ [11:53:19.0000] Again [11:53:42.0000] weinig 4 prez [11:53:51.0000] \o/ [11:54:02.0000] Ms2ger: what did he do now? [11:54:40.0000] https://bugs.webkit.org/show_bug.cgi?id=85397 [11:55:08.0000] Ms2ger: :) [11:57:49.0000] smaug++ [12:00:32.0000] old checkin comments like http://html5.org/tools/web-apps-tracker?from=61&to=62 are fun [12:04:50.0000] wait [12:05:01.0000] posts to whatwg⊙wo now get on public-whatwg⊙wo ? [12:05:06.0000] hmm [12:06:14.0000] hahaha [12:07:22.0000] Obligatory cross-posting? [12:31:47.0000] annevk: there are websocket issues? [12:33:33.0000] cross-posting to public-whatwg would be massively confusing unless public-whatwg can't be posted to [12:33:40.0000] i've disabled it until i can speak to whoever set that up [12:34:06.0000] Hixie: I set it up [12:34:30.0000] MikeSmith: so my concern is that people will post to public-whatwg and not realise they're yelling into a vacuum [12:34:35.0000] yup [12:34:36.0000] MikeSmith: can be disable posting to that list somehow? [12:34:37.0000] understood [12:34:47.0000] I will try to do that right now [12:35:00.0000] (the mirroring is fine by me in principle, though it'd be nice to get historical archives in there too) [12:37:04.0000] Hixie: systems team is planning to get the historical archives set up [12:37:15.0000] nice! [12:37:26.0000] will need to get a copy of the archives from you of course [12:37:49.0000] and I think we can't do it til later this month [12:38:12.0000] hm [12:38:18.0000] i dunno that i can get anything you can't get [12:38:20.0000] but i can look [12:38:53.0000] looks like the best i can get you is the gzipped archives here: http://lists.whatwg.org/pipermail/whatwg-whatwg.org/ [12:38:58.0000] I can't disable posting to public-whatwg, so let's leave it unsubscribed for whatwg⊙wo for now, until I can talk with the systems team about setting it up some other way [12:38:59.0000] they're in mbox format iirc [12:39:10.0000] Hixie: excellent, thanks [12:39:16.0000] MikeSmith: it's subscribed, just with mail delivery administratively disabled [12:39:21.0000] OK [12:39:38.0000] MikeSmith: happy to turn it back on as soon as we can do it in a way that doesn't lead to people's feedback being lost :-) [12:39:39.0000] I will let you know as soon as I can make time to talk with them about it [12:39:44.0000] roger [12:44:46.0000] Hixie: just minor, but it would be good to have an updated spec, especially for ArrayBuffer -> ArrayBufferView and not throwing for isolated surrogates [12:50:59.0000] k [12:51:05.0000] going through bugs now [12:51:42.0000] (it would be good in general to update all the specs i work on. not clear why websockets is more urgent than the others :-P) [13:10:26.0000] where does typed array define what an arraybufferview represents? [13:12:15.0000] Does it? [13:13:02.0000] i guess not [13:13:16.0000] this spec in general seems to be missing clear conformance requirements and definitions [13:13:35.0000] Yes [13:13:44.0000] And it tries to expose endianness [13:15:36.0000] and NaN bit patterns [13:20:46.0000] Hixie: mostly Microsoft I guess and because it's in CR at the W3C :/ [13:21:32.0000] thanks for fixing [13:36:02.0000] Heh, apparently we don't like Shelley Powers because we're sexist [13:41:27.0000] But we like Anne [13:42:55.0000] has there been a new Shelley-related incident? [13:44:29.0000] Philip`: was waiting for that :p [13:44:48.0000] othermaciej, http://lists.w3.org/Archives/Public/www-archive/2012Apr/0073.html [14:09:55.0000] document.activeElement should never be inside the shadow DOM, correct? [14:10:36.0000] correct [14:11:08.0000] Time to file a bug with Mozilla... [14:11:38.0000] Mozilla implements shadow DOM? [14:12:04.0000] I don't know if they expose it, but they seem to be using it for [14:12:41.0000] that's XBL I guess [14:12:49.0000] sounds like a pretty bad bug [14:13:07.0000] At least that's what it looks like. When gets focus by clicking on the text field portion of the element, document.activeElement is [14:13:33.0000] http://jsfiddle.net/wfkxu/ [14:13:47.0000] be interesting to poke around [14:26:36.0000] annevk: are these e-mails ones i should send you and ms2ger? http://www.whatwg.org/issues/#dom-core [14:27:11.0000] /me looks [14:27:31.0000] Hixie: I already replied to those [14:27:51.0000] Hixie: wait hmm [14:28:00.0000] Hixie: so we moved classList, not sure if it's removed from HTML [14:28:31.0000] Hixie: we have an open bug on extending DOMTokenList [14:28:53.0000] I have pointed out the DOMTokenList thing in the past on whatwg⊙wo [14:29:00.0000] not sure about classList [14:29:07.0000] annevk: k [14:29:28.0000] annevk: so do you want me to send you those as a batch and have you reply to them, or should i reply to them pointing to the dom core spec? [14:56:44.0000] Hixie: the latter I guess [14:56:50.0000] k [14:57:03.0000] Hixie: well, and maybe you should remove classList... [14:57:24.0000] but maybe you want to wait until we convince more impls to do that [14:57:31.0000] I saw a patch from sicking [14:57:33.0000] i don't agree that we should do that in the first place, so... :-) [14:58:42.0000] (because SVG className is not hte same as HTML className) [14:59:19.0000] ah yeah [14:59:24.0000] SVG will just have to change [14:59:32.0000] good luck with that [15:11:08.0000] it's really hard to decide on an email address [15:11:41.0000] o_O [15:11:54.0000] Hixie: I have issues [15:11:57.0000] :) [15:12:41.0000] annevk: It is. I can't find any way to eliminate the redundancy in “w⊙wn”. [15:14:41.0000] wilhelm: ah you already got a short domain [15:14:41.0000] i thought ian⊙hc was pretty simple, but it seems to have just made people think my name is "ian hixie" [15:14:47.0000] haha [15:15:08.0000] I think my shortest domain is quuz.org so I was thinking m or a @quuz.org [15:15:22.0000] but maybe I should go with the more verbose m @ annevankesteren.nl [15:15:32.0000] m? [15:15:36.0000] mail [15:15:56.0000] I've been trying to get hold of anne.nl but no luck [15:16:10.0000] how about annevk.nl? [15:16:16.0000] and an.ne requires a company in Niger and Niger also requires at least three letters I think [15:16:24.0000] an.ne is ugly [15:16:31.0000] Hixie: yeah that might be a good compromise [15:16:37.0000] like someone shot your name or something [15:16:38.0000] pad out with extra ns [15:16:40.0000] annnn.ne [15:16:51.0000] I'm not a big fan of "annevk" I prefer "anne" [15:17:08.0000] ah [15:17:08.0000] I like relevant country TLDs. anne*.nl is appropriate. [15:17:18.0000] yeah, mine is quite irrelevant [15:17:25.0000] I can get anne.io [15:17:27.0000] ted at oconnor.cx [15:17:29.0000] which is kind of funny [15:17:49.0000] annevk: the solution is simple: found your own country and get the .vk tld [15:18:02.0000] me⊙av [15:18:05.0000] true [15:18:13.0000] annevk: btw the advantage of something like anne⊙st rather than mail⊙as is that it makes sense as a jabber alias too [15:18:30.0000] you could just buy .anne [15:18:37.0000] it's only a few hundred grand [15:18:40.0000] I like how you use "just" [15:18:46.0000] Totally worth it. [15:18:52.0000] anne@anne [15:18:55.0000] no extensions [15:19:00.0000] dude if you did that we'd all be like "well crap, now we have to do that too" [15:19:01.0000] deal with that silly email regular expressions [15:19:16.0000] anneisaguy.com [15:19:44.0000] anne⊙nc [15:19:56.0000] actuallyaguy.com is free [15:19:58.0000] anne⊙ac [15:20:06.0000] kind of lame though [15:20:09.0000] you're name is Aguy? [15:20:12.0000] your [15:20:12.0000] your [15:20:14.0000] jesus [15:20:22.0000] haha [15:20:52.0000] notagirl.com smells of being a phishing scam [15:22:26.0000] anne.is/a.gentleman [15:22:35.0000] .is is available. :P [15:22:45.0000] And .eu is for sale, apparently. [15:23:49.0000] the whole TLD? [15:24:36.0000] Just a very limited subset. [15:27:07.0000] dreaming of the day when the password form fills in on paypal+firefox without having to click the username then the password field [15:28:53.0000] annevk: heh the only problem with my address is it has to be spelled out [15:30:00.0000] thanks for the ideas everyone, I'll ponder over it some more :p [15:30:27.0000] wilhelm: Yoda-style, "Gentleman, Anne is" [15:30:49.0000] That could work. [15:32:23.0000] Hixie: can we use your annotation script on whatwg in webapps WG? [15:32:31.0000] Hixie: would it be possible to open-source it somewhere? [15:32:51.0000] it's pretty specific to HTML, but I can make the code available if someone wants to host a fork of it somewhere else, sure [15:32:52.0000] Hixie: we would like to be able to list tests per section for example [15:32:56.0000] Hixie: excellent! [15:32:59.0000] whoever wants to have the code should e-mail me [15:33:04.0000] okay [15:33:14.0000] it's pretty hairy [15:33:19.0000] might be easier to start from scratch [15:34:04.0000] I'd love to see those annotations in the XHR2 spec [15:34:24.0000] (i'm amused that there are people in the meeting who weren't familiar with this stuff) [15:34:30.0000] (how is that even possible?) [15:35:31.0000] Hixie: how Bush got elected still boggles my mind too [15:35:51.0000] that seems qualitatively different [15:36:14.0000] I guess, though sometimes this stuff feels just as weird [15:37:53.0000] WeirdAl: yeah, all over really [15:40:09.0000] i've been saying this for years [15:42:55.0000] annevk: elected, maybe ... it's the re* part that's confusing [15:45:16.0000] it's been months since I had a tshirt idea: http://www.zazzle.com/webkit_is_awesome_mug-168431797145781438 2012-05-03 [23:28:22.0000] surely, anne⊙kn (@ pronounced as "van") [23:31:47.0000] Hixie: there's no technical reason .className couldn't be overridden for SVG elements to return something different. It's just a really confusing. But no more confusing than .className being different in HTML and in SVG [23:31:58.0000] agreed [23:32:51.0000] is .className different in HTML and SVG? [23:33:03.0000] yes [23:33:50.0000] s/is/how is/ [23:33:56.0000] string vs token list? [23:34:58.0000] string vs svg animated string value [23:35:04.0000] oh [23:35:19.0000] good lord, why would you use svg animation to animate the class? [23:35:45.0000] wasn't there discussion about changing the svg dom to make it more like the html dom? [23:36:02.0000] or is there enough legacy depending on the current svg dom to make it impossible to change? [23:55:07.0000] maybe the placeholder thing in contenteditable should be an element instead of an attribute? [23:56:01.0000] or maybe not [02:12:45.0000] "This demo uses the Audio Data API in Firefox, and Web Audio in Chrome. Please agree on a single API, browser developers!" https://developer.mozilla.org/en-US/demos/detail/audiospank [02:13:25.0000] nice idea; the first multiplayer FPS game playable by blind people that i've heard of [02:13:36.0000] (haven't tried it myself yet) [02:37:20.0000] please join browser development, web developers! [04:25:51.0000] Bah, Ms2ger and annevk aren't here. [04:30:50.0000] We're not good enough for you? [04:31:01.0000] :'( [05:14:04.0000] You called? [05:14:15.0000] (Feel free to add that note) [05:15:38.0000] $ grep -aPihc "word-wrap\s*:\s*break-word" stevef-all [05:15:38.0000] 262 [05:15:38.0000] $ grep -aEihzc "word-wrap[[:space:]]*:[[:space:]]*break-word" stevef-all [05:15:38.0000] 67 [05:16:29.0000] (67 pages out of the top 10,000 use word-wrap:break-word; total occurrences 262) [05:18:47.0000] AryehGregor: ^ [05:19:05.0000] Ms3ger, okay, thanks. [05:19:16.0000] Np [05:22:24.0000] /me goes off to class [05:35:04.0000] http://www.w3.org/TR/ttaf1-dfxp/#vocabulary-namespaces aaaaargh [05:36:05.0000] hello [05:37:10.0000] what's going on with CORS simple headers list? [05:37:37.0000] zcorpan: I believe the translation from spec speak to english of that section is "The WG are as high as a kite" [05:38:07.0000] http://dvcs.w3.org/hg/cors/raw-file/tip/Overview.html#simple-header [05:38:55.0000] "Last-Event-ID" is missed [06:35:29.0000] why is MozMutationObserver prefixed in firefox? [06:36:41.0000] MozMozMutationObserver? [06:47:14.0000] hsivonen: you around? [06:50:45.0000] zcorpan: getting kesteren.nl is hard (it's also a place name) and vankesteren.nl is owned by some web design agency [06:58:55.0000] annev⊙nn then? too ugly? :) [06:59:06.0000] too ugly [06:59:19.0000] could use kesteren.org I guess [07:04:38.0000] argh, why is firstElementChild defined on Element only? [07:05:13.0000] please make the element traversal methods available on more nodes [07:05:41.0000] really? [07:06:18.0000] you want that on DocumentFragment? [07:06:24.0000] yeah, and Document [07:06:44.0000] what is wrong with documentElement? [07:06:59.0000] nextElementSibling i want on all nodes, i think [07:07:31.0000] annevk: i need to check if foo is a document or an element instead of just using firstElementChild [07:08:43.0000] zcorpan: annevk: I've had that use case in the past too [07:09:36.0000] what kind of weird code are you guys writing? [07:10:23.0000] annevk: specifically here: https://bitbucket.org/runeh/unbose/src/9d94749c9236/src/unbose.js#cl-47 [07:10:38.0000] annevk: could have just used firstElementChild directy there [07:12:07.0000] so nextElementSibling is available on nodes other than Element? [07:12:16.0000] that code seems wrong [07:13:14.0000] afaict [07:13:15.0000] if (child.nodeType != 1 /*ELEMENT*/) { child = child.nextElementSibling; } [07:13:18.0000] makes child undefined [07:17:13.0000] annevk: if you have foo and want the next element sibling, it makes sense to just use nextElementSibling even if foo is not an element [07:17:33.0000] annevk: hmm, yea, I guess that's not tested enough. But I was going to write what zcorpan just said [07:18:29.0000] I agree with that, I'm just saying this code does not work [07:18:47.0000] annevk: yea [07:19:01.0000] ok i'll file a spec bug [07:19:48.0000] I'm not quite convinced on the others though [07:20:34.0000] annevk: which ones? [07:23:54.0000] first/lastElementChild [07:24:01.0000] oh, and childElementCount should die [07:25:01.0000] if we can kill childElementCount, that's fine with me. if we can't, we might as well make it available on the same places we make .children available [07:25:14.0000] filed https://www.w3.org/Bugs/Public/show_bug.cgi?id=16919 [07:26:15.0000] annevk: i think the rationale for this is really the same as for the new mutation methods [07:38:20.0000] jgraham: html5lib, Python 3. This is going to be fun unless we drop Py<2.6 support to explicitly have binary strings. [07:39:51.0000] Almost tempting to go Py3 native and use 3to2. [07:41:13.0000] (As then we fail hard given any Unicode/Bytes mismatch being introduced) [08:02:50.0000] gsnedders: I am happy to drop py<2.6 support and to go python 3 and use 3to2 [08:03:07.0000] In fact I'm not aware that there is another approach that sould work [08:03:10.0000] *could [08:30:06.0000] [09:01:07.0000] zcorpan: those namespaces are nothing compared to SMIL [09:02:47.0000] http://annevankesteren.nl/2006/03/smil is still a fun read [09:05:47.0000] AryehGregor: I missed that Ms3ger already ok'd a change, a change works for me, just seems overkill [09:39:42.0000] maybe I should do annevk⊙an [09:39:47.0000] that's not that bad [09:47:09.0000] http://w3.markmail.org/search/ public-html is "only" 14k messages behind www-style, despite www-style existing like forever [09:55:03.0000] good morning, Whatwg! [09:55:16.0000] Morning [10:03:37.0000] good evening Ms2ger [10:04:29.0000] So, did anything useful happen yesterday? [10:05:14.0000] yeah man, I added Event.NONE [10:05:51.0000] Mozilla implemented it first ;) [10:06:05.0000] annevk: People use that or me⊙an -- but annevk⊙an is easy to say. [10:06:24.0000] Ms2ger: haha [10:06:25.0000] Velmont, yeah, and then gmail claims he's called "me" [10:22:18.0000] annevk: :) [13:33:34.0000] argh [13:33:43.0000] w3c validator http://jigsaw.w3.org/css-validator/#validate_by_uri+with_options [13:33:53.0000] cannot be made to accept syntax like background: -moz-linear-gradient(top, #878787, #000000); [13:34:06.0000] are there any known workarounds? :) [13:34:19.0000] Moo^: is that valid CSS? [13:34:22.0000] if it is, file a bug [13:34:43.0000] MikeSmith: there is option to accept vendor prefixes [13:34:48.0000] but apparently it is ignored for values [13:34:52.0000] ah [13:35:06.0000] that is still worth filing a bug [13:36:37.0000] we have a developer who is actively working on updates to the CSS validator right now [13:36:57.0000] but not sure how long he will be doing it, so strike while the iron's hot [13:39:32.0000] MikeSmith: I stroke [13:39:50.0000] thanks [13:39:56.0000] hai [14:30:16.0000] some e-mail here suggests a mechanism whereby UAs could hash passwords, and AryehGregor brings up the need for per-user hashing [14:30:31.0000] per-user salting, rather [14:30:32.0000] is there any reason the username couldn't be used as the salt? [14:49:04.0000] Hixie: Sounds bad if you want to change the username later [14:49:18.0000] (e.g. if you use email address as username, and the user changes their email address) [14:56:45.0000] I'm now marking promoted tweets as spam [14:57:11.0000] me too [15:54:25.0000] aaaargh [15:54:36.0000] so I do the forward thing [15:54:51.0000] but Gmail keeps giving me "Delivery Status Notification (Failure)" [15:55:04.0000] while when sending anything to the address does arrive at the address [15:55:33.0000] it seems like DreamHost is blocking a particular kind of email from Gmail 2012-05-04 [17:49:36.0000] can a web worker script be loaded from a different domain? [18:56:47.0000] what should I do if I found a typo in the html spec ? send a email to whatwg⊙lwo ? [19:33:43.0000] padenot: mention it in the box at the bottom left of the spec and hit the button there [19:38:12.0000] Hixie: done, thank you. [19:39:39.0000] padenot: thank _you_! [19:39:50.0000] padenot: if you want to be acknowledged in the spec, mention your name in the bug, too [22:14:05.0000] annevk, I only asked because smaug pinged me about it. I'm okay with no spec change yet -- the relevant implementers are aware of it (at least Gecko/WebKit). [22:16:20.0000] Opera/IE too if they follow bug files against DOM which they should [22:16:33.0000] goes for every vendor really [22:16:35.0000] Hixie, MediaWiki uses the user id as the salt. As Philip` points out, the username doesn't work, unless either a) the username never changes, or b) the user is required to enter their password before the username change takes effect (bad if you want admins to be able to unilaterally change names), or . . . there are probably other things you could do. [22:17:08.0000] Also, it would be a problem if the username is not bit-for-bit what the user entered -- e.g., if whitespace is trimmed, or it's case-insensitive, etc. [22:17:24.0000] Applications tend to do various sorts of application-specific username normalization. [22:19:24.0000] Using user id as the salt only has the problem that they don't vary as much as a smallish random ASCII string -- there's probably a user 1, user 2, etc. -- so it doesn't provide as much protection as you'd want against dictionary attacks. [22:20:23.0000] Ideally you'd probably want to use a salt that's itself longer than any feasible dictionary, so that attackers can't use a dictionary to recognize the hash of even an empty-string password. [22:20:57.0000] (by "itself longer than any feasible dictionary" I of course mean "longer than could be cracked by any feasible dictionary even before you add the password") [22:22:50.0000] The latter isn't an issue for MediaWiki, though, because it uses md5(id + '-' + md5(password)), so the thing you're hashing is always long. There's not much likelihood of being able to save anything over using brute force here. [02:19:07.0000] hsivonen: ping. [07:22:18.0000] Hmm, browsers seem intent on defying me [07:23:58.0000] If I add an async script (e.g. through DOM methods), then stop sending data for the main document for a bit (e.g. by doing ob_flush(); flush(); sleep() in PHP) then I would expect the (fast-loading) async script to have run before parsing resumes [07:24:11.0000] But either I am doing it wrong or that isn't happening [07:24:22.0000] *sleep(10) [07:24:59.0000] do you get a load event for the script? [07:25:44.0000] The async script? [07:26:13.0000] yeah [07:26:53.0000] Yeah. [07:27:43.0000] Well actually in gecko it seems OK now [07:27:52.0000] Dunno if I changed something significant [07:28:41.0000] Chrome still fails [07:28:58.0000] And Opera is just weird [07:32:24.0000] Opera was weird first? [07:34:43.0000] jgraham: i wouldn't be surprised if browsers treated async scripts as if they happen to load after the page finishes loading, to reduce racy site bugs [07:36:31.0000] That could be what webkit's doing I guess [07:36:39.0000] But then the spec is a lie [07:37:12.0000] well, if the behavior is indistinguishable to scripts, it's not really [07:37:27.0000] that's defer... [07:37:44.0000] eg. if conceptually the script is stuck in a TCP receive buffer and hasn't been received by the browser until the page happened to finish loading [07:38:34.0000] well, defer does a bit more, iirc (guarantees order, and iirc guarantees scripts happen before load? not sure offhand) [07:38:56.0000] Right defer is different [07:46:14.0000] i'm surprised anyone supports async, since it seems like an interop nightmare [07:46:40.0000] (at least as it's described above--been a while since I've really squinted how it works) [09:31:16.0000] good morning, Whatwg! [09:31:42.0000] good evening dglazkov [10:18:36.0000] Anyone got any experience of dealing with online stalkers? [10:20:49.0000] gsnedders - nothing for stalkers in particular, but perhaps you can find something you can re-use from the Troll Taxonomy: tantek.com/w/TrollTaxonomy [10:22:20.0000] (If anyone wants to actually have any discussion about that with me, PM.) [10:22:27.0000] (Logged channel, etc.) [10:49:38.0000] hmm http://lists.w3.org/Archives/Public/public-whatwg/2012May/0015.html [10:49:52.0000] we should get this mailing list mess sorted [10:50:21.0000] yeah, mike is on it [10:51:08.0000] Hixie: strawman proposal: make public-whatwg⊙wo the actual list, move existing whatwg⊙wo subscribers over to that [10:51:15.0000] not happening [10:51:18.0000] OK [10:51:20.0000] :-) [10:51:23.0000] I had to ask [10:51:28.0000] we have way too much stuff pointing to whatwg@ already [10:51:34.0000] ok [10:51:49.0000] (and i'd have to have admin rights to public-whatwg to be able to ban people, etc, which the cg rules say i can't have) [10:52:50.0000] public-whatwg⊙wo shouldn't even exist if you ask me [which you didn't :)] [10:52:51.0000] this isn't a problem limited to the whatwg cg list, anyway. the editing cg has the same problem, for instance [10:53:02.0000] hober: agreed [10:53:05.0000] right. all cgs which use pre-existing infrastructure [10:53:10.0000] right [10:56:11.0000] yeah, I have pointed out to others on the team that this should be configurable [10:56:21.0000] unfortunately it is currently not configurable [10:56:26.0000] MikeSmith, Hixie: would it be feasible to, for example, subscribe public-whatwg⊙wo to whatwg⊙wo, so joining the CG puts you on the right list? [10:56:37.0000] othermaciej: done that [10:56:40.0000] we had to revert it [10:56:47.0000] oh? [10:56:50.0000] people then get duplicate messages [10:57:04.0000] and really we don't need a new mailing list [10:57:12.0000] we have a perfectly good mailing list already [10:57:14.0000] othermaciej: the other worry is people posting to public-whatwg [10:57:18.0000] right [10:57:25.0000] othermaciej: and their messages being lost in the ether [10:57:31.0000] hmm, good point [10:57:34.0000] as just happened with dougt's message [10:57:38.0000] yeah [10:57:41.0000] (i told dougt to repost to whatwg) [10:57:52.0000] (though thanks to annevk for pointing it out because looks like i never got that mail) [10:58:00.0000] (dunno why) [10:58:40.0000] I guess to really make use of the pre-existing list, joining the CG should automatically add you to the right list, and the public-whatwg@ list should not even be in the picture [10:58:59.0000] there's really no reason for anyone to be joining the CG unless they have specific patents they want to grant [10:59:07.0000] and even then -- i don't know that the FSA requires membership [10:59:32.0000] the exception would be someone like anne who wants to contribute actual text (as with his specs that he submitted recently) [11:08:07.0000] talking with systems team about it now [11:18:06.0000] will create public-whatwg-archive⊙wo for the archive mirror [11:18:33.0000] nice [11:18:33.0000] and close public-whatwg⊙wo to postings [11:18:37.0000] yay [11:18:42.0000] +1 to systeam [11:21:36.0000] whooo! [11:31:58.0000] MikeSmith: public-whatwg-archive also being closed for postings right? [11:33:17.0000] annevk: yeah, but that part I need to figure out [11:33:29.0000] because it does actually have to accept some postings [11:33:43.0000] that is, the stuff that's posted to whatwg⊙wo [11:33:56.0000] which is from various addresses of course [11:34:45.0000] well all the e-mails will look like they're coming from the whatwg list, right? [11:34:59.0000] can you just mark it moderated except for the whatwg list? [11:35:35.0000] MikeSmith: ah okay [11:36:14.0000] Hixie: yeah, I'm sure it's doable. Just need to get the attention of our mailing-list guru on the systems team [11:36:17.0000] MikeSmith: it seems it should be possible to only accept the messages that come back from whatwg⊙wo [11:36:29.0000] right [11:48:22.0000] /me finds himself replying to his own feedback on whatwg@ [11:48:24.0000] (and rejecting it) [11:49:16.0000] hey past-Hixie, you were wrong. sincerely, today-Hixie [11:50:00.0000] i didn't even realise that i was the one who wrote it until i noticed someone else replying to that e-mail saying i'd written it, because i don't pay any attention to who writes the feedback when replying [11:56:49.0000] what happened to the ability to mark log lines as important in the logs? [11:57:38.0000] bit late, but thanks to hober we now have http://w3cmemes.tumblr.com/post/22393119375 [11:57:50.0000] heh [11:58:13.0000] need to get a bit more creative [12:01:30.0000] http://w3cmemes.tumblr.com/post/22393334041 [12:02:54.0000] http://imgflip.com/i/1avf [12:03:09.0000] hehe, I'll submit that [12:14:22.0000] http://w3cmemes.tumblr.com/post/22393866467 [12:14:36.0000] ok i'll stop being a troll now [12:38:30.0000] has something changed in freenode? it started to require sasl here [12:38:39.0000] /me has no idea what sasl is [13:05:01.0000] Hixie: hahahaha [13:07:23.0000] smaug____: sasl is the ssl email auth thing? Seem to remember something like that when I set up email server. [13:11:11.0000] Velmont: not ssl [13:11:25.0000] Velmont: I do use ssl connection to irc.mozilla.org for example [13:11:40.0000] freenode started to require sasl for me [13:11:48.0000] (using a different connection than usually) [13:12:09.0000] so, I had to install sasl plugin to Chatzilla [13:13:35.0000] i've had that message when tethering before [13:23:19.0000] smaug____: use irccloud.com [13:24:14.0000] dglazkov, do you know how badly that leaks? ;) [13:24:39.0000] and irccloud is pretty horrible webapp anyway [13:24:51.0000] Ms2ger: leaks what [13:25:03.0000] All the objects [13:25:07.0000] dglazkov: it kills all the known browsers [13:25:15.0000] smaug____: works great on Chrome? [13:25:20.0000] well, in chrome is kills only the tab I think [13:25:36.0000] dglazkov: have you kept it open for a long time? [13:25:46.0000] um... you're talking with someone who's been using it for a fairly long time. [13:26:15.0000] dglazkov: you keep it open for weeks? [13:26:24.0000] no, why do I need to? [13:26:41.0000] well, to have IRC open all the time [13:27:01.0000] I know it keeps the connection open in the background sure. [13:27:04.0000] irccloud makes you always online. You don't have to have the tab open [13:27:15.0000] but if you actually *use* irc all the time [13:27:30.0000] I keep it open when I use it, sure. [13:28:04.0000] anyhow, I prefer Chatzilla :) [13:28:07.0000] it doesn't leak [13:28:18.0000] smaug____: that's fine :) I was just trying to be helpful. [13:28:42.0000] /me should continue investigating leak logs from Google Reader. [13:29:17.0000] I wonder if anyone from chrome team looks at leaky web sites [13:29:26.0000] s/looks/investigates/ [13:30:08.0000] You should all be using irssi anyway [13:30:28.0000] annevk: http://w3cmemes.tumblr.com/post/22398040394 :-P [13:30:44.0000] jgraham: yeah, irssi is an option [13:30:50.0000] but don't like its UI [13:30:55.0000] I use irssi on my N9 [13:30:56.0000] smaug____: yep. ask simonjam, anniesullivan, or tonyg on #webkit [13:31:09.0000] smaug____: for some reason they are all offline atm [13:31:30.0000] dglazkov: Maybe their irccloud tabs died :p [13:31:42.0000] dglazkov: just wondering if you see Google Reader to leak in Chrome too [13:32:04.0000] but I'll ask them later [13:32:19.0000] smaug____: no clue. I've been spec-writing for the last 4 months, so I don't know what is real or not anymore. [13:32:26.0000] http://lists.w3.org/Archives/Public/www-style/2012May/0125.html [13:32:27.0000] heh [13:33:40.0000] Pretty disappointing responses so far. bz seems convinced that prefixes are needed despite the fact that they are presumably screwing over Mozilla hard on mobile [13:33:58.0000] Although his points are not unreasonable [13:34:25.0000] bz's points rarely are unreasonable [13:35:54.0000] too true [13:37:43.0000] you can usually make changes if only one browser supports it [13:37:46.0000] jgraham: needed ever, or needed for the same extent they are applied currently (til CR) [13:37:59.0000] if you wait five years though and everyone else supports the features too with prefixes [13:38:02.0000] well then you're fucked [13:38:21.0000] othermaciej: Closer to "needed ever" [13:38:47.0000] I have only read fragments of that thread but I think I might sort of agree with his position [13:39:07.0000] Hixie: guess I did troll, my point took about 25min to sink in [13:39:27.0000] I would say properties should be prefixed until a condition is met such as: (a) spec taken up as work item by the WG; (b) there are at least two implementations [13:39:36.0000] (not necessarily complete ones) [13:40:03.0000] maybe even only (a) [13:40:38.0000] That is still going to leave us with content that will only work in a single implementation even when other implementations have the feature [13:40:46.0000] Although possibly less of it [13:41:00.0000] i like how the htmlwg is now talking about how they are the group that "owns" stuff [13:41:35.0000] as if that was how it worked :-) [13:42:52.0000] :P [13:42:52.0000] don't tell glazman [13:45:24.0000] (it will also leave authors writing content that will break if changes are made to the unprefixed semantics because they will include both and put the unprefixed version last) [13:46:50.0000] New bug: "I don't like public data" [13:46:51.0000] Hah [13:53:49.0000] i need images for [13:54:09.0000] "chair did not know HTML5 includes XML syntax" [13:54:17.0000] and [13:54:18.0000] "going through the charter line by line; because that's how we look productive" [13:59:26.0000] tantek: is the neato from you? [13:59:35.0000] Soo funny :D [14:00:18.0000] Condescending Wonka++ [14:05:13.0000] anne: http://w3cmemes.tumblr.com/post/22400217421 [14:05:52.0000] :-P [14:05:56.0000] Don't you all have work to do? Reviewing charters? :) [14:07:02.0000] there's enough acidity in these memes to burn holes through steel. [14:07:20.0000] I have another idea: "working on clever memes? don't you have work to do?" :P [14:07:26.0000] Hixie: oh god [14:08:33.0000] dglazkov, Web Components? [Condescending Wonka] Yeah, that will go well [14:08:41.0000] Anne laughed out loud :| doesn't fit with the meeting, hard to hold it. [14:09:06.0000] Ms2ger: heey. [14:09:24.0000] /me gives Ms2ger a hurt look [14:09:39.0000] /me pats dglazkov on the back [14:09:48.0000] :) [14:09:49.0000] We all have silly ideas at some point ;) [14:10:11.0000] one of the recent memes forced me to leave the room for a laugh break. i'm not saying which one. [14:19:31.0000] the meme meeting backchannel .. breaking new ground [14:20:57.0000] do web workers support CORS? [14:21:05.0000] MikeSmith: that's the only thing that tells me that there's a meeting. [14:21:23.0000] MikeSmith: I had completely blown off htmlwg meet. I was worried for my health. [14:22:00.0000] bencc: not yet [14:22:03.0000] bencc: but it is on the cards [14:22:14.0000] besides, isn't HTML5 done already? What's there to meet about. [14:22:29.0000] bencc: though actually the alternative suggestion is just to run the script in the other origin [14:22:54.0000] bencc: i haven't added this to the spec yet mostly because i'm not sure whether doing so would cause the w3c trouble (they're trying to publish a snapshot) [14:23:13.0000] bencc: and i'm not sure whether to just stop updating the w3c copy or wait or slow down the w3c process [14:23:18.0000] Hixie: how can I run the script in the other origin? [14:23:41.0000] bencc: right now you can't, but the idea would be that you can launch a worker and it runs in the worker's script's origin, if it opts in to doing so [14:23:42.0000] Hixie: agreed. cors support for webworkers will be the last update though :) [14:23:54.0000] there are other things suggested also [14:24:06.0000] e.g. spawning a thread from a data: url or text string of code [14:24:20.0000] or using workers in appcache [14:24:23.0000] dglazkov: In the absence of anything technical, one can always talk Process [14:25:06.0000] Hixie: all suggestions will be very wellcome [14:25:08.0000] jgraham: true. [14:26:17.0000] dglazkov: we did actually get people to agree on some of the open issues [14:26:21.0000] at least, the set of people here [14:27:00.0000] othermaciej: that's great to hear [14:27:38.0000] for example, it was generally agreed that replicating the ARIA syntax requirements for every ARIA attribute in the HTML5 spec was not a good idea [14:28:46.0000] Ohrly [14:36:14.0000] othermaciej: http://w3cmemes.tumblr.com/post/22402024765 ! [14:37:25.0000] :-p [14:37:48.0000] It's funny, because it's eating your soul [14:46:19.0000] Hixie: http://w3cmemes.tumblr.com/post/22402668355 [15:08:35.0000] suddenly lots of longdesc memes [15:09:02.0000] longdesc the new black [15:09:46.0000] hmm continued forking of addEventListener by Mozilla [15:09:47.0000] meh [15:09:52.0000] well dougt [15:10:38.0000] wow who did http://w3cmemes.tumblr.com/post/22404077991? that's an elaborate meme! [15:10:59.0000] annevk: hm? [15:11:19.0000] Hixie: creating more events that when registered using addEventListener will cause them to be dispatched [15:11:32.0000] Hixie: see WHATWG thread on proximity and such [15:12:13.0000] the ones i saw are just dispatched always [15:12:16.0000] at least in principle [15:12:36.0000] it's just a UA optimisation that they aren't using battery life when there's no listener [15:12:55.0000] not the same as orientation where adding the listener causes an event to be fired that wouldn't normally be fired [15:13:17.0000] the meeting now is hilarious [15:13:37.0000] for about thirty minutes there's some discussion about whether or not the spec should suggest that authors should not or must not do something [15:13:52.0000] or whether maybe UAs should be forbidden to implement something useful because not everyone supports it [15:14:25.0000] especially the emphasis that is put on how these statements in the spec are actually going to matter in reality [15:14:41.0000] like authors would read that and like, oh "i should not" better think about this again [15:15:11.0000] /me brought the popcorn [15:15:35.0000] more like forty minutes [15:15:48.0000] AND ONGOING [15:17:52.0000] Hixie: the moment the pattern exists it's likely to be reused though elsewhere [15:18:00.0000] Hixie: but mkay [15:19:16.0000] what pattern? [15:19:24.0000] *So* much talk and so little action. [15:19:26.0000] register for weird event [15:19:29.0000] event dispatches [15:19:45.0000] from now on i'll communicate [15:19:48.0000] that isn't what dougt proposed here [15:19:48.0000] in memeable lines [15:20:14.0000] Hixie: I made it. Not really elaborate, just different. [15:20:30.0000] Velmont: it's like three memes in one! [15:20:35.0000] Velmont: or did you find an existing template? [15:20:45.0000] Hixie: http://dougturner.wordpress.com/2012/03/22/device-proximity-sensor/ seems like it is? [15:21:00.0000] Hixie: No, just threw it into Gimp. [15:21:09.0000] Velmont: i go back to elaborate! [15:21:43.0000] annevk: that is black-box indistinguishable from just saying that "a deviceproximity event fires whenever the proximity changes" [15:22:03.0000] annevk: note in particular that he doesn't say that a new event should fire when you listen, even if the proximity doesn't change [15:23:31.0000] <- skeptic [15:27:14.0000] Hixie: in your response to the template element parse thread [15:27:24.0000] Hixie: you gave two examples where the result will differ [15:27:24.0000] annevk: send feedback on the list to make sure he doesn't mean to fire an event when you attach [15:28:16.0000] Hixie: are you saying that it's weird because the resulting fragments are very different due to the first element being parsed? [15:29:07.0000] Hixie: but isn't typical in the html parser? as in depending on what element we encounter, we'll drop some nodes or insert elements automatically [15:29:55.0000] arv: http://dvcs.w3.org/hg/url/rev/93306cbfbc62#l1.95 0_0 [15:30:08.0000]

http://instagr.am/p/KOLQO0g9Vd/ [15:32:00.0000] MikeSmith: fixing [15:32:51.0000] MikeSmith: looks like HTML XML-serializing [15:33:02.0000] anolis bug? [15:33:40.0000] rniwa: not nodes that came before, usually [15:33:58.0000] Hixie: oh I see what you're saying [15:34:07.0000] rniwa: and the few cases where we do mess with nodes that came before, we do it only because the input was seriously broken and we do it only because we were forced to for web compat [15:35:47.0000] Hixie: yeah. did you see annevk's response on 4/16 though? [15:36:05.0000] give me some text to search for? [15:36:23.0000] "You cannot look at various elements and make a decision" [15:36:28.0000] "E.g. if your first element is there will not be any other elements." [15:36:31.0000] <rniwa> and [15:36:36.0000] <rniwa> "Defining a new top-level insertion mode while retaining compatible behavior might be an interesting exercise in parser complexity, but it's not clear there is a benefit (use cases?) and feasibility has not been demonstrated (consider handling <p><td><p>, <tr><p><td>, ...). The more we can define in terms of the existing parser, the better it is for developers. The behavior will be more predictable and there will be less quirks t [15:36:36.0000] <rniwa> learn." [15:36:44.0000] <Hixie> the first was sufficient :-) [15:37:30.0000] <rniwa> Hixie: that argument convinced me that the proposed approach is better than allowing any elements to appear [15:37:33.0000] <Hixie> it's hard for me to evaluate without knowing what cases we're trying to handle, really [15:37:37.0000] <rniwa> Hixie: but you made a really good point as well. [15:37:49.0000] <Hixie> but i'm pretty sure we never want to change the result of earlier parsing based on later characters [15:38:02.0000] <rniwa> Hixie: rafael listed all cases in one of his emals [15:38:14.0000] <rniwa> Hixie: see his point on 4/25 [15:38:23.0000] <Hixie> same thread? [15:38:25.0000] <rniwa> Hixie: search with "rp, rt => HTMLRubyElement" [15:38:27.0000] <rniwa> Hixie: yup [15:38:41.0000] <rniwa> Hixie: http://lists.w3.org/Archives/Public/public-webapps/2012AprJun/0334.html [15:38:53.0000] <Hixie> that doesn't list the actual use cases we want to parse [15:39:02.0000] <Hixie> just gives some examples of the results [15:39:08.0000] <rniwa> Hixie: ah i see [15:39:18.0000] <Hixie> (for example, i assume anything to do with <frameset> we don't really care about the result) [15:39:35.0000] <rniwa> Hixie: he does list some cases at the end [15:39:46.0000] <rniwa> Hixie: but i suppose you're looking for more comprehensive list? [15:40:10.0000] <Hixie> the ones he lists at the end include something to do with <frameset>, so i assume they're just examples of the results of the proposal and not use cases [15:40:40.0000] <Hixie> i could be wrong :-) [15:40:47.0000] <Hixie> but i'd be surprised if one of the use cases involves framesets! [15:44:00.0000] <Hixie> must. not. use. tantek's photo. for a meme. [16:57:30.0000] <annevk> http://www.netmagazine.com/news/call-greater-diversity-web-industry-121948 2012-05-05 [17:20:01.0000] <hober> annevk: ok, things are wired to twitter now [17:20:23.0000] <annevk> sweet [17:21:00.0000] <annevk> hober: do one manual tweet maybe with the last posted? [17:22:13.0000] <hober> yeah, ok [17:27:14.0000] <annevk> :) [17:29:32.0000] <annevk> hober: add http://28.media.tumblr.com/tumblr_m3ivb8Ldac1rvsbh9o1_400.jpg as icon everywhere? [17:30:20.0000] <hober> aye aye [17:38:50.0000] <tabatkins_> If I want to propose something vaguely related to the 'unload' event, is that best done in whatwg or public-webapps? [17:39:31.0000] <annevk> whatwg [17:39:36.0000] <tabatkins_> Kk, cool. [20:57:08.0000] <cbright6063> lovely 2012-05-06 [15:20:57.0000] <othermaciej> TabAtkins: I respect your position re prefixing but I think you may be suffering from serious status quo bias and anchoring effects 2012-05-07 [23:16:26.0000] <nesta_> good morning! [23:55:52.0000] <jgraham> TabAtkins: Assuming othermaciej was talking about http://lists.w3.org/Archives/Public/www-style/2012May/0215.html, my response is haha, yeah right [23:56:59.0000] <jgraham> Tying unprefixing to the release of a testsuite just seems like a way to prevent it happening [23:58:38.0000] <jgraham> It also creates bad incentives for writing tests since the goal of the tests will be to make the Process transition happen as fast as possible, not to make implemntations that are basically about the same interoperate on edge cases [23:59:52.0000] <jgraham> /me hasn't read the whole thread yet though [00:03:24.0000] <jgraham> and re: the followup on the thread, authors currently assume rough interop anyway when they blindly add -*-whatever properties. And they also assume rough interop between the - often mostly seperate-code - -webkit- implementations [00:03:46.0000] <jgraham> So you are suggesting a bar that is utterly at odds with reality [00:07:14.0000] <jgraham> AryehGregor: If you are there, this might be faster than email. I didn't understand your email to public-html-testsuite [00:07:31.0000] <AryehGregor> jgraham, which e-mail to public-html-testsuite? [00:07:39.0000] <AryehGregor> Did I make one recently? [00:07:41.0000] <AryehGregor> I don't recall. [00:07:56.0000] <AryehGregor> Oh, that one. [00:08:56.0000] <AryehGregor> jgraham, http://dvcs.w3.org/hg/editing/raw-file/583bf580775b/selecttest/collapseToStartEnd.html <-- Works. [00:09:06.0000] <AryehGregor> Because I have: <script src=http://w3c-test.org/resources/testharness.js></script> [00:09:13.0000] <AryehGregor> If I had: <script src=/resources/testharness.js></script> [00:09:16.0000] <AryehGregor> Then it wouldn't work. [00:09:27.0000] <AryehGregor> Because that doesn't exist on dvcs.w3.org. [00:09:31.0000] <jgraham> Oh I see [00:09:45.0000] <jgraham> It's specific revisions of the *tests* that you want to link to [00:09:51.0000] <jgraham> Or open [00:09:54.0000] <AryehGregor> (By sheer coincidence, I was pointing out on Mozilla's Bugzilla that my use of w3c-test.org here makes my tests not importable into Gecko without a separate manual step.) [00:10:13.0000] <AryehGregor> Well, perhaps ideally the test suite should include whatever testharness.js was current at the time it was written. [00:10:14.0000] <jgraham> Yeah, it doesn't work for us either, for the same reason [00:10:18.0000] <AryehGregor> But at least it should work. [00:10:30.0000] <AryehGregor> find -type f -exec sed 's!http://w3c-test.org!!' {} + [00:10:35.0000] <AryehGregor> That works fine. [00:10:52.0000] <jgraham> It works fine, but makes updating things a pain [00:11:09.0000] <AryehGregor> Do you auto-update? In Gecko it's manual, in principle. [00:11:32.0000] <jgraham> No, but basically it's such an effort that we often just don't do it [00:11:41.0000] <jgraham> Which is bad and I want to change [00:11:57.0000] <AryehGregor> Gecko updating of imported test suites is totally scripted. [00:12:02.0000] <AryehGregor> We have Ms2ger to thank for that. [00:12:15.0000] <AryehGregor> (as well as the fact that we support testharness tests at all) [00:12:25.0000] <AryehGregor> Although, updating the list of expected fails is still somewhat tedious. [00:12:29.0000] <AryehGregor> I'll probably want to work on that. [00:13:05.0000] <jgraham> The problem for us, in general, is that we might have local patches to a testsuite (not so much testharness.js testsuites, since they were designed to avoid some of the problems we have encountered) [00:13:21.0000] <AryehGregor> Version them in a DVCS and rebase. :) [00:13:35.0000] <AryehGregor> (we don't have local patches to any testsuites yet, but we haven't imported many either) [00:13:47.0000] <jgraham> Yes, that's the ideal world [00:14:03.0000] <jgraham> For legacy reasons, however, all our tests are on svn [00:14:31.0000] <jgraham> sometimes people have tried horrendous things like putting git repos inside svn [00:14:34.0000] <AryehGregor> Well, there's your problem . . . [00:15:03.0000] <jgraham> Yeah, I know [00:15:55.0000] <jgraham> I know how to fix the problem in an abstract way, I just need to make it work in a concrete way [00:52:49.0000] <tabatkins_> European morning! [00:53:04.0000] <jgraham> Is that another way of saying "bad morning"? [00:53:38.0000] <jgraham> or "Commie socialist bastards morning", perhaps? [00:54:10.0000] <tabatkins_> I'm in Europe, so clearly I need to differentiate this from normal morning. [00:54:22.0000] <tabatkins_> And by "normal", I of course mean "west-coast american". [00:54:47.0000] <jgraham> Pretty sure that nothing about the west coast of america is normal [00:54:51.0000] <charlvn> tabatkins_: how is west-coast american normal? [00:55:04.0000] <tabatkins_> charlvn: That should be self-evident. [00:55:05.0000] <charlvn> yeah exactly jgraham [00:55:17.0000] <charlvn> tabatkins_: i must be missing it ;) [00:55:34.0000] <tabatkins_> charlvn: Clearly you're not a citizen of Imperial America. [00:55:41.0000] <charlvn> tabatkins_: nope! lol [00:55:54.0000] <charlvn> The Backstroke of the West [00:56:07.0000] <othermaciej> tabatkins_: good euro-morning [01:22:08.0000] <zcorpan> arun++ for fixing fileapi :-) [01:44:44.0000] <Velmont> I'm back [01:45:52.0000] <Velmont> Oh, lost internet. Was going to write; I'm back on the mother-continent. So good euro-morning to all. [01:48:19.0000] <jgraham> Help, Velmont has been invaded by the Glazkov-mind-worms [01:49:26.0000] <zcorpan> woah. https://www.w3.org/Bugs/Public/show_bug.cgi?id=16945 [01:50:07.0000] <Velmont> jgraham: WAT? [01:57:38.0000] <jgraham> zcorpan: So, is he wrong or is the spec wrong? And are browsers following the spec? [01:57:49.0000] <zcorpan> jgraham: just commented in the bug [01:58:20.0000] <zcorpan> spec is wrong, but nobody follows the spec, it seems :-) [01:58:35.0000] <jgraham> Oh, I wonder if the tests are right, somehow? [01:59:06.0000] <jgraham> Otherwise how did everyone miss this? [01:59:06.0000] <zcorpan> is there a test for this case? [01:59:12.0000] <jgraham> I don't know [01:59:58.0000] <jgraham> If it was just that the spec was so obviously wrong that everyone did something different, I think someone at least would have reported the issue [02:00:35.0000] <zcorpan> yeah that's a bit weird [02:03:11.0000] <zcorpan> i don't see a test for that particular state transition [02:15:21.0000] <annevk> anyone know why Steam does not let you play games in offline mode? [02:15:40.0000] <annevk> kind of sucked on the flight back [03:10:09.0000] <Dashiva> annevk: There's supposed to be an offline mode for steam, but you have to enable it while online or something [03:11:27.0000] <zcorpan> LOL. Please go online to go offline. [03:27:32.0000] <Philip`> I assume the idea is that "go offline" means "lock my account so no other machine can log in as me while I'm offline, then go offline" [03:28:05.0000] <Philip`> to prevent many individuals sharing a Steam account and all playing at once [03:28:24.0000] <Philip`> (which is normally prevented by having the online service check you're only logged in once) [03:34:17.0000] <jgraham> AryehGregor: Thinking about it, W3C probably don't want you running tests directly on the DVCS server, so you could see that as a feature [03:35:15.0000] <jgraham> Hence w3c-test.org being an entirely seperate domain [03:55:02.0000] <Philip`> Maybe there should be a rev583bf580775b.w3c-test.org for every Hg revision [07:54:27.0000] <jgraham> OK, reading comprehension fail [07:54:44.0000] <jgraham> Why can't I get the script-inserted-external-script opt out to work? [07:54:51.0000] <jgraham> +raciness [07:56:41.0000] <jgraham> "In addition, whenever a script element whose "force-async" flag is set has a async content attribute added, the element's "force-async" flag must be unset." [07:57:49.0000] <jgraham> So if I create a script, set the async content attribute, then remove it again, I should be left with a script that has neither async nor force-async set, right? [08:11:52.0000] <jgraham> Oooh [08:12:11.0000] <jgraham> I see where I am wrong [08:12:15.0000] <jgraham> Nevermind me [12:43:33.0000] <jgraham> With the unload thing, is it a good assumption that XHR is the only use case? [12:43:58.0000] <jgraham> Won't peoplw want to do potentially-laggy IO to the local system on site exit? [13:43:36.0000] <Ms2ger> A copy of WordPerfect? What's that? [13:45:40.0000] <Hixie> well i usually say Microsoft Word, but i wanted to vary it a bit [13:45:49.0000] <Hixie> and i couldn't think of any other word processors :-P [13:59:34.0000] <jgraham> Wordstar! [14:00:18.0000] <Ms2ger> It's even in the spec! [14:00:52.0000] <Philip`> OpenOffice.org Writer? [14:01:00.0000] <Philip`> Maybe not the catchiest name [14:01:29.0000] <Ms2ger> LibreOffice? [14:07:53.0000] <jgraham> AbiWord? [14:10:15.0000] <jgraham> Oh look Wikipedia has a list http://en.wikipedia.org/wiki/List_of_word_processors [14:11:57.0000] <Ms2ger> Of course [14:16:42.0000] <Hixie> jgraham: i thought of lots of them after writing "WordPerfect" :-P [15:23:41.0000] <Hixie> anyone understand https://www.w3.org/Bugs/Public/show_bug.cgi?id=15645 ? [15:26:15.0000] <Hixie> lordy, https://www.w3.org/Bugs/Public/show_bug.cgi?id=15704 is going to be fun [15:29:25.0000] <tabatkins_> Hixie: Actually, "o.value = +a.value + +b.value;" is even shorter. ^_^ [15:31:03.0000] <Hixie> yeah but that acts weird if the value is e.g. '0xa' [15:31:28.0000] <tabatkins_> Okay, true. [15:34:20.0000] <tabatkins_> So, throwing an idea out before I post it to the list. [15:35:10.0000] <tabatkins_> We discussed in the SVGWG today about pulling SVG further into HTML. Specifically, allowing SVG elements to also work when in the HTML namespace. (Or, equivalently, adding all the SVG elements to HTML with the same meaning and rendering.) [15:35:27.0000] <tabatkins_> SVG is totally down with this. Will it work? [15:36:41.0000] <Hixie> what would it mean? [15:36:45.0000] <Hixie> e.g. what coordinate space would one use? [15:37:06.0000] <Hixie> what's the use case? [15:37:46.0000] <tabatkins_> Baby steps. First, imagine we're still in the current situation where you need an <svg> ancestor to establish your rendering mode. [15:38:01.0000] <tabatkins_> The benefit here is just that you can use createElement() rather than createElementNS. [15:38:09.0000] <Hixie> oh, just moving <Svg> to the html namespaces [15:38:11.0000] <Hixie> sorry, i misunderstood [15:38:12.0000] <tabatkins_> And that three of the four element collisions become meaningless. [15:38:42.0000] <tantek> so we can add xpath to the scrap pile? [15:39:06.0000] <tantek> XPath: because <a href> was too simple. [15:39:43.0000] <Hixie> tabatkins_: sounds like a painful transition, and it would require very close coordination between the groups, which is likely to be hard given that html hasn't really proven to be able to coordinate even within its own groups [15:39:52.0000] <Hixie> tabatkins_: but in principle it seems like what we should have done long ago [15:40:14.0000] <tabatkins_> Okay, so no killer problems besides, you know, actually writing things correctly? [15:40:17.0000] <jamesr> are there any element name collisions? [15:40:19.0000] <tantek> Tabatkins, perhaps you should propose SVG assimilation for the next HTML WG charter. [15:40:27.0000] <tabatkins_> jamesr: style, script, a, font. [15:40:34.0000] <jamesr> some of those may be in use [15:40:55.0000] <tabatkins_> jamesr: style, script, and a are *almost* identical in the two. Merging wouldn't be too painful. [15:41:03.0000] <tabatkins_> font is vastly different. Not sure how best to handle it. [15:41:07.0000] <Hixie> tabatkins_: well i wouldn't expect it to be a smooth transition, certainly [15:41:20.0000] <tabatkins_> Simple way is that createElement('font') just always makes an HTMLFontElement. [15:41:41.0000] <Hixie> tabatkins_: solving the problem of createElementNS() might be done more easily in other ways than this though [15:41:45.0000] <Hixie> tabatkins_: if that's the main problem [15:41:59.0000] <tabatkins_> Nah, it's just one (annoying) problem. [15:42:13.0000] <tabatkins_> Agree that createElementNS itself could just be fixed by letting it take "svg" as a string or something. [15:42:49.0000] <tabatkins_> But it means that <template> and DocumentFragment.innerHTML no longer need to worry about parsing SVG specially. [15:43:10.0000] <Hixie> why do they need to parse SVG specially anyway? [15:43:37.0000] <Hixie> oh right it's this whole parse mode stuff that's being discussed in webapps [15:43:45.0000] <tabatkins_> So that "frag.innerHTML = '<g><rect /></g>';" works the right way. [15:43:49.0000] <Hixie> how do you know what context node to use, etc [15:43:50.0000] <Hixie> well [15:43:54.0000] <Hixie> you'd still have that problem [15:44:06.0000] <tabatkins_> Why? [15:44:15.0000] <Hixie> because you'd still need to go into foreign mode [15:44:19.0000] <tabatkins_> Why? [15:44:23.0000] <Hixie> to handle the /? [15:44:29.0000] <Hixie> the /> i mean [15:44:30.0000] <Hixie> not /? [15:44:53.0000] <tabatkins_> Ah. That's not quite true. Instead just special-case the SVG elements to make /> meaningful. [15:45:01.0000] <Hixie> yikes [15:45:07.0000] <tabatkins_> Meh. [15:45:23.0000] <Hixie> i'm still not sure we've 100% dealt with the fallout of foreign mode [15:45:31.0000] <tabatkins_> The HTML parser contains worse. ^_^ [15:45:32.0000] <Hixie> i'm certainly not up for trying to rewrite all that [15:45:45.0000] <tabatkins_> I'm definitely up for trying. [15:45:53.0000] <Hixie> not without a really REALLY big payoff [15:46:18.0000] <Hixie> i guess my feeling here is that overall the cost is likely to be way higher than the benefit, but that might just be because i don't see what the benefits are [15:46:26.0000] <Hixie> but i see very high costs [15:46:30.0000] <tabatkins_> So, the payoff is that this makes it easier to *then* do seamless mixing without the <svg>/<foreignObject> barrier. [15:46:42.0000] <Hixie> i don't really follow why that's a win though [15:46:46.0000] <tabatkins_> The rendering issue is then pretty easy from a CSS angle. [15:46:47.0000] <Hixie> or what it would even really mean [15:47:03.0000] <Hixie> why would you want a <rect> in the middle of a <p>? [15:47:05.0000] <Hixie> or whatever? [15:47:32.0000] <tabatkins_> Why do you want an <svg><rect /></svg> in the middle of a <p>? [15:47:41.0000] <Hixie> same reason you'd want an <img> there [15:47:46.0000] <tabatkins_> Exactly. [15:47:46.0000] <Hixie> because you have an image [15:47:50.0000] <Hixie> that's not the same [15:47:55.0000] <Hixie> an <Svg> introduces a replaced element [15:48:04.0000] <Hixie> <rect> introduces... not sure what exactly [15:48:16.0000] <tabatkins_> As would a <rect>, once I define the rendering model. [15:48:23.0000] <othermaciej> tabatkins_: might be nice if it had always worked that way, but now it would create a chameleon namespace issue [15:48:26.0000] <Hixie> when would you want _just_ a rect? [15:48:29.0000] <Hixie> that seems very odd [15:48:34.0000] <othermaciej> sucks to have an element that can be in two different namespaces [15:48:50.0000] <othermaciej> see dbaron's well-known comment on the topic [15:48:52.0000] <tabatkins_> Hixie: If you, say, had some <pattern>s defined earlier in the doc, and are painting it with them? [15:48:53.0000] <tantek> like <a> ? [15:49:04.0000] <tabatkins_> othermaciej: I don't recall dbaron's comment. [15:49:18.0000] <othermaciej> <a> is annoying for a different reason (namespace collision if you ignore the namespace) [15:49:23.0000] <Hixie> mixing at that level seems like it would be a huge departure from svg's semantics while simultanesouly being kinda counter to the whole content/presentation split [15:49:33.0000] <Hixie> svg is media-specific, html is not [15:49:41.0000] <Hixie> it's fine to have well-defined places to integrate them [15:49:52.0000] <Hixie> but i don't really see what it means to have a <path> in the middle of a <label> [15:49:56.0000] <tabatkins_> SVG is exactly as media-specific as <img>. I don't understand why you believe <img> is okay but <rect> isn't. [15:50:16.0000] <Hixie> <img> or fine, so is <svg>, the point is that <rect> is like having five of the pixels of a PNG without the rest of it [15:50:20.0000] <othermaciej> tabatkins_: http://lists.w3.org/Archives/Public/www-archive/2009Feb/0065.html [15:50:42.0000] <tabatkins_> More important, though, is being able to mix HTML into SVG so you can get better handling of text and stuff, without <foreignObject> pain. [15:50:55.0000] <Hixie> why is <foreignObject> painful? [15:53:12.0000] <tabatkins_> Because it's an extra element that doesn't *do* anything. [15:53:29.0000] <Hixie> it introduces a viewport [15:53:33.0000] <Hixie> that's pretty important :-) [15:54:42.0000] <tabatkins_> Nah, that's already done by nesting <svg> if you need a new viewport, or by the root <svg> if you don't. [15:57:11.0000] <Hixie> it introduces a viewport with the dimensions and position where you want the nested HTML. We could say you could use <svg> instead, sure, but my point is that it does something, not nothing. [15:57:14.0000] <tabatkins_> Looking over dbaron's comment, it looks like his concerns are mostly only a problem if you (a) actually care about namespaces, and (b) don't fully integrate the DOMs of the two languages. [15:57:41.0000] <Hixie> it also does the important job of keeping the visual markup segragated from the media-independent markup, allowing us to have sane conformance criteria [15:58:26.0000] <tabatkins_> I'm still not understanding that objection. A <rect> is visual in the same way an <img> is, so the same conformance criteria apply. A <text> is media-independent in the same way a <p> is, so ditto. [15:58:56.0000] <Hixie> (we could certainly define what happens if we allowed mixing svg and html together, e.g. we could say shapes aren't in flow and so margins collapse between the <p>s in <p><circle/></p>, but that seems like it'd be hugely confusing for no benefit) [15:59:05.0000] <Hixie> <rect> isn't the same as <img> [15:59:14.0000] <Hixie> <rect is the same as the pixels in the PNG in the <img> that happens to form a rectangle [15:59:27.0000] <Hixie> > ^ [15:59:48.0000] <shepazu> Hixie: that's a pretty sexist conclusion [15:59:51.0000] <tabatkins_> Not if the <rect> is hosting a <pattern>, for example. [16:00:02.0000] <Hixie> it'd be like putting the 2D canvas on all HTML elements [16:00:14.0000] <tabatkins_> Are you under the impression that <rect> is just for drawing solid-color rectangles? [16:00:26.0000] <shepazu> where by "sexist", I mean "something I don't agree with" [16:00:32.0000] <Hixie> tabatkins_: i am familiar with <rect> [16:01:10.0000] <tabatkins_> Hixie: Then I don't understand why you keep implying that a <rect> (or generally, an SVG drawing element) can't be an "image" by itself. [16:01:36.0000] <shepazu> I wouldn't rather we not have to use <foreignObject> [16:01:52.0000] <shepazu> er… s/wouldn't/would/ [16:02:04.0000] <Hixie> tabatkins_: you can certainly form an image from a single <rect>, or a single <rect> with lots of supporting patterns and gradients and colour stops, or indeed a single path, or indeed many other things [16:02:13.0000] <Hixie> tabatkins_: just like you can form a bitmap from a collection of pixels [16:02:16.0000] <Hixie> tabatkins_: but so what? [16:02:38.0000] <Hixie> tabatkins_: just because you can have a <select> with a single <optgroup> doesn't mean we should allow <optgroup> anywhere in the DOM [16:03:11.0000] <shepazu> (you can also make an image using CSS) [16:03:27.0000] <Hixie> tabatkins_: there are many things about SVG that I would change if we were starting over, but getting rid of the clear boundaries of where the image starts and ends isn't one of them [16:03:28.0000] <tabatkins_> Interestingly, we *do* allow display:table-row anywhere in the document. [16:03:56.0000] <Hixie> if we were talking about putting SVG into CSS 'content' or some such, i would be totally fine with it [16:03:56.0000] <tabatkins_> Because it has a clear meaning, and you can infer the context. [16:04:01.0000] <Hixie> because CSS is not media-independent [16:04:10.0000] <Hixie> it's the presentation layer, like SVG [16:04:41.0000] <Hixie> it makes perfect sense in the CSS layer to talk about having an arrow or a pattern or a line here or there or whatever [16:04:45.0000] <tabatkins_> Dude, quit pulling out that argument. It's *wrong*. HTML has several presentational elements. [16:05:16.0000] <Hixie> i think while you consider that argument to be wrong we are unlikely to make progress in this discussion, since it is the core premise of much of my work :-) [16:05:20.0000] <tabatkins_> So trying to argue that adding variants of them is somehow anti-semantic is simply nonsensical. [16:06:00.0000] <tabatkins_> If it's ok to refer to SVG in an <img>, it's equally okay to embed SVG in an <svg>, and it's equally okay to embed a <rect> directly (from a semantic perspective). [16:06:17.0000] <Hixie> i've attempted to explain why i disagree [16:06:52.0000] <tabatkins_> Yes, but that particular reason is nonsense, as it's inconsistent with the current (reasonable) state of HTML. [16:07:53.0000] <Hixie> telling me that it's nonsense is not going to convince me [16:08:02.0000] <Hixie> but i respect that that is your opinion [16:08:11.0000] <tabatkins_> othermaciej: As far as I can tell, dbaron's comment hinges on you (a) caring about namespaces in general (rather than importing them for the purpose of *eliminating* them for practical purposes), and (b) not fully integrating the two languages. [16:08:45.0000] <othermaciej> tabatkins_: would support for svg elements in the svg namespace be removed? [16:08:50.0000] <tabatkins_> Hixie: You are asserting that <p>foo <rect/></p> is somehow wrong, but <p>foo <svg><rect /></svg></p> is okay. You can't justify that. [16:09:11.0000] <tabatkins_> othermaciej: That would be cool, but would greatly magnify the transition pain. [16:09:21.0000] <othermaciej> tabatkins_: if not, then I believe dbaron's criticisms apply [16:09:29.0000] <othermaciej> tabatkins_: if yes, it's probably a nonstarter from a compat perspective [16:09:46.0000] <tabatkins_> othermaciej: I agree they apply. I don't think they're as important as when dbaron raised them. [16:10:12.0000] <tabatkins_> (They may not even have actually been important when they were raised. I don't know what dbaron thought about namespaces in general in 2006.) [16:10:53.0000] <Hixie> tabatkins_: i am asserting that <p>foo</p> <line .../><line .../> <p>bar</p> should not be conforming because it is almost certainly going to occur only when authors are intentionally making their HTML files media-specific and that that is counter to the purpose of a semantic markup language. [16:11:18.0000] <othermaciej> I think all of dbaron's criticisms still fully apply to the svg-in-html-namespace proposal [16:11:25.0000] <othermaciej> (unless support for the svg namespace is dropped) [16:11:32.0000] <tabatkins_> Hixie: But that *precise* markup is okay if you wrap the <line>s in an <svg> element? [16:11:33.0000] <othermaciej> implementations will have to support the same element in both namespaces [16:11:59.0000] <othermaciej> svg even has the additional issue of namespace collisions, where svg:a != html:a and svg:font != html:font [16:12:19.0000] <Hixie> tabatkins_: i think it's fine to have a well-defined short list of ways to embed media-specific material, ideally in a manner that allows for fallback in other media, as that makes it clear where the media-specific content's boundaries lie. [16:12:41.0000] <Hixie> tabatkins_: (i'm not 100% happy that <svg> in HTML doesn't have well-defined fallback) [16:12:50.0000] <tabatkins_> SVG doesn't allow for fallback. (Ninja'd!) [16:14:02.0000] <tabatkins_> othermaciej: Well, dbaron didn't object along any of those lines when we discussed it over dinner. ^_^ [16:14:15.0000] <tabatkins_> He may have forgotten his previous position, of course. [16:14:19.0000] <othermaciej> could be [16:14:47.0000] <Hixie> i think there is a huge qualitative difference between <svg><line .../><line .../></svg> and <div><line .../><line .../></div>. The former clearly indicates to authors where the boundaries of their image lies. The latter makes it way harder to distinguish. [16:15:06.0000] <Hixie> i just don't see the problem with requiring <svg>, frankly [16:16:20.0000] <tabatkins_> Hixie: It's a meaningless wrapper element, morally equivalent to a wrapper <div>. ^_^ [16:17:13.0000] <Hixie> well, no, it introduces the coordinate space, defines a viewport, and various other things, but even if it were technically identical, it isn't _morally_ identical, and that's my whole argument [16:17:35.0000] <othermaciej> implementing svg without the <svg> wrapper would be slightly annoying, because of the issues pixie mentions [16:18:05.0000] <othermaciej> <svg> defines a coordinate space and a viewbox, without which other svg elements are meaningless [16:18:07.0000] <tabatkins_> That's just a CSS issue, actually. Not difficult. [16:18:20.0000] <othermaciej> you'd essentially have to add an "anonymous box" for the svg element [16:18:22.0000] <tabatkins_> You can default to a coordinate space of the containing block. [16:18:31.0000] <othermaciej> like anonymous table rows or the like [16:20:00.0000] <Hixie> i just don't see the advantage [16:20:17.0000] <Hixie> you'd almost always want the wrapper anyway to give the coordiante space and viewbox and various other things [16:21:27.0000] <othermaciej> anything is technically possible with enough effort [16:21:48.0000] <othermaciej> but I wonder also what the advantage is, beyond saving a few characters for svg start and end tags [16:24:16.0000] <tabatkins_> Avoiding namespace annoyances for the common case, and mixing HTML into SVG without the decently larger cost of <foreignObject> start and end tags. [16:24:38.0000] <Hixie> just rename <foreignObject> to <html> [16:24:43.0000] <tabatkins_> (Particularly since, in proper SVG, you need to provide the html and body tags.) [16:24:58.0000] <Hixie> and define it as taking "flow content" [16:25:27.0000] <Hixie> (or rename it <body>) [16:25:32.0000] <Hixie> (or <doc>) [16:30:18.0000] <tabatkins_> Hm. That doesn't fix the annoyance of the *NS functions, or simplify <template> parsing, though. [16:34:24.0000] <Hixie> we should fix the *NS thing using the new API we figured out but that never went anywhere [16:34:35.0000] <Hixie> <template> parsing is a separate issue that we need to resolve in general [16:36:33.0000] <tabatkins_> <template> parsing is either "manually provide a context element" or "infer a context element, almost certainly in the way that Raf is suggesting". [16:38:49.0000] <roc> eliding <foreignObject> would be much easier than eliding <svg> [16:40:08.0000] <tabatkins_> roc: I think it requires adding x/y/width/height presentational attributes to every HTML element. Kosher? [16:41:06.0000] <Hixie> and every MathML element [16:41:11.0000] <roc> not necessarily. Use their intrinsic widths and heights only, and use transforms to position them [16:41:15.0000] <tabatkins_> (Well, technically we could do it within @style, once we properly merge all the SVG attributes into CSS properties.) [16:41:25.0000] <Hixie> (and every unknown XML element if anyone wants to embed non-HTML non-MathML CSS-styled XML) [16:41:42.0000] <tabatkins_> Hixie: I'm not sure if it's reasonable to have MathML outside of a <math>. [16:41:43.0000] <roc> you could say it only works for HTML elements. [16:42:06.0000] <tabatkins_> Given that MathML's rendering model is much different than the existing CSS ones. [16:42:16.0000] <tabatkins_> (While SVG's layout model is just a slight variant on abspos.) [16:43:38.0000] <Hixie> tabatkins_: i think solving the "<foreignObject> problem" is as easy as just renaming it <doc> or introducing a new element <doc> that works like <fO> [16:44:18.0000] <Hixie> no reason to define new CSS stuff [16:44:36.0000] <Hixie> frankly though, who is hand-authoring SVG-embedding-HTML other than me? [16:44:45.0000] <Hixie> i.e. is it really that big a problem? [16:45:15.0000] <tabatkins_> Well, I certainly am. [16:45:29.0000] <tabatkins_> And for a lot of things, it's definitely easy to hand-author. [16:53:50.0000] <Hixie> if it's just you and me, i think we can punt the issue :-) [16:54:01.0000] <Hixie> if it's not, it may be worth creating a shorter name for that element [16:55:59.0000] <tabatkins_> It's not just us. ^_^ SVG is gradually actually being used by webdevs, as support finally reaches useful levels and people realize what it can do. [16:56:41.0000] <tabatkins_> CSS is also gradually integrating more into SVG, with properties that let you "plug in" SVG elements for advanced functionality (generally not the sort of thing you'll probably make in an SVG image editor). [16:59:08.0000] <tabatkins_> And a substantial fraction (maybe even a majority? I'd have to count) of the new features accepted for SVG2 are about increasing human-authorability. [16:59:48.0000] <tabatkins_> Adding features that are easy to handle if you're a computer and don't mind repeating yourself, but are difficult to write (and impossible to read) if you're a human. 2012-05-08 [17:03:03.0000] <jamesr> Hixie, <foreignObject> is kind of too broken to hand author stuff today [17:03:28.0000] <jamesr> tons of crap is broken with it, so not much content exists using it, so UAs aren't motivated to fix i [17:03:28.0000] <jamesr> t [17:05:26.0000] <Hixie> what's broken? [17:06:27.0000] <tabatkins_> I'm going to run with roc's statement that it seems fine to just elide <foreignObject> for HTML elements. [17:08:13.0000] <roc> most of the things that are broken in <foreignObject> are just UA bugs [17:08:33.0000] <Hixie> well they're certainly not going to go away if we make it more subtle :-) [17:13:40.0000] <Hixie> oh jeez, a sudden influx of i18n bugs [18:04:28.0000] <jamesr> roc, agree re: they are UA bugs. the rest of my statement still applies in that case [18:54:36.0000] <roc> I think Gecko's foreignObject support is pretty good [18:55:05.0000] <roc> our main limitation is that we don't have accelerated compositing for the contents of <foreignObject> yet but we're working toward it [20:02:22.0000] <annevk> 'I am sorry that an overt adherence to "process" blinds some people to positive gains made when people come together and talk to each other, rather than assuming nothing but malevolent actions' coming from John Foliot, that's an instant classic [20:21:55.0000] <annevk> yay, seems my email address transition is more or less working [20:35:00.0000] <annevk> apart from public-html-testsuite that is [20:44:10.0000] <annevk> why is there no in:spam has:label in Gmail? [20:44:30.0000] <annevk> or label:* [20:56:05.0000] <Hixie> 208 [20:56:07.0000] <Hixie> er [21:01:07.0000] <annevk> Already Reported? [21:04:46.0000] <Hixie> line 208 [21:24:11.0000] <annevk> merging SVG with HTML keeps coming up, but each time it's a less just try it and see what breaks rather than thought through [21:24:59.0000] <annevk> merging SVG and CSS is more interesting I think; similarly to how SMIL got in CSS [21:30:09.0000] <Hixie> agreed [21:34:56.0000] <Hixie> jesus the more i have to deal with these frameworks the more i hate them [21:35:09.0000] <Hixie> (io.socket.js being the main culprit here today) [21:35:20.0000] <Hixie> or socket.io or whatever it's called [21:43:18.0000] <annevk> from Asa Dotzler: "not going to block k9o on un-prefixing but if there are full-screen capabilities missing that we need to have a successful app ecosystem, we'll block on those." [21:43:38.0000] <annevk> I guess a proprietary app ecosystem is now fine? [21:46:17.0000] <annevk> Hixie: btw, do you have a POP3 account for ian⊙hc that Gmail imports? [21:46:24.0000] <Hixie> no [21:46:29.0000] <Hixie> i just forward all e-mail directly [21:46:33.0000] <Hixie> ian⊙hc is just an alias [21:46:35.0000] <rniwa> annevk: moving to new email address? [21:46:49.0000] <annevk> Hixie: how did you setup your native client to be able to email from ian⊙hc? [21:47:00.0000] <annevk> rniwa: yeah, annevk⊙an [21:47:15.0000] <Hixie> annevk: why would that be a problem? [21:47:38.0000] <annevk> Hixie: e.g. when trying to do that on my iPhone it asked for a password [21:48:00.0000] <Hixie> password for what? [21:48:01.0000] <annevk> but maybe I'm looking at the wrong thing [21:48:09.0000] <rniwa> annevk: SMTP or POP3 i suppose? [21:48:32.0000] <rniwa> annevk: which SMTP/POP3/IMAP server are you using to send those emails? [21:48:39.0000] <rniwa> annevk: of gmail? or of annevk.nl? [21:48:49.0000] <annevk> rniwa: using Gmail at the moment to email from annevk⊙an [21:49:50.0000] <Hixie> oh i don't e-mail from gmail [21:50:20.0000] <annevk> ah, I guess I need to setup some kind of SMTP account? [21:50:50.0000] <Hixie> honestly if you just want to use gmail your best bet might just be to get a hosted gmail account [21:50:58.0000] <Hixie> on annevk.nl [21:51:15.0000] <Hixie> so you can just point your mx records at gmail and so on [21:51:59.0000] <annevk> I'd rather not have another Google account [21:54:41.0000] <annevk> Hixie: anyway, if hixie.ch is just an alias, how do you send email from there? [21:55:06.0000] <Hixie> pine doesn't care where you say you send mail from, it just happily puts the from address on there [21:55:34.0000] <Hixie> and the dreamhost smtp servers are fine with it because they know it's one of my dreamhost addresses, i presume [21:55:35.0000] <rniwa> annevk: which mail client do you use? [21:55:59.0000] <Hixie> i occasionally e-mail from gmail, and gmail seems fine with it too, didn't have to do anything special there [21:56:00.0000] <rniwa> annevk: i think you normally setup a separate account with gmail's credential for username/password [21:56:22.0000] <annevk> Hixie: oh, I had to authenticate it there [21:56:50.0000] <annevk> Hixie: are you sure you're not importing a DreamHost POP3 account into Gmail? [21:57:19.0000] <Hixie> there is no dreamhost pop3 account to forward [21:57:20.0000] <Hixie> so yes [21:57:25.0000] <annevk> rniwa: I haven't really decided yet which email client to use; I'm using Opera at the moment; not sure how good it is at handling multiple accounts [21:57:34.0000] <Hixie> ian⊙hc is just a forwarding alias [21:57:38.0000] <Hixie> there's no inbox behind it [21:57:50.0000] <rniwa> annevk: http://support.google.com/mail/bin/answer.py?hl=en&answer=22370 [21:57:50.0000] <Hixie> same as *@spam.hixie.ch [21:58:03.0000] <rniwa> /me finally found a manual on this :D [21:58:25.0000] <rniwa> /me hates page redesigns that moves URLs around :( [21:59:54.0000] <rniwa> annevk: honestly, the best mail client for gmail is gmail itself. i've tried several mail clients that try to support gmail features [21:59:59.0000] <rniwa> annevk: but none of them support labels well. [22:00:12.0000] <rniwa> annevk: it's fine as long as you have only one label at a time [22:00:35.0000] <rniwa> but as soon as you start adding 2-3 labels, it gets all confused :( [22:00:57.0000] <annevk> rniwa: ah k [22:02:21.0000] <rniwa> annevk: i highly recommend "preview pane" in labs [22:02:31.0000] <annevk> maybe Gmail is efficient enough [22:02:53.0000] <rniwa> annevk: yeah, i think the only reason you want to use mail clients to see emails offline [22:03:35.0000] <rniwa> i suppose you would do that much more often than i do as a spec. author though... [22:04:25.0000] <annevk> rniwa: does preview pane require a minimum width? [22:05:58.0000] <rniwa> annevk: you can do vertical split as well. [22:06:01.0000] <annevk> Hixie: I guess the problem is that I don't have a DreamHost SMTP setup as far as I know [22:06:12.0000] <rniwa> annevk: (it defaults to horizontal split) [22:06:13.0000] <Hixie> mail.annevk.nl [22:06:19.0000] <Hixie> is your smtp server [22:06:21.0000] <Hixie> i believe [22:06:37.0000] <Hixie> oh i guess if you don't have an account at dreamhost for your mail you wouldn't have an account on it though [22:07:04.0000] <annevk> right :) [22:07:20.0000] <Hixie> can you create a bogus account? [22:07:20.0000] <rniwa> annevk: setup a local SMTP server :D [22:07:40.0000] <annevk> rniwa: I meant that it doesn't split now when enabled [22:07:47.0000] <rniwa> oh :) [22:07:49.0000] <annevk> Hixie: yeah I could do that [22:07:50.0000] <rniwa> annevk: oh that's weird :\ [22:08:11.0000] <annevk> or find a better email client that allows you to configure the From address [22:08:22.0000] <rniwa> annevk: maybe it doesn't work on Opera :\ [22:08:23.0000] <rniwa> ? [22:08:26.0000] <annevk> because per the page from rniwa Gmail is fine with it [22:08:29.0000] <annevk> rniwa: ah could be [22:09:35.0000] <Hixie> on his iphone? :-P [22:09:36.0000] <Hixie> as if apple would allow that :-P [22:09:37.0000] <rniwa> annevk: fwiw, i used to run smtp server locally to send my emails. [22:10:00.0000] <rniwa> not sure if that's practical anymore though given all the sender id nonsense [22:10:46.0000] <annevk> rniwa: same problem in Safari [22:12:11.0000] <rniwa> annevk: odd [22:12:25.0000] <rniwa> annevk: did you select "horizontal split"? [22:12:40.0000] <rniwa> annevk: next to < > buttons on the upper right corner [22:12:48.0000] <rniwa> there is a button with 3 lines [22:13:12.0000] <rniwa> if you click the downward arrow mark there, you can select "no split", "horizontal split", and "vertical split" [22:13:53.0000] <rniwa> annevk: maybe it's "no split" by default. although that'll be extremely silly [22:14:19.0000] <annevk> oooh [22:14:31.0000] <annevk> rniwa: that is the default [22:14:36.0000] <annevk> thanks [22:14:36.0000] <rniwa> :( [22:15:00.0000] <annevk> works fine in Opera btw [22:17:39.0000] <rniwa> annevk: great. [22:17:45.0000] <rniwa> oh my... my chrome is on rampage :( [22:17:49.0000] <rniwa> GPU accleration [22:17:53.0000] <rniwa> :\ [22:21:16.0000] <AryehGregor> Wait a sec. How can a Google+ e-mail in Gmail allow me to add someone to a circle directly in the e-mail? Do e-mails from Google+ get special privileges in Gmail? [22:21:18.0000] <AryehGregor> /me smells antitrust. [22:24:11.0000] <annevk> AryehGregor: you can do the same from Google search results, no? [22:24:34.0000] <AryehGregor> I don't know, maybe. But at least that's dedicated UI. This was part of the e-mail. [22:24:50.0000] <AryehGregor> I wouldn't have remarked on it if it were an extra button at the side of the e-mail or something. [22:32:25.0000] <roc> I think emails from anyone that Google+ knows about get special Gmail UI [23:03:21.0000] <annevk> so http://support.google.com/mobile/bin/answer.py?hl=en&answer=138740 suggests I go to m.google.com/sync on iOS to configure the custom From header, but that page says my device is not supported [23:05:21.0000] <annevk> oh [23:05:53.0000] <annevk> per help forum it's because the UI language is not English, which I never even set to Dutch to begin with, but I guess that's some IP address sniffing [23:05:56.0000] <annevk> boohoo [23:10:19.0000] <rniwa> /me hates lang. setting per IP address sniffing [23:10:41.0000] <rniwa> every time I go to a foreign country, i need to figure out a way to get back to english website :( [23:10:55.0000] <rniwa> i don't know who decided it's a good idea. [23:12:15.0000] <annevk> anyway, I can now email from my iPhone as annevk⊙an and I don't have a custom/DreamHost SMTP setup yet [23:12:34.0000] <rniwa> nice. [23:43:06.0000] <Ms2ger> jgraham, yt? [00:05:33.0000] <annevk> MikeSmith: so public-html-testsuite is still sending emails to my old address and I think that's the last list [00:06:00.0000] <annevk> MikeSmith: I had unsubscribed for a few days now though so something seems broken [00:08:47.0000] <[tm]> annevk: away from my PC but will check it when i get back [00:09:47.0000] <[tm]> i think the reason is that nobody had sent a message to the list, so the subscriber info had not been updated [00:10:31.0000] <annevk> I thought so too, but then last night now emails arrived in my annevk⊙oc inbox [00:12:35.0000] <annevk> and I have tried unsubscribing about four times it seems per my sent history [00:17:00.0000] <[tm]> ok [00:17:21.0000] <[tm]> I'll try to get it figured out tonight [00:31:34.0000] <Ms2ger> annevk, as usual, please don't interpret "Asa Dotzler" as "Mozilla" ;) [00:54:41.0000] <zcorpan> TabAtkins: woah! let's not change the parser around foreign content to be a radically different design... [01:00:45.0000] <jgraham> OK, is this a conspiracy? Just when I want to talk to Ms2ger and AryehGregor, they both leave :( [01:05:23.0000] <zcorpan> also, let's fix <svg>.innerHTML = '<g/>' *first* [02:37:15.0000] <annevk> while trying to find the Hypertext CG (not a Community) mailing list I found http://lists.w3.org/Archives/Public/public-backplane-comments/ [02:37:18.0000] <annevk> takes you back [02:52:53.0000] <zcorpan> [tm]: you know what would be nice? memorable shortlinks to all the specs in http://platform.html5.org/ [02:54:35.0000] <kennyluck> That sounds cool indeed. [02:56:08.0000] <zcorpan> [tm]: btw, i don't see a link to selectors4 [02:56:46.0000] <annevk> you mean stuff like platform.html5.org/html and such? [02:57:39.0000] <zcorpan> annevk: yeah, but ideally shorter :) [02:58:03.0000] <kennyluck> we should have reserved tr.org. http://tr.org/html5 and such [02:59:41.0000] <annevk> kennyluck: w3.org is pretty short [03:00:01.0000] <annevk> kennyluck: you could e.g. have w3.org/r/xxx or r.w3.org/xxx [03:00:09.0000] <jgraham> sp.ec/html ? [03:00:23.0000] <zcorpan> jgraham: wfm [03:00:49.0000] <annevk> .ec requires 3 characters I think [03:01:11.0000] <jgraham> Yeah [03:01:20.0000] <annevk> it's also expensive [03:03:17.0000] <jgraham> Indeed. websp.ec probably isn't worth it then :( [03:05:30.0000] <zcorpan> can we use just html5.org/html etc? [03:07:46.0000] <annevk> zcorpan: seems like a pain to maintain [03:09:12.0000] <zcorpan> annevk: why? [03:10:31.0000] <annevk> zcorpan: because it's at the root of the site so I would be the bottleneck [03:12:43.0000] <zcorpan> hmm. would it be possible to have a script somewhere that [tm] can fiddle with that creates redirect rules for the root, or some such? [03:13:54.0000] <zcorpan> oh selectors is under css [03:14:18.0000] <annevk> not without giving mike access to all of html5.org, which is fine, but currently it's connected to most of my other domains [03:26:16.0000] <annevk> so I'd prefer a new domain or a subdomain [03:26:37.0000] <annevk> but I'm not sure it's really worth it [03:32:11.0000] <Philip`> I thought nobody was meant to call it HTML5 nowadays anyway [03:32:29.0000] <Philip`> so html5.org is an archaic domain name [03:34:20.0000] <annevk> Philip`: nostalgic [03:34:45.0000] <annevk> Philip`: also, html.org was way expensive [03:41:24.0000] <zcorpan> html60.org? [03:47:11.0000] <annevk> ah yeah, html.org owners rejected a USD 12.000 offer and will consider everything above 20.000 [03:48:43.0000] <annevk> oh hmm [03:48:49.0000] <annevk> html.org is now in the hands of w3.org [03:49:06.0000] <annevk> so maybe they put some of that Member money into domain buying fun? [03:49:25.0000] <smaug____> that would be odd [03:49:44.0000] <annevk> also appears to be using GoDaddy which is somewhat disappointing [03:50:02.0000] <annevk> smaug____: http://whois.domaintools.com/html.org shows it's owned by the W3C now [03:50:32.0000] <Ms2ger> jgraham, so [03:50:40.0000] <Ms2ger> There's two things I'd like to do [03:50:40.0000] <annevk> and I have an emailed exchange with the previous owner about the price [03:51:05.0000] <Ms2ger> One is { timeout: Infinity } (hmm, maybe that even works right now) [03:51:40.0000] <Ms2ger> The other is telling a test "I've given up on you, if you get any more results, don't tell me about them, because I don't care" [03:52:23.0000] <Ms2ger> Because I'm afraid of test results ending up reported for the next test [03:52:38.0000] <Ms2ger> (ISTR that happening with mochitests) [03:53:33.0000] <zcorpan> annevk: you should have bought it and sold it to the w3c :-P [03:54:23.0000] <zcorpan> maybe w3c want to buy html5.org as well? [03:54:43.0000] <zcorpan> for ONE MILLION DOLLARS [04:04:37.0000] <zcorpan> "first version of Web Sockets" http://lists.w3.org/Archives/Public/public-webapps/2012AprJun/0557.html [04:07:04.0000] <annevk> quick, what's our position? [04:07:12.0000] <annevk> /me throws up a coin [04:08:06.0000] <jgraham> Ms2ger: I think my solution does that [04:08:17.0000] <jgraham> Depending on what you mean by "test" [04:09:00.0000] <jgraham> So in your testharnessreport.js you would set explicit_timeout to true [04:09:10.0000] <jgraham> That would be line timeout:infinity [04:09:13.0000] <jgraham> *like [04:09:27.0000] <jgraham> Then when you want to give up on a test you would call timeout() [04:10:02.0000] <jgraham> That would set the status of any tests without results to "Not Run", and cause the page to try and report its results in the normal way [04:10:28.0000] <jgraham> and prevent anything further happening [04:10:57.0000] <kennyluck> TabAtkins, yt? [04:12:08.0000] <zcorpan> what does <script defer> do if the script is created after window.onload? [04:14:03.0000] <jgraham> zcorpan: From memory, that won't ever get run (per spec) [04:14:18.0000] <jgraham> But my memory is not that good [04:14:25.0000] <jgraham> /me adds it to a list of thinsg to test [04:16:24.0000] <zcorpan> is that what ie does? [04:18:36.0000] <jgraham> Dunno, I would need to change computers to find out [04:20:08.0000] <zcorpan> me too [04:20:36.0000] <zcorpan> but i'm curious so maybe i'll check :) [04:21:45.0000] <jgraham> http://software.hixie.ch/utilities/js/live-dom-viewer/saved/1513 shows the script running in Gecko and WebKit [04:22:45.0000] <jgraham> Oh, but that test is wrong [04:22:55.0000] <jgraham> Needs to be parser-inserted [04:43:39.0000] <tabatkins_> kennyluck: Here. [04:45:43.0000] <kennyluck> TabAtkins, I am wondering if you can list http://lists.w3.org/Archives/Public/www-style/2012Apr/0152 in the list of CSS3 V&U issue list. I have a patch for Gecko pending working group resolution on this → https://bugzilla.mozilla.org/show_bug.cgi?id=751805 [04:46:16.0000] <zcorpan> jgraham: http://software.hixie.ch/utilities/js/live-dom-viewer/?saved=1514 ie9 runs the script [04:46:20.0000] <kennyluck> I noticed that in your Syntax draft, url() can't be escaped, but I guess David wants a working group resolution. [04:47:03.0000] <jgraham> zcorpan: Yeah, because defer only has an effect on parser-inserted scripts [04:47:26.0000] <zcorpan> oh [04:47:32.0000] <jgraham> So the question is whether it is possible to create a parser-inserted script after load fires [04:48:57.0000] <jgraham> I thought http://software.hixie.ch/utilities/js/live-dom-viewer/saved/1515 might be close, but I haven't checked what document.close actually does [04:56:29.0000] <Ms2ger> jgraham, a timeout() function? Does that exist in code or only in your mind? :) [05:04:55.0000] <odinho> In HTML5 spec, is several dt's after another supposed to OR-ed or AND-ed? [05:06:37.0000] <jgraham> Ms2ger: Only in my mind [05:06:41.0000] <annevk> it's a relationship between <dt> and <dd>, so OR [05:06:53.0000] <annevk> but maybe we should document that somehow [05:06:56.0000] <odinho> annevk: OK, like I thought. [05:06:58.0000] <jgraham> Ms2ger: But it would simply be exposing the internal Tests.prototype.timeout [05:07:00.0000] <annevk> other glenn was confused by this too [05:07:02.0000] <odinho> annevk: Yes, I tried finding it in http://dev.w3.org/html5/spec/introduction.html#structure-of-this-specification [05:08:17.0000] <annevk> at some point I'm going to write "Boilerplate" [05:08:34.0000] <annevk> with Boilerplate Conformance, Terminology, etc. [05:08:46.0000] <annevk> and then other specs just say "See boilerplate" [05:08:52.0000] <annevk> plus changes [05:09:33.0000] <odinho> annevk: Would be A+. Your html5.org in itself (just linking a few links to very relevant stuff) is really useful . [05:12:37.0000] <annevk> it's getting kind of annoying to copy that stuff all over and then we find something worthy of clarifying and boom, 10 specs need to be updated and kept in sync [05:13:35.0000] <zcorpan> annevk: or make Anolis insert the boilerplate [05:14:10.0000] <MikeSmith> zcorpan: I'm happy to set up shortlinks wherever -- under platform.html5.org if that's where people would like them [05:15:50.0000] <annevk> zcorpan: a) that doesn't make it clear it's boilerplate b) that doesn't make additions to the boilerplate clear [05:16:38.0000] <zcorpan> annevk: fair enough [05:17:34.0000] <jgraham> hsivonen: AFAICT http://software.hixie.ch/utilities/js/live-dom-viewer/saved/1515 is a bug in gecko [05:17:53.0000] <Niloy> is there any plans to standardise webkit's css reflection? [05:18:21.0000] <annevk> Niloy: best to look through www-style [05:18:30.0000] <Niloy> okay thanks [05:19:16.0000] <zcorpan> MikeSmith: cool. only problem is finding a good way to make the urls short enough to be worthwhile :) [05:19:31.0000] <Ms2ger> Niloy, yes-ish [05:19:38.0000] <Ms2ger> (That's all I know about it, though) [05:23:05.0000] <Ms2ger> jgraham, I'd much appreciate the code, too ;) [05:23:23.0000] <MikeSmith> zcorpan: could do what annevk said earlier, about making them html.org/foo urls instead; but I currently don't have perms to do that -- would require some changes from annevk first I guess [05:24:20.0000] <Ms2ger> MikeSmith, you mean html5.org? [05:25:25.0000] <MikeSmith> yeah [05:27:32.0000] <odinho> If it's apache, just having a .htaccess with the redirects there would be enough I guess. Someone at the F2F used that for a "personal shortening service", don't remember who. :P [05:27:39.0000] <jgraham> Oh, W3C owns html.org? Who knew? [05:27:56.0000] <MikeSmith> is node. isEqualNode(foo) supported across browsers? [05:28:14.0000] <annevk> jgraham: I did since earlier today :p [05:28:15.0000] <MikeSmith> jgraham: I didn't know. Are you sure it does? [05:28:28.0000] <jgraham> MikeSmith: Well whois says that [05:28:33.0000] <MikeSmith> wow [05:28:50.0000] <annevk> I asked Ted about it on www-archive [05:28:51.0000] <annevk> fwiw [05:29:26.0000] <annevk> it's a) registered by GoDaddy which is bad and b) I'm kind of curious if W3C really paid that much [05:29:48.0000] <MikeSmith> I doubt W3C would pay a bunch of money for it [05:30:07.0000] <annevk> MikeSmith: I think isEqualNode() is supported yes (not sure about all edge cases) [05:30:16.0000] <MikeSmith> OK [05:30:24.0000] <MikeSmith> /me didn't know [05:30:38.0000] <annevk> MikeSmith: yeah, so I tried getting this domain January 2011; guy asked USD 20.000 [05:30:46.0000] <MikeSmith> wow [05:30:58.0000] <MikeSmith> that's nuts [05:31:14.0000] <annevk> he said he had an offer for 12.000 but wanted more [05:31:30.0000] <annevk> with a typo lol "We will consider all serious propositions abone $20,000" [05:31:33.0000] <MikeSmith> maybe Ted ended up claiming it through trademark or something [05:31:38.0000] <MikeSmith> heh [05:31:48.0000] <TabAtkins> kennyluck: Sorry, mirc wasn't binging me for mentions. [05:32:20.0000] <annevk> MikeSmith: ah yeah [05:32:35.0000] <kennyluck> TabAtkins, nah, no problem. [05:32:42.0000] <TabAtkins> I intend to resolve that so that it's completely impossible. [05:33:09.0000] <TabAtkins> In other words, so that url() doesn't allow it (because it's all parsed as a URL), and \url() is an unknown function. [05:34:21.0000] <TabAtkins> image() will allow attr(), because it takes a <string> type value (not STRING tokens, like url()). [05:34:58.0000] <kennyluck> TabAtkins, I am in favor of that for what it's worth. The whole question is whether \url() can be considered a <url>. [05:35:17.0000] <annevk> is \u the same as u? [05:35:26.0000] <annevk> because then yes [05:35:54.0000] <kennyluck> I think WebKit as well as the formal grammar in CSS2.1 disallow this. Gecko allows this, so is a test case in the CSS2.1 test suite. How weird... [05:36:41.0000] <annevk> kennyluck: sounds like a bug in the grammar [05:36:58.0000] <kennyluck> This has been raised multiple times throughout the history of www-style, but I guess there was never a clear resolution. [05:37:09.0000] <annevk> kennyluck: if the grammar is not {u}{r}{l}'(' ... ')' or something close to that [05:38:16.0000] <annevk> allowing @document-\url ... but not background:\url() seems really weird [05:38:36.0000] <annevk> now removing escapes from what should be literals... [05:38:46.0000] <annevk> I would support that, but it seems a little late [05:39:09.0000] <annevk> especially considering the current dynamics of the CSS WG [05:39:10.0000] <kennyluck> annevk, David is in favor of what you think I guess. I think TabAtkins is against. I am probably against. [05:40:09.0000] <kennyluck> In any case, I need to fix the grammar or the test suite. [05:40:13.0000] <kennyluck> s/I/We/ [05:48:37.0000] <annevk> kennyluck: dbaron? pointer? [05:49:05.0000] <annevk> kennyluck: and why would anyone be against removing complexity? compat concerns? [05:49:16.0000] <kennyluck> annevk, https://bugzilla.mozilla.org/show_bug.cgi?id=751805#c3 [05:49:57.0000] <annevk> odinho: fwiw, the question in that Origin header bug is about the Origin header not being required by HTML [05:50:03.0000] <annevk> odinho: because HTTP is not required [05:50:29.0000] <annevk> odinho: so you could opt to not include the Origin header and it's not entirely clear to the submitter of the bug what the requirements with respect to CORS are then [05:50:38.0000] <TabAtkins> kennyluck: No, \url() shouldn't be a <url>. It's an unknown function. You only get a <url> from URL tokens, or from the image() function. [05:51:18.0000] <annevk> TabAtkins: why url() not work like any other syntax construct in CSS? [05:51:25.0000] <annevk> would /\ [05:51:57.0000] <annevk> either you handle escapes everywhere or you don't [05:52:06.0000] <TabAtkins> Because how does "\url(foo bar ! /* baz */)" work? It definitely *doesn't* act like "url(foo bar ! /* baz */)". [05:52:14.0000] <zcorpan> TabAtkins: http://software.hixie.ch/utilities/js/live-dom-viewer/saved/1516 opera and gecko apply it [05:52:46.0000] <odinho> annevk: If it's never sending CORS requests, then they will always fail I guess. So in No CORS-mode you'll always get taint, in the CORS modes you'll always fail. -- I find the question a bit strange. [05:52:48.0000] <kennyluck> so does IE, but IE is in general just broken concerning CSS escape sequences... [05:53:12.0000] <odinho> s/CORS requests/an Origin header/ [05:53:27.0000] <annevk> odinho: HTML CORS requests are not dependent on supporting CORS or something [05:53:34.0000] <annevk> odinho: the dependency chain is somewhat weird [05:53:41.0000] <annevk> it'll be fixed once we have fetch v2 [05:53:44.0000] <annevk> but we don't [05:55:49.0000] <zcorpan> TabAtkins: i don't follow what the problem is with "\url(foo bar ! /* baz */)" [05:56:17.0000] <TabAtkins> zcorpan: If you're not invoking the special URL token parsing, it looks like FUNCTION IDENT IDENT DELIM COMMENT CLOSE-PAREN. [05:56:48.0000] <TabAtkins> And you can lose data there unless you're preserving comments and exact whitespace. [05:57:05.0000] <TabAtkins> Even worse is what happens if you don't close the paren. [05:57:16.0000] <TabAtkins> close the comment, i mean. [05:57:37.0000] <TabAtkins> Then you get different results no matter what, because the URL token doesn't recognize a comment inside of itself. [05:57:42.0000] <TabAtkins> But a FUNCTION token would. [05:58:24.0000] <zcorpan> but url(foo bar ! /* baz */) isn't valid anyway, is it? [05:59:44.0000] <TabAtkins> No, I believe that's valid. [05:59:56.0000] <TabAtkins> and corresponds to the url "foo bar ! /* baz */". [06:00:02.0000] <TabAtkins> Which isn't a valid url, obviously, but still. [06:00:30.0000] <kennyluck> Well, space isn't valid in unquoted url() for what it's worth. [06:00:34.0000] <TabAtkins> url(foo bar ! /* baz) is invalid, but it's well-defined and guaranteed to end at the ) character. [06:00:36.0000] <zcorpan> browsers don't seem to think it's valid [06:00:37.0000] <TabAtkins> Oh, right, sorry. [06:00:40.0000] <TabAtkins> That's true. [06:00:45.0000] <kennyluck> Seriously, I don't buy any argument here. My patch which changes Gecko to match WebKit has only 10 lines. We just need to agree on something. [06:00:56.0000] <TabAtkins> So url(foo/*baz*/). [06:01:13.0000] <TabAtkins> Yeah, it's simple in any direction. [06:02:20.0000] <TabAtkins> I mean, I suppose I could make the tokenizer look for escaped u, r, and l to invoke the url token parsing. It just doesn't seem necessary. [06:02:40.0000] <annevk> but that's what happens everywhere else [06:02:45.0000] <TabAtkins> Oh, and here's a more realistic example: url(http://example.com/*) [06:02:56.0000] <annevk> it would be highly inconsistent not to look for escapes [06:03:04.0000] <zcorpan> TabAtkins: in opera/gecko, url(image?/**/) and \url(image?/**/) both load the image, while url(image/**/) and \url(image/**/) both don't [06:03:11.0000] <TabAtkins> This is an invalid url token that ends at the ). It's an unclosed comment if it's a function. [06:03:47.0000] <zcorpan> (in live dom viewer) [06:03:51.0000] <annevk> what happens for url(x()x) [06:04:03.0000] <annevk> or url([)]) [06:04:34.0000] <annevk> (whether or not it follows the generic error handling) [06:04:58.0000] <TabAtkins> Let's see what url([)]) does... [06:05:59.0000] <zcorpan> TabAtkins: http://software.hixie.ch/utilities/js/live-dom-viewer/saved/1518 gets applied in opera/gecko, so they don't tokenize a comment there [06:09:06.0000] <zcorpan> TabAtkins: i think i'd go with majority and make the tokenizer look for escaped url :-) [06:09:43.0000] <TabAtkins> Putting together a decently comprehensive suite real quick. [06:09:44.0000] <kennyluck> Anyway, we had http://lists.w3.org/Archives/Public/www-style/2007Dec/0215 from annevk, http://lists.w3.org/Archives/Public/www-style/2010Jul/0499 from Bjoern Hoehrmann, http://lists.w3.org/Archives/Public/www-style/2011Apr/0680 from fantasai all about this for whoever is interested... [06:10:27.0000] <annevk> haven't changed my mind on this in 4+ years it seems [06:11:11.0000] <TabAtkins> Yeah, I'll do whatever's reasonable and common. Just trying to figure out what it is, comprehensively. [06:11:32.0000] <TabAtkins> I did my initial spec from testing in webkit only. ^_^ [06:12:36.0000] <kennyluck> TabAtkins, well, what you do is matching CSS 2.1 no there is no blame for you. [06:12:36.0000] <kennyluck> s/no/so/ [06:14:47.0000] <TabAtkins> That's true too. Doesn't excuse sticking with 2.1 if it's clear that reality diverges, though. [06:15:20.0000] <annevk> I don't think I ever got a reply to that email btw [06:15:23.0000] <TabAtkins> The correct answer is "nobody should ever touch the parser itself ever again". [06:15:30.0000] <annevk> CSS WG violated W3C Process, news at eleven [06:15:36.0000] <annevk> oh, film at eleven, doh [06:17:45.0000] <Ms2ger> TabAtkins, hmm, once you have a decently comprehensive suite, you can unprefix! [06:19:16.0000] <TabAtkins> Ms2ger: FINALLY [06:20:19.0000] <annevk> "PREFIX ALL THE FEATURES" [06:28:55.0000] <TabAtkins> All right, everyone non-webkit agrees. [06:29:15.0000] <TabAtkins> Escaping the u,r, and l characters are kosher, and it still invokes the URL token parsing. [06:29:30.0000] <TabAtkins> So I'll fix that in the spec. Kenny, feel free to patch accordingly. [06:30:08.0000] <TabAtkins> That still means that \url(attr(foo)) won't do what that thread wanted, because you get the URL token parsing. [06:30:43.0000] <TabAtkins> But like I said, image() handles that fine, because it parses as a FUNCTION instead. [06:31:24.0000] <kennyluck> TabAtkins, which means I need to patch CSS2. (my patch changes Gecko to match WebKit) [06:31:41.0000] <TabAtkins> Well, I'll patch CSS2.1. [06:31:50.0000] <TabAtkins> Don't match Webkit. [06:32:07.0000] <zcorpan> patch webkit! :-P [06:32:15.0000] <TabAtkins> Yes! [06:32:35.0000] <kennyluck> TabAtkins, please do. I already add my mail to the long list of mails about this I mentioned above. [06:32:44.0000] <Ms2ger> TabAtkins, only if you have proposed text and a test case ;) [06:34:11.0000] <kennyluck> Does anyone have an opinion on whether \u+1234 should be accepted as UNICODE-RANGE or not? [06:34:59.0000] <TabAtkins> Ms2ger: I'm writing up an email with proposed text now. [06:35:24.0000] <TabAtkins> kennyluck: What's the support? [06:35:36.0000] <TabAtkins> That's all I really care about. ^_^ [06:36:41.0000] <kennyluck> TabAtkins, does any browser support 'unicode-range' already? Otherwise it can't be tested... [06:37:23.0000] <TabAtkins> Yeah, I think that's supported pretty widely in @font-face. [06:37:32.0000] <TabAtkins> I know that at least *some* browsers support it. [06:37:39.0000] <zcorpan> does anyone *use* unicode-range? [06:37:47.0000] <TabAtkins> Can't answer that. ^_^ [06:38:02.0000] <zcorpan> it seemed like a misfeature last time i looked at the spec [06:39:36.0000] <TabAtkins> Not my place to say, within the context of css3-syntax. [06:41:58.0000] <kennyluck> "The CSS code above worked for me in Chrome and Safari on Mac OS X 10.6. but didn’t work in Firefox or Opera. Nor did it work in IE9 on Windows7" [06:42:35.0000] <kennyluck> If that's true, I am pretty sure WebKit follows the CSS 2.1 grammar so there's no support. [06:42:59.0000] <kennyluck> But well, I don't know if people will think that's inconsistent. [06:47:57.0000] <TabAtkins> kennyluck: You can just use the CSSOM to query the property on the font-face rule, and see if it's been dropped or not. [06:48:52.0000] <TabAtkins> It would be pretty dumb if every single browser was opposite (in different ways) in whether they support escapes in URL and UNICODE-RANGE. [06:50:45.0000] <Ms2ger> It would be appropriate for the web [07:13:54.0000] <TabAtkins> Oh, heh, we took the escaped-url as editorial, and then just forgot to make the edit. [07:14:21.0000] <TabAtkins> (It's not editorial, of course. No idea why they thought that.) [07:17:35.0000] <TabAtkins> Ah, yup, and the change was covered by uri-015; it's just that the spec doesn't match the test. [07:24:08.0000] <TabAtkins> So, annevk, no W3C Process was violated. Just forgetfulness. ^_^ [07:26:06.0000] <zcorpan> ah, so that's how not to violate the process! [07:28:31.0000] <TabAtkins> 1. Make promises to edit something. [07:28:32.0000] <TabAtkins> 2. [07:28:34.0000] <TabAtkins> 3. Profit! [07:29:40.0000] <zcorpan> yeah, we'll put in longdesc. promise. [08:20:24.0000] <jgraham> This is the second time today that Ms2ger has quit just as I want to talk to him :( [08:20:56.0000] <jgraham> AryehGregor: https://gist.github.com/2636212 [08:50:48.0000] <annevk> TabAtkins_: mwaha [10:22:18.0000] <Hixie> TabAtkins_: do you still think we should add .naturalOrientation to <img> to expose a particular facet of the EXIF data? [10:41:41.0000] <annevk> Hixie: trying to get emails back to 1000? [11:01:44.0000] <Hixie> annevk: got to hit 0,0,0 sometime this year! [11:02:47.0000] <annevk> you still want to proof your abandoned timeline? [11:03:10.0000] <annevk> because I think it might have included some assumptions about feature freeze too :) [11:05:29.0000] <annevk> http://w3cmemes.tumblr.com/post/22659085815 ref? [11:09:11.0000] <annevk> MikeSmith: is sysreq⊙wo the place to bug with TLS issues? [11:09:53.0000] <MikeSmith> annevk: yup [11:10:23.0000] <annevk> ooh [11:10:26.0000] <annevk> site-comments is public [11:10:30.0000] <annevk> I'll use that [11:10:35.0000] <MikeSmith> yeah [11:10:39.0000] <MikeSmith> that's better [11:27:10.0000] <MikeSmith> annevk: your old address is removed from the testsuite list now [11:27:41.0000] <MikeSmith> I had you in the list under "Non auto removable" [11:28:02.0000] <jgraham> annevk: You want an example of an example where where no-one remembers whether the example is an example of an in joke or not? [11:29:03.0000] <annevk> MikeSmith: ah thanks [11:29:35.0000] <annevk> jgraham: too much xzibit? [11:30:27.0000] <annevk> fast ij is fast http://lists.w3.org/Archives/Public/site-comments/2012May/0004.html [12:05:25.0000] <Hixie> annevk: not especially, i just want to actually get the feedback to zero because that's _always_ the goal. :-) [12:06:14.0000] <WeirdAl> yeah, those annoying commenters ;) [13:18:17.0000] <Hixie> abarth: what's the status of http://wiki.whatwg.org/wiki/Meta_referrer ? [13:21:00.0000] <othermaciej> Hixie: it's in WebKit (though only shipping in Chrome and we might turn it off for Safari) and Mozilla at least at one point allegedly said they'd implement [13:21:29.0000] <Hixie> k, guess i'll wait a bit longer then [13:21:37.0000] <othermaciej> Hixie: I'm not keen on the existence of the "always" value [13:22:15.0000] <othermaciej> Hixie: do you have an opinion? [13:22:33.0000] <Hixie> what's the use case for 'always;? [13:22:42.0000] <othermaciej> I do not know [13:22:49.0000] <Hixie> trackback, according to the page [13:22:55.0000] <Hixie> "A blog hosted over HTTPS might wish to link to a blog hosted over HTTP and receive trackback links" [13:23:02.0000] <Hixie> i assume it means "send trackback links" [13:23:30.0000] <Hixie> it is kinda sad that moving to https kills referer [13:24:09.0000] <othermaciej> the reason for it doing so is historical and weird, but it's created an expectation of referer privacy for at least some users [13:24:41.0000] <othermaciej> so for example some people promote using google https search to get better privacy on the search queries used to reach a site [13:24:58.0000] <Hixie> yeah [13:24:59.0000] <othermaciej> use of the "origin" value does not seem to break that assumption, but "always" would [13:25:02.0000] <Hixie> (not sure that actually works btw) [13:25:22.0000] <othermaciej> there may well be back channels for all I know [13:26:27.0000] <annevk> Google Analytics and Google Webmaster tools provide some info in how users get on your site, including search queries [13:26:48.0000] <annevk> and links on search redirects are often redirects [13:26:56.0000] <annevk> search results :) [13:28:25.0000] <othermaciej> I've heard rumors there is interest in using the "origin" meta referrer value for SSL search but I haven't heard of anyone wanting to use "always" [13:28:27.0000] <jgraham> I think people at Opera might have had concerns about "always" too. Although it was very informal discussion, and "people at Opera" might have just been "me" or something [13:28:48.0000] <othermaciej> so I would guess search providers care about whatever perceived privacy is gained by not sending the referrer [13:29:26.0000] <annevk> but they care about origin so they can promote their existence? [13:29:44.0000] <othermaciej> apparently, yes [13:30:14.0000] <othermaciej> or at least, they'd like content providers to know what traffic is being sent by their search engine, even if they don't know the exact query in any given case [13:30:14.0000] <annevk> because other than that (or CORS usage which does not apply here) it seems like a waste of bandwidth [13:31:01.0000] <othermaciej> the desire to send origin-as-referrer seems reasonable to me, though I'm in no position to explain anyone's business case for it [13:31:13.0000] <annevk> I guess the query is less and less exposed anyway with the search as you type stuff [13:32:01.0000] <Hixie> personally i'd be fine with not ever sending referer, if we send origin [13:32:05.0000] <Hixie> origin is useful for security [13:32:21.0000] <Hixie> my understanding is that "SEO" people want to know what keywords are used to get to their site [13:32:46.0000] <Hixie> (origin can also be a security/privacy problem itself, of course) [13:32:52.0000] <Hixie> (e.g. leaking intranet host names) [13:33:39.0000] <annevk> not sending Referer will likely break a number of sites unfortunately [13:33:44.0000] <Hixie> jesus, the number of ways that the <object> element's algorithm can be invoked has become ludicrously long [13:33:51.0000] <annevk> that use it to determine whether or not to display an image and such [13:35:06.0000] <Hixie> anyone know if <embed> elements that are display:none also get disabled, like <object> elemnets? [13:35:37.0000] <Hixie> oh look at that [13:35:45.0000] <Hixie> the test that roc was using to test <object> actually tests <embed> [13:35:49.0000] <Hixie> new question... [13:36:06.0000] <Hixie> anyone know if <object> elements that are display:none also get disabled, like <embed> elements? :-) [13:42:44.0000] <jgraham> Hixie: I seem to recall it matter for compat. [13:42:50.0000] <jgraham> That doesn't help much though :) [14:15:39.0000] <jamesr> Hixie, in WebKit the behaviors are the same [14:15:47.0000] <Hixie> that's what i went with too [14:15:49.0000] <jamesr> Hixie, for <object> vs <embed> re display:none [14:15:49.0000] <Hixie> more or less [14:15:49.0000] <jamesr> Hixie, for <object> vs <embed> re display:none [14:15:49.0000] <Hixie> more or less [14:15:52.0000] <Hixie> yeah [14:15:56.0000] <jamesr> queue a task, eh? [14:16:01.0000] <jamesr> which task source? some new one? [14:16:17.0000] <Hixie> same one as for other things that trigger the plugin on/off code [14:16:18.0000] <jamesr> (wondering if ordering matters vs other tasks) [14:16:32.0000] <Hixie> the DOM manipulation task source. [14:16:43.0000] <Hixie> i guess that's bad [14:16:51.0000] <Hixie> hm [14:17:22.0000] <jamesr> i'm not sure we are super careful about that (making sure stuff is in the right task source) [14:17:28.0000] <jamesr> not sure anyone is, to be honest. how tested is it? [14:17:51.0000] <Hixie> not [14:17:53.0000] <Hixie> :-) [14:18:14.0000] <jamesr> normally you just toss everything into one queue and it all just works [14:18:33.0000] <jamesr> where did you stick this new text? [14:19:10.0000] <Hixie> for <embed> it's in the "potentially active" definition and there's an open bug on making sure it is triggered on a task and not sync [14:19:15.0000] <jamesr> i may want to jack some of it for a css animations/transitions proposal [14:19:21.0000] <jamesr> what section #? [14:19:48.0000] <Hixie> and for <object> it's before the huge algorithm [14:19:49.0000] <Hixie> uh [14:20:01.0000] <Hixie> 4.8.3 and 4.8.4 apparently [14:20:12.0000] <jamesr> found it (ctrl-f for the win) [14:20:42.0000] <jgraham> No, ctrl-f for find is a horrible key combination [14:21:16.0000] <jgraham> The Opera/Firefox setup is much nicer [14:21:18.0000] <jamesr> actually i don't see this. are your changes on the whatwg.org version yet? [14:21:32.0000] <Hixie> should be [14:21:40.0000] <Hixie> that's the first version that gets regenned [14:21:51.0000] <Hixie> yup [14:22:04.0000] <Hixie> http://html5.org/tools/web-apps-tracker?from=7099&to=7100 is the diff [14:22:54.0000] <jamesr> aha [14:23:22.0000] <Hixie> actually i guess using the dom task source is ok, since we're "guaranteed" that the layout will be redone when the event loop next spins [or i guess when the next time you call an attribute that has to do it synchronously], so it's a well-defined time [14:23:45.0000] <jamesr> Hixie, "chnges" in the diff [14:23:49.0000] <Hixie> thanks will fix [14:24:35.0000] <Hixie> oh oops, i missed <embed> in the last checkin [14:25:58.0000] <jamesr> explains that 2012-05-09 [17:00:27.0000] <Hixie> so some of this feedback about responsive images suggests we should send differnet images based on bandwidth [17:00:34.0000] <Hixie> any ideas how to do that? [17:00:49.0000] <Hixie> (i mean at a high level, not syntax levle) [17:01:13.0000] <Hixie> should we just have two categories, "i am bandiwdth-constrained" and "i am willing to try for the biggest thing you have"? [17:02:53.0000] <mpt> "This is taking too long, please give me the rest of the image at the next lowest resolution" [17:04:52.0000] <Hixie> yeah but what if the user then comes out of the tunnel? [17:05:13.0000] <Hixie> what if it took a long time because the user started on wifi but the wifi router got hit by a dog? [17:07:44.0000] <jsbell> FWIW, at <previous company> with a non-Web client we used JPEG2000 progressive images. The client would just terminate the request when it had enough pixels to display, but could always do an HTTP Range request for more later on, if e.g. the user zoomed in. [17:10:08.0000] <jsbell> (More specifically, the client would request only the first 1k of data, and reason about how much more to request based on current conditions. This require the rendering pipeline and image cache to be fairly chatty.) [17:20:12.0000] <tantek> Hixie, my understanding is that the folks who've been exploring responsive images have thought through a lot of the use cases etc. You'll likely answer your questions better with some web research (unless folks here have specific URLs to such research)) [17:20:39.0000] <tantek> and yes, I agree, figure out the high-level mechanics of it first, worry about syntax later. [17:21:20.0000] <Hixie> some of the folks who've been exploring responsive images have sent zillions of e-mails on the topic to the whatwg list [17:21:25.0000] <Hixie> so i'm pretty well covered in terms of research [17:21:54.0000] <tantek> email != research in my experience. that's part of the problem. [17:22:09.0000] <Hixie> i mean, they did research, and e-mailed it to the list [17:22:10.0000] <tantek> have any sent URLs to wiki pages of research? [17:22:37.0000] <Hixie> (which is the best way to do things in the whatwg world) [17:26:47.0000] <tantek> Hixie, it works best for you, because you've chosen to treat the whatwg list as your inbox / task list. In every other way, I've found email lists (including WHATWG) to be more of a support forum, and rarely useful for much else. Though to be fair WHATWG has a decent S/N. [17:27:16.0000] <Hixie> well the help⊙wo list is a support forum by design, sure [17:27:26.0000] <Hixie> the whatwg⊙wo list is a spec feedback list by design [17:27:40.0000] <Hixie> having promised to reply to all e-mail goes a long way towards enforcing that, i think :-) [17:28:02.0000] <Hixie> (still reading it all, but so far the focus has been on image dimensions and pixel density, i haven't seen much about the network bandwidth side of responsive design, either on the threads or on the web.) [17:32:36.0000] <tantek> that being said, good to hear that you're working on responsive images - I think it will be a significant improvement to the adaptability of the web platform. [17:33:06.0000] <tantek> (and it's content and belongs in markup, rather than presentational CSS) [17:35:24.0000] <Hixie> well it belongs in both [17:35:31.0000] <Hixie> depends on what kind of image we're talking about :-) [18:41:00.0000] <abarth> Hixie, I didn't read the whole discussion above, but I don't know of anyone who is using "always" [18:41:20.0000] <abarth> Hixie: the "origin" mode seems to be the most popular [20:28:59.0000] <Hixie> lol [20:29:26.0000] <Hixie> so i've now sent enough e-mails to w3cmemes' e-mail address that google+ recommended it as a friend [20:29:35.0000] <Hixie> so i tried to add it as a friend [20:30:12.0000] <Hixie> and that sent an e-mail to w3cmemes' e-mail address asking it to join google+ [20:30:18.0000] <Hixie> which has now appeared on the tumblr -_- [20:53:58.0000] <tantek> Hixie - LOL [21:00:53.0000] <kennyluck> LOL [22:17:18.0000] <AryehGregor> jgraham, thanks. [00:07:38.0000] <zcorpan> defaults write -g ApplePressAndHoldEnabled -bool false <-- aaaaaaaah, that feels so much better [00:16:59.0000] <jgraham> AryehGregor: Not committed yet; if you have time to do a quick code review on the gist I would be happy [00:22:41.0000] <hober> zcorpan: yeah, I have (ns-set-resource nil "ApplePressAndHoldEnabled" "NO") in ~/.emacs [01:04:54.0000] <hsivonen> rafaelw_: Sorry about the delay in reply. I changed my filters to put public-webapps under my WHATWG label instead of my catchall W3C label from now on so that they don't get lost among the volume of the long tail of W3C lists I subscribe to [01:06:55.0000] <hober> Hixie: heh. deleted. [01:08:28.0000] <jgraham> hsivonen: I don't really like the idea of shoving four methodsWithLongNames on document [01:11:16.0000] <jgraham> If we believe that people prefer jQuery-style APIs to legacy-DOM-style APIs then that clearly isn't going to be popular [01:11:53.0000] <jgraham> So it will probably be wrapped in a way that reintroduces per-library magic and regexps [01:16:58.0000] <hsivonen> jgraham: Please suggest shorter names on the list. I think it makes more sense to call a single method than to first call a factory method and then set a property on the object returned by the factory method. [01:21:50.0000] <jgraham> Well the best API I can come up with is document.parse(string, ["auto"|"html"|"svg"|"mathml"|"xml"]) [01:22:27.0000] <jgraham> Where auto does some sort of magic, and the other options allow you to opt-out of the magic [01:25:12.0000] <zcorpan> tabatkins: would it be ok to have <template context=svg>? [01:26:33.0000] <zcorpan> why do we need "xml"? what would it do? [01:28:50.0000] <zcorpan> do we want contextless xml fragment parsing? [01:36:21.0000] <tabatkins> zcorpan: Maybe? It's not *horrible*, but I'd prefer to avoid it if possible. [01:36:47.0000] <tabatkins> Maybe if it was an optional switch, so we could use it to switch on <template><a>...</></> stuff. [01:37:27.0000] <tabatkins> But still by default use the "first start tag" thing. [01:38:02.0000] <tabatkins> Then we can just say "conflicts resolve in favor of HTML" without guilt. [01:51:39.0000] <AryehGregor> jgraham, I'm not sure about it until Ms2ger gets back to us on why he wants explicit timeouts instead of no test-wide timeouts at all. [01:53:28.0000] <AryehGregor> window.getLastError? Seriously? Unix's errno was so amazingly successful and non-error-prone that people need to reinvent it in JS? [01:54:19.0000] <AryehGregor> Oh, it only needs to work from window.onerror handlers. That would make some sense, although the API could be improved upon. [01:55:19.0000] <jgraham> And Ms2ger has just quite. OF course. [01:55:25.0000] <jgraham> *quit [01:57:31.0000] <jgraham> If you wanted to do that getLAstError thing, why wouldn't you just make it a property of the event? [02:01:20.0000] <AryehGregor> What event? This is window.onerror. [02:01:32.0000] <AryehGregor> It's not an actual event handler. :) [02:08:51.0000] <jgraham> Oh. Well something that passes the error object to the "event" handler then [02:09:13.0000] <jgraham> Not some global function that doesn't work in some situations [02:12:22.0000] <AryehGregor> What do you mean? [02:12:40.0000] <AryehGregor> window.onerror is the only thing that gets called for uncaught JS exceptions, right? [02:12:54.0000] <zcorpan> AryehGregor: it is a an actual event handler, but there's no actual event :-) [02:14:17.0000] <jgraham> AryehGregor: I mean if you have some function f that is called in response to an exception e, e is an argument of, or a property of an argument of, f [02:14:49.0000] <AryehGregor> jgraham, that would be nice, but it's not true for window.onerror, is it? [02:15:54.0000] <zcorpan> what you're suggesting is proposal #2 [02:18:18.0000] <jgraham> zcorpan: So it is [02:18:29.0000] <jgraham> No idea why the author thinks it's less preferable [02:18:41.0000] <jgraham> It seems obviously better to me [02:19:00.0000] <jgraham> Although if IE uses e = this, that seems like an interesting alternative [02:22:07.0000] <zcorpan> this is window in IE9 [02:22:49.0000] <odinho> I also liked the #2 better. [02:23:40.0000] <jgraham> zcorpan: Oh well in that case I guess #2 is best [02:24:46.0000] <zcorpan> WDYT about #3 (just the stack in an argument)? [02:26:57.0000] <jgraham> Seems more logical to make TC39 add the stack as a property of the event [02:27:08.0000] <jgraham> of course they will probably reject that idea [02:27:29.0000] <jgraham> Because they don't like the idea of functions being able to know their callstack [02:28:04.0000] <jgraham> s/event/Error/ [02:37:50.0000] <odinho> If everyone is doing it anyway, well, they don't have much to say then I guess :P [02:43:23.0000] <annevk> hober: you should delete the tweet too [02:43:43.0000] <hober> annevk: ok [02:44:09.0000] <annevk> hober: also, welcome to Europe, discussing prefixes? :) [02:45:00.0000] <hober> not yet, thank $deity [02:45:17.0000] <hober> right now we're talking about compositing & blending [02:45:26.0000] <jgraham> Pretty sure that prefixes are evidence that $deity is null [02:45:52.0000] <annevk> jgraham: only if you assume he's good [02:46:01.0000] <hober> jgraham: or that $deity is (not null) && vengeful [02:46:54.0000] <annevk> minutes in memespeak http://w3cmemes.tumblr.com/post/22708671971 [02:46:56.0000] <annevk> /me approves [02:58:56.0000] <hober> more minuting... http://w3cmemes.tumblr.com/post/22709255224 [02:59:38.0000] <gsnedders> odinho: Eh, there's plenty of stuff everyone supports but is unspec'd. [03:00:01.0000] <gsnedders> jgraham: wrt stack/stacktrace, I think it was decided to wait to see what impls decide to do for PTC, given TCO is required. [03:01:24.0000] <jgraham> gsnedders: You appear to have started speaking in tounges [03:03:47.0000] <jgraham> Or to put it more pleasantly, you have given in to glossolalia [03:05:49.0000] <gsnedders> s/PTC/proper tail calls/ [03:05:56.0000] <gsnedders> s/TCO/tail call optimization/ [03:06:17.0000] <jgraham> Required by what? [03:06:26.0000] <gsnedders> /me has obviously spent too much time dealing with compilers [03:06:36.0000] <jgraham> Not that I am against TCO [03:06:40.0000] <gsnedders> jgraham: The spec, as it's trivially black-box observable (running out of stack) [03:07:09.0000] <jgraham> Well I guess I should have got TCO, which might have given a clue for PTC [03:07:26.0000] <jgraham> Required by which spec? [03:07:29.0000] <gsnedders> ES6 [03:08:27.0000] <jgraham> So the tl;dr summary is "ES6 will introduce tail call optimisation (which may screw with stack traces)" [03:09:35.0000] <gsnedders> Yeah. [03:10:36.0000] <gsnedders> A large part of why Java has never had TCO is the fact the spec requires stacktraces for errors, which effectively mean having a stack. [03:10:39.0000] <gsnedders> *means [03:11:19.0000] <jgraham> Guido has the same reasoning for python (debuggability > recursion) [03:12:08.0000] <jgraham> But doesn't relying on TCO mean that your stack must have the form (roughly) A B C [D E F]* [03:12:47.0000] <jgraham> Or [03:15:11.0000] <gsnedders> jgraham: Any function with a PTC will not be on the stack after the tail call. [03:15:14.0000] <gsnedders> Or what? [03:15:32.0000] <jgraham> gsnedders: It won't be on the stack, obviously [03:15:55.0000] <gsnedders> /me needs to pack for Lkpg though [03:16:19.0000] <jgraham> I was just wondering if the stack that you would get in the absence of tail calls has to be regular in a way that would allow you to display something quite useful as a traceback [03:16:33.0000] <jgraham> even when you do have tco [03:16:52.0000] <jgraham> But I'm not sure that I'm right [03:17:03.0000] <jgraham> gsnedders: When do you arrive? [03:17:04.0000] <gsnedders> Well, you have no idea where the function was directly called from [03:17:09.0000] <gsnedders> jgraham: Late. Really late. [03:17:17.0000] <jgraham> OK [03:17:27.0000] <jgraham> To where? [03:17:31.0000] <gsnedders> Du Nord [03:17:42.0000] <jgraham> That's in Paris? [03:17:44.0000] <gsnedders> Flying into NYO [03:17:54.0000] <jgraham> Ah OK [03:18:01.0000] <gsnedders> From EDI [03:18:10.0000] <jgraham> That was the question [03:18:17.0000] <gsnedders> Ah. [03:18:21.0000] <jgraham> You have checked that there is actually a bus, right? [03:18:25.0000] <gsnedders> Yes. [03:18:47.0000] <gsnedders> Gets in at 23:15, IIRC [03:18:51.0000] <gsnedders> (the bus) [03:18:55.0000] <jgraham> OK [03:19:55.0000] <gsnedders> Staying at Hotell Du Nord, on Repslagaregatan (sp?) [03:20:04.0000] <zcorpan> ok so looking at http://kangax.github.com/es5-compat-table/non-standard/ and removing all columns except opera, ie9, ff12, sf5 and ch7-10, and removing lines where there are more "No" than "Yes", gives: [03:20:07.0000] <zcorpan> function statement , function "name" property, function "caller" property, function "arguments" property, __proto__ , __defineGetter__ , __defineSetter__ , const, RegExp "lastMatch", RegExp.$1-$9, String.prototype.substr, String.prototype.trimLeft, String.prototype.trimRight, String.prototype.anchor, String.prototype.big, String.prototype.blink, String.prototype.bold, String.prototype.link, Octal literals, error "stack" [03:21:04.0000] <gsnedders> Quite a few of those are in ES6 [03:21:46.0000] <zcorpan> what's missing? [03:22:40.0000] <gsnedders> Off hand, function properties, __define{G,S}etter__ (note IE doesn't support this), Error.stack [03:22:49.0000] <gsnedders> IE doesn't support Error.stack either [03:23:20.0000] <zcorpan> what about function statement? (http://kangax.github.com/nfe/#function-statements ) [03:23:44.0000] <gsnedders> Believe they're in, though no idea what the semantics are [03:25:39.0000] <zcorpan> where's the es6 spec? [03:26:33.0000] <gsnedders> Not everything that's been agreed to be added is in the spec yet, but http://wiki.ecmascript.org/doku.php?id=harmony:specification_drafts [03:26:50.0000] <gsnedders> Also still not agreement on __proto__ semantics [03:27:30.0000] <gsnedders> Noteably what us and JSC have impls of now are slightly different to what's in the spec, and in IMHO cleaner [03:28:46.0000] <zcorpan> i don't see anchor() anywhere [03:29:23.0000] <gsnedders> "Not everything that's been agreed to be added is in the spec yet" [03:29:44.0000] <zcorpan> k [03:34:31.0000] <gsnedders> jgraham: What's the weather like? :P [03:34:44.0000] <jgraham> Today? Lousy [03:34:54.0000] <jgraham> Rain [03:35:06.0000] <gsnedders> How warm? [03:35:17.0000] <jgraham> I haven't worked out how to get the temperature to display in unity yet [03:35:56.0000] <jgraham> http://www.temperatur.nu/ullstamma.html [03:35:58.0000] <gsnedders> Should I plan to wear enough not to be frozen given snow everywhere, or enough to not be frozen given a sane spring temperature? [03:36:23.0000] <gsnedders> Okay. [03:37:38.0000] <jgraham> In other UI rants, who thought that hiding the menu headings in unity was a good idea? It reduces discoverability and increases aquisition time [03:38:02.0000] <jgraham> /me decides to blame mpt [03:38:32.0000] <jgraham> I can only imagine it is part of some long-term plan to make application designers stop using menus [04:03:22.0000] <jgraham> Is there some way to tell from script if a CSS background image has finished loading? [04:03:59.0000] <annevk> load it through new Image and time that? [04:05:17.0000] <jgraham> Well for my purposes I could rely on timing [04:05:26.0000] <jgraham> But I would rather not [04:05:33.0000] <jgraham> (makes tests more unstable) [04:06:08.0000] <annevk> unless something changed in the last couple of months CSSOM does not expose much in this area [04:06:31.0000] <jgraham> Yeah and we still don't have drawElement in the 2D context [04:06:58.0000] <jgraham> Timing it is, I guess [04:11:50.0000] <jgraham> Next question: is it me, or is http://www.whatwg.org/specs/web-apps/current-work/#critical-subresources a lie? [04:12:08.0000] <jgraham> In particular, it seems that background images in CSS do delay the load event [04:12:15.0000] <jgraham> /me hasn't tried anything else yet [04:33:27.0000] <annevk> jgraham: even background images that don't apply at the moment? [04:33:57.0000] <annevk> jgraham: e.g. @media (min-width:10000px) { body { background:url(trololol.jpg) } } [04:34:14.0000] <annevk> jgraham: because if browser load those they're being silly [04:34:18.0000] <annevk> browsers* [04:43:19.0000] <zcorpan> jgraham: yeah they delay the load event (if they're loaded). but no spec mentions that, iirc [04:43:41.0000] <zcorpan> same with web fonts [04:48:30.0000] <jgraham> The HTML spec specifically says that they don't [04:48:59.0000] <jgraham> Fun fun [04:49:29.0000] <annevk> again, not all of them do [04:49:44.0000] <jgraham> annevk: example? [04:49:58.0000] <annevk> jgraham: see above? [04:50:23.0000] <jgraham> annevk: I haven't tried media queries, but display:none doesn't seem to stop the delay [04:50:54.0000] <annevk> selectors that don't apply? [04:51:16.0000] <jgraham> Haven't tried that [04:52:35.0000] <jgraham> But how do you tell that they don't apply? [04:53:02.0000] <annevk> #test { ... } [04:53:25.0000] <annevk> there was this site once that had about a 100 different background images depending on weather conditions and time of day [04:53:38.0000] <jgraham> Hmm, that doesn't seem to block the load event [04:53:41.0000] <annevk> all linked from the stylesheet and the markup would change [04:54:00.0000] <annevk> some browsers did load them all, they quickly changed that [04:54:06.0000] <jgraham> Doesn't that just give lag when things change? [04:54:31.0000] <annevk> prolly [04:54:49.0000] <annevk> but downloading a 100 images and keeping them around when only one is shown is kind of wasteful [04:55:04.0000] <jgraham> So basically onload is tightly coupled to layout in a way that is totally undefined? [04:55:46.0000] <annevk> well layout is not defined [04:55:57.0000] <annevk> so yes, that would follow from that [04:58:24.0000] <annevk> heycam: any news yet on when you're going to reply to my email? [05:00:40.0000] <heycam> annevk: ah which email is it again? I'll try to reply now :) [05:00:45.0000] <heycam> (otherwise I'm off on leave for a week) [05:02:17.0000] <annevk> heycam: http://lists.w3.org/Archives/Public/public-script-coord/2012AprJun/0137.html yay [05:02:38.0000] <annevk> heycam: vacation time? [05:02:53.0000] <heycam> annevk: once I hop on this plane and arrive back home, yeah [05:08:34.0000] <heycam> annevk: replied, not sure if you're looking for a deeper answer [05:09:18.0000] <heycam> annevk: mm this airport wifi might be blocking my smtp connection [05:09:40.0000] <annevk> is the answer "seemed to made sense"? [05:09:59.0000] <heycam> pretty much. for things like say insertBefore(node, undefined), that needs to be have like insertBefore(node, null) [05:10:07.0000] <heycam> I think it was just an oversight that I fixed [05:10:33.0000] <annevk> why does that need to behave the same? [05:10:44.0000] <heycam> because that's how implementations behave [05:10:51.0000] <annevk> I'm asking as someone internally wanted to make it throw [05:10:54.0000] <heycam> the alternative is to throw an exception [05:10:54.0000] <annevk> right okay [05:11:01.0000] <heycam> I see [05:11:25.0000] <heycam> well, passing undefined to something that would normally expect null seems fine [05:11:44.0000] <annevk> anyway, this makes sense to me, I'll ask him if he wants to pursue this further [05:11:45.0000] <heycam> if you were writing something in JS you'd be more likely to do a test like if (child) [05:11:48.0000] <heycam> ok cool [05:11:52.0000] <heycam> sorry for the delay [05:13:27.0000] <heycam> (the mail will probably arrive once I'm back home with decent network) [05:15:05.0000] <annevk> k [05:21:23.0000] <annevk> i love this one http://w3cmemes.tumblr.com/post/22670112919 [05:21:35.0000] <annevk> i wonder where that face is from [05:24:26.0000] <michel_v> a spanish politician, IIRC [07:31:28.0000] <jzaefferer> I'm working on a custom tooltip implementation for jQuery UI, and were running into problems related to ARIA and removing the native tooltip. We set the title to an empty string to disable the native tooltip, but that causes problems with screenreaders. I've found this Chromium ticket, suggesting that event.preventDefault on mouseover should prevent the native tooltip: http://code.google.com/p/chromium/issues/detail?id=42549 There's no mention of th [07:34:49.0000] <jzaefferer> (apparently my message was too long? here's the last part again) [07:34:50.0000] <jzaefferer> There's no mention of that in the spec: http://dev.w3.org/html5/spec/global-attributes.html#the-title-attribute - is that something that could or should be added? [08:26:04.0000] <mpt> jgraham, Mark Shuttleworth [08:27:09.0000] <mpt> jgraham, Ubuntu 12.10 will have an option to show menu titles all the time -- though why anyone would want that option turned off, I don't know [08:27:56.0000] <jgraham> mpt: He decided that UI personally? Interesting… [08:28:14.0000] <jgraham> Great, that sounds like something to look forward to [09:18:49.0000] <dglazkov> good morning, Whatwg! [09:27:41.0000] <smaug____> good evening [10:10:58.0000] <zcorpan> Hixie: could you have a look at https://www.w3.org/Bugs/Public/show_bug.cgi?id=16635 ? it should be blocking the <template> problem everyone's trying to solve [11:45:35.0000] <jgraham> Can't we make whatwg.org/html the default version of the spec yet? [11:45:43.0000] <jgraham> Uh [11:45:49.0000] <jgraham> s/default/single page/ [11:59:28.0000] <annevk> jgraham: use whatwg.org/c [12:02:43.0000] <kennyluck> Are there now more people in favor of the single page version? [12:11:00.0000] <Ms2ger> jgraham, please, no :) [12:12:27.0000] <annevk> /me always uses whatwg.org/C for multi-page [12:24:21.0000] <jgraham> PF [12:25:26.0000] <annevk> jgraham: hmm? [12:25:38.0000] <jgraham> Uh, network problems [12:26:11.0000] <jgraham> Ms2ger: Browsers cope with it a lot better than a few years ago [12:26:16.0000] <jgraham> and it is way more useful [12:28:15.0000] <Ms2ger> So, what did you want to ask me ten hours ago? [12:29:41.0000] <Ms2ger> (re timeouts / testharness.js) [12:30:19.0000] <jgraham> Ms2ger: https://gist.github.com/2636212 [12:30:34.0000] <jgraham> Is my proposal-as-code [12:31:38.0000] <Ms2ger> What's load_test_attr on line 12? [12:31:53.0000] <jgraham> gsnedders: You are at zcorpan's desk btw [12:32:27.0000] <jgraham> Ms2ger: Oh, copy and paste detris [12:34:00.0000] <Ms2ger> Do you want to do the clearTimeout in Tests.prototype.set_timeout? I guess you do [12:34:38.0000] <jgraham> Yeah, [12:35:00.0000] <Ms2ger> Velmont, exception ping ;) [12:36:39.0000] <Ms2ger> jgraham, is setup({ timeout: foo, explicit_timeout: true }) silly, and should it throw? [12:38:47.0000] <jgraham> Ms2ger: Yeah, it probably could [13:20:31.0000] <jsocol> hey Hixie et al [13:20:44.0000] <jsocol> so I have an example here: https://gist.github.com/2648133 and I've been reading through the parsing spec [13:21:01.0000] <Ms2ger> Hi again! [13:21:04.0000] <jsocol> and it seems like going from "end tag open state" to "tag name state" is a mistake [13:21:28.0000] <Ms2ger> Intentional, I'm pretty sure [13:21:29.0000] <jsocol> because "tag name state" can go to "before attribute name state" or "self-closing start tag state" [13:22:03.0000] <jsocol> and the definition of an end tag doesn't allow either of those states [13:22:30.0000] <Ms2ger> What do you mean by "allow"? [13:22:34.0000] <jsocol> so it seems like there should be an "end tag name state" that is basically the same as "tag name state" but never goes to "before attribute name state" or "self-closing start tag state" [13:22:39.0000] <jgraham> You can have end tags in attributes in the parser [13:22:43.0000] <jsocol> Ms2ger: http://dev.w3.org/html5/spec/syntax.html#end-tags [13:22:55.0000] <Ms2ger> That's you mistake [13:22:58.0000] <Ms2ger> your* [13:22:58.0000] <jsocol> there can be trailing whitespace after the tag name, but no other characters [13:23:05.0000] <jgraham> They are parse errors and ignored, but the tokenizer has to deal with them [13:23:18.0000] <Ms2ger> The authoring requirements and the actual parsing have no meaningful relation to eachother [13:23:24.0000] <jgraham> jsocol: The requirements on UAs are entirely unrelated to the requirements on authors [13:23:55.0000] <Hixie> yeah that's not a bug, that's intentional [13:24:02.0000] <Hixie> </foo bar baz> is treated as a single end tag [13:24:04.0000] <jsocol> let me rephrase [13:24:04.0000] <Hixie> for compat reasons [13:24:14.0000] <jsocol> oh compat reasons? [13:24:19.0000] <jsocol> I figured there was a reason [13:24:25.0000] <Hixie> (the "bar baz" part turns into "attributes", that are then dropped) [13:24:28.0000] <Ms2ger> It's a magic phrase [13:24:31.0000] <Hixie> pretty much everything the parser does is for compat reasons :-) [13:24:42.0000] <Ms2ger> Hixie says "compat reasons", everyone shuts up [13:24:47.0000] <jgraham> jsocol: The whole parser is basically a big tangle of compat constraints [13:25:19.0000] <Hixie> (btw you may find http://whatwg.org/html to be a better reference) [13:25:33.0000] <jsocol> the unfortunate consequence of using the same "tag name state" is that the parser does something non-obvious for authors [13:25:39.0000] <Hixie> (it's mostly the same text but is more likely to remain up to date) [13:25:47.0000] <jgraham> Hixie: Unrelatedly, when one spins the event loop waiting for a condition, it is only supposed to observe the condition at the end of tasks, right? [13:26:24.0000] <Ms2ger> jsocol, well, you've got an error [13:26:26.0000] <Hixie> jgraham: it checks for the conditions continually, but queues a task when it is met [13:26:35.0000] <jgraham> Hixie: Huh? [13:26:38.0000] <Ms2ger> jsocol, there are multiple ways to fix it up [13:27:05.0000] <Hixie> jgraham: which part of my statement was confusing? [13:27:24.0000] <Ms2ger> jsocol, I'm not sure if one of the DOMs you expected is more obvious than what we had to go with [13:27:42.0000] <jgraham> So if I have a script that temporarily causes a condition to be met, it thinks the condition has been met, even if it is never true at the end of the event loop? [13:27:43.0000] <jsocol> Ms2ger: the "correct" way is obviously not to have the error. but either of those seems entirely more obvious, as a web developer [13:27:44.0000] <Ms2ger> Then again, I've read some of the parsing algorithm... [13:27:56.0000] <jsocol> than a missing > cutting off an entire paragraph [13:28:10.0000] <Ms2ger> So, my definition of "obvious" may differ from the one used by sane peopl [13:28:11.0000] <Ms2ger> e [13:28:14.0000] <Hixie> jsocol: it's definitely not intuitive, i agree [13:28:25.0000] <Hixie> welcoem to html :-) [13:28:31.0000] <jsocol> heh :) [13:30:25.0000] <jsocol> at the risk of being "that guy" it seems like two new states would allow the parser to do something much more intuitive with unfinished end tags [13:30:35.0000] <jsocol> though there might be other analogous states needed elsewhere [13:31:06.0000] <jgraham> Hixie: For example, if I have a document with one inline resource loading which is delaying the load event and I run a script (e.g. from DOMContentLoaded) that stops that load and starts a new load, should onload be delayed by the new load? [13:31:17.0000] <jgraham> Or can that not happen for some reason? [13:31:50.0000] <Ms2ger> jsocol, the final remark in http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2010-December/029409.html is relevant :) [13:32:32.0000] <jgraham> (the only reason I can think that it might not be possible is if the "delay a load event" always happens in resposne to something async, so you always reach the end of the event loop with nothing blocking the load at least once) [13:34:23.0000] <jsocol> Ms2ger: heh ;) [13:36:14.0000] <jsocol> Ms2ger: to be fair, though, this isn't about aesthetics so much as Postel's law and trying to minimize the impact of a small error [13:36:48.0000] <Ms2ger> jsocol, the HTML parser is a textbook example of how Postel's law is fundamentally broken [13:37:47.0000] <jsocol> I don't think it's broken so much as "really hard" [13:38:30.0000] <jsocol> Ms2ger: anyway, politically, if I bring this up on the list, it'll just get shot down? I don't want to waste my time. [13:38:38.0000] <Ms2ger> Well [13:38:48.0000] <Ms2ger> Hixie already shot you down, pretty much :) [13:39:00.0000] <jgraham> Alternatively the HTML parser is a textbook example of how postel's law is necessary [13:39:20.0000] <jgraham> jsocol: The requirements on UAs here aren't going to change [13:39:42.0000] <jsocol> jgraham: fair enough. [13:39:46.0000] <jsocol> thanks for all your time [13:39:58.0000] <jgraham> In general we are *very* conservative with parser changes now that the main browsers are interoperable and the universe didn't end [13:41:57.0000] <Hixie> yeah i wouldn't expect the parser to change [13:42:00.0000] <Hixie> that ship has sailed [13:42:07.0000] <gavinc> Is DOM4 on the same 2014 time line for recomendation? [13:42:11.0000] <Hixie> i mean, not for intentional things like this [13:42:21.0000] <gavinc> sorry, same HTML5 timeline [13:42:27.0000] <Hixie> it will still change for bugs of course :-) [13:42:28.0000] <Ms2ger> gavinc, there's a timeline in the new WebApps charter [13:42:37.0000] <Hixie> gavinc: this may be the wrong channel for asking about the HTML5 timeline :-P [13:42:39.0000] <jgraham> People care about those timelines? [13:42:42.0000] <Ms2ger> It may or may not have any relation to reality [13:43:03.0000] <Hixie> jgraham: i think as written (from memory) it would not delay onload, but let me check [13:43:12.0000] <Ms2ger> gavinc, and the timeline for HTML5 is still rec in 2022 [13:43:16.0000] <gavinc> people writing another w3c recomendation which would like to refer to DOM 4 do indeed care [13:43:19.0000] <Ms2ger> Whatever the W3C may claim [13:43:41.0000] <jgraham> People writing other recommendations should do the right thing and reference DOM4 [13:43:42.0000] <Hixie> Ms2ger: unless they violate the process [13:43:45.0000] <Hixie> Ms2ger: which is likely to occur [13:43:49.0000] <Ms2ger> Which they will, of course [13:43:59.0000] <jgraham> It is a myth that you can't do that and advance your own recommendation [13:44:20.0000] <Ms2ger> jgraham, it's irrelevant what you reference anyway [13:44:28.0000] <Ms2ger> Also, meant for gavinc [13:44:42.0000] <Ms2ger> The keys are so close together [13:44:47.0000] <gavinc> It's a myth that W3C staff likes mentioning anytime I try to refrence something from Webapps or HTML WG ;) [13:45:10.0000] <Hixie> gavinc: tell them to get teh fuck out of the way [13:45:16.0000] <Hixie> gavinc: isntead of harming the web [13:45:22.0000] <jgraham> Well yeah, but it would be better to reference the thing that people should actually look at rather than hope people will realise you were just playing Process games [13:45:36.0000] <gavinc> I'm aware! [13:46:01.0000] <Hixie> gavinc: bureaucracy is opt-in. They can't make you do anything you don't want to do. [13:46:39.0000] <jgraham> Also, I am reasonably sure this particular requirement isn't really baked in anywhere [13:46:52.0000] <gavinc> Yes, but I'm hurting the web anyway working on RDF recomendations ;) [13:46:56.0000] <Ms2ger> Oh [13:47:01.0000] <Ms2ger> In that case, go away ;) [13:47:12.0000] <Hixie> lol [13:47:28.0000] <Hixie> ah well if you're working on RDF, then yeah, you definitely shouldn't go to REC until HTML5 is in REC in 2022. :-P [13:47:46.0000] <Hixie> (TR/ page is such a waste of time) [13:48:03.0000] <gavinc> Hey hey, at least XMLLiteral and a new HTML literal use the DOM as their value space [13:56:32.0000] <Hixie> can anyone sanity check this idea for me?: http://junkyard.damowmow.com/506 [13:57:05.0000] <Hixie> the algorithm is a bit busted [13:57:10.0000] <Hixie> but ignore that [13:57:50.0000] <gavinc> wouldn't it make sense to use the RFC uri template mechanisum too? [13:57:59.0000] <Hixie> lordy [13:58:08.0000] <Hixie> that seems like overkill [13:58:30.0000] <jgraham> Oh my. That's horribly ugly [13:58:41.0000] <Hixie> you reckon? [13:58:51.0000] <a-ja> question re: <dialog> -- i see there's a declarative way to close a dialog, but is there a declarative (i.e. scriptless) way to "open" a dialog? [13:59:02.0000] <Hixie> a-ja: there is not [13:59:14.0000] <a-ja> hrm [13:59:24.0000] <Hixie> what's your use case? [13:59:30.0000] <gavinc> also the pixal density is in terms of what? what exactly is 1? [13:59:45.0000] <Hixie> jgraham: ("you reckon?" being an honest question, not sarcasm, for the record) [13:59:53.0000] <Hixie> gavinc: 1 = 96dpi [14:00:15.0000] <Hixie> gavinc: device pixels per CSS pixel [14:00:17.0000] <jgraham> Inventing a whole templating microsyntax for a specific case of image loading seems pretty ugly, yes [14:00:42.0000] <a-ja> Hixie: e.g. opening a modal nav menu on mobile [14:00:47.0000] <Hixie> jgraham: it's not really a new syntax, it's just your regular substitution mechanism [14:00:55.0000] <gavinc> Hixie: oh, like -moz-device-pixel-ratio got it [14:01:20.0000] <jgraham> Hixie: From a UA point of view it is a new syntax, and needs special processing rules [14:01:29.0000] <Hixie> jgraham: i don't really care what the substitution syntax is if you have a better idea [14:01:47.0000] <Hixie> jgraham: but as far as i can tell the implementation is just three lines of code -- replace this string with this string, three times [14:02:22.0000] <Hixie> jgraham: very similar to addProtocolHandler()'s syntax [14:02:24.0000] <gavinc> Hixie: if one was going to introduce one, using the URI templates syntax wouldn't be a bad idea [14:02:26.0000] <jgraham> Hixie: You need to parse the src versions attribute and deal with invalid values [14:02:44.0000] <Hixie> jgraham: yeah, that's basically the "sizes" attribute with an extra axis [14:02:56.0000] <Hixie> gavinc: that's more complicated than just three substitutions [14:03:03.0000] <jgraham> and deal with unexpected % values in the filename template and so on [14:03:24.0000] <Hixie> nah you jus do the same as addPRotocolHandler() [14:03:33.0000] <Hixie> replace %w with width value if present. [14:03:36.0000] <Hixie> ditto h, d [14:03:45.0000] <Hixie> then resolve and fetch [14:04:07.0000] <jgraham> So you can't have the literal string %w in the image filename [14:04:19.0000] <Hixie> it would be invalid anyway [14:04:21.0000] <jgraham> and it only solves the use case where the only thing you care about is the physical image size, which people claimed wasn't enough [14:04:34.0000] <Hixie> i guess %d should be %r or something [14:04:40.0000] <Hixie> since we can't use a-f [14:04:50.0000] <Hixie> jgraham: how so? [14:05:02.0000] <webben> What's the declarative way to close a dialog? [14:05:08.0000] <jgraham> Hixie: What do you mean? [14:05:19.0000] <Hixie> jgraham: what doesn't it handle? [14:05:24.0000] <Hixie> (other than bandwidth isues) [14:05:42.0000] <Hixie> webben: <form method=dialog> <input type=submit> </form> [14:05:47.0000] <Hixie> webben: iirc [14:05:50.0000] <webben> ta [14:06:23.0000] <jgraham> Well bandwidth was one that people mentioned. I'm not sure if there were other device properties people cared about [14:06:33.0000] <jgraham> I would need to read the use cases again [14:07:02.0000] <webben> <form method=dialog action="#fragment-id-of-dialog"><input type="submit"></form> might work for opening [14:07:25.0000] <Hixie> there were various device properties people mentioned, e.g. input touch vs keyboard, but as far as image selection goes only bandwidth, available space, and pixel density were requested, at least in what i've read so far [14:07:30.0000] <Hixie> (still reading) [14:08:02.0000] <a-ja> webben: what Hixie said.....plus there some extra handling for input image coordinates [14:08:09.0000] <Hixie> gavinc: (do you have a link to the latest uri templating stuff so i can confirm or contradict my fears? i can't find a relevant link on google.) [14:08:32.0000] <gavinc> Hixie: http://tools.ietf.org/html/rfc6570 not -that- horrifying [14:08:54.0000] <gavinc> Not saying implement the whole thing [14:08:58.0000] <gavinc> but use it's syntax [14:08:58.0000] <Hixie> a-ja: i don't have an a priori objection to finding a way to open a dialog declaratively, but i'm not sure i really understand how it would work [14:08:58.0000] <gavinc> but use it's syntax [14:08:58.0000] <Hixie> a-ja: i don't have an a priori objection to finding a way to open a dialog declaratively, but i'm not sure i really understand how it would work [14:09:14.0000] <gavinc> on the other hand, implementing the whole thing would be generally useful ;) [14:09:15.0000] <Hixie> gavinc: if we used its syntax, you wouldn't be able to mix this stuff with templates later [14:09:42.0000] <Hixie> /me really isn't sold on uri templates in general, fwiw [14:09:42.0000] <gavinc> Hixie: Yeah, but that happens with tons of template syntax [14:09:43.0000] <Hixie> /me really isn't sold on uri templates in general, fwiw [14:10:05.0000] <Hixie> i really don't see any advantage to just making this look like uri templates, it would make people think it was related [14:10:09.0000] <Hixie> which is just confusing [14:10:13.0000] <Hixie> and would cause all kinds of trouble [14:10:14.0000] <Philip`> /me wonders if he's missed why you'd ever want to pass parameters inside the filename, instead of doing face.jpg?w=600&h=200&d=1 [14:10:23.0000] <Hixie> e.g. people asking for more of it to be implemented [14:10:29.0000] <Hixie> Philip`: static CDNs [14:10:47.0000] <Hixie> Philip`: but note that the proposal above supports query parameters too [14:11:11.0000] <Hixie> just do "face.jpeg?w=%w&amp;h=%h&amp;d=%r" [14:13:12.0000] <gavinc> "my%5Fface%5%w%5%h%5%r" gets a bit special [14:13:43.0000] <Hixie> you don't have to escape _ [14:13:55.0000] <Hixie> not sure what your %5s are [14:14:06.0000] <gavinc> err forgeting to type F [14:14:44.0000] <webben> Hixie: I guess in a-ja's case, menu items to open submenus would actually be dialog submit forms opening the appropriate dialog for the submenu. [14:15:01.0000] <webben> the submenu dialog would have a close form [14:15:08.0000] <Hixie> webben: ah, yeah, if you want a button to open another dialog... [14:15:20.0000] <a-ja> Hixie: i'm thinking link (possibly wrapping a button) to set a href'ed dialog's open attribute. [14:15:23.0000] <webben> so you could pop open or close the entire menu tree without any JS. [14:15:37.0000] <Hixie> yeah [14:15:46.0000] <Hixie> that wasn't really the use case i had in mind when designing this [14:15:49.0000] <Hixie> but it's an interesting use case [14:16:30.0000] <Hixie> if you have any examples (ideally with screenshots) of sites doing things like this, it would be great to document them (the screenshots) on the wiki and mail a link to the list [14:16:32.0000] <gavinc> So if your passing the value of src-template into somethat that does uri quoting... the UA would need to see face-%25w-%25h%40%25d.jpeg as being the same? [14:16:35.0000] <a-ja> Hixie: something that'd look like menu button at http://addons.mozilla.org/en-US/android/ [14:16:53.0000] <Hixie> gavinc: i don't understand the question [14:16:53.0000] <gavinc> or is src-template not really an URI so you can't really do quoting? [14:16:53.0000] <Hixie> gavinc: i don't understand the question [14:17:35.0000] <Hixie> a-ja: i don't see a button labled "menu" on that page [14:17:50.0000] <gavinc> if src-template does quoting than the replacement is not just %w but %25w and if it doesn't allow quoting it's a bit funky? [14:18:02.0000] <Hixie> what do you mean by "does quoting"? [14:18:17.0000] <gavinc> Can you use % quoted strings in it? [14:18:22.0000] <Hixie> it works the same as addProtocolHandler() [14:18:33.0000] <Hixie> the value is taken as a literal [14:18:38.0000] <Hixie> you do the substitutions [14:18:42.0000] <Hixie> then you resolve the URL [14:18:44.0000] <gavinc> ah [14:18:45.0000] <Hixie> then you fetch the image [14:18:45.0000] <gavinc> okay [14:18:57.0000] <a-ja> Hixie: top right.....maybe it looks different with UA-sniffing [14:19:19.0000] <gavinc> Hixie: So just have to make sure on a producers end that %w doesn't get quoted like % would otherwise [14:19:21.0000] <Hixie> a-ja: what browsers should i be testing with? [14:19:42.0000] <Hixie> gavinc: it's not a URL, right [14:19:50.0000] <a-ja> i'm using FF android nightly [14:19:54.0000] <Hixie> it's a pattern that created a URL [14:20:05.0000] <Hixie> a-ja: url to install that? :-) [14:20:07.0000] <gavinc> Hixie: producers need to make sure that it's quoted LIKE a URL except in some special way [14:20:23.0000] <Hixie> a-ja: got it [14:20:29.0000] <a-ja> k [14:23:55.0000] <gavinc> Hixie: also generally, what's the advantage of a template vs just providing a URL for each of the sizes? [14:24:08.0000] <a-ja> Hixie: looks pretty much same on stock browser (gingerbread) [14:24:15.0000] <benvie> there's no "official" .idl for html5 aside from what can be made from concatenating together the chunks from the actual spec itself right? [14:24:15.0000] <benvie> just want to make sure I'm not missing something obvious [14:24:15.0000] <benvie> there's no "official" .idl for html5 aside from what can be made from concatenating together the chunks from the actual spec itself right? [14:24:15.0000] <benvie> just want to make sure I'm not missing something obvious [14:24:23.0000] <gavinc> eg.. src-600x200x2 src-600x200x1 src-200x200x1 [14:26:25.0000] <Hixie> right, got that installed [14:26:27.0000] <Hixie> let's look at this page now [14:26:58.0000] <gavinc> Hixie: or src-sized="200x200x1:blah.jpg" [14:27:32.0000] <Hixie> benvie: correct [14:27:43.0000] <Hixie> gavinc: the browser needs to know what the values are so it can pick the right one [14:28:00.0000] <Hixie> a-ja: that menu doesn't seem modal [14:28:13.0000] <Hixie> a-ja: just looks like a regular menu to me [14:28:16.0000] <annevk> gavinc: which draft are you editing? [14:28:23.0000] <Hixie> a-ja: for which i'd recommend <menu>, not <dialog> [14:28:36.0000] <annevk> oh RDF [14:28:38.0000] <annevk> :) [14:29:51.0000] <gavinc> annevk: sigh ... thanks :P http://dvcs.w3.org/hg/rdf/raw-file/default/rdf-turtle/index.html [14:30:19.0000] <annevk> jgraham: the protocol handler stuff has substitution [14:30:39.0000] <gavinc> Hixie: Right, so if it can read the sizes from an attribute, why can't it read the sizes AND the URL from an attribute? [14:30:52.0000] <annevk> I should read everything first, all my questions appear addressed [14:30:53.0000] <gavinc> Hixie: What does the template gain? [14:31:01.0000] <annevk> or maybe just continue playing Portal 2 [14:31:49.0000] <benvie> Thanks. I've included the resulting IDL in a thing I made and included references to the source for all IDLs included, so I wanted to make sure I was referencing the right thing. Resulting JSON is here https://github.com/Benvie/idl-for-javascript/blob/master/json/html5.json [14:33:30.0000] <annevk> gavinc: you already reference HTML5, HTML5 references DOM4, problem solved [14:34:26.0000] <gavinc> annevk: Yes, I shall make this argument some more :D [14:34:49.0000] <gavinc> but yay no more required XML-C14N! [14:35:29.0000] <annevk> now you just need to get rid of that RDF reference and it might start to look sane :p [14:37:04.0000] <a-ja> Hixie: i could see using menu....but it wouldn't have same flexibility (positioning/fullscreen/background) that dialog has. that case can actually be done w/o script now (cept for aria) [14:37:34.0000] <a-ja> with overlays [14:37:49.0000] <Hixie> gavinc: some people want urls like just banner-200 banner-640 banner-2800 banner-4960 [14:38:02.0000] <Hixie> gavinc: others want urls like banner.jpeg?w=200&h=400 [14:38:13.0000] <Hixie> gavinc: others want urls like banner⊙2 [14:38:14.0000] <gavinc> Hixie: ... yes... I'm saying PUT the WHOLE URL in the attribute [14:38:33.0000] <Hixie> gavinc: oh and repeat the common part over and over? [14:38:48.0000] <Hixie> gavinc: and repeat the dimensions twice, once for the url and once for the value for the browser to know what it is? [14:39:03.0000] <gavinc> Hixie: src-600x200x2="mine_600_200_2.jpg" [14:39:21.0000] <gavinc> Hixie: Yep, 'cause what if there isn't a common part? [14:39:30.0000] <Hixie> gavinc: that seems really verbose [14:39:35.0000] <gavinc> Hixie: or it's just 600x200x2="larger.jpg" [14:40:20.0000] <annevk> Hixie: do we need anything more than x1 and x2 in practice? [14:40:24.0000] <gavinc> Hixie: src-600x200x2="higher-res.png", 200x100x1="tiny.gif" [14:40:33.0000] <Hixie> annevk: different widths [14:40:55.0000] <gavinc> Hixie: If it's templated you have to use the same image format for each size [14:40:55.0000] <annevk> Hixie: hmm [14:41:08.0000] <Hixie> gavinc: seems better to push the people using those kinds of filenames towards filenames that use a pattern, than force authors who do use a pattern to not gain anything from doing so [14:41:18.0000] <Hixie> gavinc: why? [14:41:58.0000] <gavinc> Hixie: well, or you have to not use file extentions, which people seem rather loth to give up [14:42:21.0000] <Hixie> gavinc: or they can just use whatever file extensions they like, the browsers ignore them anyway :-P [14:42:42.0000] <gavinc> Hixie: Yes, but that way lies crazy [14:42:51.0000] <annevk> at this point not really [14:43:02.0000] <annevk> determining the image type from the file signature is extremely reliable [14:43:07.0000] <gavinc> Hixie: Saving a .png to my desktop that was an image/jpeg is a bit funky [14:43:10.0000] <annevk> and way better than using Content-Type [14:43:43.0000] <gavinc> Hixie: Just saying, there isn't a NEED for templating. Can do exactly the same thing by just requiring the full URL [14:45:11.0000] <llrcombs> so, is CORS an official standard now? [14:45:18.0000] <llrcombs> The xkcd.com sysadmin wants to know [14:45:47.0000] <llrcombs> it was a last call draft expiring on may 1st; I dunno if there were complaints or not [14:46:04.0000] <othermaciej> that depends on what counts as an "official standard" [14:46:12.0000] <othermaciej> it is implemented in many browsers [14:46:21.0000] <othermaciej> it is officially on the W3C Recommendation track [14:46:24.0000] <llrcombs> accepted as standard by W3C, I think [14:46:31.0000] <othermaciej> it has not yet reached the final "Recommendation" state [14:46:31.0000] <annevk> xkcd uses XHTML 1.1 [14:46:41.0000] <annevk> sent as text/html [14:46:44.0000] <othermaciej> on the other hand, nearly nothing you actually use is a W3C Recommendation [14:46:47.0000] <annevk> doesn't seem they care much about standards :p [14:47:06.0000] <llrcombs> othermaciej: I'm not the sysadmin, I'm asking because he wants to know on another server [14:47:33.0000] <annevk> llrcombs: CORS is implemented by all browsers, big sites use, it's not going to change [14:47:43.0000] <annevk> llrcombs: xkcd.com using it would further cement it [14:48:05.0000] <Hixie> gavinc: sure, there's many ways we can make this more verbose and annoying :-P [14:48:09.0000] <annevk> llrcombs: as far as W3C games go, it might become a Candidate Recommendation soonish, depending on how soon I get around to making some edits [14:48:24.0000] <Hixie> gavinc: some of the proposals on the thread even involve multiple new elements and fallback hierarchies [14:50:09.0000] <othermaciej> llrcombs: well, I gave you all the relevant info I have, I am not sure what you or he would consider to be an official standard though [14:50:49.0000] <llrcombs> othermaciej: well, I'll accept it as standard at this point; not sure if he will. I've passed on the info, though [14:50:53.0000] <othermaciej> llrcombs: if his real question is "should I use it", then the answer is "totally yes, it's in multiple browsers and highly interoperable" [14:51:19.0000] <othermaciej> the chances of it being removed from the standards track or changed incompatibly are very low [14:51:51.0000] <annevk> Hixie: http://lists.w3.org/Archives/Public/www-style/2012Feb/1103.html is what's proposed for CSS which is much simpler [14:52:08.0000] <annevk> Hixie: it doesn't address your width use case and dunno if the CSS WG bikeshedded it into something more complex [14:52:47.0000] <annevk> llrcombs: is he also in charge of the HTML? because sending XHTML as text/html is doubtful as far as following standards goes [14:53:09.0000] <llrcombs> annevk: he says he does that because it's better supported by some browsers [14:53:16.0000] <llrcombs> not sure how [14:53:19.0000] <llrcombs> lemme just copy [14:53:22.0000] <annevk> [citation needed] [14:53:47.0000] <annevk> anyway, I don't care, awesome comics are readable either way [14:53:49.0000] <llrcombs> <davean> llrcombs: the standard specifies sending XHTML as text/html is acceptable <davean> llrcombs: And it works better and we should be satisfying the compatability rules [14:55:43.0000] <annevk> http://www.w3.org/TR/xhtml-media-types/#text-html [14:56:23.0000] <annevk> I guess if you want to use XHTML and not HTML and still somehow work with IE you do want text/html indeed [14:56:39.0000] <webben> XHTML 1.1 as text/html? 2009 wants its boring debates back. [14:56:48.0000] <annevk> webben: more like 2004 [14:56:52.0000] <llrcombs> I do believe it satisfies appendix A, but I'm not sure. If you really care, ask at #xkcd on foonetic [14:56:57.0000] <othermaciej> Hixie: I think an <img> attribute that takes similar syntax to the proposed CSS image-set would be a fine solution [14:57:04.0000] <gavinc> and make it Ployglot, 'cause making it Ployglot is so much fun [14:57:37.0000] <annevk> when <template> is there XML is over [14:57:39.0000] <othermaciej> Hixie: it can be in addition to src, so fallback works [14:58:04.0000] <othermaciej> Hixie: and presumably the same thing for any other content attributes that reference an image, not sure what the full set of those is (at least <video poster> I guess) [14:58:07.0000] <annevk> assuming <template> is going to be widely adopted and used [14:58:36.0000] <llrcombs> what's <template>? [14:58:40.0000] <annevk> othermaciej: <link rel=icon> [14:58:57.0000] <annevk> othermaciej: oh wait that already has a syntax [14:59:14.0000] <gavinc> othermaciej: I was unaware of image-set! Yes! That on <img>! That's clearly what I meant ;) [14:59:21.0000] <othermaciej> annevk: that has a syntax to represent different sizes, and I suspect scale as separate from size is irrelevant [14:59:53.0000] <annevk> llrcombs: feature primarily for "web components" [15:00:48.0000] <othermaciej> annevk: (because icons are most likely rendered at a visible size chosen by the UA independent of intended scales of the underlying images) [15:01:07.0000] <llrcombs> link to spec? [15:01:58.0000] <annevk> llrcombs: http://dvcs.w3.org/hg/webcomponents/raw-file/tip/explainer/index.html [15:02:15.0000] <othermaciej> Hixie: anyway you should definitely consider <http://lists.w3.org/Archives/Public/www-style/2012Feb/1103.html> as part of input for this, it even suggests an HTML syntax, though we have only implemented the proposed CSS syntax in WebKit so far [15:02:22.0000] <llrcombs> that's helpful. I'll read up later. Thanks! [15:35:32.0000] <Hixie> othermaciej: image-set? [15:36:12.0000] <Hixie> othermaciej: oh the 1x 2x thing? [15:36:15.0000] <othermaciej> Hixie: I'm not sure what you are asking about image-set, but my answer is probably "yes" (assuming it is a yes/no question) [15:36:32.0000] <Hixie> othermaciej: my question was "what is image-set", but your later link cleared that up [15:36:40.0000] <Hixie> othermaciej: people also want to be able to handle different available widths and heights [15:36:44.0000] <Hixie> widths mainly [15:36:53.0000] <Hixie> e.g. banner on narrow window vs banner on fullscreen tablet [15:38:06.0000] <othermaciej> Hixie: I see - not sure how to handle that one [15:38:27.0000] <othermaciej> Hixie: CSS media queries can handle things like screen size of course, but don't apply to content images [15:38:36.0000] <othermaciej> replicating the full CSS media query syntax would be a bother [15:39:24.0000] <Hixie> othermaciej: yeah [15:39:27.0000] <othermaciej> I suspect people serving different image widths for different screen widths often want to do additional adaptation to the layout, so a technique for dealing with just the images may have less value [15:39:38.0000] <Hixie> othermaciej: what i'm proposing (See junkyard link earlier) is close to what image-set is doing [15:39:52.0000] <Hixie> well, the layout can be adapted in css [15:40:03.0000] <Hixie> the point is there are some content-level images that would also need to be adapted [15:40:09.0000] <Hixie> and it'd be sad to have to have the css replace the content-level images [15:41:12.0000] <othermaciej> Hixie: forcing a fixed naming convention seems like a significant limitation with little upside [15:41:29.0000] <othermaciej> Hixie: images for different resolutions might not even be served from the same host [15:41:53.0000] <Hixie> that seems unlikely [15:42:18.0000] <othermaciej> Hixie: for the width thing, I guess you can set the layout width of the <img> with a CSS media query and then use your algorithm to pick the source, otherwise it seems unlikely to be helpful [15:42:57.0000] <othermaciej> Hixie: I suspect Apple Web properties will in many cases want to vary by device pixel ratio but not by available screen width, does your proposal provide for that? [15:43:24.0000] <Hixie> yes [15:43:26.0000] <Hixie> see link above [15:43:35.0000] <othermaciej> I looked at http://junkyard.damowmow.com/506 and it's not obvious [15:43:53.0000] <Hixie> it uses width, height, and pixel density [15:43:55.0000] <othermaciej> do you have to specify the width and height and put it in the image name even if you don't want to vary on it? [15:44:19.0000] <Hixie> no, you could just use foo@%r.jpeg to use foo⊙1 or foo⊙2 [15:44:34.0000] <othermaciej> and what do you put in src-versions? [15:44:41.0000] <Hixie> i guess you do have to specify a dimension in the sec-versions attributes [15:44:43.0000] <Hixie> src-versions [15:45:26.0000] <othermaciej> I think authors for the "resolution adaptation only" use case will not want to mention dimensions in the markup [15:45:33.0000] <Hixie> yeah, that's valid [15:45:56.0000] <Hixie> in many of hte examples i saw, there were many images with filenames that had all the same data as the dimensions [15:46:06.0000] <Hixie> it seems really sucky to require those authors to duplicate all that data [15:46:11.0000] <othermaciej> and I think that use case will be more common than the width adaptation use case [15:46:29.0000] <Hixie> i think they're both common enough to be addressed directly [15:47:13.0000] <othermaciej> I might have a non-representative sample of developer requests [15:47:43.0000] <othermaciej> but I have heard lots of requests for resolution adaptation and none for width adaptation (on an image-by-image basis) [15:47:49.0000] <othermaciej> not even on the CSS side [15:48:05.0000] <othermaciej> but my point was only that width adaptation should not make resolution adaptation alone more complicated [15:48:10.0000] <othermaciej> (more complicated to use) [15:49:14.0000] <othermaciej> we do have some limited testing/deployment experience with image-set in CSS and authors seem to like it [15:49:38.0000] <othermaciej> in some cases authors objected to a fixed filename pattern solution, though I don't know that we ever proposed an adaptable name template as you have it [15:50:18.0000] <othermaciej> in current deployment I think there are cases where the 1x name is undecorated and the 2x name has @2x appended before the extension [15:52:50.0000] <Hixie> yeah [15:53:09.0000] <Hixie> i'd love to support a way to support separate names while still not requiring duplication [15:55:59.0000] <othermaciej> I think you'd need to support author choice of either a list or the template thing to support both [15:56:12.0000] <othermaciej> I can't think of an obvious way to unify other than that [16:01:03.0000] <Hixie> yeah [16:01:05.0000] <Hixie> me either [16:01:08.0000] <Hixie> doing both seems lame [16:02:20.0000] <gavinc> the list covers all the uses cases but requires some duplication, the template doesn't support as many use cases [16:03:30.0000] <Hixie> right [16:03:43.0000] <Hixie> supporting more use cases isn't always a win if the result is less usable overall [16:04:02.0000] <Hixie> often the key to language design is figuring out which use cases to forsake [16:05:42.0000] <annevk> the only thing I have seen suggested thus far is resolution [16:06:02.0000] <Hixie> a lot of the e-mails on whatwg discussed image dimensions [16:06:10.0000] <annevk> oh, I guess some have suggested width as well, but I'm not sure that's quite as valid [16:06:10.0000] <Hixie> in fact that was discussed long before pixel density [16:06:26.0000] <gavinc> annevk: I mean for example one image on one server and the other image on another, the template method can't support that at all [16:06:29.0000] <Hixie> (the concern being a page that displays on a tiny mobile display and a huge desktop display) [16:07:27.0000] <annevk> tiny mobile display does not really exist anymore though [16:07:44.0000] <Hixie> o_O [16:08:29.0000] <Hixie> i can fit about 45 phone screens the size of my phone in the space of my desktop's screen :-P [16:08:44.0000] <annevk> most desktop sites render fine on my phone [16:09:04.0000] <annevk> and pixel wise they are not too different, although I guess that might change [16:10:18.0000] <gavinc> annevk: err, at css pixal size they (read iPhone) are still small [16:11:29.0000] <annevk> sure, but you also hold it closer to you, so it sort of works out, especially combined with zooming [16:13:20.0000] <annevk> it would be interesting to know how many sites try to accomplish the different image for mobile and desktop but the rest remains the same today [16:13:33.0000] <annevk> because it seems a rather obscure case [16:14:19.0000] <Hixie> dude if you send my 1280x720 phone a 2560x1440 picture, you're wasting a lot of pixels [16:14:38.0000] <annevk> whereas providing high resolution images without messing with bandwidth usage for lower resolution devices seems like something everyone wants to do [16:14:42.0000] <Hixie> but if you send my 2560x1440 desktop a 1280x720 picture, you're gonna find it doesn't fit my screen [16:15:13.0000] <Hixie> (both are 96dpi-equivalent displays) [16:17:14.0000] <annevk> that's what often ends up happening when browsing reddit and it doesn't matter that much :) [16:19:17.0000] <Hixie> a lot of web authors seem to disagree about it not mattering [16:21:35.0000] <Hixie> hober: (did you ever mention http://lists.w3.org/Archives/Public/www-style/2012Feb/1103.html on whatwg btw?) [16:26:40.0000] <Hixie> anyone know off-hand what ascii punctuation characters aren't allowed anywhere in a URL? [16:29:14.0000] <Hixie> damnit, commas are valid in URLs [16:33:48.0000] <Hixie> http://junkyard.damowmow.com/507 [16:34:22.0000] <Hixie> othermaciej, annevk, gavinc, hober: ^ [16:35:05.0000] <annevk> why do you pick the widest if they're all too wide and not the least wide? [16:35:42.0000] <Hixie> oversight [16:36:10.0000] <othermaciej> Hixie: seems good in general, though I'm not totally sure all details of the algorithm are right (why drop ones w/ no width when sorting by width?) [16:36:19.0000] <annevk> also, if you split on comma first, you can allow omitting everything [16:36:28.0000] <Hixie> othermaciej: i don't know what to do with the other ones [16:36:34.0000] <Hixie> annevk: can't split on comma, urls can have commas [16:36:43.0000] <Hixie> annevk: so have to parse more carefully [16:37:47.0000] <annevk> "," is a reserved character afaik [16:38:06.0000] <gavinc> yeah, "," is reserved [16:38:35.0000] <Hixie> it's a sub-delim, allowed in various places in urls [16:38:49.0000] <Hixie> e.g. anywhere in the path [16:38:51.0000] <annevk> yeah sorry [16:38:58.0000] <gavinc> mmm, yeah and a complex grammar to find them likely isn't worth while [16:39:33.0000] <Hixie> easy enough to walk the string bit by bit [16:39:34.0000] <annevk> but just like spaces sometimes have to be escaped, you could require it to be escaped here, but more complicated parsing works too I guess [16:39:46.0000] <annevk> I wonder how the image-set parsing is defined [16:39:50.0000] <Hixie> url(...) [16:40:00.0000] <annevk> oh right [16:40:28.0000] <gavinc> could require , [16:40:30.0000] <gavinc> err [16:40:47.0000] <gavinc> comma followed by whitespace [16:40:49.0000] <gavinc> but nah [16:41:07.0000] <annevk> Hixie: overall looks much better than the previous proposal [16:41:30.0000] <Hixie> this is basically hober's proposal slightly extended [16:41:38.0000] <Hixie> i can't claim credit :-) [16:42:23.0000] <annevk> hober's asleep, if you act quick the internet will never know [16:42:28.0000] <Hixie> hah [16:42:34.0000] <annevk> speaking of which, nn everyone [16:42:44.0000] <Hixie> nn [16:49:58.0000] <Hixie> ok i guess i'll spec this out tomorrow [16:50:04.0000] <Hixie> need to speak to hober about naming of the attribute [16:50:08.0000] <Hixie> set="" is kinda lame [16:53:20.0000] <Hixie> bbiab. if anyone wants to bikeshed the set="" attribute name post it here 2012-05-10 [18:20:56.0000] <gavinc> yay, DOM4 HTML datatype for RDF preposed http://lists.w3.org/Archives/Public/public-rdf-wg/2012May/0222.html [18:22:43.0000] <gavinc> proposed too [20:06:45.0000] <MikeSmith> Hixie: w3c systems team tells me that they're having trouble getting hypermail to recognize the format of the online whatwg archives [20:06:55.0000] <MikeSmith> perhaps due to missing Received headers [20:06:56.0000] <Hixie> weird [20:06:59.0000] <Hixie> ah maybe [20:07:05.0000] <Hixie> can't they fake them or something? [20:07:07.0000] <MikeSmith> they are wondering if you have the original mobx [20:07:10.0000] <MikeSmith> maybe [20:07:12.0000] <Hixie> i have nothing [20:07:16.0000] <MikeSmith> OK [20:07:33.0000] <Hixie> sorry :-( [20:08:03.0000] <MikeSmith> no problem [20:08:10.0000] <MikeSmith> I'm sure we'll get it figured out [20:32:50.0000] <kennyluck> Hixie, why don't you just make a mbox out of your whatwg folder? [20:32:51.0000] <kennyluck> (Or whoever has all the mails) [20:32:52.0000] <kennyluck> Hixie, why don't you just make a mbox out of your whatwg folder? [20:32:52.0000] <kennyluck> (Or whoever has all the mails) [20:33:02.0000] <Hixie> i certainly don't have all the e-mails [20:33:08.0000] <Hixie> i delete them once i've replied to them :-) [20:33:13.0000] <Hixie> dunno if anyone else does [20:33:45.0000] <othermaciej> Hixie: I agree that "set" is not the best attribute name [20:34:13.0000] <othermaciej> I was thinking maybe src-set or srcset or something if the parallel to image-set is actually worthwhile [20:34:29.0000] <Hixie> yeah [20:34:34.0000] <Hixie> srcset might work [21:46:08.0000] <othermaciej> hmm I guess this applies to <input type=image src> and <video poster>, is it also required for the case of using <object> to reference an image (e.g. to be able to have a structured textual alternative)? [21:47:57.0000] <othermaciej> also in theory <command icon> but who knows when/if anyone will implement <command> [22:05:01.0000] <tantek> yay for another microsyntax! [22:05:28.0000] <tantek> othermaciej, you mean <object data> ? [22:06:56.0000] <othermaciej> tantek: yes, <object data> in the specific case of using it to refer to an image (though I guess you don't even know that up front and it's not clear what scale factor would mean for plugin content or an HTML document) [22:10:18.0000] <Hixie> othermaciej: i dunno that i'd worry about those [22:12:31.0000] <othermaciej> it certainly seems useful for poster frames [22:12:38.0000] <othermaciej> there is no obvious alternative in that case [22:13:08.0000] <othermaciej> for <input type=image> I guess you could use a client-side image map instead as an arguably preferable alternative [22:13:12.0000] <tantek> othermaciej - why limit to images? [22:13:22.0000] <tantek> (for <object data> that is) [22:13:58.0000] <tantek> at different scales you might decide say a raster vs. a vector graphic may make more sense. [22:14:18.0000] <tantek> and if it's a raster graphic, like SVG, it might also be interactive, if it happens to be an interactive medium [22:14:36.0000] <othermaciej> tantek: the semantic is that it both selects based on device pixel ratio, and scales by that factor [22:14:57.0000] <othermaciej> I can see how you might use a vector image as one of a set, but <img> with an SVG source caters to that [22:15:14.0000] <othermaciej> I'm not sure how it would make sense to select one of several HTML files based on device scale [22:15:17.0000] <tantek> as Hixie was pointing out, sometimes bandwidth is a selection factor [22:15:27.0000] <tantek> which may also be true for a static image vs. an interactive image [22:15:37.0000] <tantek> where interactive image may be SVG, or may be an iframe [22:15:39.0000] <Hixie> othermaciej: for poster, shouldn't we just use the highest res image? i mean, the user might go fullscreen, etc [22:15:53.0000] <othermaciej> I am skeptical that bandwidth-based selection can be defined or implemented sanely [22:16:03.0000] <Hixie> yeah me to [22:16:03.0000] <Hixie> yeah me to [22:16:04.0000] <Hixie> o [22:17:10.0000] <tantek> heck, I'd even offer up network reliability as another axis [22:17:19.0000] <tantek> and both of those are real-world design problems [22:17:27.0000] <tantek> as anyone with AT&T in SF knows [22:17:41.0000] <tantek> or that goes to conferences with open/free wifi like SXSW [22:17:56.0000] <tantek> both are trivial examples of both bandwidth and reliability problems/challenges [22:18:30.0000] <othermaciej> Hixie: you could always use an over-res image, but then you have to set an explicit size when you might otherwise not need to; and also it might be wasteful of bandwidth for pages that embed many videos but expect the user not to play most [22:18:41.0000] <othermaciej> Hixie: I concede that it is a less common case than <img> [22:19:18.0000] <Hixie> yeah, worth keeping in mind, certainly [22:19:31.0000] <tantek> the one example of this kind of thing in the original image had nothing to do with dimensions or pixel density. it had to do with bandwidth. lowsrc [22:19:35.0000] <othermaciej> tantek: network bandwidth, latency and reliability are all useful things to know, but they are hard to define or compute in a sane way [22:19:36.0000] <Hixie> i'm a little reluctant to go all-in and define a solution for everything at once [22:19:37.0000] <tantek> original img [22:20:05.0000] <tantek> Hixie, I didn't claim to have or know of any kind of easy/simple solution. [22:20:27.0000] <othermaciej> tantek: the one case where there's an attempt to use them over the network is for adaptive video streaming, which intrinsically deals with the fact that these properties change unpredictably over time [22:20:28.0000] <Hixie> i was responding to othermaciej about poster=, data=, etc [22:21:20.0000] <othermaciej> if srcset works out it would not be a huge deal to later define <video posterset> or <input type=image srcset> [22:21:27.0000] <othermaciej> (or whatever the name ends up being [22:21:27.0000] <Hixie> yeah [22:21:28.0000] <othermaciej> (or whatever the name ends up being [22:21:28.0000] <Hixie> yeah [22:23:09.0000] <alystair> what's the core issue being discussed here? [22:23:21.0000] <Hixie> responsive web design, subsection images [22:25:22.0000] <alystair> Isn't responsive design mostly a solved CSS ordeal? Also sprites? [22:25:23.0000] <alystair> Isn't responsive design mostly a solved CSS ordeal? Also sprites? [22:26:02.0000] <tantek> alystair - two things. 1. except when images are content rather than just presentational, and 2. many/most of the CSS solutions fail to avoid downloading all variants anyway. [22:26:28.0000] <tantek> so no, responsive design is not really "solved". there are some hacks. [22:31:12.0000] <alystair> seems like something that could be solved at a level below html? eg. some sort of header being sent and server can shoot back other files? (if we are discussing the high density display issue) [22:36:35.0000] <tantek> alystair, the general pattern of "some sort of header being sent and server can shoot back other files" is AFAIK called CONNEG AKA "content negotiation" and has largely been decided to be a near complete failure. [22:37:02.0000] <tantek> Hixie, for the set attribute, why not simply re-use CSS style declaration syntax? [22:40:40.0000] <tantek> e.g. putting urls inside a url() function, and then using ; delimited property:value declarations [22:40:51.0000] <tantek> group with { } as necessary [22:53:04.0000] <othermaciej> tantek: I presume in part to align with http://lists.w3.org/Archives/Public/www-style/2012Feb/1103.html [22:55:43.0000] <tantek> othermaciej - note that Ted advocates url(foo-lowres.png) rather than just a naked foo-lowres.png [22:56:00.0000] <tantek> (in the CSS variant) [22:56:17.0000] <tantek> doing so in the set="" attribute variant makes parsing simpler for that also. [22:57:16.0000] <othermaciej> tantek: yeah, I think the CSS variant should let you drop url(), but point taken [23:01:43.0000] <tantek> it does feel like the "set" syntax is just a "light" syntax for writing mediaqueries [23:25:01.0000] <othermaciej> tantek: kind of, yes - in practice using media queries to do the image resolution thing turns out to be too awkward [23:25:11.0000] <tantek> indeed [23:25:17.0000] <othermaciej> at the CSS level it's just syntactic sugar [23:25:24.0000] <tantek> or vinnegar [00:02:45.0000] <niloy> firefox applying same origin policy for web fonts is very annoying, doesnt work well with CDN [00:41:16.0000] <odinho> yay for getting something speced like srcset. [00:42:15.0000] <zcorpan> srcset? [00:43:33.0000] <hober> zcorpan: <img srcset="foo.jpg 2x" src=bar.jpg> [00:43:54.0000] <odinho> zcorpan: http://junkyard.damowmow.com/507 [00:43:55.0000] <hober> speaking of which, Hixie: I don't remember if I mentioned it on whatwg [00:44:39.0000] <hober> Hixie: I'll post to whatwg next week (CSS F2F this week). [00:44:47.0000] <odinho> odinho == Velmont, btw, -- I found out I could just use this name from work and Velmont from home :P [00:44:58.0000] <Hixie> hober: don't bother, i'm about to reply to the relevant thread subsubing your mail [00:44:59.0000] <hober> Hixie: I don't think you need the w & h bits. [00:45:01.0000] <Hixie> subsumingh [00:45:04.0000] <Hixie> subsuming [00:45:12.0000] <hober> Hixie: otherwise this looks a lot like what i have written up for html [00:45:22.0000] <hober> I was going with srcset [00:45:24.0000] <Hixie> yeah, that's not a coincidence :-) [00:45:38.0000] <hober> for a while i had src-set, but hyphens are weird in html attr names [00:45:42.0000] <Hixie> yeah [00:45:52.0000] <Hixie> anyway i expect to spec that tomorrow [00:46:01.0000] <Hixie> in other news [00:46:17.0000] <Hixie> i thought position:absolute;left:0;right:0;margin:auto;width:auto; would shrink wrap [00:46:20.0000] <Hixie> am i wrong? [00:46:22.0000] <odinho> hober: earlier prop was different http://junkyard.damowmow.com/506 -- then someone said it should look more like your earlier proposal on csswg. [00:46:46.0000] <hober> Hixie: i'll have to think about the algorithm you have here, to see how it matches what i've been working on [00:47:26.0000] <hober> odinho: oh, jebus. src-template. ewwwwh. :) [00:47:36.0000] <Hixie> hey! [00:47:38.0000] <Hixie> it's not THAT bad [00:47:40.0000] <odinho> :P [00:47:41.0000] <Hixie> :-P [00:47:48.0000] <Hixie> jeez [00:48:00.0000] <hober> fwiw hyatt wanted to do something like that [00:48:08.0000] <hober> though i think his idea was less ugly :) [00:51:18.0000] <hober> tantek othermaciej: I think in CSS we should simply allow any <image> in the image-set syntax, which would let you have "foo.png" or url(foo.png) etc. etc., but only used url() in examples to keep things simple [00:57:32.0000] <hober> Hixie: i went ahead and sent my drafted-quite-a-while-ago proposal to the whatwg list. :) [00:58:26.0000] <Hixie> hober: k, i'll take a look. i hope you won't be offended if my reply doesn't reference it since i've already written it based on your www-style mail :-) [00:58:37.0000] <hober> yeah, no problem. :) [01:00:31.0000] <hober> i guess i should get around to writing the epic "why <picture> is a terrible idea" email [01:01:12.0000] <jgraham> Hixie: BTW did you answer my question about spinning the event loop? [01:01:20.0000] <odinho> Better preemtively buy some popcorn then. [01:01:27.0000] <Hixie> jgraham: yes [01:01:33.0000] <Hixie> jgraham: you didn't like my answer though [01:01:35.0000] <hober> odinho: :) [01:02:21.0000] <jgraham> Hixie: You were going to check on the followup about load events [01:02:40.0000] <Hixie> jgraham: oh, right. yeah, i looked it up, and i was correct in my description. [01:03:06.0000] <Hixie> jgraham: the way the spec is phrased, as soon as there's nothing left blocking, the task is queued to continue on [01:03:17.0000] <Hixie> jgraham: even if something else blocking is immediately introduced [01:07:55.0000] <jgraham> OK, so you can't implement "spin the event loop until /condition/" with a list of functions that run each time the event loop is run, they have to be run… well actually I don't know what the smaller atomic unit iun the spec is [01:09:01.0000] <jgraham> s/implementation/hypothetical implementation/ [01:10:14.0000] <jgraham> Oh, I didn't say implementation [01:10:17.0000] <jgraham> Sigh [01:10:26.0000] <jgraham> So what is the atomic unit? [01:13:33.0000] <Hixie> if any of the definitions rely on anything being atomic, and the thing in question isn't described as happening "atomically", they're broken [01:14:09.0000] <Hixie> short of that, the atomic unit is something like CPU cycle. [01:14:12.0000] <jgraham> Hixie: I still don't understand then. Sorry if I am being thick :) [01:14:18.0000] <Hixie> planck time [01:14:54.0000] <jgraham> Well I can't check if /condition/ is true once per CPU cycle [01:15:11.0000] <Hixie> turns out, for all the conditions, you don't need to [01:15:14.0000] <Hixie> unless i made a mistake [01:15:40.0000] <jgraham> OK, so the concrete case I was looking at was stuff blocking the load event [01:15:59.0000] <Hixie> right [01:16:15.0000] <Hixie> so you can implement that using a counter [01:16:21.0000] <Hixie> and whenever you change it to zero, set a flag [01:16:56.0000] <jgraham> OK [01:17:16.0000] <jgraham> Well if that's how it's supposed to work, it is possible to make it work that way [01:17:49.0000] <jgraham> But it wasn't how I imagined the model from what it says in the spec [01:17:51.0000] <Hixie> yeah, that's a different question [01:17:56.0000] <Hixie> it's what the spec says [01:18:02.0000] <Hixie> whether the spec is right, i dunno [01:18:25.0000] <Hixie> you shouldn't imagine things, just read what the spec says :-P [01:18:43.0000] <jgraham> The spec isn't always 100% clear :p [01:18:50.0000] <Hixie> in other news, shimming <form method=dialog> is surprisingly harder than it looks [01:19:03.0000] <Hixie> jgraham: file a bug or send mail if there's something i can clarify, always happy to try to do that [01:20:05.0000] <jgraham> Hixie: Sure. Thanks for being patient :) [01:20:18.0000] <Hixie> np, sorry for writing a crappy spec :-) [01:20:59.0000] <Hixie> we really still have no way for JS to access the form data set on submission? [01:21:08.0000] <Hixie> who's running this show, a clown? [01:23:23.0000] <jgraham> Heh [02:52:56.0000] <annevk> gavinc: btw, for canonical stuff, you can prolly use DOM's isEqualNode() [02:52:57.0000] <annevk> gavinc: btw, for canonical stuff, you can prolly use DOM's isEqualNode() [02:59:51.0000] <annevk> Hixie: people did ask about accessing the contents of FormData [03:00:00.0000] <annevk> Hixie: I wasn't quite sure whether that would be the correct solution [03:22:52.0000] <jgraham> hsivonen_: Do you know anything about how gecko decides whether to block onload for CSS resources? [03:54:27.0000] <hober> annevk: you might enjoy http://lists.w3.org/Archives/Public/www-style/2012May/0409.html [03:56:13.0000] <annevk> hober: it's kind of unbelievable it even has to be discussed given how many people implemented it and how often it's used already [03:57:31.0000] <hober> i know, right? [03:57:36.0000] <hober> it's *crazy* [04:00:10.0000] <annevk> and the alternative is not exactly easier to author [04:00:48.0000] <annevk> (understatement) [04:36:36.0000] <odinho> Meh, you got me into the black hole of reading www-style, -- reading vendor prefixes discussion. Quite interesting but not really what I was doing. [04:37:24.0000] <michel_v> save yourself, mammal [04:38:38.0000] <odinho> I really like many of the responses, -- but I'm mostly clicking Florian's links, which I know will be good - and also reading Henri Sivonen's now. [04:39:30.0000] <annevk> reading www-style is a good way to worsen one's jetlag [04:39:40.0000] <odinho> annevk :| [04:44:07.0000] <annevk> I am curious though how that prefix discussion went [04:44:29.0000] <hober> it hasn't happened yet [04:44:31.0000] <hober> in person anyway [04:44:41.0000] <hober> the most recent thread was actually pretty good [04:44:49.0000] <odinho> Ahh. Guess we'll hear it first on w3cmemes [04:45:11.0000] <hober> i blame othermaciej for the improved light-to-heat ratio on the subject [04:45:59.0000] <annevk> :) [04:46:23.0000] <odinho> hober: The thread florian started? [04:59:55.0000] <hober> odinho: yeah [05:01:06.0000] <odinho> hober: Yea, that was what I was reading. Seems to have gotten quite much support. Which is great. [05:02:08.0000] <odinho> Anywayz. So, DOMString takes \0 with no problem, -- but doesn't really print it out in console etc. -- But I guess this is perfectly legal then: IDBObjectStore.createIndex("a\0b") ? Since it takes DOMString as argument (name of the index). [05:04:35.0000] <gsnedders> odinho: Yeah, unless the spec special-cases it. [05:05:06.0000] <zcorpan> AryehGregor: <link sizes> has "any" for vector images [05:05:29.0000] <odinho> gsnedders: OK, just wanted to double check so that I don't allow crazystuff that shouldn't be allowed. [05:06:10.0000] <AryehGregor> zcorpan, why would you bother specifying more than one size at all if a vector size is available? Just for fallback? I ignore the case where you want significantly different (not just rescaled) images for different resolutions, although that's a real scenario -- for larger icons you might want more detail. [05:06:29.0000] <smaug____> annevk: what is the latest spec for appcache [05:07:40.0000] <smaug____> I guess http://www.whatwg.org/specs/web-apps/current-work/multipage/offline.html#offline [05:09:33.0000] <zcorpan> AryehGregor: dunno [05:27:51.0000] <annevk> smaug____: yup [06:02:55.0000] <jgraham> /me wishes for a single-page CSS spec [06:07:16.0000] <hober> jgraham: http://dev.w3.org/csswg/css-style-attr/ is probably as close as you're going to get [06:17:48.0000] <Velmont> lol, I'm credited as you in the webapps minutes, hober. [06:18:48.0000] <odinho> Hmm. Not so easy having two irc's open I see. I write in the wrong window :P [06:19:29.0000] <hober> Velmont: link? [06:21:39.0000] <odinho> hober: http://www.w3.org/2012/05/02-webapps-minutes.html#item10 (I'm Velmont from work) [06:21:58.0000] <odinho> hober: Stuff you said there was not said by you, but by me ;-) [06:22:29.0000] <odinho> timeless: So ... should probably update the minutes there. Don't know how to do that. [06:23:37.0000] <hober> several of the things attributed to me should be, though [06:23:47.0000] <hober> you should be more specific when you email the request to update the minutes [06:23:49.0000] <odinho> hober: Yes, just in the IndexedDB place :-) [06:24:11.0000] <hober> i would be very surprised if i got minuted saying anything about indexeddb :) [06:25:58.0000] <jgraham> Velmont: You could just be odinho here too. That would help my little brain :) [06:34:33.0000] <odinho> jgraham: I mean, -- Velmont is my private user. -- odinho is work user. I just shouldn't use the Velmont user in the #whatwg channel, it'll be less confusing :P [06:34:59.0000] <AryehGregor> Live DOM Viewer is down for me. :( [06:35:47.0000] <jgraham> AryehGregor: WFM [06:36:05.0000] <AryehGregor> Error 324 (net::ERR_EMPTY_RESPONSE): The server closed the connection without sending any data. [06:37:19.0000] <jgraham> Oh, maybe it loaded from a cache for me [06:37:24.0000] <jgraham> It's down now [07:05:05.0000] <hsivonen_> jgraham: I know nothing about that [07:11:34.0000] <lkjhl> hi [07:11:42.0000] <lkjhl> what this is about ? [07:14:42.0000] <lkjhl> annevk hi [07:17:09.0000] <limbu> heeeellllllllloooooooooo [07:17:53.0000] <zcorpan> limbu: see http://whatwg.org/ [07:30:41.0000] <limbu> okay got it [07:30:47.0000] <limbu> your a web compnay [07:32:53.0000] <hober> nope [07:33:10.0000] <odinho> He left anyway. [08:07:01.0000] <Wilto> Oh what hello did someone say "images with disparate sources?" [08:07:54.0000] <hober> Wilto: yeah [08:08:19.0000] <Wilto> Mentioning images three times summons me like a marginally more obnoxious version of Beetlejuice. [08:08:26.0000] <Wilto> How goes it, hober? [08:09:25.0000] <hober> good good. survived the hour-long css3 grid layout discussion. now, autogeneration of super/subscript glyphs in css3 fonts. [08:09:30.0000] <hober> you? [08:09:51.0000] <Wilto> That sounds like an action packed, thrill-a-minute kinda day. [08:10:26.0000] <hober> you know it! [08:10:47.0000] <Wilto> Not bad, man, not bad. Psyched to see the respimg topic being broached; about to have my Boston-themed shirt design go up on unitedpixelworkers.com [08:11:11.0000] <hober> o rly [08:11:18.0000] <Wilto> aww yiss [08:11:22.0000] <hober> /me is from the Boston area [08:11:53.0000] <Wilto> I know, man—I seen your hat on Twitter. [08:11:57.0000] <Wilto> http://wil.to/bos-pxl.png Shh. [08:12:06.0000] <Wilto> I am already way off topic in here. [08:12:21.0000] <hober> ooooh, nice. /me might have to pick that up. [08:12:41.0000] <hober> /me needs another t-shirt like he needs a hole in the head [08:13:37.0000] <Wilto> I hear you; I’m already swimming in Pixelworkers stuff. [08:13:44.0000] <Wilto> Oh, so, original question: [08:14:24.0000] <Wilto> I’m gonna shoot an email to the respimg CG and ask them to join in the discussion on the list. Any qualms from anybody? [08:16:40.0000] <hober> if you think that would be useful [08:17:38.0000] <tabatkins> Wilto: like I mentioned on the list, the RespImg proposal is really *not* meant for the use-case that @srcset is solving. [08:17:56.0000] <Wilto> That’s why I wanted to ask first, yeah. [08:18:20.0000] <Wilto> A subset of the CG’s discussion has been around the resolution problem specifically—I didn’t know if getting those folks involved might be helpful. [08:21:46.0000] <tabatkins> It might be useful to discuss what sort of decisions you can make with the explicit resolution information. [08:21:53.0000] <tabatkins> The obvious one is retina screens. [08:22:16.0000] <tabatkins> And the second obvious one is deciding differently based on your knowledge of recent bandwidth. [08:22:41.0000] <Wilto> Right. I mean, we’ve all got the shiny new iPads front-of-mind. [08:23:14.0000] <Wilto> There’s been a lot — a _lot_ — of discussion around the idea of giving users the ability to opt out of high-resolution images regardless of bandwidth, but that’s more a subject for UAs. [08:28:00.0000] <jgraham> 10:13 < Wilto> I am already way off topic in here <-- I see you din't read the topic closely enough [08:28:37.0000] <Wilto> The "sense of logic" part? I wouldn't know logic if it bit me, man. [08:53:46.0000] <tabatkins> Wilto: I don't think that (users opting out of high-res) is something that we need to worry about from a spec perspective at all. It's a UA thing. [08:54:07.0000] <Wilto> tabatkins: Yeah, with you there. [08:54:10.0000] <hober> yup [09:05:37.0000] <rafaelw_> hsivonen: you around? any chance we can attempt to settle on a solution for DocumentFragment.innerHTML? [09:06:15.0000] <annevk> I don't think we need MQs in markup though if we get srcset [09:06:25.0000] <annevk> or the <picture> design or whatever [09:12:47.0000] <Wilto> annevk: I mean, if srcset is limited to DPI we would. [09:14:17.0000] <Wilto> I think srcset as it’s being pitched now is a great solution, it’s just a solution to a very specific issue. It doesn't really solve the greater "responsive images" problem. [09:16:03.0000] <annevk> Hixie made it work for width/height too [09:16:13.0000] <annevk> he just hasn't posted it yet but it was discussed here yesterday [09:16:45.0000] <odinho> Wilto, here ya go: http://junkyard.damowmow.com/507 [09:16:59.0000] <Wilto> Thanks, odinho. [09:18:49.0000] <Wilto> Ah, yeah, we had something along these lines in one of our early drafts. [09:19:24.0000] <Wilto> I mean, at this point, I'm almost ready to say "please just give us any solution." [09:19:36.0000] <Wilto> But that looks like an absolute nightmare for authors. [09:19:50.0000] <odinho> http://junkyard.damowmow.com/506 hixies first draft [09:20:08.0000] <Wilto> Ah man. [09:20:22.0000] <Wilto> I just... was this just done in a vacuum? [09:21:05.0000] <odinho> Nah, don't think so. I've followed the respimg stuff, -- but it's too much talk and too little action ;-) Time to get it moving. [09:21:07.0000] <Wilto> This is a solution, and I absolutely don't fault anyone for throwing ideas out there, but... we've been talking about this stuff publicly for almost a year now. We've worked through a lot of this. [09:21:22.0000] <odinho> I really like hixies 507 proposition. [09:21:25.0000] <Wilto> There's the thing. I literally have no idea what we can do beyond https://github.com/Wilto/respimg [09:22:50.0000] <Wilto> It would be obnoxious if I kept bombarding you guys with this stuff. [09:23:20.0000] <Wilto> But speaking as an author, having worked through a nearly identical idea with a group of other authors, 507/508 would be just gross to work with. [09:23:58.0000] <Wilto> In what way is it better than <picture>, seeing as it breaks with every convention for specifying alternate sources already built into markup? [09:24:16.0000] <Wilto> Not a rhetorical question there; I can absolutely be convinced. [09:24:24.0000] <annevk> it's way simpler? [09:24:25.0000] <odinho> New tag is heavy stuff. [09:24:33.0000] <odinho> what annevk said :-) [09:24:36.0000] <annevk> and yeah, you really don't want new elements for images [09:24:38.0000] <Wilto> It’s more markup, but it’s _familiar_ markup. [09:24:45.0000] <Wilto> annevk: Why? [09:25:07.0000] <Wilto> How was this scheme decided on for video, and why would a variance in media markup make sense? [09:25:22.0000] <odinho> Wilto: We are not designing HTML from scratch. [09:25:40.0000] <Wilto> No, of course not. [09:25:49.0000] <odinho> Wilto: In that scheme, doing <picture> like that might in fact be the way it was done. -- But we're not there. [09:26:33.0000] <othermaciej> Wilto: <img> can't have children, and a new element with children specifying the sources would have a hard time with degrading gracefully [09:26:55.0000] <Wilto> othermaciej: It would degrade the same as <video>, <canvas>, etc. [09:27:06.0000] <odinho> othermaciej: I think the idea was that you'd have a <img> child in there that would be the default. [09:27:10.0000] <Wilto> Specifying fallback markup, ignored by browsers that do support it. [09:27:11.0000] <qubodup> hi, I was guided here on #html [09:27:23.0000] <othermaciej> Wilto: if you had a new element analogous to <video> (<picture> I guess?) would, in addition to source-type elements, need to have an <img> child for UAs that don't support it, then also a way to have a non-image textual equivalent [09:27:39.0000] <qubodup> I am wondering why a white space char appears inside an <a href> in one case but not in another http://ompldr.org/vZG96bg [09:27:39.0000] <Wilto> othermaciej: Yep, exactly. [09:27:40.0000] <odinho> But I so prefer the simplicity of srcset="logo-hd.png 2x" [09:27:43.0000] <othermaciej> so your minimum complexity for just doing the retina display 2x scale thing would be a lot higher [09:28:25.0000] <othermaciej> instead of <img src="foo.png" srcset="foo.png 1x, foo⊙2p 2x" alt="The world's greatest foo"> [09:28:57.0000] <Wilto> So this is more or less decided, then. [09:29:30.0000] <Wilto> Based on short-term conversation and a few quick drafts, in a total vacuum, independent of all the conversation that has happened in the Community Group. [09:29:33.0000] <odinho> Wilto: Heh, if we're having a discussion we need to actually discuss :-) [09:29:54.0000] <othermaciej> you would have to do <picture><source src="foo.png" media="max-device-pixel-ratio: 1"><source src="foo⊙2p" media="min-device-pixel-ratio: 2"><img src="foo.png>The world's greatest foo</picture> [09:30:36.0000] <othermaciej> (or whatever the exact syntax would be) [09:30:38.0000] <Wilto> othermaciej: I’m not saying it’s pretty, but at least it uses already established conventions. [09:31:04.0000] <othermaciej> <video> <source> also has a first-match selection algorithm [09:31:14.0000] <othermaciej> while srcset as proposed currently attempts to do best-batch [09:31:39.0000] <Wilto> I see no benefit of 507 over the new element for users, and despite being _less characters_ the syntax is going to be completely alien to developers. [09:31:39.0000] <annevk> Wilto: conventions are not everything [09:32:01.0000] <annevk> Wilto: simplicity counts [09:32:09.0000] <odinho> Wilto: I can't speak for others, but I've been following the CG. -- And seeing the proposals both hober and hixie came with they seem to have seen it as well. [09:32:15.0000] <Wilto> Brevity and simplicity are not the same thing. [09:32:29.0000] <Wilto> Where have they been in the conversations, then? [09:32:30.0000] <annevk> Wilto: there's a lot more parameters for video/audio unfortunately that make them more complex, though in retrospect I'm not sure <source> was such a good idea [09:32:36.0000] <othermaciej> the syntax is pretty similar to proposed CSS image-set, which is implemented in WebKit and has been used by developers without undue confusion [09:32:53.0000] <Wilto> I can’t help feeling like the CG was a little sandbox for the developers to play in while the decisions were made without us, seeing the progress presented here. [09:33:09.0000] <odinho> Yes, it's basically image-set ported to html. [09:33:23.0000] <annevk> Wilto: I didn't know there was a CG [09:33:26.0000] <paul_irish> just to provide context, he's corralled and led a group of the best mobile web developers, created a CG, isolated a solution (from many), fought for and won consensus within the group, wrote a draft spec and proposed it. Basically he's done the thing standards folks really want "authors" to do. Which is why this this feels so defeating. [09:33:30.0000] <othermaciej> I don't even know if there is a decision [09:33:40.0000] <othermaciej> I'm certainly not the decider [09:34:16.0000] <othermaciej> I just gave some input, and I'm explaining to you guys now what I think the tradeoffs among different options are [09:34:20.0000] <Wilto> Thanks, paul_irish. I’m tryin’, at least. [09:34:39.0000] <othermaciej> I did not know there was a CG for this specific topic, nor was I aware that it had produced a draft spec [09:34:41.0000] <annevk> paul_irish: not sure we'd want people to create their own groups, not sure there has been a decision either [09:34:48.0000] <othermaciej> pointers? [09:35:14.0000] <Wilto> othermaciej: Admittedly, this is my effort at winging a spec. I am no spec author. https://github.com/Wilto/respimg [09:35:21.0000] <Wilto> Just a way of getting all the details in one place. [09:35:51.0000] <Wilto> The community group is here. http://www.w3.org/community/respimg/ [09:36:22.0000] <Wilto> If you're in the mood for prose, some of the details leading to the group's creation and efforts are detailed at http://www.alistapart.com/articles/responsive-images-how-they-almost-worked-and-what-we-need/ and http://www.netmagazine.com/features/state-responsive-images [09:36:57.0000] <Wilto> Interest from the public crashed the CG database servers when the group was formed. [09:37:09.0000] <othermaciej> Wilto: it seems like a neat proposal, one thing I am not clear on is the specific algorithm to be used for selecting among the <source> elements [09:37:16.0000] <othermaciej> last one whose media query matches wins? [09:37:24.0000] <Wilto> If no one’s heard of my group -- considering how big my soapbox is -- it sure sucks to be your average developer with something to say. [09:38:00.0000] <othermaciej> was it announced to any of the existing groups working on standardization of HTML? [09:38:02.0000] <annevk> well it's pretty easy, submit proposals to whatwg⊙wo [09:38:06.0000] <divya> yes it was. [09:38:07.0000] <Wilto> I did. [09:38:10.0000] <annevk> alright [09:38:18.0000] <divya> annevk: it was submitted [09:38:21.0000] <odinho> Wilto: As I said, I've followed it after I saw it on whatwg. [09:38:24.0000] <annevk> then I'm pretty sure it was considered [09:38:28.0000] <divya> huh [09:38:34.0000] <divya> and then dismissed? [09:38:36.0000] <Wilto> Yeah. No accusations here, odinho. I’m just discouraged, here. [09:39:04.0000] <Wilto> annevk: An email would’ve been nice. [09:39:06.0000] <annevk> divya: well no, I assume Hixie will write one of his emails replying to everyone who gave input on the topic [09:39:22.0000] <odinho> AFAIK he started on that yesterday. [09:39:28.0000] <odinho> Wilto: The email is not yet sent. [09:39:37.0000] <othermaciej> was rasping mentioned on whatwg@ prior to May 10th? [09:39:42.0000] <othermaciej> er, "respimg" [09:39:47.0000] <othermaciej> I can't find any mention [09:39:48.0000] <annevk> Wilto: if you emailed the list you'll get one :) [09:40:08.0000] <Wilto> annevk: Hixie responded that CSS solved this problem, and then I didn’t hear from him further. [09:40:31.0000] <Wilto> Look, my tone isn't great here, and for that I genuinely apologize. [09:40:40.0000] <annevk> Wilto: I get the frustration :) [09:40:41.0000] <othermaciej> and I can't find any mentions at all on public-html [09:40:45.0000] <Wilto> But, man, this is kind of a beatdown. [09:40:46.0000] <odinho> hober heard that hixie was writing a respimg reply, -- and so he remembered he had an unsent email about it, and sent that. [09:40:53.0000] <divya> othermaciej: http://www.w3.org/Search/Mail/Public/search?type-index=public-html&index-type=t&keywords=picture+element [09:41:10.0000] <Wilto> Interestingly, othermaciej, it first came up in 2007 and a few times since. One sec; I have previous discussion documented. [09:41:21.0000] <divya> othermaciej: it is under 'prior art and discussion' [09:41:34.0000] <Wilto> Actually, yeah, there are a few links at https://github.com/Wilto/respimg#5-prior-discussion [09:42:01.0000] <othermaciej> none of those emails seem to mention the respimg community group [09:42:07.0000] <othermaciej> that's the bit I'm wondering about [09:42:18.0000] <annevk> Wilto: anyway, I suggest giving it another day / two days until Hixie posted his rationale [09:42:34.0000] <Wilto> This… seems fundamentally flawed. [09:42:39.0000] <othermaciej> you can't expect people to be aware of your CG's discussions if you didn't tell them about the CG [09:42:42.0000] <annevk> Wilto: and if you disagree with his arguments say so, this is not exactly set it stone [09:42:54.0000] <annevk> in* [09:43:24.0000] <Wilto> othermaciej: It was; I can dig through the list for mentions. I mostly documented discussion on the threads prior to the CG. [09:43:48.0000] <othermaciej> and if you want to make a spec that might get implemented in browsers, it's to your advantage to make sure at least some browser implementors participate in your group [09:43:59.0000] <Wilto> othermaciej: They have. [09:44:05.0000] <necolas> othermaciej: the respimg CG took down the whole server, such was the initial interest. i think people knew it was there [09:44:35.0000] <divya> othermaciej: i dont even know anymore. W3C says create community group, get partners in. [09:44:35.0000] <annevk> necolas: W3C systems team might have known, I never heard about that :) [09:44:37.0000] <divya> so it was done. [09:44:43.0000] <divya> so now what then. [09:44:48.0000] <necolas> annevk: that's surprising [09:44:49.0000] <divya> everything that was asked of developers was done [09:45:01.0000] <divya> especially with the superhuman effort of Wilto to put them all together. [09:45:02.0000] <necolas> annevk: considering the massive interest in the dev community [09:45:06.0000] <Wilto> Yeah. Like divya says: I wasn’t given a process. I was given a Wordpress blog and a hearty "good luck." [09:45:12.0000] <othermaciej> I can't find any mention of the string "respimg" in my public-html folder, and only one mention on May 10th (today) in my whatwg folder [09:45:45.0000] <divya> othermaciej: so we need to alert public-html with respimg? [09:45:58.0000] <Wilto> Yeah, is that what it comes down to? [09:46:11.0000] <othermaciej> well the fact that you didn't is probably the reason many folks working on the html spec itself have never heard of it before [09:46:19.0000] <Wilto> In fairness, at no point did I see it fit to say "our community group still exists." [09:46:24.0000] <othermaciej> if they had heard of it, they might have participated [09:46:27.0000] <divya> othermaciej: who *are* working on the spec? [09:46:40.0000] <divya> all we know is hixie [09:46:43.0000] <annevk> CGs work if you have all the relevant stakeholders there; just having a CG and a bunch of people participating does not necessarily create a standard [09:46:45.0000] <necolas> othermaciej: the problem is that the WG's dont engage the general developer community, even when they set up CG blogs for them [09:46:46.0000] <divya> how are we supposed to know who all to work? [09:46:51.0000] <divya> Who *are* the stackholders. [09:46:55.0000] <annevk> and if you want to change HTML, it goes via the WHATWG [09:47:00.0000] <Wilto> othermaciej: The point of the community groups is to connect developers with those involved standards process. [09:47:01.0000] <divya> this spec was shown to implementors in Opera, Google [09:47:02.0000] <Wilto> in the* [09:47:06.0000] <divya> and whoever showed interest. [09:47:19.0000] <divya> (i am using the word spec loosely) [09:47:35.0000] <Wilto> No arguments there, divya. It’s a... "Pseudospec." [09:47:47.0000] <Wilto> Just key implementor details in one semi-formal place. [09:47:48.0000] <divya> at NO POINT did anyone mention pinging whatwg about the existence of the CG [09:47:57.0000] <divya> or anything similar. [09:48:01.0000] <annevk> divya: not sure who you talked to then... [09:48:11.0000] <annevk> divya: but you can't extend HTML in a vacuum [09:48:27.0000] <divya> annevk: i sent it to the opera core mailing list [09:48:28.0000] <necolas> right back at ya :P [09:48:31.0000] <divya> zcorpan even commented [09:48:42.0000] <annevk> necolas: sure man, but 90 people is less than 1500 [09:48:45.0000] <Wilto> I sent it to Chrome’s standards list, as well. [09:48:50.0000] <necolas> there's obviously been a communication breakdown [09:48:59.0000] <divya> yeah lets just move on from this. [09:49:02.0000] <necolas> hopefully everyone can learn lessons from it [09:49:03.0000] <divya> the point is now what. [09:49:10.0000] <Wilto> Agreed. [09:49:21.0000] <annevk> I recommend emailing the proposal to the WHATWG list at least [09:49:22.0000] <necolas> because im sure most of the whatwg actually want to have happy, engaged developers involved [09:49:25.0000] <othermaciej> independent of the communications breakdown, it seems like there are two useful things: [09:49:42.0000] <necolas> and developers dont want to feel pushed out and told off for not doing things as they should be done [09:49:55.0000] <othermaciej> 1) send the proposal (not just a link to the CG) to whatwg@ [09:50:04.0000] <othermaciej> 2) send a list of use cases it's meant to address [09:50:31.0000] <othermaciej> 3) if Hixie puts something in the spec that does not meet all your use cases, point out which ones are not addressed [09:50:55.0000] <othermaciej> in general you will have a better time if you focus more on use cases and don't sweat it too much if the final syntax is not exactly what you wished for [09:51:26.0000] <annevk> necolas: agreed, if there's anything we can improve in http://wiki.whatwg.org/wiki/FAQ that would be useful to know [09:51:36.0000] <annevk> I have to go now unfortunately [09:51:41.0000] <othermaciej> I guess that was 3 things [09:51:49.0000] <Wilto> I genuinely don’t care about the pattern itself, beyond the fact that—of the many markup patterns we discussed—this is the one for which most developers involved voiced a preference. [09:51:58.0000] <Wilto> As long as the use cases are, as you said, met. [09:52:06.0000] <Wilto> 3 makes me very uneasy, however. [09:52:39.0000] <othermaciej> I can see that <picture> can in theory do lots of things that imgset can't, what I'm wondering about is which of those things are specifically important to web developers for particular use cases [09:52:49.0000] <divya> Wilto: i think the least we can do is #1 and #2 [09:52:53.0000] <Wilto> I assumed that, as someone who has spent a year discussing this with many developers and browser representatives, there would be a general overtone of “oh, good, welcome to the discussion—you’ll likely have insight here." [09:52:56.0000] <divya> lets worry about #3 when it comes to that. [09:53:13.0000] <Wilto> It sounds more like I have a chance to say "oh hey waitasec" after it has been specced. [09:53:19.0000] <Wilto> Guess so, yeah. [09:53:51.0000] <othermaciej> Wilto: you are definitely welcome to discuss! [09:53:55.0000] <tantek> othermaciej - perhaps every new CG created should be announced to whatwg? [09:54:03.0000] <othermaciej> I suggest that you put your list of use cases front and center [09:54:20.0000] <othermaciej> tantek: well, if it is planning to do something HTML-related then yes [09:54:27.0000] <Wilto> Again, guys, I genuinely apologize for my tone here. [09:54:50.0000] <Wilto> I've just owned this topic for a very long time. It's... Not un-stressful. [09:55:12.0000] <tantek> othermaciej - I think the problem with that is that CG are started to solve a problem / use-case, and that "do something HTML-related" is a specific type of solution that may or may not be clear from the outset. [09:55:33.0000] <tantek> hence my suggestion that might as well announce them all to whatwg, since they *might* do something HTML-related [09:55:57.0000] <necolas> Wilto: the dev community appreciates your massive efforts to bridge the gap [09:56:21.0000] <divya> well we certainly do Wilto. [09:56:28.0000] <Wilto> Thanks, guys. [09:57:42.0000] <tantek> /me reiterates his position that email lists tend to function more as support forums in practice than for any actual development (which has shifted in practice to IRC + wikis, or email just for snapshot announcements of proposals) [09:58:31.0000] <othermaciej> tantek: good point - but they might also do something Web API related or CSS related (or perhaps in rare cases SVG-related) [09:58:42.0000] <tantek> The whatwg list only seems to "work" because Hixie uses it as his personal inbox. [09:59:16.0000] <tantek> othermaciej - those groups seem more willing to accept outside consensus as a starting point for discussion than whatwg. [09:59:50.0000] <othermaciej> I think different people are comfortable with different communication modes [10:00:39.0000] <tantek> othermaciej - differences in communication comfort doesn't dispute my conclusion about email = support forum in practice. [10:00:57.0000] <shepazu> tantek: for reference, the SVG WG works via email and telcon more than IRC [10:01:16.0000] <tantek> cont'd re other groups vs. whatwg: whatwg appears have to have developed more of an NIH type culture (as evidenced by the responsive images example). [10:01:46.0000] <tantek> shepazu, I can see "has worked", but I'm not convinced that occurs much anymore. [10:02:12.0000] <tantek> even CSS has switched more to many things being quickly hashed out in IRC first [10:02:13.0000] <othermaciej> I don't know if this is an example of NIH, people were just genuinely other ware of the other proposal and other venue [10:02:31.0000] <othermaciej> might have sent the discussion in a different direction had they known [10:02:50.0000] <shepazu> tantek: I'm not trying to convince you, I'm telling you how the SVG WG works, and you can choose to believe otherwise if it fits your model better… Audio WG is the same way, btw [10:02:57.0000] <othermaciej> otoh whatwg seems to only consider proposals that have actually been implemented as particularly strong precedent, not just ones that have been discussed [10:03:05.0000] <tantek> othermaciej - but that doesn't matter, once made aware that there has already been a lot of work done, it's worth evaluating that rather than a few day brainstorm in progress. [10:03:31.0000] <tantek> othermaciej - that has the weakness of preferring browser implementers over developers / authorability. [10:03:51.0000] <tantek> which is an inversion of the the respective HTML design principles' community priorities. [10:04:25.0000] <othermaciej> tantek: I don't think anyone has told the proponents of the other proposal to go home or ignored their input, the most recent actual discussion about this was prior to knowing about their work [10:04:34.0000] <othermaciej> at least, that applies to any discussions I've been involved with [10:05:08.0000] <tantek> from Wilto's apparent frustration, it sounds like he feels like he is being effectively ignored. correct me if I'm wrong Wilto. [10:05:15.0000] <othermaciej> tantek: the design principle priorities are about who benefits, not who is most qualified to decide - just as asking users what they want isn't usually the best way to do HI design, asking clients what they want isn't necessarily the best way to do API design [10:05:23.0000] <Wilto> Speaking as a developer completely new to any "pitching standards" processes, I assumed there was an inherent connection between the CGs and those involved in this process. [10:05:25.0000] <tantek> othermaciej - sure [10:05:29.0000] <othermaciej> I think Wilto expected that everyone should have already known about his CG and taken it as precedent [10:05:44.0000] <Wilto> I didn't think it was necessary — or even appropriate — to point it out on several occasions. [10:06:19.0000] <tantek> Wilto, that being said, nothing is stopping you or the CG from bringing your proposal directly to browser implementers (e.g. filing bugs in their respective bug databases) [10:06:28.0000] <Wilto> tantek: I have, yeah. [10:06:28.0000] <tantek> if you think you have a solid consensus proposal [10:06:40.0000] <tantek> if the browser implementers implement it, then whatwg will likely spec it as implemented. [10:07:05.0000] <othermaciej> fwiw, at Apple we discussed the tradeoffs between a <picture>-like design vs the srcset-style design and thought the latter was better [10:07:10.0000] <tantek> so that's another path that's always available, you don't have to coordinate with whatwg (though you're likely to get a better spec/technology/adoption if you do) [10:07:21.0000] <tantek> othermaciej, good to know [10:07:40.0000] <othermaciej> in large part due to experience with media queries being too verbose for the related use cases in CSS [10:07:49.0000] <Wilto> I mean… I _want_ to coordinate with the WHATWG. [10:08:04.0000] <othermaciej> clients we talked to really wanted something more concise, thus the CSS image-set() proposal [10:08:22.0000] <tantek> Wilto, cool, then do you understand what is requested of you re: documenting use-cases? [10:08:24.0000] <othermaciej> the feedback we have gotten is that it's more usable [10:08:59.0000] <tantek> othermaciej - it sounds like the question is not just one of syntax/usability, but does it solve the use-cases. [10:09:14.0000] <tantek> thus I agree that some documenting of use-cases, e.g. on a wiki, is an important next step [10:09:16.0000] <othermaciej> I agree [10:09:30.0000] <annevk> tantek: fwiw, announcing all CGs on the WHATWG list is bad for S/N [10:09:41.0000] <tantek> annevk - are there that many CGs? [10:09:42.0000] <Wilto> tantek: I believe so, yes. [10:09:52.0000] <annevk> tantek: yes [10:09:53.0000] <tantek> and if CGs don't affect the web platform, then what are they doing at W3C? [10:09:58.0000] <Wilto> And I do appreciate the guidance in here, sincerely. [10:10:08.0000] <othermaciej> there are a huge number of CGs and there are indeed ones that do not affect the Web platform [10:10:47.0000] <othermaciej> Apple folks' opinion on <picture> vs srcset would probably be different if we knew <picture> handled some important use cases that can't practically be done w/ srcset [10:11:48.0000] <Wilto> othermaciej: Is Apple comfortable with implementing a "look-ahead" of sorts on the img tag? [10:12:05.0000] <timeless> odinho / hober : if you guys have corrections [10:12:13.0000] <othermaciej> Wilto: I'm not totally sure what you mean by "look-ahead" [10:12:14.0000] <timeless> i'm working on webapps minutes now [10:12:25.0000] <Wilto> The word I've gotten from a few WebKit devs is that it wouldn't fly; the src of the img will be prefetched in any case, leading to a redundant request. [10:12:54.0000] <Wilto> Which would very much defeat the purpose of any new markup pattern. I mean, otherwise, we'd just script something. [10:14:48.0000] <Wilto> Or use CSS pattern to swap the source, using some of the proposals out there already. Namely: http://www.w3.org/TR/css3-values/#attr [10:15:39.0000] <othermaciej> Wilto: I'm still not sure what you mean by look-ahead [10:16:10.0000] <othermaciej> I think in either the picture-style solution or the srcset-style solution, we'd only load the selected image [10:16:33.0000] <Wilto> That's likely not the right term. I mean... In terms of prefetching, would the browser be able to say "this is an image. Before fetching the src, I should see if there are alternate sources specified." [10:16:41.0000] <Wilto> On paper, I'd think "yeah, of course it can." [10:17:17.0000] <othermaciej> I believe it is possible with either the srcset syntax or the <picture> syntax, with a browser-native implementation at least [10:17:44.0000] <Wilto> If so, it would have to prefetch an appropriate source. I don't know the internals there; would the client information be available at that time? Would they want to kick in that logic during the prefetching process? [10:17:50.0000] <timeless> annevk: gah. i'm ~500 conversations behind on www-style [10:18:04.0000] <Wilto> Hm. See, we kept pitching ideas based on modifying img, and they were all promptly shut down for that reason. [10:19:25.0000] <Wilto> A new element would afford UAs an opportunity to build in more advanced logic. They... generally do not seem eager to bolt things onto <img>, whether that's by choice or necessity. [10:19:36.0000] <othermaciej> prefetch would have to be smarter to avoid accidentally prefetching the src [10:20:02.0000] <othermaciej> and prefetching an image based on width selection may well be impossible to do correctly [10:20:08.0000] <Wilto> Right. Which, on paper, seems fine. I don't know though; I have to go on the word of the browser devs we've been speaking to. [10:20:11.0000] <othermaciej> but prefetching the right one based on resolution is practical [10:20:39.0000] <Wilto> I haven't heard anything indicating that width-based sources would be an issue. [10:21:22.0000] <webben> Has anyone proposed a @defer attribute or something for iframe (would be useful for implementing non-blocking ads, widgets, videos etc)? [10:24:37.0000] <othermaciej> width-based sources (at least as proposed in srcset) depend on the available width, which is not known until layout time, so can't be done at prefetch time [10:25:42.0000] <Wilto> Yeah, see, that's what I was worried about. [10:26:11.0000] <Wilto> We'd effectively be saying "if this <img> has an alternate source, disable prefetching on it." I'm certain that won't fly. [10:26:15.0000] <othermaciej> so you could choose to do no preloading at all in that case [10:26:32.0000] <othermaciej> not really an issue for the 1x vs 2x resolution use case though [10:26:34.0000] <Wilto> No one has voiced the same concerns about <picture>, however. I assume they're afforded more flexibility in implementation with a blank canvas. [10:27:27.0000] <othermaciej> I do not know if it is possible to evaluate arbitrary media queries at preloading time [10:27:29.0000] <Wilto> (we tried <canvas> for workarounds, too. Ba-dum tss.) [10:33:19.0000] <timeless> anyone from the webapps f2f around to help me? [10:38:47.0000] <tantek> timeless what's up? [11:11:07.0000] <dglazkov> rafaelw_: I'd suggest starting a new thread with the new-best solution and getting ayes/nays on it [11:12:01.0000] <rafaelw_> ok [11:26:30.0000] <timeless> tantek: trying to fix up the minutes [11:30:31.0000] <tantek> timeless ok [11:55:20.0000] <timeless> /me sighs [11:55:30.0000] <timeless> Chrome is such a crappy browser for simple <find-forward> tasks [11:55:53.0000] <timeless> Obviously having scrolled to an anchor using <find>, i want to search from the _top_ of the document when i find again [11:56:00.0000] <timeless> and i obviously want to search letter at a time [11:58:36.0000] <timeless> odinho: ok... [11:58:41.0000] <timeless> odinho: read through everything [11:58:41.0000] <timeless> ... but when we send new comments [11:58:41.0000] <timeless> ... it's because we got to a new point [11:58:46.0000] <timeless> odinho: so far it's things that aren't really defined [11:58:46.0000] <timeless> ... and nitpicking, making things easier to read [11:58:57.0000] <timeless> hober: i hope the two of you are ok w/ that correction [12:03:51.0000] <timeless> /me goes back to looking at the minutes and the minute scripts [12:26:24.0000] <Ms2ger> gavinc, you may be interested in the compromise webperf has for its DOM reference: https://dvcs.w3.org/hg/webperf/raw-file/tip/specs/UserTiming/Overview.html#DOM3Core [12:32:58.0000] <timeless> Ms2ger: what's w/ the random link/no-link behavior? [12:33:09.0000] <timeless> Available at http://www.ietf.org/rfc/rfc2119.txt [link] [12:33:19.0000] <timeless> The latest version of DOM4 is available at http://www.w3.org/TR/dom/ [notlink]. [12:33:35.0000] <timeless> This version of the ECMAScript Language is available from http://www.ecma-international.org/publications/standards/Ecma-262.htm [notlink]. [12:33:47.0000] <Ms2ger> "Incompetence" [12:33:53.0000] <timeless> aha [12:33:59.0000] <timeless> well, i'm happy [12:34:12.0000] <timeless> my scripts can now distinguish Paul_Cotton form Paul_Kinlan [12:34:19.0000] <timeless> at least, if Paul_Kinlan speaks... [12:34:35.0000] <timeless> (the script is still pretty stupid in the case when paul kinlan doesn't speak) [12:34:43.0000] <timeless> s/form/from/ [12:47:01.0000] <dglazkov> timeless: file a bug? I'll cc finnur, who can fix it [12:47:21.0000] <eighty4> anyone happen to know if you can redefine how contenteditable inserts <br>s and <div>s and so on? [12:52:14.0000] <Ms2ger> I'm sure AryehGregor does [12:57:06.0000] <gavinc> Ms2ger: thanks [12:57:37.0000] <Ms2ger> Np [13:11:40.0000] <timeless> dglazkov: that'd require me to find the bug database and credentials [13:11:50.0000] <timeless> sorry, my top priority right now is minutes from last week [13:16:28.0000] <timeless> tantek: <dt id="T21-51-45">21:51:45 [chaals]</dt><dd>tantek: deoms are awesome. scope is broader than I ahd understood. How broad is the scope intended to be?</dd> [13:16:43.0000] <timeless> i presume "deoms" should be "DCOM" ? [13:17:08.0000] <timeless> /me isn't sure and hasn't gotten around to running the results through a spell-checker [13:18:08.0000] <tantek> no deoms is a typo transposition of demos [13:18:34.0000] <timeless> ok [13:18:39.0000] <tantek> but sure, related topics mentioned were DCOM, DSOM etc., I think that was earlier [13:19:07.0000] <tantek> and CORBA for that matter [13:19:26.0000] <timeless> /me nods [13:19:37.0000] <timeless> lemme send this thing through a spell checker [13:21:57.0000] <timeless> kam = kamos = komoroske = Alex_Komoroske:, right? [13:21:58.0000] <tantek> https://twitter.com/jackrusher/status/199851560665235457 [13:23:42.0000] <tantek> except that I don't think the #webintents folks know of CORBA - they're simply reinventing it (as well as OpenDoc/OLE) by walking down the same paths. it's like some odd 20 year generational cycle. [13:24:25.0000] <timeless> well [13:24:30.0000] <timeless> i'm vaguely involved in intents [13:24:41.0000] <timeless> and i've dealt w/ corba and ole, and i have memories of opendoc [13:25:34.0000] <timeless> what prevents someone from claiming that the embeddable Gmaps apis violate OpenDoc/OLE patents? [13:25:37.0000] <tantek> timeless - you seem to be vaguely involved with many web things :) [13:25:44.0000] <timeless> yes [13:26:01.0000] <timeless> OT: did you get around to rebooting your mac? [13:26:06.0000] <tantek> I'm not very familiar with embeddable Gmaps APIs - you'd have to talk to someone who understands both technologies [13:26:08.0000] <timeless> /me just upgraded PlayBook and phone os's [13:26:15.0000] <tantek> no [13:26:16.0000] <tantek> I don't reboot very often. [13:26:26.0000] <timeless> well... to the end user, you end up w/ a google map inside someone else's web page [13:26:39.0000] <timeless> you can interact w/ the google map pretty much the way you would if it weren't embedded [13:26:47.0000] <timeless> which is more or less what OpenDoc/OLE allow [13:27:12.0000] <timeless> Intents most of the time will not actually give you that [13:27:15.0000] <tantek> problem with OpenDoc/OLE is it was a huge set of technologies [13:27:18.0000] <tantek> some of which was patented [13:27:27.0000] <timeless> it's mostly just giving you a function that can return a value [13:27:30.0000] <tantek> so you can't prove anything by the negative [13:27:35.0000] <timeless> it's closer to CORBA than OLE/OpenDoc [13:28:09.0000] <timeless> hrm [13:28:14.0000] <timeless> put slightly differently [13:28:16.0000] <tantek> timeless - from the WebApps f2f there is no "mostly just" with webintents. [13:28:22.0000] <tantek> it's an ever growing scope [13:28:31.0000] <tantek> I wish you luck with attempting to limit it [13:28:38.0000] <timeless> are the OLE/OpenDoc patents older than the MPEG1 patents? :) [13:28:44.0000] <tantek> I've already decided on the specific focus I'm solving with web actions. [13:29:00.0000] <tantek> timeless - if you're curious, you can do your own patent search. [13:29:16.0000] <tantek> I'm not familiar with MPEG1 patents (nor do I have any interest in being) [13:29:18.0000] <timeless> i'm pretty sure my lawyers would yell at me :) [13:29:22.0000] <tantek> exactly [13:29:31.0000] <tantek> so it's futile to attempt to reason by the negative [13:29:41.0000] <timeless> /me sighs [13:29:50.0000] <timeless> well, i'm not trying to really reason [13:29:54.0000] <timeless> just paint a broad stroke [13:30:04.0000] <timeless> i have a feeling that ole/opendoc could be older than mpeg1 [13:30:23.0000] <timeless> and we've sort of reached a point where mpeg1 patents were offering to expire [13:45:56.0000] <Velmont> timeless: Thanks for fixing that :) [13:47:25.0000] <rniwa> the whole discussion about arabizi was funny.... [14:34:10.0000] <timeless> Velmont: thanks for pointing it out [14:34:20.0000] <timeless> Testing [14:34:20.0000] <timeless> rniwa: I like to write tests [14:34:30.0000] <timeless> ... if we have a tool that could go through the tests [14:34:30.0000] <timeless> <inserted> schepers: [14:34:30.0000] <timeless> rniwa: the CSS Tool called "Shepard" (by plinns) [14:34:38.0000] <timeless> -- help? [14:34:53.0000] <shepazu> huh? [14:35:19.0000] <timeless> chaals inserted "shepazu:" before "Shepard" [14:35:25.0000] <timeless> i'm not sure why [14:35:35.0000] <timeless> was he trying to say that you were speaking? [14:35:43.0000] <timeless> i.e. did you say "the CSS Tool called" ? [14:36:36.0000] <timeless> /me decides that shepazu probably did say that (and not rniwa) [14:37:10.0000] <shepazu> timeless: yes, I did mention Shepard [14:37:38.0000] <timeless> ok, great, thanks [14:37:50.0000] <timeless> note that the final minutes won't say schpers [14:38:02.0000] <timeless> it says that so i can test a scribe speaker-present tool [14:49:35.0000] <paul_irish> shepazu: where is this tool? i'm interested. [14:50:35.0000] <paul_irish> http://test.csswg.org/shepherd/ [14:50:56.0000] <paul_irish> http://wiki.csswg.org/test/css2.1/contribute [14:53:23.0000] <shepazu> paul_irish: it's where you just pointed [14:53:29.0000] <shepazu> :| [14:53:29.0000] <paul_irish> :) [14:54:40.0000] <timeless> ok, i've sent off minutes for webapps [14:54:46.0000] <timeless> i'll try to deal w/ html tomorrow [14:54:53.0000] <timeless> maybe. i have other things to worry about tomorrow [14:54:58.0000] <timeless> they're in www-archive for now [14:55:12.0000] <timeless> if people ( Velmont, shepazu , odinho , anne, ...) could review them [14:55:14.0000] <timeless> i'd appreciate it [15:04:50.0000] <Velmont> timeless: Velmont === odinho [15:06:08.0000] <rniwa> timeless: sorry, i was away from keyboard [15:06:14.0000] <rniwa> timeless: did you need me for something? [15:37:23.0000] <cbright6062> sigh. the busyness of work. [15:56:34.0000] <smaug____> seriously, linking to dart code for use cases for the web... [16:19:36.0000] <tabatkins> smaug____: Dart is almost JS. Seems legit. [16:20:38.0000] <smaug____> Dart code is like Java [16:20:44.0000] <smaug____> as much webby [16:35:29.0000] <arv> smaug____: The Dart library uses the same DOM APIs as JS. [16:36:08.0000] <smaug____> arv: the API is quite different [16:36:28.0000] <smaug____> and Java uses the same APIs too [16:36:30.0000] <arv> smaug____: it is irrelevant which language someone uses when the underlying capabilities is just DOM4 [16:36:48.0000] <arv> smaug____: yes, and java has the same issue regarding context less parsing [16:36:55.0000] <smaug____> actually, there isn't any webidl spec for Dart [16:37:23.0000] <arv> smaug____: so? [16:37:38.0000] <smaug____> and my main point is that it is very un-webby and silly to demonstrate anything on a research language and not using JS [16:37:47.0000] <arv> smaug____: there isn't one for cs eitehr [16:37:57.0000] <smaug____> arv: where is the webidl spec for Dart? [16:38:08.0000] <arv> smaug____: see above. the langiuage is irrelevant [16:38:42.0000] <smaug____> ok. I use Perl next time in my examples :/ [16:39:08.0000] <arv> sgtm 2012-05-11 [23:08:49.0000] <AryehGregor> <eighty4> anyone happen to know if you can redefine how contenteditable inserts <br>s and <div>s and so on? <-- What behavior do you want? [23:17:08.0000] <AryehGregor> eighty4, currently IE/Opera wrap everything in <p> on Enter, Firefox inserts <br>, WebKit uses <div> for line breaks but very recently added support to opt in to <p> instead (defaultparagraphseparator command). I'm currently working on a patch for Gecko to more closely match IE/Opera behavior, which is what the editing spec requires. [23:17:36.0000] <AryehGregor> So once that patch lands, the most recent version of everything will do <p> (at least if you opt in with defaultparagraphseparator, for WebKit). [23:18:33.0000] <AryehGregor> Although the exact places they create the separator will still vary -- e.g., "foo\nbar" in an empty editable div should create "<p>foo</p><p>bar</p>" per spec, but in WebKit probably produces "foo<p>bar</p>", browsers might leave useless trailing <br>'s, etc. [23:18:37.0000] <AryehGregor> Any questions? :) [23:19:03.0000] <AryehGregor> (there are a lot more details here, FWIW, I'm just talking about the most basic case; what happens if you hit Enter in an <li>, <h#>, etc. is a different story) [00:33:05.0000] <jgraham> zcorpan: Compile errors don't have a stack either [00:36:10.0000] <zcorpan> jgraham: indeed [00:36:22.0000] <zcorpan> jgraham: but they have a column number [00:39:12.0000] <jgraham> Weren't you proposing passing in both? [00:41:07.0000] <zcorpan> yes. but he suggested dropping column and replacing it with Error [00:41:35.0000] <zcorpan> and putting column on Error [00:41:50.0000] <hober> /me catches up on the backlog [00:42:11.0000] <hober> not being in pacific time can be inconvenient [00:43:22.0000] <jgraham> zcorpan: That doesn't sound very compatible [00:44:09.0000] <zcorpan> jgraham: column is a new thing, though [00:44:49.0000] <jgraham> hober: You are living in the silicon valley bubble :p [00:45:00.0000] <hober> :) [00:45:08.0000] <jgraham> zcorpan: No one supports it? [00:45:15.0000] <zcorpan> right [00:45:18.0000] <jgraham> I see [00:45:28.0000] <zcorpan> maybe ie10, dunno [00:45:37.0000] <zcorpan> ms proposed it [00:46:10.0000] <jgraham> Well I think passing in the error object somewhere (when it exists) is a good idea [00:46:23.0000] <hober> Wilto othermaciej tantek et al.: sorry I missed your discussion last night. [00:46:56.0000] <hober> i'll write up my thoughts on <picture> and why i prefer <img srcset> for the Retina/non-Retina use case soon, but i'm at the css f2f and don't have the time for it right now [00:47:25.0000] <hober> [i had the time to "write up" <img srcset> because i had already written the email, just hadn't sent it. :)] [00:48:26.0000] <hober> Wilto: I was aware of the CG and the various conversations online that preceded it (on ALA, the etherpad, the june 2007 thread on public-html, etc.) [00:48:40.0000] <zcorpan> i prefer an attribute because it makes the processing much simpler. <source> is hard [00:49:06.0000] <gsnedders> zcorpan: Is there any reason why we can't just pass in the exception value? [00:49:18.0000] <hober> Wilto: and reviewing all of that most definitely influenced the design of my proposal [00:49:32.0000] <gsnedders> I mean, we can always create SyntaxError objects aside from the eval case [00:49:41.0000] <zcorpan> gsnedders: no reason, i'm just trying to figure out what information he wants [00:50:00.0000] <hober> Wilto: <img srcset> isn't a panacea; it's intended to solve some of the problems identified under the umbrella term "responsive images" [00:50:01.0000] <zcorpan> gsnedders: oh you mean for compile errors? [00:50:13.0000] <zcorpan> gsnedders: yeah we could create an exception object for that case [00:50:21.0000] <gsnedders> zcorpan: Nah, I mean in general, and suggesting how we handle the compile-error case [00:50:36.0000] <hober> zcorpan: indeed [00:52:21.0000] <hober> Wilto: anyway, i'm sorry if you interpreted my email as ignoring the hard work that the cg has done [00:54:52.0000] <hober> Wilto: what happened was quite the opposite; i went over many of the discussions (both in the cg, on the whatwg and html wg mailing lists, and in the wider community prior to the cg's formation) while working on the problem [01:41:18.0000] <smaug____> hmm, whether or not un-prefix MutationObserver now [01:41:29.0000] <smaug____> perhaps I should test the implementations some more [01:41:53.0000] <annevk> if you unprefix it now and there's edge case bugs you should still have time to fix them [01:41:54.0000] <smaug____> or anyone else willing to test the implementations? ;) [01:42:15.0000] <smaug____> annevk: true [01:42:30.0000] <annevk> it's more about whether you trust the general design enough I guess [01:42:40.0000] <zcorpan> UNPREFIX ALL THE THINGS! [01:42:52.0000] <annevk> but given that there's now blog posts advocating people to use it, I'd prefer we just went with it [01:42:59.0000] <smaug____> it is perfect design. design by me (and others) ;) [01:43:02.0000] <Ms2ger> Hah [01:43:10.0000] <Ms2ger> It's in a spec, surely it isn't perfect [01:43:27.0000] <smaug____> that is true. should we take it out from the spec :) [01:43:48.0000] <jgraham> Didn't the link to "spec" in the blogpost go to an email? [01:44:05.0000] <annevk> yeah so no problems there [01:44:15.0000] <smaug____> jgraham: in mozilla hacks= [01:44:19.0000] <smaug____> s/=/?/ [01:44:23.0000] <annevk> yup [01:44:31.0000] <smaug____> jgraham: I posted a comment about that [01:44:37.0000] <smaug____> with the right link [01:44:50.0000] <jgraham> Oh, I saw it on planet mozilla, so no comments [01:45:04.0000] <smaug____> I just posted my comment [01:45:55.0000] <smaug____> annevk: jgraham: but in general Opera doesn't see anything hugely wrong with the API ? [01:45:58.0000] <smaug____> /me should ask MS [01:46:43.0000] <Ms2ger> Land it [01:46:46.0000] <Ms2ger> Oh, other MS [01:49:50.0000] <jgraham> smaug____: I haven't studied it in detail. I thought annevk was following it more closely. But I asked some of our developers and I think their opinion was roughly "it can't be worse than mutation events" mixed with a little "we are probably stuck with mutation events, so is having an extra API here a good idea" [01:49:57.0000] <jgraham> +? [01:50:12.0000] <annevk> yeah [01:50:39.0000] <annevk> though also combined with if we can get rid of mutation events that'd be really awesome... [01:51:08.0000] <smaug____> jgraham: well, we at least should try to get rid of mutation events [01:51:33.0000] <annevk> I'd hate having to define them in DOM [01:51:55.0000] <smaug____> MutationObserver is 5-10 faster than MutationEvents (both in Gecko and Webkit), and it is 'safe' [01:52:14.0000] <jgraham> Well sure, if we can kill mutation events there will be rejoicing in the streets of Oslo [01:52:17.0000] <smaug____> I would assume mutation events have caused crashes also in Opera [01:52:23.0000] <smaug____> :) [01:52:39.0000] <jgraham> (or Linköping, but everyone thinks Opera == Oslo) [01:53:01.0000] <smaug____> (Hmm, have I been in Linköping) [01:53:36.0000] <smaug____> (probably just passed by) [01:58:37.0000] <gsnedders> (Probably just passed by, it's not like anything ever happens here.) [02:01:09.0000] <smaug____> (I've been often in places where nothing really happens. Silicon Valley is one such place.) [02:02:06.0000] <jgraham> gsnedders: You were talking just last night about Judas Priest playing here! [02:02:27.0000] <jgraham> If you want to see 60 year old men in bondage gear, this is the town for you! [02:03:08.0000] <gsnedders> Well, by Lkpg's scale, that's an unprecidented event. [02:03:23.0000] <gsnedders> I never said in more ordinary terms it was notable. [02:44:27.0000] <Ms2ger> "we are close to closing the XSL WG. The XSL FO WG is going to be closed in the next 6 months." [02:44:29.0000] <Ms2ger> Hear, hear [02:49:39.0000] <jgraham> UPDATE W3C.groups SET status='inactive' WHERE name LIKE 'X%'; [03:08:43.0000] <charlvn> lol jgraham [03:33:49.0000] <zcorpan> tabatkins: hsivonen's proposal makes <set> and <image> svg elements [04:21:02.0000] <tabatkins> zcorpan: Yeah, misread. Sorry, hsivonen! [04:31:15.0000] <benvie> seeing all the svg property interfaces blink into existence 5-10 at a time when setting one property due to mutation observers and svg's late init is pretty funny [04:43:03.0000] <tabatkins> SVG has late init? [05:34:11.0000] <jarek> btw, would it be suitable to use mutation observers for implementation of undo manager in SVG editor? [05:35:18.0000] <jarek> I have noticed that SVG-edit is registering undo operations manually which is a bit complicated [05:40:28.0000] <zcorpan> hey everyone! check out the new DOM Parsing spec at http://www.w3.org/TBD [06:11:02.0000] <kennyluck> Hmm.. I don't get that joke. [06:13:44.0000] <hober> kennyluck: it's no joke [06:14:01.0000] <hober> https://www.w3.org/Bugs/Public/show_bug.cgi?id=11204 [06:16:30.0000] <jgraham> hober: Well it is. Just the other kind of joke [06:17:00.0000] <hober> jgraham: indeed. [06:17:44.0000] <zcorpan> jgraham: what's the idea with setup()? what's supposed to happen when it fails? [06:19:00.0000] <jgraham> zcorpan: In theory the idea is that you put any pre-test setup in that function and testharness.js will report to a higher layer if it fails [06:23:45.0000] <zcorpan> yeah, that's what i thought. it just doesn't seem to work [06:24:03.0000] <zcorpan> my tests still time out even if setup throws [06:24:54.0000] <jgraham> Oh. Well it might be broken in some way I guess [06:27:53.0000] <kennyluck> lol [06:30:48.0000] <jgraham> zcorpan: If you put the test somewhere I can see it I will look [06:56:09.0000] <Wilto> hober: So you’re saying that imgset does not preclude <picture>. [06:56:30.0000] <Wilto> Despite there being some overlap in the concerns they address? [07:00:53.0000] <Wilto> …No, actually, these address the same concerns. [07:01:45.0000] <Wilto> In different ways, but the same core issues. Just, one uses a syntax arrived upon through the consensus of the Community Group, countless blog posts, comments—hell, _tweets_—and one was arrived at after a cursory glance at the CG. [07:02:09.0000] <Wilto> I’m not some especially irritating kid with an opinion here, guys. [07:02:17.0000] <Wilto> I’m representing the developers’ consensus. [07:03:04.0000] <Wilto> So when you say “this syntax is better for authors,” here, in a vacuum, I am here to inform you that we considered something just like this already and nobody liked it. [07:03:06.0000] <hober> I think "a cursory glance" is an unfair characterization of the work that went in here. [07:03:52.0000] <Wilto> Take it or leave it, sure, but don’t pretend the developer consensus is the reasoning behind this proposed syntax when _getting that consensus has been my job_. [07:03:55.0000] <jgraham> Wilto: You would be much more convincing with citations [07:04:13.0000] <jgraham> Specifically to citations to problems with the syntax [07:04:25.0000] <jgraham> Rather than making an argument from authority [07:05:29.0000] <Wilto> Sure. Happy to collect those for you. [07:06:25.0000] <hober> I don't understand what you think I'm pretending. [07:06:45.0000] <Wilto> You can understand where I’m irritated, when the CH was posed as a place to give developers a voice in the creation and adoption of standards, and the _good news_ is that people had a look at it before starting to make decisions. [07:06:47.0000] <hober> (i'm not being intentionally obtuse, btw, i really didn't understand that last line) [07:07:13.0000] <hober> I certainly understand your irritation. [07:07:37.0000] <hober> That said, how some people described a community group isn't something I have any control over [07:07:48.0000] <Wilto> No, of course not. [07:08:14.0000] <Wilto> But like… you saw the number of folks involved in the discussion in the CG, the ALA article, the .net article, likely some of the chatter on Twitter, etc. etc. [07:08:34.0000] <Wilto> Even if the others missed it somehow, man, you’ve seen that this is a big deal. [07:08:49.0000] <Wilto> And no small amount of work has been done. [07:09:23.0000] <hober> is anyone saying that that work hasn't been done? [07:09:27.0000] <Wilto> Hell, not even asking me to get involved—someone could have mentioned that these discussions were kicking off in IRC. [07:09:29.0000] <Wilto> No, not at all. [07:09:42.0000] <Wilto> I’m saying that despite all that, the discussion kicked off in a vacuum. [07:10:04.0000] <hober> (sorry if i'm coming across as short, btw; it's day 3 of the css wg and it's getting kinda punchy in here. good thing the room is padded.) [07:10:10.0000] <Wilto> You’re not working with developers on this. You’re “considering” us. [07:10:24.0000] <Wilto> We planned on working with the WHATWG. [07:10:41.0000] <hober> aren't we working together right now? [07:11:01.0000] <hober> i don't understand "the discussion kicked off in a vacuum" [07:11:11.0000] <Wilto> Hixie already “proposed a solution.” [07:11:13.0000] <hober> there have been many threads on whatwg@ (and elsewhere, obv.) about this [07:11:31.0000] <Wilto> Which is fine, don’t misunderstand. [07:11:47.0000] <hober> yeah, and when he described it i recalled that i had a pending draft email about that, so i dug it up and hit send [07:12:26.0000] <hober> i'm not at the machine that i usually use for email, so it didn't get property threaded as a reply on one of the existing threads [07:12:39.0000] <Wilto> But that’s where I get irritated. So he just whipped that up, and I’m guessing from the responses in general that he had no idea the CG existed, and hasn’t paid much attention to the internet at large talking about respimg. [07:13:14.0000] <Wilto> But that syntax is “better for authors?” What I’m getting at is I’ve, y’know, _asked_ the developers. [07:13:18.0000] <hober> i don't see how you can assert that he "just whipped that up" [07:13:19.0000] <jgraham> Wilto: I don't think getting irritated is going to help you at all [07:13:34.0000] <jgraham> If you have concrete objections to that syntax [07:13:36.0000] <jgraham> make them [07:13:39.0000] <hober> i don't know how much time he spent on it, so i avoid aserting that he spent a lot of time or not much time [07:13:44.0000] <jgraham> Cite the developers [07:13:53.0000] <Wilto> jgraham: Yeah, guess I’m spinning my wheels here. [07:14:15.0000] <Wilto> jgraham: Before I do, are there any citations that Hixie’s proposal is preferred by developers? [07:15:16.0000] <zcorpan> an attribute would be better for *implementors* (because it makes the processing orders of magnitude simpler (which in turn means fewer bugs)) [07:15:37.0000] <Wilto> zcorpan: Citations? [07:16:02.0000] <Wilto> zcorpan: By which I mean, has this been discussed with implementers? [07:16:11.0000] <jgraham> Wilto: zcorpan is an implementor [07:16:18.0000] <zcorpan> Wilto: i've QA'd <video><source> for opera, i know it is orders of magniture more complex than an attribute would be [07:17:09.0000] <jgraham> (it's rather easy to understand why; if you have multiple elements, you can get things like <script> in the middle that add or remove things in real time, whereas attributes can be processed atomically) [07:17:18.0000] <Wilto> Cool. The order of operations for benefit is “users -> developers -> implementors” though, yeah? [07:17:29.0000] <jgraham> Yes [07:17:43.0000] <Wilto> I’m not dismissing what you’re saying, mind. Again, I’ve been working with browser devs right along, and this is important stuff. [07:17:43.0000] <kennyluck> Wilto, I would strongly suggest you write an email to whatwg⊙wo, even just a mail with a pointer to the discussion on CG. No body can read every mail on every mailing list. [07:17:45.0000] <zcorpan> yes. but if the implementors get riddled with bugs, that's not useful for developers either [07:18:07.0000] <jgraham> If you can prove that multiple elements addresses use cases better than a single attribute, that makes a difference [07:18:30.0000] <Wilto> Has this been discussed with reps from any other UAs? [07:18:48.0000] <jgraham> But yeah, all things being equal, avoiding interoperability problems before they have the chance to occur is a win for everyone [07:19:15.0000] <Wilto> Opera doesn’t do anything—or have anything planned—in the way of image prefetching, correct? [07:19:16.0000] <hober> Wilto: when you say "browser devs", do you mean implementors? dev rel? other browser-affiliate people? [07:19:35.0000] <Wilto> Implementors and a few devrel folks. [07:19:44.0000] <jgraham> Wilto: We generally don't talk about the roadmap [07:19:48.0000] <jgraham> Except when we do [07:19:59.0000] <zcorpan> jgraham: you got the meme wrong :-P [07:20:16.0000] <jgraham> zcorpan: I don't know all the memes [07:21:04.0000] <gsnedders> Object.getPrototypeOf(xOriginWindow) doesn't throw anywhere… [07:21:27.0000] <Wilto> jgraham: Well, that’s a pretty major factor in implementation across all the browsers. [07:21:40.0000] <zcorpan> Wilto: i haven't discussed this really until now. but i'm happy to write down the complexity differences between an attribute and elements sometime next week [07:22:06.0000] <Wilto> zcorpan: I’d like to see that, yes. [07:22:07.0000] <jgraham> Wilto: It doesn't seem like it's that complicated really [07:22:15.0000] <jgraham> Wilto: You can DNS prefetch [07:22:35.0000] <jgraham> But you can't speculatively grab the resources in a bandwidth-efficient way [07:22:47.0000] <jgraham> Because it depends on layout which one you pick [07:23:01.0000] <Wilto> jgraham: I can’t speak to that complexity, but I know it has been a major factor in discussions with Chrome. [07:23:20.0000] <hober> jgraham: in my proposal, asset choice doesn't depend on layout. [07:23:23.0000] <Wilto> I’m happy to revisit that and post the results—or the discussion itself—publicly. [07:23:32.0000] <hober> jgraham: one of the differences between hixie's and mine [07:23:36.0000] <jgraham> hober: Oh. [07:23:51.0000] <jgraham> That seems to address fewer use cases? [07:23:55.0000] <hober> yup [07:24:02.0000] <zcorpan> now, nice weekend everyone :-) [07:24:11.0000] <hober> jgraham: the perfect is the enemy of the good :) [07:24:18.0000] <hober> jgraham: baby steps and all [07:24:24.0000] <jgraham> OK, I guess you can always ditch use cases in order to get prefetching working [07:24:54.0000] <jgraham> hober: Sure, I don't know what the right compromise is, exactly [07:29:01.0000] <Wilto> The idea for a new element came about because we were looking for a way to preserve prefetching—or at least sandbox any custom prefetching behavior—without abandoning functionality. [07:30:10.0000] <jgraham> I don't understand how element vs attribute can affect prefetching. I understand that not depending on layout can affect what prefetching you can do at the expense of not meeting all the use cases [07:33:53.0000] <Wilto> jgraham: I’ll gather more information on that. [07:39:46.0000] <Wilto> I’m going to post these discussions to the Community Group; just a heads-up. [07:40:06.0000] <Wilto> And the associated proposals. [07:40:41.0000] <Wilto> And reach out to the developer community to voice their preference. Would anyone like to be the primary point of contact for that, or should I direct them to the WHATWG at large? [07:41:54.0000] <jgraham> To the WHATWG [07:42:05.0000] <hober> whatwg⊙wo is the right place. n.b. "+1" posts (posts which express a preference but do not make a novel point) are discouraged on the whatwg list. [07:42:10.0000] <jgraham> If you email the WHATWG list then nothing will get lost [07:46:13.0000] <Wilto> Sure; sounds good. [07:46:21.0000] <kennyluck> Seriouly, I don't quite buy any "theories" with bikeshedding issues and in that case voting just helps. [07:47:04.0000] <Wilto> I’ll encourage people to sound off with their +1s in another forum, since a big part of this discussion seems to be developer preference. [07:47:14.0000] <Wilto> The list may not be the right place for it, but it should be recorded. [07:50:11.0000] <kennyluck> Wilto, agreed. Hixie ought to consider a poll as "data", but I think he prefers to call it "usability research". I don't quite know the difference but he might be able to give you some hints. [07:52:35.0000] <Philip`> /me guesses the difference is that polls are usually biased towards groups with the strongest opinions and with the widest ability to convince other people to vote, whereas usability research tries to get a smaller but more representative sample [07:54:17.0000] <jgraham> If developers have a preference, they should articulate *why* they prefer one form over another [07:56:28.0000] <Wilto> Well, I’ve been taking on a firehose of developer opinions on this subject for months. Which is not so much like “herding cats” as it is “herding YouTube commenters.” [07:57:10.0000] <hober> wheee! :) [07:57:38.0000] <Wilto> I could cite all the discussions on the CG, and gather up links to offsite stuff, or I can just tell them to get involved in these discussions. You’re gonna get some +1s—and naturally a reasoned argument would be preferable—but they shouldn’t be discounted entirely either. [07:58:08.0000] <Wilto> hober: Man oh man, do folks have Opinions on the Internet. [08:43:43.0000] <karlcow> +1 to the above [08:44:51.0000] <hober> Wilto: indeed. http://xkcd.com/386 [09:10:54.0000] <dglazkov> good morning, Whatwg! [10:12:49.0000] <staydecent> <source src="... looks redundant [10:14:10.0000] <barnabywalters> staydecent: what would you suggest? [10:17:43.0000] <staydecent> Why not have all elements within <picture> be <img>? [10:18:24.0000] <barnabywalters> could that not cause serious backwards compatibility problems? [10:18:42.0000] <staydecent> ah.. right. Old browser would just ignore the <source> elements? [10:19:10.0000] <barnabywalters> yep [10:19:47.0000] <barnabywalters> that's the beauty of the <picture> element + fallback (although I hate the name. It's like calling <video> <film>) [10:19:48.0000] <Wilto> staydecent: Current browser behavior is just to find the src, and then “case closed.” Old browsers would ignore them, but so would new ones without some major parsing changes. [10:20:18.0000] <Wilto> barnabywalters: Hah, I don’t love `picture` either, truth be told. Every time we raise the issue, though, no one has anything better. [10:21:05.0000] <staydecent> Yeah. i agree I find the <picture> element much more readable for developing than the set attr. But, the naming is a little off. But again, I have no suggestions atm [10:21:18.0000] <barnabywalters> hmmm... I read <image> is an alias for <img>, so if that's true that wouldn't work [10:21:18.0000] <ShaneHudson> Somebody mentioned in the comments about calling it <image>... I think it would work, more consistant without being confusing [10:21:40.0000] <Wilto> barnabywalters: Yeah, a lot of older browsers treat img and image the same. [10:21:54.0000] <zcorpan> s/a lot of older// [10:22:00.0000] <othermaciej> barnabywalters: I think <movie> and <sound> might have been fine names for <video> and <audio> [10:22:02.0000] <Wilto> Which stands to introduce Troubles. [10:22:03.0000] <staydecent> how much older? [10:22:23.0000] <zcorpan> the html spec requires <image> to be parsed as <img> [10:22:29.0000] <Wilto> zcorpan: Is that right—current, too? Admittedly, someone else ran those tests and reported back. [10:22:30.0000] <staydecent> ah [10:22:35.0000] <barnabywalters> <sound> might have been alright, but I don't like <movie> [10:22:58.0000] <zcorpan> Wilto: current too, yeah. pages depend on it. :( [10:23:09.0000] <Wilto> “Oh, what a tangled web we weave.” [10:23:09.0000] <ShaneHudson> How about <responsive> since <picture> includes <img> inside of it, rather than replacing [10:23:33.0000] <barnabywalters> that's a bit ambiguous [10:24:19.0000] <Wilto> Yeah—and kind of hitches our horses to that term. I loves me some RWD of course, but… [10:24:43.0000] <ShaneHudson> True, it is too much of a buzzword as it is [10:25:53.0000] <staydecent> This is a bit of a braindump, but: what about <media> which could expand to have <source>...etc. and then either <img> or <video>? [10:26:22.0000] <barnabywalters> staydecent: so you'd have something like <media type="video"> [10:26:25.0000] <staydecent> Maybe that conflicts with media-queries tho. [10:26:31.0000] <staydecent> yeah [10:26:37.0000] <zcorpan> let's not complicate <video> [10:26:38.0000] <ShaneHudson> I like the idea of giving chance for <video> as well as <img> but I am not sure if there would be troubles with implementations [10:27:02.0000] <othermaciej> <video> already has media-query-based responsiveness [10:27:10.0000] <Wilto> I’m with zcorpan there. [10:27:19.0000] <othermaciej> also, the <element type=actual-element> pattern is kinda lame [10:27:33.0000] <ShaneHudson> Would it be completely stupid to have <img> as the main element with optional sources inside? <img src="default"><source></img> [10:27:34.0000] <othermaciej> it's only really useful for <input> since it falls back to a text input, which is often a reasonable choice [10:27:39.0000] <barnabywalters> yeah, agreed. Doesn't seem very stable [10:27:40.0000] <Wilto> It’s tough because this topic feels like an opportunity to address so many other, related issues. [10:27:41.0000] <zcorpan> othermaciej: i argued on the list that <source media> is the wrong solution and maybe should be dropped :-) [10:27:46.0000] <zcorpan> (for video) [10:27:46.0000] <Wilto> But we can’t, y’know? [10:27:59.0000] <othermaciej> zcorpan: I do wonder if anyone is using it [10:28:23.0000] <othermaciej> zcorpan: the original purpose was for (rather hypothetical) accessibility-related and bandwidth-related additions to media queries [10:28:38.0000] <othermaciej> <source> mainly seems useful only for content type selection [10:28:44.0000] <zcorpan> othermaciej: indeed [10:29:04.0000] <barnabywalters> that's a very good point - using source for images would be inconsistent with video [10:29:11.0000] <zcorpan> othermaciej: i think it makes more sense to use something like "adaptive streaming" with a single <source> URL to switch streams [10:29:30.0000] <zcorpan> since bandwidth and wanted resolution can change over time [10:29:34.0000] <barnabywalters> zcorpan: can you go into more detail about how that would work? [10:29:49.0000] <othermaciej> adaptive streaming does make more sense as an approach to video [10:29:57.0000] <staydecent> Would you need the attr type="video" tho? [10:29:58.0000] <othermaciej> not really as sensible for static images [10:30:51.0000] <Wilto> Let’s not get too bikeshed-y in here though, guys. [10:31:07.0000] <Wilto> The community group is great for this stuff; I’d like to keep the brainstorming in there, if that’s cool. [10:31:18.0000] <zcorpan> barnabywalters: something like apple's "adaptive streaming" thing for mpeg4, but not necessarily exactly that [10:31:52.0000] <barnabywalters> so pushing the responsibility more towards the browser manufacturers? [10:32:17.0000] <zcorpan> compared to what? [10:32:43.0000] <zcorpan> <source media> is the responsibility of the browsers too [10:32:45.0000] <barnabywalters> compared to picture where the authors decide what images they're providing and when to display them [10:32:53.0000] <ShaneHudson> It would make it a lot easier for general users defintely... it is easy to learn about <img> but <picture> (although the best soloution so far) is pretty complicated [10:33:00.0000] <zcorpan> i'm talking about <video> right now :-) [10:33:05.0000] <barnabywalters> ah, my bad :) [10:33:23.0000] <ShaneHudson> Hey chriscoyier :) [10:33:25.0000] <zcorpan> for images, adaptive streaming doesn't make much sense, as othermaciej said [10:34:04.0000] <staydecent> What about <picture> -> <imagegroup>? [10:34:22.0000] <zcorpan> i guess i should write that email about complexity with <source> vs an attribute [10:34:27.0000] <barnabywalters> staydecent: to me, imagegroup implies a group of related but distinct images [10:34:27.0000] <othermaciej> for video, you really want to change bandwidth/quality tradeoff in the middle of playback rather than loading a whole new resource [10:34:42.0000] <othermaciej> for still images, you could at most select once [10:35:01.0000] <staydecent> barnabywalters: I feel like you guys have thought on this quite well. Good point. [10:35:15.0000] <othermaciej> for resolution-based decisions, media queries it seems are not really sufficient, given the implied semantic of scaling in accordance with the scale factor used for selection [10:35:16.0000] <zcorpan> othermaciej: yeah. if you chime in and say so on the list, maybe we can kill <source media> now :) [10:35:45.0000] <zcorpan> (i think opera is the only one to have implemented it, though mozilla are apparently working on it right now) [10:35:59.0000] <staydecent> can I check some logs somewhere for alternate suggestions that have been made for <picture>? [10:36:02.0000] <barnabywalters> staydecent: TBH I prefer it to <picture>, but I don't think it's right, as such [10:36:47.0000] <ShaneHudson> I agree that <imagegroup> feels more like a list tag than alternatives [10:37:09.0000] <padenot> zcorpan: I just finished the patch to have <source media>, it is landing today, afaik [10:37:10.0000] <zcorpan> othermaciej: (subject "Re: [whatwg] <source> media attribute behavior, static or dynamic ?" on whatwg) [10:37:19.0000] <zcorpan> padenot: ah [10:37:20.0000] <padenot> zcorpan: in Gecko, that is [10:37:35.0000] <zcorpan> padenot: do you think it is a useful feature? [10:37:47.0000] <othermaciej> zcorpan: I have to think about it more before saying anything definitive [10:37:56.0000] <padenot> zcorpan: it can have its use, but I don't expect a lot of people to use it [10:38:25.0000] <zcorpan> padenot: what would it be useful for? [10:38:45.0000] <padenot> zcorpan: get scaled down video on mobile [10:39:10.0000] <zcorpan> padenot: but it's not really appropriate for that, afaict (as i said on the list) [10:39:15.0000] <padenot> zcorpan: the bandwith problem is likely to be addressed by dash [10:39:32.0000] <zcorpan> dash? [10:40:12.0000] <padenot> zcorpan: http://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP [10:40:56.0000] <zcorpan> ah. right. why isn't that a better solution to the mobile problem than media=""? [10:42:29.0000] <padenot> zcorpan: i don't know enough about dash yet to provide a good answer, i think [10:42:37.0000] <staydecent> Sorry if this is the wrong place: Where does current support lean regarding <picture> vs set=""? [10:43:06.0000] <padenot> zcorpan: don't know if we can request specifically a resolution or the like [10:44:06.0000] <zcorpan> padenot: but the browser would be able to "adapt" to a resolution that currently fits the viewing device, right? [10:44:30.0000] <jgraham> ~. [10:44:45.0000] <othermaciej> on mobile you may want to serve different size rather than different quality (these are generally separate parameters to the encoding process) [10:44:54.0000] <othermaciej> but generally this is done by UA sniffing [10:45:11.0000] <ShaneHudson> For a renaming of <picture>... what about <quality> or <version>? Too 'meta'? [10:45:31.0000] <padenot> zcorpan: I think it is reasonnable to say that, but I don't know for sure [10:45:48.0000] <barnabywalters> shanehudson: you're talking about renaming the <source> child tags? [10:46:30.0000] <barnabywalters> or the <picture> tag it's self? [10:46:40.0000] <ShaneHudson> barnabywalters: I was thinking more of <picture> to <versions> or something [10:47:21.0000] <barnabywalters> ShaneHudson: I think the root element should describe the content type. Otherwise things could get sticky later on [10:47:43.0000] <barnabywalters> e.g. for different content types, content types that don't even exist yet, etc [10:48:17.0000] <ShaneHudson> fair enough, in an ideal world it would be nice to be able to work with any content type, but that is of course not the case [10:49:31.0000] <Zunflappie> What about CSS > instead of background-image also a front-image? As a overlay over the image? <img src="small.jpg" style="front-image: url(big.jpg); media=min-400px"> [10:49:48.0000] <Zunflappie> I dont know the right syntax for inline style with media-queries, but you got the point? [10:50:01.0000] <staydecent> Zunflappie: that would load both images though? [10:50:18.0000] <Zunflappie> Yeah, but dont need javascript. [10:50:38.0000] <barnabywalters> Can I get people's opinion on this syntax: http://jsfiddle.net/aZ37N/ -- would there be any serious backward compatibility/other problems? [10:51:06.0000] <barnabywalters> the <version> is hypothetical, it's using the <img> tag as a wrapper I'm interested in [10:51:14.0000] <Zunflappie> Look good, only the <version> is strange. [10:51:35.0000] <Zunflappie> </img> isnt needed, but should be a big thing to learn [10:51:36.0000] <barnabywalters> sure, as I say, I think that needs to be given more thought as to naming/exact syntax [10:51:56.0000] <Zunflappie> Indeed, it sounds / reads good [10:52:07.0000] <barnabywalters> Zunflappie: the </img> is the whole point — using the <img> as a fallback and also the wrapper for alternate sources [10:52:30.0000] <barnabywalters> e.g. we change <img> to be a non-self-closing tag (or whatever terminology) [10:53:19.0000] <staydecent> barnabywalters: I like your proposition. Having the img element as the parent makes sense to me [10:53:32.0000] <MikeSmith> barnabywalters: there are serious backward-compatibility problems [10:53:38.0000] <Zunflappie> Euh... yeah. Not like <hr> but as <caption>. And maybe <video> also? <video src="normal.avi"><... src="large"></video> [10:53:41.0000] <barnabywalters> MikeSmith: I thought as much ;) [10:54:15.0000] <staydecent> damn, backwards compatibility is no fun. [10:54:28.0000] <MikeSmith> http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cimg%20src%3D%22Mobile.jpg%22%20alt%3D%22%22%3E%0A%20%20%20%20%3Cversion%20src%3D%22med-res.jpg%22%20media%3D%22%22%20%2F%3E%0A%20%20%20%20%3Cversion%20src%3D%22hi-res.jpg%22%20media%3D%22%22%20%2F%3E%0A%3C%2Fimg%3E [10:55:50.0000] <gorh> hi there [10:56:02.0000] <barnabywalters> MikeSmith: I'm missing something here — what is the problem you're highlighting in that page? [10:56:35.0000] <Zunflappie> There was nog closing-tag for the option (like </option>) [10:56:43.0000] <MikeSmith> your first version element becomes a sibling of the img, not a child, and the second version element becomes a child of the first one [10:57:04.0000] <MikeSmith> we can't change the parsing behavior for img [10:57:04.0000] <Zunflappie> <img src="..."><option src="..."></option></img> could work? [10:57:09.0000] <barnabywalters> ah, okay [10:57:38.0000] <Zunflappie> So both <options> <captions> are first-generation-children of <img>? [10:58:20.0000] <ShaneHudson> barnabywalters: I *really* like that syntax (<img><version></img) but I expect there would be some major problems wtih compatibility [10:58:35.0000] <barnabywalters> Zunflappie: even if we added closing tags to the source selection tags, they'd still be siblings, not children of <img> [10:58:43.0000] <jgraham> The only solution involving <img> that can work is to put all the sources in an attribute [10:59:02.0000] <jgraham> You *can't* change the parsing of <img> [10:59:14.0000] <jgraham> (or <image>) [10:59:20.0000] <Zunflappie> Such as <img src="small.jpg" medium="middle.jpg" large="giant.jpg"> ? [10:59:38.0000] <barnabywalters> jgraham: indeed. Looks like using <img> isn't likely to work [10:59:56.0000] <Zunflappie> <picture> as alternative/extra is then a good thing? [10:59:59.0000] <gorh> hé hé i see i arrived in the right discution :) [10:59:59.0000] <jgraham> Zunflappie: There is a better proposal (imho) on the whatwg list [11:00:01.0000] <ShaneHudson> It will not work as attributes [11:00:07.0000] <staydecent> Sorry to diverge a bit here, but this article is confusing: http://www.w3.org/community/respimg/2012/05/11/respimg-proposal/ -- Which group proposed the picture solution? [11:00:23.0000] <jgraham> ShaneHudson: Why not? Did you see the proposal on that whatwg list? [11:00:25.0000] <ShaneHudson> The syntax for set for example looks horrid and would be hard for even us to get to grips with, let alone beginners [11:00:41.0000] <barnabywalters> ShaneHudson: agreed. set="" is horrible [11:00:52.0000] <Zunflappie> Staydecent: thats the page why I turned here in. The <picture> looks good to me, as <img> is still there. [11:01:08.0000] <gorh> i have to say that i, too would prefer by far <picture><source ... to the attribute way, at least for the sake of my eyes ) [11:01:30.0000] <staydecent> I agree as well. I am in favour of <picture> over the set attr. Just confused as to who has proposed what. [11:01:34.0000] <ShaneHudson> Yeah I have been following <picture> for quite a while. I do not believe it is the best soloution, but certainly superior to set="" [11:01:55.0000] <barnabywalters> the tag based approach would also be easier to generate/parse in code [11:01:59.0000] <Zunflappie> Is is possible to work only with <img>'s? <picture><img src="small" media="min-width: 400px;"><img src=""><img src=""></picture> [11:02:12.0000] <ShaneHudson> Nope [11:02:14.0000] <Zunflappie> Or does that render always 3 images (so small, middle and large?) [11:02:24.0000] <staydecent> yeah for older browsers would render all 3 [11:02:24.0000] <ShaneHudson> Since that would (on older browsers at least) render all of them [11:02:57.0000] <gorh> tbh i can cope with throwin <img> tag away ;) [11:03:26.0000] <barnabywalters> gorh: not if you want to support older browsers, surely? [11:03:59.0000] <ShaneHudson> Wouldn't it be nice to start with a clean slate! [11:04:19.0000] <barnabywalters> ShaneHudson: indeed :( That'd be no fun though :) [11:04:21.0000] <staydecent> Lol, Legacy is never fun. But necessary. [11:04:39.0000] <gorh> yep sure, but this can be coped as well lest say by hiding img via css for regular users [11:05:32.0000] <gorh> but i see the point on keeping the legacy browsers in the scope [11:05:33.0000] <staydecent> gorh: but then the image would still be downloaded by the user [11:05:46.0000] <Zunflappie> Exactly. What about IE6 or 10 year old machines (not callling them phones...). <picture> is like <video>, what is a good point. But i like a 3-letter-shortcut (like IMG for IMAGE) more. <VID> maybe? [11:06:02.0000] <staydecent> gorh: a goal of responsive images is not to force a mobile phone to download more than it needs to [11:06:11.0000] <Zunflappie> Hiding with CSS doesnt prevent downloading its contents! [11:06:19.0000] <barnabywalters> Zunflappie: that's interesting, I prefer full words to the abbreviations [11:06:35.0000] <Zunflappie> So you like <division> and <italic>? [11:06:45.0000] <Zunflappie> Or what about <tr>???? [11:06:48.0000] <barnabywalters> Zunflappie: within reason :) [11:06:56.0000] <gorh> yep sure, but i say do that by css, but there are plenty of other way to do so [11:07:15.0000] <barnabywalters> okay, <table-row> or <abbreviation> are a bit of a mouthful, but I think <video> is fine [11:07:16.0000] <Zunflappie> Yeah, Server-side. Otherwise it will get a HTTP-request [11:07:28.0000] <staydecent> gorh: CSS can't prevent a device from downloading the orginal image [11:07:41.0000] <gorh> css can't [11:07:58.0000] <gorh> but don't tell me that there is no other way ) [11:08:03.0000] <Zunflappie> Javascript either..... only php/ruby on rails >> server, as the html should be altered [11:08:27.0000] <Zunflappie> And then... you must sniff the device first. That's not RESPONSIVE! [11:10:58.0000] <Zunflappie> Hereby its all said? [11:11:58.0000] <ShaneHudson> It is a shame there is not yet anyway to do responsive images without a multitude of them. Keep the original <img> tag and automatically make "responsive" [11:12:45.0000] <barnabywalters> ShaneHudson: I think someone has done that on the server side with cookies, generating a different copy of the image [11:12:52.0000] <gorh> no but infact i'm not that much attached to the <picture> tag, what annoys me the most is the attribute way [11:13:03.0000] <Zunflappie> <ShaneHudson: There is a manner with Javascript. But it takes also PHP for the correct sizes/urls [11:13:27.0000] <ShaneHudson> I am thinking browser side.. I have seen adaptive images etc. but none of them seem perfect for the job [11:13:43.0000] <ShaneHudson> Maybe it will come when we have bandwidth sniffing or something like that [11:13:50.0000] <Zunflappie> As there cant be a horse.jpg (small), horse.jpg (medium) and a horse.jpg (large)..... Or it must be in folders: .../small/horse.jpg and .../medium/horse.jpg etc [11:14:11.0000] <Zunflappie> that IO.<something) has that sniffing? [11:14:16.0000] <gorh> and, i also realise that it mainly is about code readability ... wile i should focus on the usability [11:14:23.0000] <ShaneHudson> The only one I know is foresight.js [11:14:37.0000] <Zunflappie> See http://css-tricks.com/which-responsive-images-solution-should-you-use/ >> [11:14:53.0000] <Zunflappie> Sencha.IO (based on foresight.js) [11:14:56.0000] <Jayflux> Can anyone tell me the point of the <embed> tag? all browsers seem to work fine without it when the attributes are put on the object tag instead, it also validates properly. [11:15:32.0000] <ShaneHudson> Jayflux: I believe that was to do with ie5.5? May well be wrong though [11:15:42.0000] <gorh> but, tbh, with daily uses, i prefer from far working on readable code rather than looooooong tags [11:15:53.0000] <ShaneHudson> defintely [11:16:01.0000] <Jayflux> so it may aswell be obsolete ShaneHudson ? [11:16:12.0000] <Jayflux> in terms of today usage [11:16:32.0000] <ShaneHudson> Jayflux: I believe so, but it is not an area I pay much attention to tbh [11:17:09.0000] <Zunflappie> What is 'tbh'? Probaly not The Birthday Hub.... :P [11:17:47.0000] <gorh> to be honest [11:18:56.0000] <Zunflappie> ok! When does w3c comes with a poll? [11:19:26.0000] <gorh> what i mean is that, the <picture way is more developer friendly, and the <img set...> one is more browser friendly, and i'm still no sure if i prefer or to have to deal with both syntax and compatibility, and have something easy to read [11:19:54.0000] <gorh> or to have each img tag filling lines and lines of not explicit code [11:21:06.0000] <barnabywalters> gorh: I think one of the great things about HTML is it's ease of use. It's really easy to understand, and it'd be a pity to lose that, especially on something as fundamental as an image [11:22:47.0000] <staydecent> Is there any course of action to help picture get implemented over img set? I guess being in this IRC? [11:23:07.0000] <Zunflappie> Good point barnabywalters! So always maintain current syntaxes [11:23:09.0000] <gorh> barnabywalters: i totaly agree on that this is why i like the <picture> way ) [11:23:53.0000] <Zunflappie> And more: its like the new <video> and <audio>..... make <embed> and <object> the same syntax? [11:24:04.0000] <ShaneHudson> staydecent: http://www.w3.org/community/respimg/ [11:24:08.0000] <gorh> Zunflappie: you really find the attribute way more clear ? [11:24:10.0000] <Zunflappie> As a browser supports html5, it can be doen [11:24:30.0000] <staydecent> ShaneHudson: Thanks [11:24:33.0000] <barnabywalters> Zunflappie: actually, the similarity to <video> and <audio> is something I'm uncomfortable about [11:24:39.0000] <Zunflappie> Gorh: like <img src="small.jpg" medium="middle.jpg" .......> atc? [11:25:52.0000] <gorh> as simple as that that would be a dream ) [11:25:53.0000] <Zunflappie> <video> <audio> and <picture> are all media? More then text can't video/audio/picture realign or crop. Audio is special (as it doenst need take space in a page) [11:25:58.0000] <barnabywalters> as video and audio's <source> choses different codecs based on browser compatibility, whereas proposed picture <source> choses a different size based on screen size — so very different meanings to the same syntax [11:26:09.0000] <gorh> but what about breakpoint to choose image etc etc [11:26:13.0000] <lorin> Zunflappie: xlarge jumbo gojira [11:26:36.0000] <barnabywalters> gorh: Yes, <img src="" medium="" etc is inflexible. In fact, html used to have lowsrc="" or something [11:26:52.0000] <barnabywalters> so that'd almost be a step backwards :) [11:27:02.0000] <Zunflappie> lorin: thats a point. Anno 2012 we want 3 sizes... but what if I want 4 or 5? lowsrc="" works a bit, but aint flexible. It was only a idea. [11:27:10.0000] <gorh> one tag to define 3 or more image path and beahvior ... it implies a tag too long in my opignon [11:27:27.0000] <Zunflappie> <picture> with media-queries are better.... but i hate the syntax of the queries. MIN and MAX could be better for me. [11:28:03.0000] <barnabywalters> Zunflappie: agreed, I'm not sure about using media queries as the syntax [11:28:20.0000] <barnabywalters> as they assume that small screen size = smaller bandwidth [11:28:51.0000] <ShaneHudson> which is completely wrong. My home wifi is slower than 3G lol [11:28:57.0000] <Zunflappie> Thats not a problem, but more when i resize my window [11:29:10.0000] <Zunflappie> Shane: haha, fix your Wifi then :P [11:29:36.0000] <ShaneHudson> House is too far away from the exchange sadly. But luckily here at uni I am on fibre :D [11:29:46.0000] <barnabywalters> ShaneHudson: I get that sometimes. Mine has really bad range, I have to use 3G when I'm upstairs sometimes [11:30:36.0000] <Zunflappie> Oke. I hope i haved helped the whatwg to choose for <picture> or something else. Nevermind.... but please: choose fast. [11:30:46.0000] <Zunflappie> Its 20:30... time for Top Gear (and beer and chips) [11:31:27.0000] <staydecent> Zunflappie: Haha, enjoy! [11:32:45.0000] <gorh> long live to the <picture> tag :) i'm off a well ) [11:36:19.0000] <staydecent> Anyone willing to recap the issues/limitations of the <picture> tag besides semantics? [11:37:27.0000] <ShaneHudson> http://trac.webkit.org/changeset/111637 Image-Set in webkit [11:38:37.0000] <barnabywalters> ShaneHudson: I just found a link to that too, I can't really make much sense of how it could be used though [11:39:20.0000] <barnabywalters> ah, http://lists.w3.org/Archives/Public/www-style/2012Feb/1103.html explains it better [11:39:42.0000] <ShaneHudson> Hmm yes I see, so it is like <picture> but for css backgrounds [11:40:10.0000] <ShaneHudson> The syntax of that actually looks fairly nice [11:40:43.0000] <barnabywalters> There's some good explanations of why several options wouldn't work (including the one I suggested earlier) here: https://etherpad.mozilla.org/responsive-assets [11:41:07.0000] <ShaneHudson> *bookmarked* [11:41:20.0000] <ShaneHudson> That is a very good article [11:41:30.0000] <ShaneHudson> well, document :O [11:41:43.0000] <barnabywalters> even better, as (I think) we can edit it [11:41:48.0000] <ShaneHudson> yeah [11:42:01.0000] <staydecent> is that content pretty much the same as : http://www.w3.org/community/respimg/wiki/Main_Page [11:42:32.0000] <ShaneHudson> yeah very similar [11:43:42.0000] <staydecent> Shoot, now I'm leaning towards the image-set CSS approach: http://www.w3.org/community/respimg/2012/04/08/using-css-to-control-image-variants/ [11:44:18.0000] <staydecent> But, how could that work with CMS's? [11:44:27.0000] <barnabywalters> staydecent: it looks alright for images specified through css, I can't see anything about using it in markup [11:44:48.0000] <ShaneHudson> image-set is good for backgrounds, but not for images since they are usually part of the content rather than layout, and you would not want css for each and every image! [11:45:08.0000] <staydecent> Good points [11:45:30.0000] <padenot> /b 9 [11:45:58.0000] <barnabywalters> I suspect that image-set could work nicely alongside a markup solution though [11:51:55.0000] <zcorpan> Wilto: ok i've sent an email to whatwg now. lemme know whether it makes sense [11:52:14.0000] <Matt__> Hello [11:52:25.0000] <staydecent> Just found this <picture> Polyfill: https://github.com/scottjehl/picturefill/blob/master/picturefill.js good way to test out how the proposed implementation would "feel" [11:52:36.0000] <Matt__> Sorry, hang on while I drag up my IRC knowledge from the depths of time and figure out how to change my name.... [11:52:46.0000] <ShaneHudson> Do /nick "name" [11:53:22.0000] <Matt__> Thanks Shane [11:53:33.0000] <ShaneHudson> I take it we are all agreed that set="" is the wrong way to go? [11:53:42.0000] <Matt__> Set is the wrong way to go [11:53:58.0000] <barnabywalters> ShaneHudson: agreed [11:54:06.0000] <Matt__> Seem /nick "MattWilcox" didn't work ;) [11:54:17.0000] <ShaneHudson> really? strange [11:54:26.0000] <tomasf> try without the quotes [11:54:50.0000] <ShaneHudson> I still really like the <img><version></img> idea, shame that it would break everything! [11:55:01.0000] <Matt__> Might be something to do with being on webchat.freenode.net? I don't know, I've not used IRC in a decade or so. [11:55:20.0000] <MattWilcox> Ah ha! [11:55:23.0000] <MattWilcox> Thanks [11:55:49.0000] <ShaneHudson> Heh yeah I shouldn't have included the quotes [11:56:18.0000] <staydecent> ShaneHudson: Agreed. I also am against the set attribute. [11:56:47.0000] <MattWilcox> I don't think the picture element is perfect, it needs refinement in another revision really. [11:56:55.0000] <ShaneHudson> MattWilcox: most of us seem to like the <picture> but feel it is the wrong name? [11:57:01.0000] <ShaneHudson> What would you say needs to change about it? [11:57:05.0000] <zcorpan> for people not reading the list, please see http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2012-May/035784.html [11:57:07.0000] <MattWilcox> But it's a problem not unique to picture, but to the whole "adaptive" methodology that's in HTML at the moment. [11:58:01.0000] <MattWilcox> Shane: The problem is it being too impractical for real world use *en mass* - my thoughts were written up here: http://mattwilcox.net/archive/entry/id/1081/ [11:58:32.0000] <MattWilcox> And I proposed a solution here, which was put the the RespImg group, and HTMLWG mailing list: http://mattwilcox.net/archive/entry/id/1082/ [11:59:09.0000] <staydecent> zcorpan: Thanks, that's a good writeup. [11:59:31.0000] <ShaneHudson> Hmmm, not entirely sure I like the breakpoint syntax but it certinally does make sense [11:59:42.0000] <zcorpan> staydecent: thanks [12:00:18.0000] <staydecent> zcorpan: are the issues with <picture> shared with <video>? [12:01:12.0000] <MattWilcox> Shane: I'm open to refinement and ideas, the syntax I put there was literally a first idea. [12:01:53.0000] <MattWilcox> Thing is, what I propose is something that can be factored on-top of the existing <picture> proposal if we're desperate to ship now. [12:02:41.0000] <zcorpan> staydecent: <video><source> is very complex for much the same reasons, yes, but is different (maybe even more complex) because the loading algorithm tries to load each <source> in turn, with potential for scripts to change things around at any point [12:02:42.0000] <ShaneHudson> Yeah, I get the idea and am not sure how it could be improved off the top of my head. Seems slightly "clunky" but it would only need to be written once [12:03:47.0000] <MattWilcox> Shane - yep, that's the huge benefit of that approach. And I agree that it's a bit verbose; so is <picture> itself. [12:04:10.0000] <staydecent> zcorpan: do current implementations deal with that in anyway? More so, has work been done already to deal with parsing all the source tags? [12:04:11.0000] <ShaneHudson> Maybe from a naming point of view, case changes to name, and match changes to case? [12:04:12.0000] <zcorpan> staydecent: <picture> would try to find a "best match" source, and could decide to change the source when the environment changes, so it's not the same algorithm as <video><source> at all really [12:04:29.0000] <staydecent> zcorpan: ahhh ok. that makes sense [12:04:36.0000] <MattWilcox> Yeah, perhaps. The naming itself doesn't bother me - that's implementation detail. It's the pattern I like. [12:04:36.0000] <zcorpan> staydecent: we have suffered with bugs with <video><source> because of its complexity [12:05:17.0000] <staydecent> zcorpan: in that case and out of curiousity what is your stance on picture vs img set? [12:05:36.0000] <MattWilcox> zcorpan: Yeah, but it works, right? So why not re-use it. [12:06:03.0000] <zcorpan> staydecent: as i said in the email, i think an attribute is simpler and would be a shorter path to interop [12:06:34.0000] <barnabywalters> MattWilcox: the <video><source> model has a different semantic meaning to the proposed <picture><source> [12:06:51.0000] <barnabywalters> and, as zcorpan says, there have been bugs [12:06:53.0000] <zcorpan> MattWilcox: i doubt all browsers follow the spec to the letter with <video><source> today. i know opera has bugs, and yet i think opera is *most* closely aligned with the spec [12:06:54.0000] <MattWilcox> I am not bothered about the ease with which vendors can implement a feature. That's done *once* in the life of HTML. It is FAR more important the syntax be flexible and easy to understand because millions of authors will learn and use it every day. [12:07:15.0000] <zcorpan> MattWilcox: also, as i said, we can't reuse the same code because <video><source> is not a best match [12:07:27.0000] <staydecent> zcorpan: ah didn't realise that was you. well, this changes my opinion quite a bit. Maybe Ill spend more time thinking about possible improvments to the set attr syntax [12:08:12.0000] <MattWilcox> zcorpan: hmm, not a best match? I have a feeling it's more complex an issue than it seems from my perspective. [12:08:25.0000] <staydecent> have there been any proposals in terms of bandwidth queries? Or is that something for far-off in the future? [12:08:38.0000] <MattWilcox> Bandwidth MQs have been requested for months [12:08:48.0000] <zcorpan> MattWilcox: <video><source> tries to play each <source> in turn, and stops when it has found one that plays [12:08:57.0000] <MattWilcox> But I don't know of any talk about implementation details. [12:09:04.0000] <zcorpan> MattWilcox: <picture> wants to decide which <source> is the best one, and load that [12:09:16.0000] <staydecent> Right, but the img set attr doesn't follow MQ syntax, unless I;m mistaken [12:09:20.0000] <annevk> hober: btw, does viewport width really depend on layout? [12:09:24.0000] <MattWilcox> zcorpan: ah, OK then the beaviour is quite different despite the syntax being similar [12:09:26.0000] <zcorpan> MattWilcox: (and probably do it later again when the environment has changed) [12:09:32.0000] <annevk> hober: because that seems like something that's an input to layout and is therefore already known [12:09:38.0000] <zcorpan> MattWilcox: indeed [12:09:44.0000] <annevk> hober: and can therefore be used at fetching time easily enough [12:10:12.0000] <zcorpan> staydecent: i think using MQ is also inappropriate, but for a different reason [12:10:24.0000] <MattWilcox> zcorpan: would it not be the same if <picture> was extended so that it could deal with requesting different file formats? [12:11:03.0000] <MattWilcox> zcorpan: one thing I'd have liked <picture> to adopt is some method of making it much easier for new file formats like WebP to become practically usable - by offering fallbacks. [12:11:06.0000] <staydecent> zcorpan: I agree. I feel MQ should stay within CSS. [12:11:33.0000] <zcorpan> MattWilcox: different file formats isn't the use case we're trying to solve [12:11:38.0000] <staydecent> But I can't stand the <img set> syntax as it stands here: http://www.w3.org/community/respimg/2012/05/11/respimg-proposal/ [12:11:42.0000] <MattWilcox> zcorpan & staydecent: I thing they should be moved OUT of CSS and put predomenantly into the <head> [12:12:24.0000] <ShaneHudson> staydecent: Yeah, I can understnd why they thought of it but it would be terrible to add images like that [12:12:24.0000] <MattWilcox> zcorpan & staydecent: but with CSS having the ability to over-ride as needed. See http://mattwilcox.net/archive/entry/id/1082/ for why [12:13:18.0000] <MattWilcox> I think the RespImg Group would have looked at adapting <img> more if Hicksie hadn't told members that alterations to <img> were impossible. [12:13:34.0000] <staydecent> I'm guessing they went with "600w 200h" instead of "600x200" as to not be confused with the DPI declaration "2x" [12:13:39.0000] <staydecent> ? [12:13:59.0000] <zcorpan> the problem with MQ for the bandwidth/"load the page faster" use case is that the MQ describes a state of the UA, and the author needs to get the query "right", where it is easy to get the behavior backwards (like resulting in a lower resolution image when the user zooms *in*) [12:14:05.0000] <MattWilcox> staydecent: irrelevent. However you look at it that ties adaption to pixel units. Responsive designs do not use pixel units. [12:14:07.0000] <annevk> MattWilcox: where did Hixie say that though? [12:14:44.0000] <MattWilcox> annevk: there are a couple of people trying to find that out because it has been an accepted thing over there, brought up a few times. I've not found the original source yet. [12:14:58.0000] <zcorpan> whereas with the "dpi" syntax, the author just describes information about the image, and the user agent is free to decide which image it deems is most appropriate for the user given the environment (the browser is in a much better position to know this than the author) [12:15:13.0000] <annevk> MattWilcox: changing the parsing is impossible, adding attributes is not [12:15:37.0000] <annevk> MattWilcox: maybe there was some confusion about terminology there and what people understand as to what "parsing" means... [12:15:40.0000] <MattWilcox> See http://www.w3.org/community/respimg/common-questions-and-concerns/ and the section "What about modifying the behaviour of <img>?" [12:16:20.0000] <annevk> MattWilcox: yeah that seems about parsing [12:16:28.0000] <annevk> MattWilcox: with the "look ahead" remark [12:16:30.0000] <ShaneHudson> I agree that we have to be careful about changing <img> due to older (and current) browsers... but that does not mean we cannot touch it [12:16:51.0000] <annevk> MattWilcox: doesn't really mean you can't add new attributes, in fact, one of the proposals we have does so in a backwards compatible way [12:17:12.0000] <staydecent> MattWilcox: I don't understand, aren't you using pixel units in your proposal: http://mattwilcox.net/archive/entry/id/1082/ [12:17:49.0000] <MattWilcox> I'm sure there has been some misunderstanding about it, which is incredibly frustraiting. Browser vendor representitives were asked about it, and this is a major reason why editing <img> was ruled out. [12:17:52.0000] <annevk> Wilto: btw, it would be interesting to know who encouraged you to form a CG [12:18:35.0000] <annevk> Wilto: a CG is a fine place for discussion, but if parallel discussion happens in other groups there's no real guarantee you'll be heard unfortunately [12:18:39.0000] <MattWilcox> staydecent: I'm using CSS media. That example is pixels, but there is no reason at all why I couldn't have also used %, em, vh or anything else. [12:19:15.0000] <staydecent> ahh, so how could that flexibility be used in a html attr or element? [12:19:17.0000] <ShaneHudson> annevk: The CG was the first disccusion group/website I saw on the matter.. except blog psots of course [12:19:21.0000] <MattWilcox> staydecent: that's the brilliance of MQ. I used pixels in the example because it's easier for people to figure out than introducing stuff like vh or em which they may be less familiar with using in that way. [12:20:21.0000] <annevk> oh lol, that zeldman guy sure likes drama https://twitter.com/zeldman/status/201019688946384897 [12:20:26.0000] <MattWilcox> staydecent: we're proposing MQ syntax be added to HTML. That's how the <picture> tag works. There's nothing in my proposal that's different to the main <picture> one except I abstract the query away from the code block. [12:20:48.0000] <ShaneHudson> annevk: lol [12:21:05.0000] <staydecent> So, that syntax could be added to the <img set=""> attr? [12:21:25.0000] <staydecent> *if* one decided to take that route... [12:21:51.0000] <MattWilcox> I don't know. I can't see how the MQ syntax could be cleanly added into a singular attribute. [12:21:54.0000] <ShaneHudson> Who proposed set anyway? I have not yet seen anyone who has agreed with it, and I had not heard of it before today [12:22:06.0000] <annevk> Apple [12:22:09.0000] <MattWilcox> imgset was from someone at Apple [12:22:12.0000] <ShaneHudson> Ah, that explains it lol [12:22:20.0000] <MattWilcox> Intended for CSS, and now being pushed into HTML [12:22:24.0000] <MattWilcox> It's poor in both cases [12:22:33.0000] <MattWilcox> But less poor in CSS [12:22:35.0000] <ShaneHudson> yeah the css version I mostly agree with, but html I don't [12:22:36.0000] <MattWilcox> Still poor though. [12:22:38.0000] <annevk> I think it makes sense [12:22:42.0000] <staydecent> From what I;ve gothered today: <picture> element is easier to read and write for developers, but the implementation within browsers is more complex than the set attr.. [12:23:00.0000] <MattWilcox> staydecent: I'd agree with that summary. And I think that's not a problem. [12:23:04.0000] <barnabywalters> staydecent: that's my impression too [12:23:20.0000] <zcorpan> for the record, multiple attributes would be fine from the "impl complexity" perspective [12:23:21.0000] <annevk> for the case where you just want highres images <img src=lowres set="highres 2x"> is way easier than the alternative [12:23:33.0000] <MattWilcox> As I've said: vendors only have to worry about doing hard work once. But once it's in HTML, that's millions of people stuck with that syntax, for life. [12:23:42.0000] <ShaneHudson> Agreed, browser developers are very clever (as opposed to entry-level html) and it only needs to be implemented once [12:23:52.0000] <MattWilcox> annevk: of course it is, because it's far more targeted and far less useful. [12:24:17.0000] <zcorpan> MattWilcox: developers need to live with bugs in browsers [12:24:24.0000] <MattWilcox> annevk: you're not making a fair comparison really. [12:24:27.0000] <annevk> MattWilcox: did you see the extended version that also deals with width/height or is there some other use case? [12:24:42.0000] <MattWilcox> zcorpan: bugs get fixed. syntac doesn't. [12:24:47.0000] <MattWilcox> syntax [12:24:59.0000] <annevk> MattWilcox: well you have to consider what is going to be the common case and see how that compares [12:25:03.0000] <MattWilcox> annevk: so many other use cases! [12:25:17.0000] <MattWilcox> annevk: I know the common case, it's not retina. [12:25:30.0000] <annevk> MattWilcox: because trying to boil the ocean with a single feature generally does not work very well [12:25:38.0000] <MattWilcox> annevk: Other ways people wish to implement responsive images include... [12:25:42.0000] <annevk> see <object>, XML namespaces, ... [12:26:07.0000] <MattWilcox> annvek: ... NOT via pixels (layouts are often set in flexible units, and so are pictures) ... [12:26:14.0000] <ShaneHudson> Speaking of retina... does anyone know a good way to actually test retina without a device? [12:26:21.0000] <annevk> MattWilcox: yeah so just use SVG... [12:26:52.0000] <MattWilcox> annevk: ... and responding to screen size is only done because it's a proxy for bandwidth. MQ's are aiming to include bandwidth sensors too. And then <picture> can use them. [12:27:30.0000] <annevk> how would that work if you switch networks? [12:27:42.0000] <barnabywalters> MattWilcox: yes, one of the appeals (to me) of the <picture> element is it's extensibility [12:28:01.0000] <MattWilcox> annevk: SVG is not a raster format. Seriously, it is absolutely inapropriate for the majority of use cases. You can not SVG a photo. You can only SVG mathematically drawn images. [12:28:28.0000] <shepazu> SVG ALL THE THINGS!!!!! [12:28:31.0000] <annevk> MattWilcox: I guess I misunderstood what you meant by "NOT via pixels" [12:28:47.0000] <zcorpan> MattWilcox: as i've said, the browser is in a better position to know which image is appropriate (if it has information about the available images and their resolutions) than the author writing the MQ, even if MQ gets extended [12:29:03.0000] <MattWilcox> annvek: that's an implementation detail - but the browser has a recent connection it can query the real throughput on all the time - the last requested asset :) [12:29:37.0000] <MattWilcox> acorpan: No. The designer knows what is required to meet the requirements of the design. The browser does not. [12:30:02.0000] <barnabywalters> I'm off for the moment (going to sleep on it). Bye! [12:30:09.0000] <MattWilcox> Bye :) [12:30:28.0000] <zcorpan> MattWilcox: so you care about the adaptive layout use case, basically? not the "save bandwidth"/"load page faster" use case? [12:30:39.0000] <staydecent> zcorpan: Could developers avoid altering <picture><source> elms by listening to some Event? [12:30:54.0000] <MattWilcox> annvek: and yeah, I mean using pixel in a MQ is not actually recommended - most responsive designs are not based on pixels. But the assets that get served certainly are pixel based. [12:31:15.0000] <annevk> MattWilcox: the width/height proposal is not MQ based [12:31:26.0000] <annevk> which I think is a plus really, not all of MQ makes sense here [12:31:35.0000] <MattWilcox> zcorpan: I care about both, but the saving bandwidth is done via the design. Not via automated heuristics on the browser. [12:31:45.0000] <annevk> and putting bandwidth testing in MQ seems certainly way complicated to author [12:31:52.0000] <zcorpan> staydecent: browsers need to get the edge cases right even if developers could avoid it, because developers could also not avoid it, and on the web scale all edge cases will happen, and browsers must deal [12:32:23.0000] <zcorpan> MattWilcox: ok. MQ is appropriate for the layout use case. I'm arguing it is *not* appropriate for the bandwidth use case [12:32:45.0000] <MattWilcox> zcorpan: From the perspective of a designer the process goes like this: Detect the device size -> Design something to fit this device -> Implement this design as HTML/CSS/JS [12:33:29.0000] <MattWilcox> zcorpan: all the bits that we talk about are the last bits in the stack. It's thought about backwards a lot of the time, and we end up making the tools we have dictate the designs that are applied. It's the wrong way. [12:33:57.0000] <staydecent> zcorpan: I agree with that sentiment. But with my lack of knowledge, I find I often run into bugs in my scripts by not doing something in the right order or event. [12:34:18.0000] <MattWilcox> zcorpan: the design is the final experience. Everything else is tooling to achieve the design. [12:34:53.0000] <MattWilcox> zcorpan: and of course, a good design consideres the users context (device size, connection speed, etc). [12:34:57.0000] <staydecent> zcorpan: My point being, it's already part of my practice to make sure I do things with proper timing/within proper events. [12:36:06.0000] <zcorpan> staydecent: not sure i follow how this relates to the topic at hand :) [12:36:24.0000] <jarek> "SVG is not a raster format. Seriously, it is absolutely inapropriate for the majority of use cases. You can not SVG a photo." [12:36:40.0000] <jarek> in other words: SVG is suitable for all kinds of graphics on the web except photos [12:36:52.0000] <staydecent> zcorpan: RE: "browsers need to get the edge cases right even if developers could avoid it" [12:37:33.0000] <jarek> I mean gradients, buttons, icons, overlays - all this should be done with SVG, not bitmaps (as it is today) [12:39:02.0000] <jarek> it's sad that there are so many CSS3 fanboys, but nobody cares about SVG [12:40:01.0000] <staydecent> jarek: do you have any recommneded links/libraries/examples/tutorials for creating those elements with SVG? [12:40:03.0000] <shepazu> jarek: hey, I care about SVG! and I know 18 other people who do, too [12:40:11.0000] <ShaneHudson> Do you have any examples of using SVG for those? I must admit I focus much more on css and don't do much with svg [12:40:25.0000] <MattWilcox> Jarek: I have built websites for the last 8yrs, professionally, for various clients in various sectors with various companies. I can tell you that the vast majority of designed images are not achievable in SVG. [12:40:56.0000] <shepazu> MattWilcox: that seems like a pretty broad claim [12:41:44.0000] <MattWilcox> shepazu - look on some websites. Find me 5, just FIVE content images that are vectors. Now do the same but look for photo's. [12:42:45.0000] <MattWilcox> shepazu: remember we're talking content images here, not site-design elements, which are applied via CSS. We're on about HTML. So, it's stuff that clients add to sites in their CMS's. [12:42:46.0000] <shepazu> MattWilcox: http://yourlogicalfallacyis.com/burden-of-proof [12:43:02.0000] <MattWilcox> shepazu: oh grow up. [12:43:10.0000] <jarek> MattWilcox: with filter effects you can achieve most Photoshop effects [12:43:17.0000] <shepazu> MattWilcox: http://yourlogicalfallacyis.com/ad-hominem [12:43:21.0000] <MattWilcox> shepazu: it's not about effects. It's about content type. [12:43:30.0000] <jarek> MattWilcox: also, gradient meshes are on the roadmap which will allow as to create photo-realistic artworks [12:43:43.0000] <shepazu> MattWilcox: if it can be done in Illustrator, it can be done in SVG [12:43:50.0000] <jamesr> SVG is not useful for all icons/buttons/etc [12:43:52.0000] <jarek> staydecent: they problem with SVG is that we don't even have good authoring tools. Inkscape is a mess, Illustrator... :/ [12:43:57.0000] <MattWilcox> Yes, and most <img> content can not be donein illustraitor [12:44:19.0000] <shepazu> depends on your design aesthetic [12:44:19.0000] <MattWilcox> OK, have a look at all img's on the following websites: http://www.flickr.com/ [12:44:44.0000] <MattWilcox> http://en.wikipedia.org/wiki/Motor_racing#Motor_racing [12:44:51.0000] <MattWilcox> NOT DESIGN. Content. [12:44:52.0000] <ShaneHudson> shepazu: Design aesthetic has nothing to do with this at all... this is about use cases. SVG is brilliant for a few use cases but not for most [12:44:55.0000] <MattWilcox> We are doing CONTENT images. [12:45:17.0000] <MattWilcox> http://www.appliancesonline.co.uk/ [12:45:26.0000] <MattWilcox> http://dashes.com/anil/2011/07/if-your-websites-full-of-assholes-its-your-fault.html [12:45:29.0000] <shepazu> MattWilcox: nobody was claiming that you should do raster images in SVG, jarek mentioned gradients, buttons, icons, overlays, and other UI elements [12:45:31.0000] <MattWilcox> These are just some open tabs. [12:45:38.0000] <MattWilcox> Look at the <img> in them [12:45:45.0000] <MattWilcox> Almost none are possible in SVG [12:46:05.0000] <staydecent> MattWilcox: Your ponint makes sense. SVG is a red-herring in this discussion. [12:46:18.0000] <jamesr> and you frequently want different assets for icons at different resolutions (not just a simple scale up) [12:46:19.0000] <MattWilcox> UI elements are perfect places for SVG. But UI elements are defined in CSS - they are not embedded assets. [12:46:28.0000] <MattWilcox> Agreed. [12:46:31.0000] <shepazu> diagrams, charts, infographics, etc. are also great uses of SVG [12:46:37.0000] <MattWilcox> Those are not use cases applicable to the <img> tag, which is what we are discussing [12:47:11.0000] <mdelcx> hey all [12:47:14.0000] <zcorpan> jamesr: http://my.opera.com/ODIN/blog/2009/10/12/how-media-queries-allow-you-to-optimize-svg-icons-for-several-sizes [12:47:17.0000] <shepazu> MattWilcox: well, that's what you were discussing… Jarek brought up another topic that the rest of us were discussing [12:47:37.0000] <MattWilcox> Content images are <img> - the vast majority of all <img> throughout the web are added via some form of CMS, and are *not* vector based. [12:47:51.0000] <ShaneHudson> shepazu: diagrams etc are perfect for svg, and most people that know of svg know what it is good for [12:48:25.0000] <shepazu> and for <img> content that is a UI element, or is a digaram, chart, etc., SVG is reasonable for those use cases [12:48:26.0000] <jarek> yeah, I did not mean that there should be no support for responsive images [12:49:09.0000] <MattWilcox> To clarify: SVG is perfect for any vector. But you are not going to apply those as content images ( <img> ) 99% of the time, if only because you have to be a developer to know how to get them in. I don't know of any CMS that allows you to upload a vector image and that then outputs it *as* a vector. [12:49:25.0000] <MattWilcox> Agh, it's late and I must go! [12:49:31.0000] <shepazu> I strongly believe that we need a responsive image mechanism for rasters [12:49:43.0000] <shepazu> s/believe/agree/ [12:49:50.0000] <MattWilcox> I think everyone singing of the same sheet to be honest, bar small nuances that don't matter too much [12:50:05.0000] <staydecent> Must focus on work. Good discussion :) [12:50:10.0000] <MattWilcox> Cheers all for the conversation and viewpoints :) [12:51:42.0000] <mdelcx> I've been mulling this over today... has any thought been given to adding client "metadata" at the request level? [12:52:11.0000] <mdelcx> handling the serving of assets using the application, rather than repurposing media queries and modifying the markup [12:52:18.0000] <mdelcx> (in the client) [12:53:25.0000] <mdelcx> if the client added headers to every request, it would be a snap for the application to handle it [12:53:25.0000] <zcorpan> mdelcx: so basically conneg? [12:54:00.0000] <zcorpan> mdelcx: i think someone has considered it and written an email on whatwg explaining why it's a bad idea [12:54:01.0000] <mdelcx> zcorpan: i haven't been part of the conversation, so I'll have to take a look at that WG [12:54:34.0000] <zcorpan> mdelcx: basically, it would result in more overhead for all websites, even if only a small fraction would use the information [12:55:21.0000] <zcorpan> mdelcx: and it would increase finger printing [12:55:39.0000] <adiabatic> I hear there's a new proposed syntax for responsive images or similar from Hixie that Zeldman doesn't like much. Does anyone know offhand where it might be the whatwg list archives? I'm not exactly sure what I should be plugging into Google to find it [12:55:43.0000] <mdelcx> overhead, for sure [12:56:03.0000] <mdelcx> zcorpan: fingerprinting would depend on what the client had access to send [12:56:35.0000] <mdelcx> it shouldn't be any worse than a cookie [12:57:08.0000] <zcorpan> mdelcx: any new piece of information about the user increases finger printing [12:57:33.0000] <zcorpan> mdelcx: even if one piece isn't "worse" than a cookie, taken together they easily can identify a single user [12:57:46.0000] <zcorpan> mdelcx: we want to minimize it as much as possible [12:57:56.0000] <mdelcx> eh, I don't know if some client metadata and a cookie could identify a client [12:58:01.0000] <mdelcx> but i see you point for sure [12:58:23.0000] <adiabatic> Sure, but if you can piece together enough sufficiently weird client data… [12:58:26.0000] <zcorpan> cookie isn't the only thing there is today :-) [12:58:36.0000] <mdelcx> i know :) [12:59:15.0000] <adiabatic> Incidentally, is there a way to enumerate all the fonts that a user has on a system? I think I'm unique simply based on the fonts I have installed alone [13:00:10.0000] <zcorpan> adiabatic: you can't enumerate them, but you can list lots of fonts, try to apply them and measure the text width to see if the font was applied or not [13:31:30.0000] <zcorpan> i feel like it's time for a new #csspubquiz but i can't think of anything. maybe i should start with html quizzes instead [13:39:03.0000] <benvie> re: fonts that's not even falling back on flash which *will* give you a full list of fonts [14:03:24.0000] <adiabatic> benvie: heh, one more reason why I'm glad the only Flash I have on my system is bundled with Chrome [14:06:20.0000] <zcorpan> you're glad that flash is bundled with chrome? :-P [14:47:37.0000] <benvie> but yeah the amount of vectors towards making a UID for people are multitude =D [14:57:28.0000] <annevk> btw [14:57:40.0000] <annevk> a bunch of people submitted feedback as @ replies to @whatwg [14:57:51.0000] <annevk> dunno really what to do with those [14:58:09.0000] <kennyluck> sigh [14:58:09.0000] <kennyluck> j [14:58:10.0000] <kennyluck> ust [14:58:30.0000] <annevk> suggestions welcome, I'll get some sleep meanwhile [14:58:47.0000] <kennyluck> just count them as -1 to the current proposal  [14:59:05.0000] <annevk> we don't care about that [14:59:39.0000] <annevk> this is not the IETF [15:02:22.0000] <Wilto> Some fledgeling developers might not be so familiar with IRC or mailing lists, but likely just wanted to be heard as well. [15:02:49.0000] <Wilto> I might not -1 them for trying to participate. Just saying. [15:03:56.0000] <annevk> Wilto: kennyluck was not saying that and I was not saying that either [15:04:43.0000] <kennyluck> I was saying that counting them is at least better than ignoring them at all. [15:04:50.0000] <annevk> Wilto: in particular I was asking for suggestions on how to deal with them, since we mostly use @WHATWG as a broadcast channel, we don't really have discussions over twitter with that account [15:05:30.0000] <annevk> kennyluck: but I was not suggesting to ignore them... [15:05:58.0000] <annevk> anyway, nn [15:06:45.0000] <Wilto> Oh, oh—voting against the _current_. My mistake, man. [15:07:18.0000] <Wilto> annevk: Totally redirect them to the CG. I tried to do that when people started a-bikesheddin' in here, too. [15:08:37.0000] <kennyluck> Wilto, yeah, that would work, I guess. [15:08:59.0000] <gsnedders> I guess a fair few of us use here to bikeshed in to avoid doing so on the mailing list :P [15:10:34.0000] <Wilto> It is the place for the community to get their discuss on—I mean, I don’t deny that it’s good to give people a forum in which to bikeshed to their little hearts' content. It's the internet, after all. [15:11:10.0000] <kennyluck> Wilto, I agree. We need more places like that for purely bikeshedding issues. [15:11:10.0000] <Wilto> I enjoy a good bikeshedding from time to time myself. Red, if anyone's wondering. [15:11:11.0000] <kennyluck> Fore [15:11:13.0000] <kennyluck>  [15:11:14.0000] <kennyluck>  [15:11:15.0000] <kennyluck> For [15:11:15.0000] <kennyluck> exm [15:11:43.0000] <Wilto> Yeah. It's like herding cats in there, but somebody's gotta do it. [15:11:46.0000] <kennyluck> For example, 'display: flex' vs. 'display: flexbox' for flex box is purely bikeshedding. [15:12:08.0000] <kennyluck> so are a lot of CSS stuff :p [15:12:12.0000] <Wilto> I vote `flexxx`, me. Sounds hip and edgy. [15:12:28.0000] <Wilto> Starring Vin Diesel. [15:12:40.0000] <Wilto> Okay, I'm getting punchy. You guys have a good weekend, yeah? [15:13:03.0000] <kennyluck> Wilto, you too. 2012-05-12 [05:48:39.0000] <annevk> https://twitter.com/WHATWG/status/201293054546685952 [06:37:02.0000] <bga_> Google replaces GET params to json [06:37:04.0000] <bga_> http://e2a66cn9997j2srbkidin2psq2dusjaa-a-fc-opensocial.googleusercontent.com/ps/ifr?viewParams={%22displayLink%22:true, ... ,%22FONT_FACE%22:%22normal%20normal%2012px%20'Trebuchet%20MS',%20Trebuchet,%20Verdana,%20sans-serif%22}} [06:39:43.0000] <bga_> web is losing common protocol [06:40:09.0000] <bga_> external program can not parse it completely [06:40:28.0000] <bga_> :/ [07:51:35.0000] <charlvn> bga_: i have seen similar implementations, that would not be the only one (although probably from the most major institution) [08:03:11.0000] <bga_> charlvn young devs that does not care about common standards disappoint me. RFC is law book of internet [08:34:42.0000] <annevk> bga_: what's wrong with using custom format in GET parameters? [09:08:15.0000] <kennyluck> I think asking people to join the WHATWG mailing list is futile, given that the Responsive Images Community Group actually uses the blog system for discussions... [09:08:28.0000] <kennyluck> And now the CG site is down *shrug* [09:10:50.0000] <annevk> kennyluck: you're welcome to try something else [09:11:11.0000] <kennyluck> annevk, I was going to suggest the CG site but now it's down :( [09:11:53.0000] <annevk> the CG is currently about 90 people whereas the WHATWG is 1500 [09:12:04.0000] <annevk> not sure it's helpful to require all those to subscribe to the CG... [09:16:09.0000] <Wilto> annevk, kennyluck: Nah, I wouldn't expect everyone to. Anyone deeply rooted in the issue should take some time to familiarize themselves with the work of the CG, at least. Otherwise, I'm happy to act as the CG's representative on the mailing list. [09:17:18.0000] <Wilto> It just seems like the most practical approach. [09:17:55.0000] <annevk> Wilto: I do get the feeling the CG has been misinformed about a number things, 1) purpose of CGs 2) mutability of <img> 3) prefetching logic [09:17:59.0000] <annevk> which is kind of sad :( [09:18:34.0000] <annevk> I just noticed there's even an article on A List Apart with the same (mis)information [09:18:34.0000] <Wilto> 2 and 3 may very well be. I hope to work with you guys on clarifying those points, for certain. [09:18:52.0000] <Wilto> annevk: Yeah. I wrote it, based on the information we had at hand. [09:19:05.0000] <annevk> oh and fwiw, typos are just as likely with media queries [09:19:06.0000] <Wilto> I wish more members of the WHATWG had been involved in the CG from the start, for that reason. [09:19:26.0000] <Wilto> Sure. But we’re a lot more apt to spot those. MQ are familiar. [09:19:58.0000] <Wilto> As for item 1, I was under the impression that CGs were intended to allow the community to work together on finding solutions, and contribute to the standards process. [09:19:59.0000] <annevk> I doubt most authors know media queries [09:20:23.0000] <Wilto> —Most authors catering to multiple screen sizes and resolutions will. Of course they will. [09:20:29.0000] <annevk> when they got some traction at one point they looked mostly misused :( [09:20:52.0000] <annevk> people using device-width to detect iPhone/iPad [09:20:54.0000] <Wilto> You can’t make a case that learning two syntaxes is more simple. It's fundamentally flawed. [09:21:06.0000] <jgraham> If I have the right meme, familiar thing is familiar [09:21:09.0000] <Wilto> That doesn’t seem related. [09:21:20.0000] <Wilto> But this isn’t about some developers making poor choices. [09:21:33.0000] <annevk> sure it is, it's about all developers [09:21:40.0000] <Wilto> This is about an unfamiliar syntax, leading to more points of failure, leading to a poor experience for _users_. [09:21:41.0000] <jgraham> Having a syntax that doesn't let you do insane things is good for ergonomics [09:21:47.0000] <annevk> can't just ignore the ones that do the wrong thing [09:22:04.0000] <Wilto> You’re saying this proposed syntax is completely free from potential abuse. [09:22:04.0000] <jgraham> Even if it means not reusing a more general syntax that would let you do the insane thing [09:22:15.0000] <Wilto> Let’s just assume developers can do things wrong in either case. That’s not a stretch. [09:22:30.0000] <jgraham> I'm saying it's tailored to meet the use cases [09:22:51.0000] <Wilto> And in any case, if you were to convince me personally that developers would prefer the proposed syntax, it doesn't matter. [09:23:12.0000] <Wilto> I sincerely hope you saw the comparison thread in the WG, when it was live. [09:23:15.0000] <annevk> Wilto: no [09:23:29.0000] <Wilto> Developers are almost unanimously in favor of <picture>. I lined to it in my last post on the mailing list. [09:23:30.0000] <jgraham> I haven't seen the comparison thread yet [09:23:32.0000] <Wilto> linked* [09:23:46.0000] <Wilto> Once it’s back up, it's a worthwhile read. [09:23:58.0000] <jgraham> But this isn't design-by-democracy aka design-by-committee [09:24:21.0000] <Wilto> jgraham: You're toeing a dangerous line, there. [09:24:29.0000] <Wilto> The developers should absolutely have a say in these matters. [09:24:30.0000] <jgraham> Not really [09:24:33.0000] <Wilto> And they are emphatic on this. [09:24:34.0000] <jgraham> Oh [09:24:41.0000] <jgraham> That was a reply to the first line [09:24:51.0000] <jgraham> Of course people should be free to present arguments [09:24:58.0000] <jgraham> Including ones based on ergonomics [09:25:33.0000] <Wilto> The fact remains that I am speaking for the developers I've worked with on this matter for almost a year now and every "+1" that comes along with that. [09:26:07.0000] <jgraham> Sure. But the important thing to communicate is the arguments they have presented [09:26:15.0000] <jgraham> Not how many people agreed with them [09:26:49.0000] <Wilto> This is the developer's preference, which trumps implementor convenience. And if this syntax stands to introduce more developer error, it isn't best for the user. That's the important thing, above all else. [09:27:19.0000] <jgraham> Sure, developer ergonmoics are an important consideration [09:27:23.0000] <Wilto> But discussing it in here likely isn't a convenient use of anyone's time. Keeping it on the mailing list is probably best. [09:27:28.0000] <jgraham> They are not the only consideration of course [09:27:55.0000] <Wilto> I suppose I haven't seen a clear argument made for the proposed change either, apart from "this is easier to implement." [09:28:15.0000] <annevk> it's a lot simpler to author as well [09:28:30.0000] <Wilto> Authors seem to disagree. [09:28:38.0000] <Wilto> As an author, I personally disagree. [09:29:38.0000] <annevk> it would be interesting to do some usability testing [09:29:42.0000] <Wilto> I'm sorry; there is no case to be made that this is somehow the developer preference. We can work together with that information and find a solution, or the WHATWG can choose to ignore it. [09:29:51.0000] <jgraham> Well I can see that it could be worse if you want to manipulate the srcset with script for example. Although I don't know if there is a use case for that, and it could be trivially made better with a DOM api like srcList [09:30:05.0000] <annevk> Wilto: I don't think you can claim to represent all authors [09:30:15.0000] <annevk> well you can I guess [09:30:16.0000] <Wilto> annevk: Of course not. [09:30:34.0000] <jgraham> Wilto: So far it seems that at least some of your preference might be based on miscommunication (the prefetch thing) [09:30:37.0000] <Wilto> I _assumed_ my role in this would be to share the prevailing sentiment from developers. [09:30:51.0000] <Wilto> That has nothing to do with syntactical preference, jgraham. [09:31:25.0000] <jgraham> Presumably it affects your overall preference though [09:31:49.0000] <Wilto> I suppose I don't entirely understand the "not invented here"-esque resistance I'm encountering here. [09:32:08.0000] <Wilto> I would like to think that developers stand to bring a great deal of valuable information to this discussion. [09:32:18.0000] <annevk> Wilto: it's not at all that, it's just that a design not based on <img> is way more complicated [09:32:23.0000] <jgraham> I think the problem is that you are thinking in terms of NIH [09:32:27.0000] <Wilto> annevk: For whom? [09:32:40.0000] <annevk> for everyone [09:33:16.0000] <Wilto> annevk: What’s more complicated and error prone: an especially long sentence in a language one understands, or an especially short one in a language one doesn't fully understand? [09:33:18.0000] <jgraham> Wilto: I seriously suggest not thinking of this as an us-vs-them situation, and taking the time to present the arguments that you have [09:33:49.0000] <jgraham> and being open to arguments from people who are coming to the problem afresh, or from a different perspective [09:33:53.0000] <Wilto> jgraham: That is exactly what I've wanted. Somehow, it seems that all this has been framed as a defense of the syntax that we've been working on — the burden of proof is on us. [09:34:07.0000] <Wilto> jgraham: I'd like to share what we've learned with those people. [09:34:51.0000] <Wilto> But again, all we're going to do here is dig in our heels further. It's best if we hash these things out on the mailing list. [09:35:05.0000] <jgraham> Wilto: My suggestion is that you post a mail to WHATWG presenting an alternative proposal and giving as many pros/cons relative to hober's proposl as you think are relevant [09:35:41.0000] <Wilto> jgraham: I will, for certain. That's been the plan. [09:36:04.0000] <jgraham> Great. When you do that I think we can have a more productive discussion :) [09:36:19.0000] <Wilto> I certainly hope so. [09:37:56.0000] <Wilto> And if I've come across as "attacking," I promise you all it wasn't my intention. I just want to make sure we're all on equal footing, so that we can work together on getting this solved. [09:38:31.0000] <Wilto> My written tone kinda sucks, too. That's on me, and I apologize if I've seemed confrontational here. [12:06:05.0000] <divya> /msg wilto yoyo [12:06:13.0000] <divya> Oops :) [13:54:08.0000] <Wilto> Protocol question for you guys, if anyone’s here. [13:54:53.0000] <webben> ? [13:55:25.0000] <Wilto> So, I’ve assembled all my details and use-cases and such at https://github.com/Wilto/respimg/#adaptive-image-element [13:56:02.0000] <Wilto> What would be less obnoxious: posting it wholesale to the mailing list, or preparing a quick summary with links to the important sections? [13:56:54.0000] <Wilto> I’m almost certain the latter, but didn’t know if having it all the information centralized on the list might be easier for everyone. [14:04:34.0000] <webben> dunno that either's obnoxious [14:05:02.0000] <Wilto> Well, suppose that’s fair. What’s the best way to present this information, then? [14:05:20.0000] <Wilto> Happy to serve this up in whatever way is most helpful to everyone. [14:05:54.0000] <webben> Wilto: Posting a link to the page seems fine. So does converting it to plain text and putting it in an email with a link. [14:06:12.0000] <webben> The advantage of the later is it makes it easier to quote and respond. [14:06:42.0000] <webben> Wilto: but hey, it's markdown so it's not too much trouble either way. [14:06:43.0000] <Wilto> That’s fair. Didn’t want to post a whole novel to the list if that’s frowned-upon. [14:06:45.0000] <Wilto> Thanks! [14:07:14.0000] <webben> yw [14:32:15.0000] <jgraham> Wilto: Post the whole thing, but I reccomend reordering it so that the use cases are at the top, not the proposal [14:33:22.0000] <Wilto> Yeah, that makes sense. Also, I might post it to the WHATWG wiki—some GitHub readme floating out in the aether probably isn’t the most practical thing in the world. [14:34:59.0000] <Wilto> I’ll also be making it less… spec-ish. I was just winging it based on real specs. [14:52:03.0000] <Wilto> Hixie: Sorry; would you mind setting me up a WHATWG Wiki account? [14:52:14.0000] <Wilto> Hixie: mat⊙mc [15:06:37.0000] <Hixie> done [15:09:29.0000] <othermaciej> Wilto: I believe your proposed sample solution to "3.4. High-Resolution Displays" is not complete [15:09:58.0000] <othermaciej> Wilto: you need to explicitly set a size or image-resolution for the higher-defeinition images, otherwise they will just be treated as larger, not higher resolution [15:11:33.0000] <othermaciej> (if setting image-resolution it would require a stylesheet rule embedded in a media query @-rule) [15:12:31.0000] <othermaciej> Wilto: also, it would be really useful to identify whether there are any important use cases besides resolution adaptation and available screen width adaptation [15:14:54.0000] <tantek> network bandwidth / reliability adaptation? [15:15:12.0000] <tantek> happens often in mobile use-cases [15:16:21.0000] <othermaciej> I'm sure there's lots of potential other use cases, it just happens that Wilto picked ones which are also arguably covered by <img srcset> [15:17:38.0000] <tantek> othermaciej - hence I'm contributing those specific use-cases [15:18:09.0000] <othermaciej> hopefully the use case page gets added to a wiki so it's easy for folks to extend [15:18:19.0000] <tantek> if there are other potential use cases that others care about, they can contribute them. their potential existence does not refute or diminish the significance of the actual existence of the specific use cases I provided. [15:18:25.0000] <tantek> agreed [15:19:08.0000] <tantek> document all the use-cases! [15:22:55.0000] <Wilto> Hixie: Thanks! [15:23:02.0000] <Wilto> And still a work in progress; grain of salt and all. [15:23:21.0000] <Wilto> I noticed `img set` seemed to cover specific widths— 200w, in the example markup. [15:24:34.0000] <Wilto> I’m not sure if the plans include min-width, max-width, etc. I don’t think it’s a matter of one pattern out… adapting the other, so much as it’s a matter of making sure all our bases are covered with either. If we’d be extending `set` to cover all the same things media queries do, maybe media queries are the better option. [15:25:09.0000] <Wilto> `200w` just seemed very specific, at face value. [15:26:50.0000] <Wilto> I’ve added a few more while reformatting the document to be more wiki-appropriate. I’ll have something final posted soon. [15:36:11.0000] <othermaciej> Hixie's proposed width/height semantics for imgset come along with a selection algorithm [15:36:18.0000] <othermaciej> so it doesn't mean just that one width [15:37:10.0000] <othermaciej> it picks the widest that will fit in the available space, if I recall correctly [15:37:36.0000] <Wilto> Based on the container, or the viewport? [15:37:46.0000] <Wilto> Sorry; I’ve just been going on http://junkyard.damowmow.com/507. [15:37:50.0000] <Wilto> I can leaf back through the mailing list. [15:38:13.0000] <othermaciej> "The algorithm here could be to sort the images by width, and remove all [15:38:13.0000] <othermaciej> those that are wider than the available width (except for the widest one [15:38:14.0000] <othermaciej> if they're all too wide) or that don't have a width unless none have [15:38:15.0000] <othermaciej> widths" [15:38:19.0000] <othermaciej> it depends on what "available width" means [15:38:33.0000] <othermaciej> if it means the viewport, then it does the same thing as media query width selection [15:38:42.0000] <Wilto> Right. [15:39:05.0000] <othermaciej> if it means the available layout space, then it's a huge pain to implement and would do something not achievable by media queries [15:39:12.0000] <othermaciej> I don't really know which was intended [15:40:14.0000] <othermaciej> I also don't know which is more useful to authors [15:40:28.0000] <Wilto> Oh, here’s something I’ve been turning over in my head: [15:41:19.0000] <Wilto> Should I put together use cases based on specced behavior? Obviously this is just an example, but the `monochrome` media query could _theoretically_ be used to serve a monochrome image. [15:41:40.0000] <othermaciej> the best way to put together use cases is to base them on things people actually want to do [15:41:42.0000] <Wilto> But in my experience, no browser really pays that media query any mind. [15:42:03.0000] <othermaciej> regardless of whether a given proposal supports them [15:42:16.0000] <Wilto> Yeah, I’ve been sticking to real-world examples. I mean, that certainly stands to reason. Just checking. [15:42:38.0000] <othermaciej> I doubt any substantial number of authors is interested in creating and serving separate monochrome images for monochrome displays [15:42:49.0000] <Wilto> I've worked with way too many wacky mobile browsers to believe in the phrase "in a perfect world." [15:43:04.0000] <Wilto> othermaciej: Of course. As I said, that was obviously just an example. [15:47:12.0000] <Hixie> the algorithm is actually in the spec now, fwiw [15:47:26.0000] <Hixie> the only thing i haven't specced is some mechanism for the browser to automatically flip in a new image on the fly [15:47:58.0000] <Hixie> which is hard because it means doing an async network fetch and then switch it in a stable state, which is non-trivial to spec [15:48:46.0000] <Wilto> I can only imagine. [15:49:44.0000] <Wilto> As media queries are expanded over time, can we assume that this disparate method of detecting client information will be updated in parallel? [16:23:30.0000] <Wilto> othermaciej: In example 3.4, the size of the image would be controlled through CSS. Or am I not following you? [16:28:42.0000] <othermaciej> Wilto: yes, you could control it through CSS, the point is that it won't give the right result without specific additional CSS (whereas for example the imgset proposal could handle scaling and lay out based on intrinsic size) [16:29:57.0000] <othermaciej> Hixie: when you wrote the algorithm spec how did you operationalize "available width"? [16:30:05.0000] <othermaciej> we weren't sure what was intended from the rough draft [16:30:12.0000] <Wilto> Ah, okay. [16:33:36.0000] <Hixie> othermaciej: width of the img element's containing block [16:33:41.0000] <Hixie> othermaciej: or some such [16:34:04.0000] <othermaciej> Hixie: so, that's awkward because it means you can't start loading the image (or preloading for that matter) until after you do layout [16:34:37.0000] <othermaciej> Hixie: a version based on viewport/window width would not have the same issue and could even participate in prefetching [16:35:23.0000] <othermaciej> not starting the load until after first layout would have a significant negative effect on page load performance, based on my experience with these things [16:36:44.0000] <Hixie> othermaciej: a version based on viewport width wouldn't really handle the use cases, but i'll keep that in mind [16:36:48.0000] <Hixie> gotta go [16:37:30.0000] <othermaciej> good point, to the extent that you are describing the image rather than writing a rule list [16:57:53.0000] <Wilto> othermaciej: So the higher density image is rendered within the intrinsic dimensions of the original src if no w/h values are specified, yeah? [16:59:06.0000] <othermaciej> I don't know if the w/h are supposed to affect intrinsic size of the image as Hixie is drafting it [16:59:11.0000] <othermaciej> but the resolution selection does 2012-05-13 [17:11:24.0000] <Wilto> This is… very confusing; I’m sorry. [17:12:32.0000] <Wilto> I don't claim to be any smarter than the average developer, but that in and of itself should say a great deal. [17:13:23.0000] <Wilto> I've been trying to write this markup as though it were for real use cases, and it is just inscrutable. [17:16:27.0000] <Wilto> That's speaking as a core contributor to jQuery Mobile, and having worked on the responsive BostonGlobe.com designs—tailoring things for client capabilities is kind of my thing. [17:16:35.0000] <Wilto> I'll revisit it with fresh eyes later, I suppose. [17:20:50.0000] <Wilto> Now that the Community Groups are back online: there’s no small amount of consensus on that point. http://www.w3.org/community/respimg/2012/05/11/respimg-proposal/#comments [17:41:26.0000] <abarth> is there a charter for the WHAT community group? [17:41:35.0000] <abarth> i can't find the link [17:42:39.0000] <abarth> maybe its the same as the WHATWG charter? [17:45:53.0000] <othermaciej> there is not [17:45:58.0000] <othermaciej> Community Groups don't have charters [17:46:05.0000] <othermaciej> also they can self-create an operating agreement [17:46:10.0000] <othermaciej> which is similar to a charter [17:46:19.0000] <othermaciej> but the WHATWG CG has not done so, at least so far [17:56:06.0000] <|Kellan|> Hi Everyone, I am redesigning my companies home page. We are a movie company. I am having a hard time deciding what is semantic for our photo gallery on our front page. Our page has 4 major sections. The first contains an image slider with our featured movies. The next section is a photo gallery of our 12 latest releases. 3rd section is photos of the stars we represent then finally DVD movie covers. Article doesn't seem right. I chose [17:56:06.0000] <|Kellan|> section but I am not sure it is right either. [18:22:35.0000] <abarth> othermaciej: thx [18:22:49.0000] <abarth> othermaciej: what defines the scope of a community group? [18:23:10.0000] <abarth> maybe they don't have a scope and the patent protections are on a per-document basis? [18:27:22.0000] <abarth> ah, there is a CLA and an FSA [18:44:17.0000] <othermaciej> abarth: the CLA and FSA are voluntary agreements that can be applied to any CG product, if a CG does not have an explicit operating agreement, then "whatever CG members want to work on" is about it [18:46:43.0000] <abarth> do you know of a CG that has an example of a well thought out operating agreement? [18:51:49.0000] <abarth> looks like these folks have a bunch of text at least http://www.w3.org/community/native-web-apps/charter/ [18:53:52.0000] <abarth> looks like most don't have an operating agreement (at least not in an obvious place on their pages) [19:21:15.0000] <othermaciej> I would like the WHATCG to have a well thought out operating agreement, though I am not sure if the distraction and debate likely to arise in the process of trying to create one would exceed the benefit [19:55:54.0000] <Hixie> abarth: the WHATWG CG's "charter" is the same as the WHATWG's, because it _is_ the WHATWG. The CG part is nothing but a way to provide the framework for patent licensing. [20:00:51.0000] <Hixie> othermaciej: so using the viewport dimensions instead of the containing block dimensions could work, i guess, since there's presumably a 1:1 relationship between viewport dimensions and the dimensions you'd expect the containing block to be [20:00:59.0000] <abarth> i wasn't sure if CGs needed to declare their scopes (like IETF working groups) [20:01:29.0000] <Hixie> abarth: there's a one-paragraph description, but no, they're pretty bureaucracy-free [20:07:39.0000] <othermaciej> Hixie: it depends on whether the w and h represent the image size or a desired target size [20:07:48.0000] <othermaciej> if the latter, then yeah, viewport size is totally adequate [20:08:51.0000] <Hixie> othermaciej: what's the difference? [20:09:21.0000] <Hixie> othermaciej: in the existing text, it's the size of the image. but i would change it to the max size of the viewport for the image. [20:09:25.0000] <othermaciej> well, if w and h indicate the size of the image, how would the algorithm combine that with a viewport size to pick one? [20:09:29.0000] <othermaciej> ok [20:09:43.0000] <othermaciej> yeah, it would have to be max size of the viewport or something along those lines [20:10:13.0000] <Hixie> othermaciej: the problem is that it makes it hard to reuse the src="" attribute for a default fallback -- in the existing text i can use the height="" and width="" and src="" attributes to provide a default so you don't have to repeat yourself [20:10:32.0000] <Hixie> othermaciej: but the perf issue is pretty serious [20:16:13.0000] <tantek> othermaciej - what's an example of what you would consider a good, well thought out, operating agreement? [20:17:25.0000] <Hixie> the empty string [20:17:31.0000] <tantek> LOL [20:19:13.0000] <tantek> /me is not even sure what is meant by an "operating agreement" in the context of a CG. [21:09:48.0000] <abarth> tantek: My understanding is that it's something you need to agree to in order to join the CG [21:21:33.0000] <Hixie> is there an equivalent of indexOf() on String that starts at a given index? [21:44:46.0000] <abarth> Hixie: indexOf takes a start offset [21:44:51.0000] <abarth> Hixie: it's the second parameter [21:44:55.0000] <Hixie> oh, good to know [21:45:00.0000] <Hixie> i should have looked it up! [00:39:27.0000] <tantek> /me reads logs to catch-up [00:41:19.0000] <tantek> abarth - a-ha, in that case, a good minimal CG operating agreement would be to require that all participants contribute everything via CC0 and OWFa FSA, per my recommendations here: http://tantek.com/2011/240/b1/w3c-community-groups-opportunities-suggestions-challenges#cg-opportunities - cc: othermaciej [01:32:41.0000] <AryehGregor> tantek, there are already licensing agreements required when you join any CG. [02:33:22.0000] <othermaciej> tantek: it's not clear to me from your blog post what benefits that would have over the existing CG CLA and FSA [02:37:33.0000] <othermaciej> tantek: in fact, the CG FSA and CLA look somewhat similar to the OWFa ones in their effect (including the copyright license) [08:22:52.0000] <tantek> AryehGregor, right, the W3C once again created a (yet another) custom (W3C-specific) license, I know that, duh. That doesn't dispute my post. [08:23:16.0000] <tantek> othermaciej, the advantages of CC0 and OWFa have already been well established. (just like MIT etc.) [08:23:25.0000] <Ms2ger> jgraham, do you know if people have thought about testing worker apis? [08:23:41.0000] <tantek> the biggest ones are a) standard licensing (as opposed to bespoke W3C licensing), which leads to b) more portability. [08:24:09.0000] <tantek> No spec should be beholden to any one specific standards organization. Standards organizations should be a means to the ends of producing useful specs, nothing more. [08:24:24.0000] <tantek> They should be tools/services, not "owners". [08:25:15.0000] <Ms2ger> This is the W3C we're talking about, right? [08:25:17.0000] <tantek> Thus if the services/tools fail for any reason, or if better tools/services emerge, those working on the spec should have the freedom to move the spec to different tools/services. That's what you get with standard (organization independent) licenses. [08:25:50.0000] <tantek> Ms2ger - scroll up re: CGs / licensing / operating agreement. Yes this is re: my blog post about suggestions of how to better use CGs. [08:26:01.0000] <tantek> (link in the logs ;) ) [08:26:21.0000] <Ms2ger> I know [08:26:24.0000] <Ms2ger> I'm cynical. [08:27:00.0000] <tantek> this is why for example, Annevk and I put CC0 and OWFa on the Fullscreen spec. [08:28:27.0000] <Ms2ger> I note that neither is on the spec right now. [08:29:38.0000] <tantek> Ms2ger that's because Ian Jacobs demanded they be removed before the CG could publish the spec. [08:29:50.0000] <Ms2ger> Thank you for making my point. [08:29:53.0000] <tantek> This was a reversal of the policy as understood before (that CGs were allowed to multilicense) [08:30:18.0000] <tantek> so that's in dispute. we plan on putting those licenses back. [08:30:35.0000] <tantek> but for now, just wanted to publish a draft rather than to wait for that fight to finish. [08:32:00.0000] <tantek> also, I've worked with Mozilla's lawyers on this (had them review CG CLA vs. CC0 vs. OWFa), and I can say that yes, our lawyers determined that not only was it *ok* (compatible) to multilicense specs with CG CLA, CC0, OWFa, but that it's a *good thing* to do so. I'm working on a blog post accordingly with more details. [08:33:00.0000] <tantek> So yes, I can say that it is Mozilla's position (legal reviewed) that specs are best multilicensed with standard licenses, specifically CC0 and OWFa (in addition to whatever bespoke license any particular organization, like W3C CG CLA, that's required). [08:33:09.0000] <tantek> We're not giving up on this. [08:36:58.0000] <Ms2ger> Good to hear that [08:37:27.0000] <Ms2ger> I'll believe in the results when I see them [09:31:12.0000] <Ms2ger> /me likes how the Blob constructor should ignore the first element in the array per spec [10:54:51.0000] <othermaciej> tantek: CG FSA and OWFa FSA have word-for-word identical copyright licensing terms [10:57:54.0000] <othermaciej> tantek: the material difference appears to be in patent licensing, where the CG version seems superior, and where it would not be fun to figure out if the two are agreements that it is possible to offer simultaneously (since one is a royalty-free license and the other is a non-assert promise) [11:01:28.0000] <othermaciej> (sorry about the broken grammar there, hope the point is clear) [11:09:44.0000] <Hixie> don't the copyright terms even allow relicensing, or something? so long as attribution is given? [11:10:02.0000] <Hixie> if it's gpl-compatible it doesn't really matter what the terms are exactly, it's good enough for all intents and purposes [11:10:31.0000] <Hixie> (though certainly it's my opinion, and google's opinion, that furthering license proliferation is a mistake the w3c keeps making) [11:38:16.0000] <jgraham> Ms2ger: zcorpan has written some tests We should release them, but they predate testharness.js (by a few weks iirc) [11:39:44.0000] <jgraham> *weeks [11:42:11.0000] <espadrine> Ms2ger: this Blob constructor thing is a bug, right? They mean "0 <= i < a.length", do they not? [12:22:32.0000] <Ms2ger> jgraham, mm, MS also submitted some [12:23:03.0000] <Ms2ger> I was wondering about a way to write worker tests without manually messing with postMessage [12:28:50.0000] <Ms2ger> espadrine, do they? :) [12:29:56.0000] <espadrine> ok ^^ [12:39:12.0000] <jgraham> Ms2ger: Oh, well we should release ours then and fix up the harness afterwards [12:39:30.0000] <Ms2ger> wfm [12:39:38.0000] <jgraham> Do Mozilla have any tests? :) [12:52:09.0000] <Ms2ger> Probably [12:52:39.0000] <Ms2ger> http://mxr.mozilla.org/mozilla-central/source/dom/workers/test/ [15:39:41.0000] <Velmont> tests tests 2012-05-14 [18:12:44.0000] <othermaciej> wow this srcset/<picture> thread got huge [18:14:41.0000] <othermaciej> for those discussing it previously: the CG FSA copyright license is non-exclusive (so presumably authors should be allowed to multi-license if they care to) and also identical to the OWFA copyright license (thus not really an instance of license proliferation) [18:14:54.0000] <othermaciej> it is also pretty much maximally permissive afaict, only requires attribution [18:25:10.0000] <othermaciej> tantek: one thing I'm curious about (maybe to be covered in your blog post) - how is it possible to multi-license with CC0, since CC0 places a work in the public domain and therefore presumably makes it not copyrighted and not subject to a copyright license? [18:26:09.0000] <othermaciej> I'm not a copyright expert so it's not clear to me if this is a genuine contradiction or not [19:30:46.0000] <jamesr_> any firefox mobile peeps around? [22:57:25.0000] <AryehGregor> othermaciej, if you read the text of CC0, it's really just an irrevocable royalty-free license to use the work however one pleases. So it's trivially compatible with any other license, in the sense that all other licenses are redundant, IIUC. [22:58:20.0000] <AryehGregor> Personally I have no problem with requiring attribution, but IANAL and don't know if the CG copyright license is actually compatible with other standard permissive licenses, so I'd prefer to keep CC0 in addition to be safe until I get clarification from lawyers. [22:58:31.0000] <AryehGregor> Preferably approval by the FSF or OSI. [22:58:45.0000] <othermaciej> AryehGregor: I read it - it claims to abandon copyright and place the word in the public domain [22:58:50.0000] <othermaciej> "To the greatest extent permitted by, but not in contravention of, applicable law, Affirmer hereby overtly, fully, permanently, irrevocably and unconditionally waives, abandons, and surrenders all of Affirmer's Copyright and Related Rights and associated claims and causes of action..." [22:59:10.0000] <othermaciej> it seems like if you surrender all copyright, you can't also grant a copyright license to the same work [22:59:22.0000] <othermaciej> but again, this may be a naive and uninformed understanding [22:59:42.0000] <AryehGregor> Ah, I see. You also have this: "Should any part of the Waiver for any reason be judged legally invalid or ineffective under applicable law, then the Waiver shall be preserved to the maximum extent permitted taking into account Affirmer's express Statement of Purpose. In addition, to the extent the Waiver is so judged Affirmer hereby grants to each affected person a royalty-free, non transferable, non sublicensable, non exclusive, irrevocable [22:59:42.0000] <AryehGregor> and unconditional license . . ." [22:59:42.0000] <AryehGregor> Ah, I see. You also have this: "Should any part of the Waiver for any reason be judged legally invalid or ineffective under applicable law, then the Waiver shall be preserved to the maximum extent permitted taking into account Affirmer's express Statement of Purpose. In addition, to the extent the Waiver is so judged Affirmer hereby grants to each affected person a royalty-free, non transferable, non sublicensable, non exclusive, irrevocable [22:59:43.0000] <AryehGregor> and unconditional license . . ." [23:00:18.0000] <othermaciej> yes, in jurisdictions where it is impossible to abandon copyright and place a work in the public domain, it is instead a permissive license [23:00:52.0000] <othermaciej> but holding copyright is, by my understanding, a status that is jurisdiction-by-jurisdiction [23:02:07.0000] <AryehGregor> In any event, it is accurate to say that any work licensed under CC0 can also be used under the terms of X, where X is any license whatsoever. [23:02:32.0000] <othermaciej> FSF does not seem to have examined the OWFA FSA or the W3C CG FSA [23:02:40.0000] <AryehGregor> Which is the idea of multi-licensing, right? If I release a GPL work that includes some BSD content, I can accurately say that the work is GPL-licensed, because the whole work can be used under the terms of the GPL. [23:03:12.0000] <othermaciej> I don't understand how it is possible to grant a copyright license to a work in the public domain [23:03:49.0000] <othermaciej> since I am not a lawyer, I cannot do more than (a) notice my confusion; and (b) consult an attorney [23:04:06.0000] <AryehGregor> If it's actually in the public domain, then I'd think you can't "grant" a license, no. [23:04:29.0000] <AryehGregor> But it doesn't really matter. If the work is marked as CC0, then any additional licenses are redundant and therefore harmless. [23:05:15.0000] <othermaciej> what I would expect is that they are not only redundant but ineffective [23:05:27.0000] <othermaciej> but of course public domain grants you all the rights that any possible copyright license would [23:05:48.0000] <AryehGregor> What would it mean for it be "ineffective"? [23:07:02.0000] <othermaciej> I mean that you can't grant a copyright license to a work not under copyright, as I understand it [23:08:48.0000] <AryehGregor> Right, doing so makes no sense, but is also harmless. I mean, I could write a piece of paper that says "I hereby grant othermaciej the right to vote in United States presidential elections" and give it to you. It would be stupid, and not technically correct, but in the end it makes no difference to anything. [23:22:31.0000] <tantek> othermaciej - exactly don't bother reasoning about it since you're not a lawyer. I've asked our (Mozilla's) lawyers and according to them, yes, it is totally fine for a CG to require contributors agree to all three (CC0, OWFa, CG CLA), and then multilicense specs accordingly. [23:23:04.0000] <tantek> no conflicts, and such multilicensing is compatible and preferable. [23:23:36.0000] <AryehGregor> Do they have an opinion on whether the CG CLA is GPL-compatible? [23:23:40.0000] <AryehGregor> Or BSD-compatible, for that matter? [23:23:49.0000] <tantek> I didn't ask them to evaluate that [23:24:05.0000] <tantek> we do know that CC0 is GPL and BSD compatible however. [23:28:57.0000] <AryehGregor> Yes, trivially. [23:30:28.0000] <AryehGregor> Which is why I told Ian Jacobs that I'd prefer to keep the editing spec CC0-licensed to make sure that it's GPL-compatible until I have confirmation from a lawyer that it's not an issue. [23:30:56.0000] <AryehGregor> It can definitely be useful to copy-paste spec text into source code comments. [23:34:04.0000] <tantek> agreed. [23:34:33.0000] <tantek> Frankly I don't think Ian Jacobs has any right to privately ask for CC0 multilicensing to be removed from CG specs. [23:35:11.0000] <AryehGregor> I just declined. [23:35:28.0000] <tantek> I've already posed the recommendations/questions in my blog post of many months ago, and am still waiting for answers. Until policies are officially clarified with URLs on w3.org, I think the right thing to do is to keep the CC0 and OWFa multilicensing on there. [23:37:44.0000] <tantek> BTW to be clear - I've raised the issue, and my (and Mozilla's) recommendation of using multilicensing on CG specs directly to the AB at their f2f meeting a few months ago in San Mateo (I was a one-time invited expert to the meeting). I didn't get *any* negative feedback about that course of action. [23:40:10.0000] <AryehGregor> My editing spec was CC0-licensed before it moved to a CG, and no one objected to that at any point. [23:40:33.0000] <tantek> AryehGregor, same with Fullscreen. [23:58:00.0000] <othermaciej> tantek: my plan is to ask Apple's in-house counsel since this does not appear to be an issue that can be evaluated intuitively [23:58:38.0000] <othermaciej> I do think CG spec contributors should be allowed to multi-license in whatever way they choose [23:58:58.0000] <othermaciej> not sure why the w3c would say otherwise, and it seems against the spirit of CGs to refuse [00:00:57.0000] <othermaciej> I don't personally have a problem with anyone who wants to using CC0 but I do not think I'd push for it or use it myself, as it's not clear to me that removing the attribution requirement is worth the extra license complexity [00:16:42.0000] <AryehGregor> I don't have a problem with the attribution requirement, I have a problem with using a license that's not widely recognized. [00:16:46.0000] <AryehGregor> I'd be fine with BSD. [00:17:23.0000] <hsivonen> /me thinks it was a mistake that CC killed the license variants that didn't require attribution [00:17:49.0000] <AryehGregor> CC0 doesn't require attribution. :) [00:17:50.0000] <AryehGregor> CC0 doesn't require attribution. :) [00:18:01.0000] <hsivonen> now there are workarounds like "you can attribute the wiki instead of individuals" [00:18:27.0000] <hsivonen> AryehGregor: CC0 - The best CC license. :-) [00:19:25.0000] <hsivonen> I'm saddened that the i18n group is filing bugs asking for complexity without use cases [00:19:31.0000] <hsivonen> e.g. asking for non-ISO weeks [00:19:59.0000] <hsivonen> why bother if people in non-ISO-week countries don't actually use week numbers on forms [00:21:47.0000] <AryehGregor> Who uses week numbers on forms at all? [00:22:40.0000] <hsivonen> AryehGregor: the rationale was "European business" [00:23:08.0000] <hsivonen> AryehGregor: Personally, I've spoken about week numbers in "European business" setting but haven't seen them on a Web form ever, IIRC [00:24:41.0000] <hsivonen> anyway, non-ISO weeks are so crazy that IMO we shouldn't support them even if it means that we need to drop ISO week out of "fairness" or something [00:24:50.0000] <wilhelm> I use week numbers all the time. But not <input type='week'>. [00:26:49.0000] <hsivonen> /me is unhappy about the mismatch of the usage of the word "weekend" in the U.S. and the start day of the week in calendars that are laid out one week per row [00:28:19.0000] <hsivonen> even more unhappy that Ubuntu in en-US *language* does not have a pref for making weeks start on Monday in the calendar that pops out if you click the menubar clock [00:29:18.0000] <hsivonen> /me is generally unhappy about the coupling of en-US language and calendar stuff in software [00:30:52.0000] <charlvn> hsivonen: just checked and kubuntu doesn't seem to have that option either, although i also can't say it bothers me that much [00:32:21.0000] <AryehGregor> hsivonen, the most common convention in the US is that calendars are laid out one week per row with Sunday as the first day. [00:32:25.0000] <AryehGregor> Oh, that's your point. [00:32:57.0000] <hsivonen> AryehGregor: right. [00:33:23.0000] <AryehGregor> In Israel, the weekend is Friday/Saturday, so we're good. :) [00:34:00.0000] <AryehGregor> (historically Saturday was the seventh day according to everyone, AFAIK, but I guess lately no one cares about the religious significance outside of Israel) [00:35:24.0000] <charlvn> haven't tested this but http://tuxtweaks.com/2008/12/change-the-week-start-day-in-ubuntu/ [00:36:28.0000] <hsivonen> charlvn: thanks. though I wonder if that breaks when software is updated or something [00:37:48.0000] <hsivonen> for some reason, Ubuntu has an English - Denmark locale. Maybe that would make the calendar ISO-compliant while using untranslated UI strings [00:38:17.0000] <jgraham> /me is unhappy about week numbers. [00:39:08.0000] <wilhelm> Why? [00:40:12.0000] <jgraham> Because they don't really make any sense (see: confusion about zero-point) and I never know which week it is. It always seems less ambiguous to use dates [00:40:48.0000] <jgraham> (week numbers are never used in the UK. This may affect my bias) [00:41:45.0000] <hsivonen> they never make sense unless both parties of the communication are looking at a calendar that renders week numbers *and uses Standard ISO week numbers* [00:42:49.0000] <hsivonen> example of FAIL: hotel in Portugal using ad hoc week numbers and expecting guests from parts of Europe where people expect ISO weeks to be the standard EU-wide [00:43:06.0000] <wilhelm> jgraham: They make perfect sense in countries where they are actually used. (c: [00:43:32.0000] <jgraham> I live in a country where they are actually used and it just means that I am permanently confused [00:44:01.0000] <jgraham> Or, at least, unable to work out when things are happening [00:45:09.0000] <othermaciej> do ISO weeks start on Monday? [00:45:45.0000] <hsivonen> othermaciej: yes [00:46:31.0000] <AryehGregor> Week numbers are also never used in the US. [00:46:35.0000] <othermaciej> is that the way calendars are usually laid out in Europe? (it's been a while since I've seen a non-American calendar) [00:46:38.0000] <AryehGregor> (that I've ever seen) [00:46:47.0000] <othermaciej> I've certainly not seen week numbers used for anything [00:46:59.0000] <AryehGregor> Android lays out weeks starting with Monday. It annoys me slightly. Especially since I actually live in the one country where the workweek starts on Sunday. :) [00:47:21.0000] <hsivonen> othermaciej: at least over here (Finland), calendars are laid out with Monday in the first weekday column and Sunday in the last weekday column [00:47:53.0000] <hsivonen> (typically with the *ISO* week number as the first column to the left of the Monday column) [00:48:55.0000] <hsivonen> AryehGregor: so you've moved from New York to Israel? [00:48:59.0000] <AryehGregor> hsivonen, yes. [00:49:11.0000] <AryehGregor> Currently Jerusalem. [00:49:25.0000] <othermaciej> given the way week numbers are used in europe, would it be plausible to input one in a Web form, even if it is not usually done> [00:49:26.0000] <othermaciej> ? [00:50:17.0000] <hsivonen> othermaciej: I suppose it's *plausible*, since people do agree to do stuff on week N. [00:50:35.0000] <jgraham> Yeah, although the idea of it horrifies me :) [00:50:35.0000] <jgraham> Yeah, although the idea of it horrifies me :) [01:01:03.0000] <wilhelm> Here's a Scandinavian calendar: http://www.timeanddate.no/kalender/ [01:06:43.0000] <hsivonen> whoa. Norway and Denmark take the Monday off for Pentecost [01:09:22.0000] <hsivonen> annevk, Ms2ger, Hixie: Now that cloning a node has moved out of the HTML spec, dealing with the "already started" flag can't be traced from the cloning algorithm [01:09:59.0000] <hsivonen> so basically you have to know what to look for in order to know what the cloning steps need to do to find additional normative statements beyond DOM Core [01:11:04.0000] <jgraham> Oh, that sounds extremely bad [01:11:07.0000] <annevk> it has been argued HTML and DOM should be the same spec [01:11:12.0000] <annevk> but I'm not sure if that's better [01:11:40.0000] <annevk> the normative reference crowd would not be pleased [01:12:50.0000] <annevk> hsivonen: we can probably add a note under it that points to HTML [01:12:59.0000] <annevk> hsivonen: if you think that works please file a bug [01:13:08.0000] <annevk> jgraham: "extremely" [01:13:27.0000] <annevk> jgraham: it's just a search in some other platform specs that deal with nodes, there's not that many [01:14:12.0000] <hsivonen> annevk: ok. I'll file a bug [01:15:37.0000] <jgraham> annevk: The expectation that people implementing any feature should do a full search of specs other than the one that defines the feature is unreasonable and will lead to problems [01:18:16.0000] <annevk> it kind of depends on how you implement and define the feature really, it's quite clear here that there are other requirements elsewhere [01:18:27.0000] <annevk> but I guess that's a fine argument to make against modularization [01:21:57.0000] <hsivonen> filed https://www.w3.org/Bugs/Public/show_bug.cgi?id=17044 [01:22:51.0000] <AryehGregor> The DOM spec should create a hook for the HTML spec to alert readers that it's an extension point, probably, or something like that. [01:22:56.0000] <AryehGregor> With a note pointing to HTML. [01:23:06.0000] <AryehGregor> (or we should just have mutual normative dependency, but . . .) [01:23:18.0000] <annevk> it has a hook for more than HTML [01:23:24.0000] <annevk> since it's prolly needed by SVG too [01:23:32.0000] <annevk> and the shadow stuff I guess [01:23:48.0000] <annevk> although maybe the shadow stuff should move into DOM [01:24:45.0000] <AryehGregor> Ah, I see, it does have a hook. [02:04:45.0000] <AryehGregor> Nice -- Google autosuggest noticed I live in Israel and when I typed "xhr", suggested "סיר". [02:05:02.0000] <AryehGregor> Probably an average Israeli is more likely to have forgotten to switch their keyboard layout than to actually search for "xhr". [02:20:27.0000] <odinho> AryehGregor: good it was only a suggestion. -- not always nice being on vacation and google playing smart all the time. [02:26:38.0000] <jgraham> Google is very annoying. If I want to search in English it picks Swedish and if I want to search in Swedish it picks English (I guess it is also right sometimes, but I never notice those times) [02:27:16.0000] <pcs> Google may already know you are interrested in IT (https://www.google.com/settings/ads/onweb/). Maybe the first suggestion should have been XmlHttpRequest. [04:08:26.0000] <annevk> argh [04:08:31.0000] <annevk> why did https://www.w3.org/Bugs/Public/buglist.cgi?product=WebAppsWG&component=XHR stop working? [04:08:48.0000] <[tm]> annevk: bug [04:08:59.0000] <[tm]> this is the second time I've seen this [04:09:10.0000] <annevk> someone renamed XHR to XHR 2.0 [04:09:19.0000] <zcorpan> data:text/html,<video controls style="background:lime;width:150px"> - seems safari/chrome/gecko get this wrong [04:09:35.0000] <[tm]> annevk: oh [04:09:38.0000] <[tm]> not me [04:09:47.0000] <annevk> the person who did broke a bunch of links [04:09:54.0000] <annevk> can we change it back? [04:09:55.0000] <[tm]> I don't remember doing that at least [04:09:59.0000] <[tm]> yeah [04:10:19.0000] <annevk> the same happened with DOM btw [04:10:33.0000] <annevk> I guess someone renamed it to DOM4? [04:10:39.0000] <annevk> this is really annoying [04:10:59.0000] <[tm]> https://www.w3.org/Bugs/Public/buglist.cgi?product=WebAppsWG&component=XHR is fixed [04:11:17.0000] <annevk> ta [04:11:18.0000] <[tm]> annevk: you want me to change DOM back [04:11:23.0000] <annevk> yeah please [04:11:35.0000] <annevk> I wonder what else got broken [04:12:01.0000] <[tm]> dunno [04:12:06.0000] <[tm]> what product is DOM under? [04:13:08.0000] <[tm]> I don't find a "DOM4" component, man [04:13:22.0000] <[tm]> DOM Core? [04:13:42.0000] <annevk> yeah sorry [04:13:50.0000] <annevk> that used to be just DOM [04:13:58.0000] <[tm]> OK, I'll change it back [04:14:32.0000] <[tm]> if there are any others, ping me [04:14:41.0000] <[tm]> headed out to lunch now [04:14:48.0000] <annevk> thank you! [04:25:30.0000] <annevk> so if we want to expand FormData, how exactly should we go about that [04:26:21.0000] <annevk> setting it on <form> could be done, but you cannot currently populate it from the server [04:26:32.0000] <annevk> although I guess if we define multipart/form-data better that would be doable [04:27:02.0000] <annevk> you cannot really do FormData[field] = value [04:27:10.0000] <annevk> because you can have multiple fields with the same name [04:28:31.0000] <annevk> and some values have both a value and a name [04:28:46.0000] <annevk> although I guess we could always make those File's [05:06:23.0000] <annevk> http://wiki.whatwg.org/wiki/FormData [05:06:47.0000] <annevk> proposed extensions to FormData XMLHttpRequest and HTMLFormElement [05:16:14.0000] <benvie> does Opera have a place where they put idls they use for producing interfaces that go into their releases/ [05:18:11.0000] <annevk> benvie: we don't do IDL [05:19:39.0000] <benvie> mm ok, to rephrase [05:20:15.0000] <benvie> is there a place that has a description of the DOM interfaces that Opera provides to javascript, preferably in a format that facilitates machine generation? [05:21:47.0000] <benvie> (this being for use in javascript code itself, not externally) [05:22:05.0000] <annevk> maybe http://caniuse.com/ has something like that? I don't think we produce it [05:23:22.0000] <benvie> I'm trying to expand the work I've done here to also automatically describe differences between implementations: https://github.com/Benvie/idl-for-javascript [05:24:01.0000] <benvie> right now it just produces spec based json but it's only a bit more work to also describe the differences as well [05:25:59.0000] <annevk> cool project [05:26:41.0000] <annevk> can't you just get the data for Opera by running the existing data you have through it? [05:26:51.0000] <annevk> I guess that does not give you proprietary extensions unique to Opera... [05:27:09.0000] <annevk> but I don't think we have many of those [05:27:16.0000] <benvie> yeah that's where I'm going next, generating difs based on given json description vs. what the browser shows [05:27:37.0000] <annevk> no bugs when parsing the IDL from the specifications? [05:28:20.0000] <benvie> there's a small handful of things the parser breaks at, but in all instances I've commented those out and then put the changed version that parses correctly on the next line [05:28:27.0000] <benvie> it's mostly some very new things [05:29:22.0000] <benvie> like in DOM4 for operations, the "default value" syntax has been used which the parser doesn't hndlr [05:29:41.0000] <benvie> `optional NodeFilter? filter = null` which I replace with `optional NodeFilter? filter` [05:30:03.0000] <annevk> k [05:30:25.0000] <annevk> I'm asking as I don't run the IDL through an actual checker so they may very well have bugs [05:31:04.0000] <benvie> yeah, and I'm sure with some of them there are. Many of the newer ones are generated directly in the browser by me since no official IDL exists [05:32:39.0000] <benvie> basically like `[].map.call(document.querySelectorAll('.idl'), function(node){ return node.textContent }).join('\n')` [05:32:46.0000] <benvie> but it works surprisingly well [05:36:59.0000] <annevk> what are we missing IDL definitions for? [05:39:28.0000] <benvie> let me pull up my "thing which breaks all things" and look, one moment [06:00:52.0000] <benvie> ok it's breaking too many things. Anyway, the goal wasn't primarily for that purpose. Anything that implements the specifications should be fine. It's just that IDL has been primarily used to describe both the specifications as well as the implementation differences. [06:00:53.0000] <benvie> ok it's breaking too many things. Anyway, the goal wasn't primarily for that purpose. Anything that implements the specifications should be fine. It's just that IDL has been primarily used to describe both the specifications as well as the implementation differences. [06:02:55.0000] <annevk> sorry about the ambiguous "we"; I meant what features have you covered that are not covered in any standard [06:03:45.0000] <benvie> oh, none. The purpose is to simply translate idls provided directly by w3c and whatwg as json [06:03:55.0000] <benvie> not to invent anything new [06:05:20.0000] <annevk> oh, then I do not understand what you meant by "Many of the newer ones are generated directly in the browser by me since no official IDL exists" [06:05:30.0000] <benvie> I was looking into the possibility of using the provided "official" idls from those sources as well as specs available from, for example, webkit's WebKit/WebCore/Source and mozilla's mozilla-central/dom [06:05:44.0000] <benvie> to automatically generate the differences in actual implementation [06:06:12.0000] <benvie> because the idls in those places to actually show pretty accurate where the differences are [06:07:17.0000] <benvie> so I wanted to see try and include that perspective from as many major vendors as I could while I work on this [06:07:24.0000] <benvie> there's way too many to do it manually [06:07:50.0000] <benvie> (interfaces) [07:03:13.0000] <annevk> heh [07:03:22.0000] <annevk> apparently XMLHttpRequest was authored respec at some point [07:03:26.0000] <annevk> in /\ [07:04:07.0000] <annevk> almost six years ago after it changed I wrote "remove useless distintion between methods and attributes and put them in IDL order (which also happens to make a lot of sense)" [07:06:30.0000] <timeless> rniwa: no, you were contextual reference. [07:14:08.0000] <timeless> /me beams [07:14:24.0000] <odinho> timeless: WHAT!? :D [07:14:33.0000] <timeless> list-users.pl now knows that mark must be Mark_Watson, because mark_vickers is Mark_vickers [07:15:02.0000] <timeless> -- my script for analyzing the roll call v. the speaking list grew some intelligence on Friday [07:15:07.0000] <timeless> which is rather handy [07:15:12.0000] <odinho> Oh. That's some smart action right there ;] [07:15:36.0000] <timeless> sadly, the presence of Cythia_Shelly confuses "chaals" [07:15:41.0000] <timeless> so chaals has to be changed to CMN [07:16:43.0000] <timeless> s/Cythia/Cynthia/ [07:17:13.0000] <timeless> /me is trying to confirm that "johnsim [07:17:22.0000] <timeless> " is John_Sim and not something else... [07:17:38.0000] <odinho> timeless: Can't see why? Can't it be smart about that too? Or actually print out its guesses when you do ,guesstimate . Then being able to correct it. Odin_Hørthe_Omdal is odinho [07:17:44.0000] <timeless> he's John Simmons [07:18:02.0000] <timeless> well [07:18:04.0000] <timeless> i can cheat [07:18:13.0000] <timeless> i can do c...m...n..._(chaals) [07:18:16.0000] <timeless> and then it'll work [07:18:20.0000] <timeless> /me could do that [07:18:28.0000] <timeless> yes, it learned to do that last week too [07:18:49.0000] <timeless> and if there are two people w/ the same () tail, it generally decides that it isn't a nick but a something else, e.g. *_(Samsung) [07:19:35.0000] <timeless> well, err [07:19:39.0000] <timeless> it's /supposed/ to get that right [07:19:40.0000] <timeless> it's /supposed/ to get that right [07:19:45.0000] <timeless> /me pokes it with a stick [07:21:11.0000] <timeless> /me cries [07:21:30.0000] <odinho> hah [07:22:11.0000] <timeless> well, Mark_Vickers as a speaker causes mark to be filed right [07:22:21.0000] <timeless> but Mark_Vic_ as a speaker just results in the script giving up [07:22:37.0000] <timeless> it says that both mark and Mark_Vic_ are ambiguous for either mark [07:22:49.0000] <timeless> it isn't terrible, but it isn't ideal either [07:24:23.0000] <timeless> basically it has two passes, and for ambiguous speakers, if a person is taken in the first pass, they're not considered as candidates in the second pass [07:24:57.0000] <timeless> but the logic for handling people w/ first+last names isn't wonderful, and it doesn't handle all the magical ways people munge their names :/ [07:25:20.0000] <jgraham> /me wonders if timeless realises that, except in cases of internet-drama, he is the only person that reads the logs for 10 years (at which point hsivonen or matjas reads them and proves that the "original point" of some feature isn't whatever someone is claiming) [07:25:44.0000] <timeless> heh [07:25:56.0000] <timeless> jgraham: well, ArtB reads them [07:26:09.0000] <timeless> although he tends to read them because he expects lawyer-drama [07:27:55.0000] <jgraham> Alwyer drama is too scary for me [07:27:59.0000] <jgraham> *lawyer [07:28:12.0000] <jgraham> I have to cower behind the sofa [07:28:20.0000] <jgraham> Like Doctor Who [07:59:39.0000] <matjas> zcorpan: I thought D at first, but I guess it’s A? [08:00:08.0000] <matjas> does `<!--` give a free pass for _one_ `</script>` occurence, or what? [08:01:00.0000] <annevk> http://a11ymemes.tumblr.com/post/23033757039/a-crying-woman-with-her-hand-over-her-face-says search for "duplicated" in search [08:01:04.0000] <annevk> that's pretty horrific [08:01:11.0000] <zcorpan> matjas: mohahaha :-P [08:01:16.0000] <annevk> s/search/source/ [08:01:54.0000] <timeless> annevk: that's a scribe.pl s/// ? :) [08:02:36.0000] <annevk> timeless: it's for humans [08:02:51.0000] <timeless> as opposed to perler-s. ok :) [08:02:53.0000] <matjas> zcorpan: looking forward to the correct solution… i have some note-adding to do [08:03:19.0000] <timeless> hey, was there a Rich other than Richard Schwerdtfeger? : [08:03:23.0000] <annevk> it's C I think [08:03:35.0000] <annevk> but I never followed the crazy <!-- in <script> thingie closely [08:03:35.0000] <annevk> but I never followed the crazy <!-- in <script> thingie closely [08:03:57.0000] <annevk> timeless: Opera's richt was not there at least [08:04:07.0000] <timeless> data:text/html,<script><!--</script>hi [08:04:25.0000] <timeless> annevk: right, that's helpful [08:05:12.0000] <timeless> http://pastebin.mozilla.org/1638348 [08:05:22.0000] <annevk> hmm C seems wrong [08:05:23.0000] <annevk> :( [08:07:12.0000] <Philip`> "richt" is begging for a biscuit-based joke [08:07:31.0000] <timeless> ? [08:08:37.0000] <jgraham> You probably have to be English [08:08:38.0000] <zcorpan> annevk: :-) [08:08:44.0000] <jgraham> Which richt is! [08:08:52.0000] <matjas> zcorpan’s current status: http://i.imgur.com/Emi0l.gif [08:09:00.0000] <Philip`> Hmm, maybe it's a culture-specific association [08:09:02.0000] <Philip`> http://www.nicecupofteaandasitdown.com/biscuits/previous.php3?item=31 [08:09:05.0000] <zcorpan> quite! [08:09:16.0000] <jgraham> Also https://twitter.com/#!/hallvord/status/200953940433584128 <- this seems to be undefined per spec [08:13:26.0000] <jgraham> Philip`: That page is clearly wrong to prefer milk chocolate digestives over dark chocolate ones however [08:14:35.0000] <jgraham> Although the behaviour facebook relies on seems to be impossible to spec [08:15:11.0000] <annevk> jgraham: just takes longer [08:15:32.0000] <jgraham> heh [08:18:22.0000] <jgraham> /me wonders where the term "breakpoint" related to media queries has come from [08:18:31.0000] <jgraham> It seems very confusing [08:26:56.0000] <timeless> Philip`: thanks [08:27:04.0000] <karlcow> jgraham: I see it here at least but for pages http://www.w3.org/TR/css3-page/#allowed-pg-brk [08:28:12.0000] <timeless> hober: so.... [08:28:24.0000] <jgraham> karlcow: That seems entirely different [08:28:30.0000] <timeless> to make my script happy, you'll appear as Edward_O_Connor_(ted)_(hober) [08:29:03.0000] <karlcow> jgraham: yes. what I said for pages. :) [08:29:09.0000] <jgraham> karlcow: AFAICT people are using it to mean "point at which the layout transitions due to a different media query taking effect" [08:36:49.0000] <annevk> http://trac.webkit.org/changeset/116384 <intent> ... [08:40:04.0000] <odinho> flagged off!? [08:46:05.0000] <timeless> http://trac.webkit.org/changeset/116384/trunk/Source/WebCore/html/HTMLTagNames.in [08:46:24.0000] <timeless> http://trac.webkit.org/changeset/116384/trunk/Source/WebKit/chromium/src/FrameLoaderClientImpl.cpp [08:46:31.0000] <timeless> #if ENABLE(WEB_INTENTS_TAG) [08:47:26.0000] <annevk> odinho: oh [08:48:39.0000] <jgraham> Flagged off is good [08:49:14.0000] <odinho> jgraham: OK. Not always easy to know. Depends on how you parse it :] [08:50:09.0000] <timeless> odinho: you could read the diffs [08:50:14.0000] <timeless> it was pretty easy to find.. [08:50:41.0000] <jgraham> odinho: The rule is that "flagged off" is good, "fucked off" is bad [08:50:41.0000] <odinho> timeless: Yeah, -- I did that after posting the comment :P [08:56:30.0000] <annevk> TabAtkins_: when looking for links to prefetch, you're not going to parse CSS [08:57:28.0000] <jgraham> annevk: Why not? I mean you parse HTML [08:58:37.0000] <annevk> TabAtkins_: also the whole notion of making HTML links dependent upon statements in CSS seems kind of insane [08:58:38.0000] <annevk> TabAtkins_: also the whole notion of making HTML links dependent upon statements in CSS seems kind of insane [08:58:49.0000] <jgraham> That I agree with [09:00:15.0000] <timeless> +1 [09:01:23.0000] <Philip`> ("-o-link"?) [09:04:56.0000] <Ms2ger> (I have heard claims that support was dropped) [09:06:49.0000] <dglazkov> good morning, Whatwg! [09:56:29.0000] <tantek> hsivonen, jgraham, I don't particular care for ISO-week numbers either, nor for odd EU non-ISO-week numbers. They're both poorly designed. I prefer ordinal dates (2012-135) and different weeks altogether: http://newcal.org [09:58:14.0000] <tantek> othermaciej, from my understanding talking to lawyers, CC0 is basically a more internalized / international-aware version of the MIT license. If you're going to ask your lawyers anyway, please consider asking them if CC0 is close enough to MIT that you can use either or both (since apparently you can already use MIT). [10:00:28.0000] <tantek> And for those that have looked, the W3C CG CLA is basically a W3C-specific version of OWFa (same lawyer(s) worked on both). I'm annoyed that W3C didn't just use the standard OWFa by reference. Creating yet another bespoke license doesn't benefit anyone (see Hixie's point about license proliferation etc.). [10:00:31.0000] <annevk> tantek: MIT requires acknowledgment [10:00:43.0000] <annevk> tantek: because of "Copyright (c) <year> <copyright holders>" [10:00:58.0000] <annevk> tantek: which you are required to include in redistribution [10:01:14.0000] <annevk> tantek: CC0 has no such thing, especially in countries that acknowledge Public Domain [10:01:15.0000] <annevk> tantek: CC0 has no such thing, especially in countries that acknowledge Public Domain [10:03:09.0000] <tantek> annevk - in practice I see no reason for such explicit attribution requirements in standards. In practice when using a standard people link to it which IMHO is sufficient. [10:04:00.0000] <annevk> I agree that standards should be Public Domain [10:04:28.0000] <annevk> see http://annevankesteren.nl/2012/02/standards-red-pill [10:10:45.0000] <timeless> sorryhrm [10:11:06.0000] <timeless> does weinig ever get minuted as sam? [10:11:50.0000] <annevk> timeless: could be, also, I don't think there was any other Sam present... [10:12:14.0000] <Ms2ger> Heh, nice [10:12:23.0000] <Ms2ger> ToUint32([10]) == 10 [10:12:32.0000] <timeless> annevk: rubys [10:13:17.0000] <annevk> timeless: not during WebApps afaik [10:13:38.0000] <timeless> ah [10:13:41.0000] <timeless> this is html5 [10:13:45.0000] <timeless> i already did webapps [10:13:53.0000] <timeless> minutes needing to add one line for tantek [10:13:56.0000] <annevk> timeless: weinig was not present during HTML [10:14:14.0000] <timeless> ok, good [10:14:24.0000] <timeless> so sam=rubys for html [10:14:33.0000] <annevk> timeless: also, this is taking too much of your time man, you could do something useful, like write a spec or some tests :) [10:14:49.0000] <Yuhong> annevk: When reparsing invalid XHTML as HTML, consider *disabling scripting* to prevent XSS attacks. [10:15:40.0000] <annevk> Yuhong: you can provide feedback for Opera's browser via https://bugs.opera.com/wizarddesktop/ [10:16:04.0000] <annevk> Yuhong: I haven't really been in the loop much on the XHTML -> HTML thing [10:17:15.0000] <Yuhong> annevk: But you did create XML5, and I hope the XML Error Correction CG considers XSS attacks. [10:18:44.0000] <timeless> http://pastebin.mozilla.org/1638460 [10:20:40.0000] <annevk> Yuhong: best to email the CG then [10:22:50.0000] <Yuhong> annevk: BTW, on the Encoding Standard, are you really asking MS to add support for JIS X 0212? [10:23:14.0000] <Yuhong> annevk: AFAIK, the Encoding Standard is a superset of what IE supports already. [10:28:52.0000] <timeless> /me sighs [10:28:58.0000] <timeless> <richard...> blah [10:28:59.0000] <Ms2ger> http://tranbot.net/html5/ [10:29:01.0000] <timeless> rich: bleh [10:29:04.0000] <timeless> Ted: eh [10:29:08.0000] <timeless> Rich: blek [10:29:16.0000] <timeless> so, clearly all 3 rich*'s are the same :) [10:32:15.0000] <divya> Ms2ger: where is this from? [10:33:43.0000] <Ms2ger> The landfill [10:33:51.0000] <Ms2ger> Or the W3C bugzilla, if you prefer [10:36:01.0000] <Ms2ger> "I hope the abstract to be ,and still think about let the webpage be easy ,no more effect ,no use of graphic , to serve the goal of information bass" [10:36:02.0000] <Ms2ger> "I hope the abstract to be ,and still think about let the webpage be easy ,no more effect ,no use of graphic , to serve the goal of information bass" [10:36:44.0000] <divya> wattt [10:37:31.0000] <Ms2ger> Same source [10:39:16.0000] <timeless> Ms2ger: is that a valid korean translation? [10:40:26.0000] <Ms2ger> You think I speak Korean? [10:44:37.0000] <Ms2ger> "thats very interesting. i would like to know if that cross implied notice ever subsides?" [10:49:14.0000] <annevk> Ms2ger: what are you doing? [10:49:31.0000] <Ms2ger> Triaging bugs [10:50:11.0000] <annevk> for HTML? [10:50:13.0000] <Ms2ger> Yeah [10:50:18.0000] <annevk> ah [10:50:30.0000] <annevk> I guess I stopped getting those emails now I'm no longer subscribed to public-html [10:50:40.0000] <Ms2ger> "maxlength should probably be complicated with minlength" [10:51:16.0000] <Ms2ger> Also, most bugs seem to end up in Other Hixie specs [10:54:17.0000] <timeless> in html-wg, is MC always michaelC? [10:55:12.0000] <timeless> Ms2ger: that's probably because other-hixie-specs has fewer gardeners [10:55:29.0000] <Philip`> It could be MC Hammer [10:55:47.0000] <timeless> /me thanks Philip` for his constructive input [10:56:12.0000] <timeless> ah, it is michaelC [10:56:48.0000] <padenot> q/b 2 [10:57:49.0000] <timeless> ?? [11:06:39.0000] <annevk> matjas: fwiw, you need to point to the parser section to answer parsing questions authoritatively [11:07:28.0000] <matjas> annevk: thanks. so ABNF is non-normative? [11:07:57.0000] <matjas> also, I’m *still* not sure, so it’s far from authoritative :') [11:08:54.0000] <annevk> matjas: well, it might be normative for authors, haven't looked at it [11:09:02.0000] <TabAtkins> For what, CSS? [11:09:09.0000] <annevk> matjas: but that requirements on authors don't explain what will happen in practice [11:09:16.0000] <annevk> s/that/ [11:09:23.0000] <annevk> TabAtkins: no [11:09:37.0000] <TabAtkins> Oh, I see the context. [11:10:10.0000] <annevk> matjas: I like your twitter tagline btw :) [11:11:24.0000] <matjas> \o/ [11:12:19.0000] <divya> ++ [11:18:20.0000] <othermaciej> annevk, tantek: to clarify, MIT license requires attribute if you redistribute the text or derivative works thereof, not if you simply "use" a standard in the common-sense terms (in case anyone was confused by "when using a standard" [11:23:57.0000] <annevk> othermaciej: I guess I was [11:24:20.0000] <Ms2ger> http://www.w3.org/TR/html5/semantics.html#the-root-element < nice page [11:25:30.0000] <annevk> o_O Ms2ger reading TR/ [11:25:39.0000] <annevk> or is this the bug reports again? [11:25:40.0000] <Ms2ger> I blame i18n [11:25:43.0000] <Ms2ger> Yeah [11:25:53.0000] <annevk> maha [11:31:02.0000] <Ms2ger> /me closed 63 bugs [11:33:28.0000] <jarek> why there is no official SVG validator? [11:33:28.0000] <jarek> why there is no official SVG validator? [11:33:36.0000] <jarek> http://jiggles.w3.org/svgvalidator/ tells me to go away [11:34:07.0000] <jarek> what was wrong with "previous SVG validator" that it was pulled down? [11:56:13.0000] <jarek> should SVG authoring tool output SVG files with a doctype? [11:57:50.0000] <jarek> I know it is in the spec, but according to https://jwatt.org/svg/authoring/#doctype-declaration discourages this [12:20:14.0000] <TabAtkins> jarek: No, just start with <svg>. [12:20:24.0000] <TabAtkins> Doctype isnt' required for XML - it's just useless noise for SVG. [12:20:36.0000] <jarek> TabAtkins: what about "<?xml version="1.0"?>"? [12:20:49.0000] <TabAtkins> Doesn't do anything in any important reader. [12:21:17.0000] <TabAtkins> Start your doc with <svg xmlns="whatever the ns is" viewBox="0 0 foo bar"> and you're golden. [12:21:36.0000] <jamesr> the xmlns= is pretty stupid [12:21:47.0000] <TabAtkins> Correct! [12:21:51.0000] <jamesr> start with <!DOCTYPE html><svg viewBox=""> and be more golden, right? [12:21:57.0000] <TabAtkins> But every time I say "Let's just merge it into HTML", people complain at me. [12:22:03.0000] <TabAtkins> jamesr: Won't work in <img> then. ;_; [12:22:40.0000] <jarek> xmlns attribute seems to be required by Chrome [12:22:46.0000] <jamesr> yes you need it [12:22:52.0000] <jamesr> i'm saying you shouldn't, but today you definitely gotta have it [12:23:31.0000] <Hixie> if the choice is between DOCTYPE magic or xmlns magic, i think xmlns magic is the least evil choice [12:23:50.0000] <Hixie> in other news, holy whatwg thread batman [12:23:58.0000] <TabAtkins> Next thought - change the root element. <svg2> puts you in the HTML ns by default. [12:24:02.0000] <TabAtkins> Woo! [12:24:19.0000] <jarek> there is also version="1.1" baseProfile="full" magic... [12:24:20.0000] <jarek> there is also version="1.1" baseProfile="full" magic... [12:24:30.0000] <jamesr> Hixie, why do i gotta choose one? [12:24:40.0000] <TabAtkins> jarek: Those are also useless noise. Ignore them. [12:25:11.0000] <TabAtkins> jarek: I have no idea if standalone viewers care, but if you're doing something for the web, *all* you need is the xmlns on the <svg> root. [12:25:16.0000] <Ms2ger> jamesr, <!DOCTYPE html><svg doesn't work in XML, fwiw [12:25:28.0000] <TabAtkins> Ms2ger: That's the point. [12:25:30.0000] <jamesr> i generally don't care 'bout that [12:28:02.0000] <jamesr> is it possible for two frames that are different origin from the root frame to get a synchronous JS connection to each other? [12:28:32.0000] <Hixie> jamesr: because if you give 2 you're just wasting your time [12:28:42.0000] <TabAtkins> Can't someone mutate their origin so they can communicate? [12:29:29.0000] <annevk> jamesr: the main problem with your SVG snippet is the DOM it creates [12:29:31.0000] <jamesr> can they? i dunno [12:30:48.0000] <annevk> jamesr: if their effective TLD is the same [12:32:16.0000] <annevk> hehe http://w3cmemes.tumblr.com/post/23051147674 [12:34:04.0000] <jamesr> annevk, what defines effective TLD? i'm guessing this is something like a parent frame creates a child frame, then modifies its document.domain to point somewhere else? [12:37:20.0000] <annevk> jamesr: two iframes on x.y.org and y.y.org set document.domain to y.org [12:37:47.0000] <annevk> jamesr: HTML and http://publicsuffix.org/ have details on how document.domain operates [12:38:20.0000] <annevk> it also applies to cookies, but I believe abarth didn't put it in the cookie draft (or had to take it out because the IETF doesn't like reality) [12:53:23.0000] <abarth> annevk: its in the cookie spec [12:53:31.0000] <jamesr> annevk, thanks - once they do that, how do they get a ref to each other? [12:53:32.0000] <abarth> annevk: its just not required [12:54:29.0000] <TabAtkins> jamesr: The containing page has to pass the refs over, I *think*. Though maybe a SharedWorker from one can be talked to by the other? [12:54:47.0000] <TabAtkins> But I don't think you can send any objects to a SharedWorker that allow sync JS communication. [12:55:01.0000] <abarth> jamesr: the sort answer is yes [12:55:01.0000] <abarth> jamesr: the sort answer is yes [12:55:09.0000] <abarth> jamesr: frames can crawl the frame hierarchy [12:55:15.0000] <abarth> regardless of origin [12:55:15.0000] <abarth> regardless of origin [12:55:22.0000] <abarth> parent.frames[3].contentWindow [12:55:29.0000] <jamesr> aha! thanks [12:55:54.0000] <Ms2ger> abarth, parent.frames[3] is the window already [12:56:01.0000] <abarth> right :) [12:56:06.0000] <annevk> parent[3] should work too [12:56:13.0000] <Ms2ger> Right [12:56:32.0000] <annevk> .frames is just a pointer back to itself [13:06:25.0000] <Ms2ger> Can someone at Opera look at DSK-363698? [13:08:06.0000] <annevk> sweet http://lists.w3.org/Archives/Public/public-whatwg-archive/ [13:08:21.0000] <jgraham> Ms2ger: Yep, we all can. Don't you feel lame now? [13:08:57.0000] <Ms2ger> :) [13:09:08.0000] <Ms2ger> Ooh [13:12:38.0000] <jgraham> Anyway, I can get that fixed [13:13:30.0000] <Ms2ger> jgraham++ [13:15:11.0000] <jgraham> Well I'm not really volunteering to do the work ;) Although I guess it's not too hard to do that either. [13:16:39.0000] <othermaciej> annevk: nice [13:21:31.0000] <TabAtkins> Jeez, whatwg looks positively dead compared to www-style. [13:46:00.0000] <tantek> isn't this the WHATWG's primary failure mode? (e.g. respimg) http://w3cmemes.tumblr.com/post/22831818920/the-css-wgs-primary-failure-mode [13:47:45.0000] <TabAtkins> tantek: That doesn't usually cause the WHATWG to slow down too much. [13:47:51.0000] <othermaciej> I believe the two different proposals in that area are getting a reasonable evaluation of their pros and cons, at least in on-list discussion [13:48:29.0000] <othermaciej> what more is expected, to avoid the failure mode? [13:48:31.0000] <annevk> tantek: afaict the two proposals have radically different trade offs despite some people initially claiming it's just syntax [13:48:55.0000] <tantek> annevk - good to know [13:49:16.0000] <tantek> presumably use-cases will be used to determine trade-off priorities [13:49:24.0000] <othermaciej> yes, the proposals are significantly different operationally, not just syntactically [13:49:32.0000] <TabAtkins> Indeed. The fact that some people aren't listening to the "MQs are horrible for doing resolution negotiation" is annoying. [13:50:12.0000] <tantek> resolution negotiation sounds like something you do with yourself at the end of the year. [13:50:42.0000] <jgraham> Yeah, it seems like some people have a preconceived idea that they want to use media queries [13:52:32.0000] <othermaciej> I prefer to think of it in terms of pros and cons rather than "one idea is obviously bad" [13:52:48.0000] <TabAtkins> Oh, no, neither is bad. They solve different things. [13:52:55.0000] <jgraham> Well it's not obviously bad [13:53:05.0000] <jgraham> But it's not obviosuly good either [13:53:05.0000] <jgraham> But it's not obviosuly good either [13:53:08.0000] <TabAtkins> It's just annoying that some people aren't listening to the "idea A *cannot* solve problem B, stop trying to make it do so" refrain. [13:53:09.0000] <TabAtkins> It's just annoying that some people aren't listening to the "idea A *cannot* solve problem B, stop trying to make it do so" refrain. [13:53:23.0000] <Ms2ger> ap, the spec is right, don't be a jerk [13:53:24.0000] <othermaciej> it's not clear to me what the full list of use cases is, and whether there is actually a relevant use case addressed by one and not the other [13:53:35.0000] <jgraham> It feels like some people think that it is obviously good and so haven't been very receptive to technical issues [13:53:51.0000] <jgraham> that might suggest otherwise [13:54:49.0000] <ap> Ms2ger: what makes it "right"? screwing the engine that already had the argument optional for another one that mis-implemented same behavior? [13:54:53.0000] <othermaciej> well, much of the pro-<picture> discussion has been based on perceived clarity of the syntax, without clearly acknowledging the difference in operational effect [13:54:53.0000] <othermaciej> well, much of the pro-<picture> discussion has been based on perceived clarity of the syntax, without clearly acknowledging the difference in operational effect [13:55:06.0000] <othermaciej> ap, Ms2ger: which bit of spec is this? [13:55:14.0000] <ap> othermaciej: https://bugs.webkit.org/show_bug.cgi?id=78887 [13:55:31.0000] <ap> othermaciej: specifically, comment 3 [13:55:38.0000] <jgraham> tantek: The fact that WHATWG avoids that failure mode might be due to strong technical leadership aka at the end of the day Hixie makes some choices and they generally aren't so bad that everyone pushes back. But when they are he generally adjusts the spec. [13:57:47.0000] <othermaciej> Ms2ger: do you have specific evidence that the spec is right besides what's in the www-dom thread (which I have not read yet) [13:58:19.0000] <jgraham> FWIW my feedback on that is that importNode is never used so consistency there isn't a strong consideration [13:58:40.0000] <jgraham> Also, we are screwed because Microsoft will probably refuse to change IE10 [13:59:43.0000] <Ms2ger> othermaciej, it's just sad that WebKit's official policy appears to be that a spec that doesn't match WebKit is wrong by definition [14:00:07.0000] <othermaciej> Ms2ger: I think al's argument for why the spec is wrong is more sophisticated than "doesn't match WebKit" [14:00:27.0000] <othermaciej> as I understand his argument, it says that "all older browsers used to default to false or throw, so making the new behavior default to true is unwarranted and dangerous" [14:01:18.0000] <Ms2ger> Oh, it's not just here [14:01:34.0000] <Ms2ger> He makes that argument every time a spec doesn't match WebKit [14:02:22.0000] <othermaciej> well, I was looking at this specific argument, which seems prima facie reasonable to me [14:03:59.0000] <jgraham> I tend to agree. Irrespective of the general attitude of webkit developers in general or in the specific, in this case their argument seems right [14:04:03.0000] <othermaciej> read the discussion thread now [14:04:03.0000] <othermaciej> read the discussion thread now [14:04:19.0000] <jgraham> s/general attitude/attitude/ [14:04:37.0000] <othermaciej> I think I agree with ap's position in this case, it doesn't seem that he expressed his position in a jerky way, and it's not clear why the spec should trump WebKit and IE behavior [14:04:53.0000] <othermaciej> Support Existing Content principle should trump "editors think this other way is more elegant" [14:05:13.0000] <othermaciej> if editors don't agree with that, then it seems fine for implementations to ignore the spec on that point [14:06:28.0000] <annevk> so in WebKit all arguments were already optional [14:06:37.0000] <annevk> including for e.g. createNodeIterator [14:07:02.0000] <annevk> where the NodeFilter constant argument defaulted to 0 because that's what undefined got turned into [14:07:09.0000] <annevk> which incidentally makes it useless [14:08:17.0000] <annevk> in any event, I don't disagree with "Support Existing Content" but we generally do not follow that principle if browsers disagree wildly [14:08:49.0000] <othermaciej> if there's specific cases where changing the default has been shown to create minimal compat risk, and the win to changing it is large, then it seems reasonable to change [14:09:03.0000] <othermaciej> it doesn't seem like either of those conditions has been established in this case [14:10:02.0000] <othermaciej> I think supporting WebKit-specific content and Trident-specific content still trumps "this other way is more elegant" [14:10:15.0000] <Wilto> I trust everyone saw the `picture` use cases and polyfills posted to http://wiki.whatwg.org/wiki/Adaptive_images , yeah? [14:10:38.0000] <Wilto> Just figured I’d throw that out there. [14:11:09.0000] <annevk> othermaciej: I guess we disagree on that then [14:11:15.0000] <othermaciej> given that Gecko-specific or Presto-specific content is unlikely to break in this case, since content almost never relies on an exception being thrown, and content narrowly targeting either of those engines is more rare [14:11:32.0000] <annevk> Wilto: did you post it to the mailing list? [14:11:36.0000] <Wilto> annevk: Yessir. [14:11:40.0000] <TabAtkins> Wilto: Yeah for me, at least. [14:12:02.0000] <annevk> Wilto: I mean I did see it, but not everyone follows the wiki closely [14:12:34.0000] <othermaciej> annevk: would you claim that, for example, mobile-targeted content that has only been tested in WebKit-based browsers does not count as content worth supporting, by definition? [14:12:35.0000] <TabAtkins> Wilto: What does "full-bleed" mean? [14:12:39.0000] <Wilto> annevk: Yep, no worries. It was posted to the mailing list as well; just wanted to be sure it got around. [14:13:08.0000] <Wilto> Oh, 100% width. [14:13:13.0000] <Wilto> Effectively. [14:13:17.0000] <TabAtkins> Ah, kk. [14:13:24.0000] <annevk> othermaciej: no [14:13:43.0000] <Wilto> Gotta bolt; just wanted to check in. Thanks, guys! [14:13:44.0000] <othermaciej> annevk: then what part of my statement do you disagree with? [14:13:54.0000] <TabAtkins> Wilto: On the wiki page, I don't see good descriptions of why you want to make <img> responsive, in particular - that is, why CSS background images aren't appropriate. [14:14:02.0000] <othermaciej> annevk: I assume what you disagreed with was 'I think supporting WebKit-specific content and Trident-specific content still trumps "this other way is more elegant"' [14:14:41.0000] <TabAtkins> Wilto: I'll take the question elsewhere if you're leaving. Thanks. ^_^ [14:15:49.0000] <annevk> othermaciej: I don't really see it as a black/white thing; on top of that there's no WebKit/Trident-specific content found here [14:16:55.0000] <othermaciej> annevk: if there's no content anywhere that relies on the default, then there's less risk (still risky to have different defaults in different browsers), but I don't see anyone making that case [14:17:19.0000] <Wilto> Oh—good catch, TabAtkins. I’ve been steeped in this stuff for so long that I can’t help making a few assumptions. I’ll add that for certain. [14:19:23.0000] <annevk> othermaciej: Firefox 13 seems to be making that case [14:21:01.0000] <jgraham> Not if it is failing on mobile-targeted content bu no one is noticing due to the general unhealthiness of that ecosystem [14:23:51.0000] <othermaciej> annevk: I do believe it makes that case for desktop-targeted cross-browser content [14:59:53.0000] <Velmont> So. The Common Crawls dataset, 50TB, 5 billion web pages. -- Are we sometimes using searches through that to see how much people are using stuff we'd want to change? [15:03:06.0000] <zcorpan> matjas: https://twitter.com/zcorpan/status/202156793277857794 [15:37:53.0000] <zcorpan> Hixie: you need to fire 'error' when the fetch fails (srcset) [15:37:54.0000] <zcorpan> Hixie: you need to fire 'error' when the fetch fails (srcset) [15:38:37.0000] <Velmont> zcorpan: Where are you reading? [15:39:08.0000] <zcorpan> Velmont: http://www.whatwg.org/specs/web-apps/current-work/?slow-browser#attr-img-srcset [15:41:16.0000] <Velmont> zcorpan: Cool. Sending me to the benchmark page :P [15:41:38.0000] <zcorpan> at least without running scripts :-) [15:41:39.0000] <Velmont> Oh, it was not as slow as before now. Opera must have fixed a bug or something. [15:41:41.0000] <TabAtkins> Hixie: Why are you making up new units for that? [15:47:57.0000] <Hixie> zcorpan: why? [15:48:07.0000] <Hixie> TabAtkins: they're not units, just ways to distinguish which is present [15:48:28.0000] <zcorpan> Hixie: you fire load if it's successful. why not error? [15:48:38.0000] <Hixie> zcorpan: nothing happens if it's not successful [15:48:46.0000] <Hixie> zcorpan: so why bother telling anyone? [15:49:01.0000] <Hixie> zcorpan: (think of a case like you've gone offline and then resize the window) [15:49:36.0000] <Hixie> (i originally wanted to use another event if the image changed) [15:49:47.0000] <zcorpan> Hixie: hmm. dunno. maybe it's better to just emit a message to the error console... [15:50:12.0000] <Hixie> that's certainly reasonable [15:52:47.0000] <othermaciej> is the use of "breakpoint" in that one proposal a common term of art in web design? [15:52:51.0000] <othermaciej> or design in general? [15:53:11.0000] <TabAtkins> othermaciej: I've heard it commonly, though this could be due to a twitter bubble around me. [15:53:12.0000] <othermaciej> I keep imagining stopping in gdb [15:53:23.0000] <Velmont> Hixie: Actually, why having w and h at all? I mean, do you mean you want to be able to change the size of the picture (and thus it's ratio) and not only resolution? writing 800w and 1600w is certainly nice syntax, as the 0.-numbers can get unweildy. But someone can do <img src=pic.jpg width=800 height=600 srcset="mypic⊙2j 2x, strangepic.jpg 200w 200h"> then. What would that actually entail? [15:53:53.0000] <Hixie> i don't understand the question [15:54:08.0000] <TabAtkins> Velmont: That would mean that you'd only consider strangepic at all when the screen is smaller than 200px wide and tall. [15:54:12.0000] <Hixie> but maybe it's better if i just reply to the zillions of e-mails already on the list [15:54:19.0000] <Hixie> rather than starting yet another conversation here :-) [15:54:28.0000] <Velmont> Hixie: hehe :] I'm in those. [15:54:47.0000] <TabAtkins> The 0w/0h stuff filters the list, and the UA then decides among the rest which to download based on its own heuristics and the presented resolution data. [15:55:14.0000] <TabAtkins> From a layout perspective, resolution only has an effect on the intrinsic size of an image. [15:55:42.0000] <Velmont> TabAtkins: Yes, -- I'm asking if it would be allowed to change height×width ratio in srcset. [15:55:43.0000] <TabAtkins> If you set @width and @height, sending a 32dpi versus a 192dpi image is identical, except that one has more detail. [15:56:03.0000] <TabAtkins> Velmont: It doesn't change the ratio. You might misunderstand what the w/h numbers are doing. [15:56:10.0000] <Velmont> Ah. [15:56:29.0000] <TabAtkins> They're equivalent to using a max-width/height Media Query. [15:56:32.0000] <Velmont> It's not talking about the picture. It's actually a viewport mediaquery hint? [15:57:16.0000] <Velmont> Hmmm. Okay. I thought normal.jpg (800x600), then hd⊙2 2x, hd⊙2 800w 600h would be equal. [15:57:31.0000] <Velmont> That's making that usage clearer, yes. :P [15:57:59.0000] <TabAtkins> Is there a reason for the strange spacing in your last line? [15:58:10.0000] <TabAtkins> (I'm wondering if I'm missing some characters that should be there.) [15:58:29.0000] <Velmont> TabAtkins: Heh, no... I really wanted to \n\tEXAMPLEHERE [15:59:08.0000] <TabAtkins> Ah, kk. [16:01:18.0000] <TabAtkins> Hixie: Yeah, just posting to the list would be good. I'd want to see if there are good use-cases for MQ-in-general, or if the explicit restriction to the max-width/height MQs are good enough. [16:01:32.0000] <TabAtkins> I suspect they'd be good enough, based on what I've heard of use-cases so far. [16:03:58.0000] <Velmont> Is even having both max-height max-width really necessary? Having one number (the largest) and treating it as a bounding-box would solve the cases I think of. [16:04:29.0000] <Velmont> The smartphone probably wants to fetch one it can show in portrait and landscape anyway. [16:04:37.0000] <TabAtkins> I dunno! [16:05:14.0000] <Velmont> And I guess people are not using 2048x200 pictures on a real device. -- Or maybe they are. Hmm, headers might be. [16:05:42.0000] <TabAtkins> Yup, that's a use-case. [16:05:49.0000] <Velmont> I should think before I speak. [16:05:53.0000] <TabAtkins> ^_^ [16:05:54.0000] <Velmont> Will sound so much wiser. [16:06:02.0000] <Velmont> :D [16:07:27.0000] <zewt> fetching two images for every picture is expensive on latent, bandwidth-quotad devices, so i'm guessing few phones would want to do that anyway (at least today) [16:13:19.0000] <Velmont> zewt: Fetching two images? No, I was only talking about actually fetching one that will fit both portrait and landscape. So if you're in portrait, you overestimate, and pick a slightly bigger picture that'll also use all your pixels in landscape. Hence you only need to fetch that image once, in case the user switches to landscape. [16:13:39.0000] <Velmont> zewt: So you are actually hindering an extra image fetch here. [16:15:30.0000] <Velmont> zewt: I should've written "UA wants to fetch _one_ image that it is big enough to use in both portrait and landscape". [16:28:05.0000] <zewt> gluh [16:28:11.0000] <zewt> it needs to not be possible for sites to prevent you from pasting into form fields [16:28:24.0000] <zewt> sites that stop you from copying and pasting your email address into confirm fields need to burn in a ditch [16:29:13.0000] <TabAtkins> Yus. [16:29:41.0000] <zcorpan> http://xkcd.com/970/ [16:29:47.0000] <zewt> "we will deliberately waste your time" [16:35:59.0000] <TabAtkins> Even worse are those that do this for your password confirmation field. I don't *ever* type my passwords if I can avoid it - they get generated by a hasher. [16:38:32.0000] <zcorpan> or where there are restrictions on the password such that you can't pick anything sane, like with bredbandsbolaget [16:39:13.0000] <zewt> in general password rules just make everyon capitalize their password and add "1" [16:39:27.0000] <zewt> they're a pretty good sign that whoever's running the site doesn't think things through [16:39:28.0000] <zewt> they're a pretty good sign that whoever's running the site doesn't think things through [16:39:46.0000] <TabAtkins> Yes. [16:39:57.0000] <richwild> Yes1 [16:39:59.0000] <zcorpan> which, iirc, requires a length between 6 and 8 chars, requires at least one uppercase letter, and at least two numbers, and no punctuation and no non-ascii [16:40:18.0000] <zewt> well non-ascii is asking for trouble, heh [16:40:36.0000] <TabAtkins> max-length password requirements just *infuriate* me. Out of all the things you could require, that's the one that's not defensible in *any* way. [16:40:55.0000] <zewt> (re: normalization) [16:41:06.0000] <Philip`> TabAtkins: Maybe they're storing your password in a fixed-width database field [16:41:07.0000] <TabAtkins> Unless they're not hashing your password before storing (which would make it fixed-length). [16:41:39.0000] <TabAtkins> Philip`: See what I just said. ^_^ [16:41:43.0000] <zcorpan> yeah i couldn't even use a normal word, capitalize it, and add two numbers at the end, because it ended up being longer than 8 chars [16:42:15.0000] <TabAtkins> My password generator, at least, has the ability to *almost* satisfy those requirements. [16:42:48.0000] <TabAtkins> I can force it to generate an 8-char password with at least one uppercase letter, number, and punctuation. I just can't do *2* numbers, so I'd have to hope it happens to generate two. [16:42:57.0000] <Philip`> Banks here tend to ask for random subsets ("Enter the 2nd, 5th and 8th characters of your password" etc) so they can't be storing just a hashed password [16:43:25.0000] <TabAtkins> Philip`: Those banks are idiots. :/ [16:43:49.0000] <TabAtkins> I mean, if you just ask for three characters, that's a pretty decent chance of randomly guessing correctly. [16:43:57.0000] <richwild> What do you think they must be storing? [16:44:04.0000] <jamesr_> TabAtkins: they store a hash of each character! [16:44:11.0000] <TabAtkins> jamesr_: Brilliant! [16:44:11.0000] <Philip`> TabAtkins: They ask for 3 digits of the 4-digit PIN too [16:44:17.0000] <TabAtkins> ... [16:45:05.0000] <Philip`> (I assume they're just trying to reduce the effectiveness of keyloggers by not making you type in all your login details at once, so attackers would have to log multiple logins) [16:49:17.0000] <TabAtkins> Shrug. The right solution to that is a fob that provides OTPs. :/ [16:49:36.0000] <Hixie> which, ironically, is also somewhat common in europe [16:49:56.0000] <Hixie> natwest has both sent me hardware and requires me to log in as Philip` describes [16:52:00.0000] <TabAtkins> lolwut [16:53:32.0000] <Philip`> I think they require the card-reading hardware device just for relatively high risk operations, like setting up a new payee [16:53:49.0000] <Hixie> i don't think i've ever had to end up using it [16:54:03.0000] <Philip`> presumably on the assumption that most people would be unwilling to put up with the inconvenience of having to use it just for checking their balance or paying regular bills [16:54:42.0000] <Hixie> it'd be a hell of a lot less inconvenient to swipe my card and get a number than it is to have to decode my password to figure what damn digit they want [16:59:56.0000] <jamesr> zewt, in multi-monitor setups you have to pick (fairly arbitrarly) a device to vsync to 2012-05-15 [17:03:12.0000] <jamesr> the flow control situation is interesting for offscreen stuff [17:39:05.0000] <zewt> jamesr: well, you either have to explicitly say which element to use (which only webkit supports, unless that page is out of date), or sync vsync together (no idea if that happens in practice) [17:39:24.0000] <jamesr_> everything in a page is vsync'd together [17:39:56.0000] <zewt> not if the page spans monitors [17:40:03.0000] <jamesr_> doesn't matter [17:40:11.0000] <jamesr_> everything within a page is rendered at the same time [17:40:23.0000] <zewt> it's not rendering, it's backbuffer flipping [17:40:49.0000] <jamesr_> that's not a concern of the web platform. the backbuffer is prepared for everything in the page at the same time [17:40:49.0000] <zewt> rendering happens when it happens (with webgl) [17:40:57.0000] <zewt> not with webgl it isn't [17:41:12.0000] <jamesr_> sure it is [17:41:17.0000] <zewt> it isn't? heh [17:41:30.0000] <jamesr_> draw ops in webgl to the webgl backbuffer happen whenever [17:41:49.0000] <jamesr_> those are resolved/flipped/copied into the rest of the system as an independent step (not observable from the web directly) [17:41:58.0000] <zewt> then it flips to the front buffer on monitor vsync, which depends on which monitor the element happens to be on (unless it spans monitors, in which case all bets are off, most likely) [17:42:15.0000] <jamesr_> no [17:42:16.0000] <zewt> the point is that you want to begin rendering right after vsync, and that depends on the monitor [17:42:25.0000] <jamesr_> it has to go to at least one intermediate buffer [17:42:45.0000] <jamesr_> and nothing at all depends on which monitor an element lands on [17:43:11.0000] <zewt> these are all implementation details; what matters is when you want to begin rendering, which is as close after vsync as possible [17:44:06.0000] <jamesr_> yes. in the multimonitor case, just pick some monitor and go [17:44:29.0000] <jamesr_> i've heard of some non-web systems picking the monitor with the largest area of intersection with the window [17:44:34.0000] <zewt> not some monitor; you want to pick the monitor the element is on, whenever possible [17:44:39.0000] <jamesr_> no [17:44:47.0000] <zewt> can you say something more informative than "no"? heh [17:44:53.0000] <jamesr_> it doesn't make any difference what monitor the element is on [17:44:53.0000] <zewt> why would you ever not do that? [17:44:59.0000] <jamesr_> because individual elements are not flipped independently [17:45:03.0000] <jamesr_> the whole buffer is prepared at once [17:45:08.0000] <jamesr_> and then made available to every monitor [17:45:08.0000] <jamesr_> and then made available to every monitor [17:45:12.0000] <zewt> how would you know that? heh [17:45:50.0000] <zewt> it's an implementation detail, and a good implementation would sync as close to each monitor as possible [17:46:01.0000] <jamesr_> i know that because i'm an implementor? [17:46:10.0000] <zewt> of every implementation ever? [17:46:27.0000] <jamesr_> an implementation as you describe would be infeasible [17:46:59.0000] <zewt> how so? you flip the region associated with the webgl context (which in a great many cases is the entire window) on vsync for the monitor it's on [17:47:34.0000] <zewt> (clearly someone at webkit agrees, if their requestAnimationFrame takes an element) [17:47:59.0000] <jamesr_> see this is where it's hard to take you seriously. i wrote the element param for webkit's RAF, and the spec text, and deleted that code [17:48:14.0000] <jamesr_> you can't cite a nebulous "someone at webkit" against me when that someone is me [17:48:17.0000] <zewt> okay i don't feel like a condescending conversation right now [17:48:18.0000] <zewt> later [17:48:52.0000] <jamesr_> it really gets my goat when people say "implementations do X or could easily do X" when that's clearly not true [17:49:23.0000] <zewt> uh huh [22:50:08.0000] <AryehGregor> "Wait until any invocations of this algorithm that had the same method context, that started before this one, and whose timeout is equal to or less than this one's, have completed." [22:50:13.0000] <AryehGregor> "equal to"? [22:50:47.0000] <AryehGregor> Oh, I guess everything runs single-threaded, so it's not a deadlock. [23:14:25.0000] <matjas> in which cases must `>` be escaped in HTML? I can only think of unquoted attribute values [23:15:06.0000] <AryehGregor> matjas, this should have all the rules: http://www.whatwg.org/specs/web-apps/current-work/multipage/syntax.html#writing I think you're probably right. [23:17:02.0000] <AryehGregor> Maybe in <textarea> if preceded by "</textarea", and likewise for <title>? [23:17:18.0000] <AryehGregor> That's not relevant if you're really asking "which characters do I have to escape?" and are already escaping <. [23:21:18.0000] <AryehGregor> There are other cases where ">" isn't allowed, but it can't be escaped in those cases. [23:21:40.0000] <AryehGregor> Like attribute names, or certain configurations within <script>/<style>/comments. [23:24:39.0000] <matjas> AryehGregor: thanks for confirming. that’s indeed the use case here, escaping HTML (and `<` is being escaped) [00:29:56.0000] <Hixie> phew [00:30:07.0000] <Hixie> finally caught up with the tsunami [00:38:18.0000] <Ms2ger> /me didn't hear about a tsunami in the Bay Area [00:38:31.0000] <Hixie> the responsive tsunami [00:38:38.0000] <Hixie> been replying to that thread for like a week now [00:38:41.0000] <Ms2ger> Oh, have fun with that :) [00:38:43.0000] <Hixie> every night i'm almost done [00:38:46.0000] <Hixie> i come back the next morning [00:38:49.0000] <Hixie> and it's run away again [00:38:53.0000] <Hixie> but i finally caught up! [00:39:41.0000] <Ms2ger> I would suggest fixing some bugs instead ;) [00:40:06.0000] <Hixie> bugs are on hold til i deal with mail [00:40:16.0000] <Hixie> which is a far bigger pile right now [00:40:42.0000] <Hixie> though i see from the chart that you've done a good amount of work on the bugs yourself! [00:40:45.0000] <Hixie> nice! [00:41:09.0000] <Ms2ger> Saddening how much junk accumulated, really :) [00:41:15.0000] <Hixie> heh [00:49:42.0000] <ShaneHudson> How do we go about removing srcset from the living spec? Both this irc channel and the community group were unanimous that srcset is the wrong way to go. [00:50:19.0000] <matjas> are the spec short URLs documented anywhere? if not I’ll take a stab at writing it up [00:50:50.0000] <matjas> e.g. http://www.whatwg.org/specs/web-apps/current-work/multipage/embedded-content-1.html#attr-img-srcset can be http://whatwg.org/html/embedded-content-1.html#attr-img-srcset or even http://whatwg.org/html#attr-img-srcset [00:51:48.0000] <ShaneHudson> For which one? [00:52:22.0000] <matjas> ShaneHudson: the multi-page and one-page versions of the HTML living standard [00:52:56.0000] <ShaneHudson> Ah, I have only seen the multi page so far [00:53:24.0000] <matjas> Hixie: heads up, http://www.whatwg.org/specs/ is missing a </strong> [00:59:38.0000] <ShaneHudson> this is the first time I have got involved with the spec, so if there is anything I should be doing to help, let me know please! [01:07:47.0000] <odinho> ShaneHudson: "Both this irc channel and the community group were unanimous that srcset is the wrong way to go.". Really? I've not seen that in this IRC-channel. And I've been here and discussed it. :-) [01:09:07.0000] <ShaneHudson> Yes, I asked everyone about 5pm GMT two days ago I believe [01:09:24.0000] <ShaneHudson> about 20 people said that they did not like srcset, and nobody said otherwise [01:09:38.0000] <odinho> ShaneHudson: What you do is reply to Hixie's email that came in just half an hour ago. [01:09:53.0000] <ShaneHudson> I am still reading it, it is extremeley long [01:10:21.0000] <odinho> ShaneHudson: Well, popularity contests is not the best way for specs to be made. Technical discussion and merit, however, is. [01:11:18.0000] <ShaneHudson> True of course, but have you not seen the discussions on the community group? [01:11:18.0000] <ShaneHudson> True of course, but have you not seen the discussions on the community group? [01:11:33.0000] <[tm]> annevk: I need to get the URL spec ready for FPWD publication, but having some trouble figuring out what anolis switches I need to flip to get it W3C-styled [01:11:54.0000] <kennyluck> ShaneHudson, you can join the W3C HTML WG and write a change proposal. [01:11:55.0000] <kennyluck> ShaneHudson, you can join the W3C HTML WG and write a change proposal. [01:12:05.0000] <ShaneHudson> I will reply to Hixie's email later today, so that I do not rush it [01:12:39.0000] <odinho> ShaneHudson: Well, I get lots of email. -- I have not read it lately (I did when it was announced some time ago). Anyway, we'll be hearing it on WHATWG later, which has a much bigger audience and where most people read and listens in. [01:12:44.0000] <Ms2ger> [tm], have a look at the DOM4 Makefile ;) [01:12:48.0000] <ShaneHudson> kennyluck: Ah I am able to join the WG? Was not sure how exclusive it was, I was under the impression I could only join the CG [01:13:17.0000] <Ms2ger> ShaneHudson, joining the HTML WG is generally a waste of time [01:13:31.0000] <odinho> ShaneHudson, kennyluck: No need joining HTML WG... Better to just take it up in WHATWG where it's discussed. [01:14:00.0000] <Ms2ger> If you've got technical arguments for your suggestions, you will find the WHATWG list is enough for you [01:14:59.0000] <odinho> Ms2ger: Heh, implictly saying what the other one is used for? [01:15:57.0000] <kennyluck> ShaneHudson, well, this is a step by step instruction → http://www.paciellogroup.com/blog/2011/12/how-you-can-join-the-w3c-html5-working-group-in-4-easy-steps/ [01:16:20.0000] <kennyluck> I am not going to say if that's a waste of time or not. Other people will tell you. :p [01:16:44.0000] <odinho> ShaneHudson: I'm seconding Ms2ger by the way. [01:16:49.0000] <jgraham> Please don't fork the discussion into to a third group [01:17:16.0000] <kennyluck> That I agree. THough writing a change proposal is another thing. [01:17:32.0000] <Ms2ger> A bigger waste of time? :) [01:17:42.0000] <kennyluck> No idea. *shrug* [01:18:02.0000] <odinho> Hehe, I think we can discuss technical on the list first. [01:18:28.0000] <jgraham> Writing a change proposal isn't the best way to get the spec changed [01:18:36.0000] <ShaneHudson> kennyluck: Thank you for the link, I will bookmark it but for now will take the advice of Ms2ger and odinho [01:18:40.0000] <kennyluck> I am simply answering "How do we go about removing srcset from the living spec?" question, and I am a bit sad that no one in this channel has mentioned a, well, way. [01:18:49.0000] <jgraham> The best way is to present convincing technical arguments that it is wrong [01:19:36.0000] <jgraham> kennyluck: I assume the goal is not "remove srcset" per-se but "enable a good design for adaptive image loading" [01:19:38.0000] <remysharp> anyone know if img@srcset define in the spec to look for the srcset **before** trying to download the img@src url? (posted before, not sure it made it through the intertubes) [01:19:39.0000] <odinho> change proposal: http://w3cmemes.tumblr.com/post/22414849805 [01:19:53.0000] <ShaneHudson> It is going to be very hard to go against hixie since he is obviously understands far more of browser development than I do. But as a spec that every developer should stick to, srcset is not the right way to go. [01:19:57.0000] <jgraham> remysharp: What do you mean? [01:20:13.0000] <kennyluck> jgraham, I am not familiar with this topic but I am just answering "How do we go about removing srcset from the living spec?". [01:20:13.0000] <kennyluck> jgraham, I am not familiar with this topic but I am just answering "How do we go about removing srcset from the living spec?". [01:20:28.0000] <remysharp> jgraham: if srcset makes it in to browsers, I mustn't request the img@src first by default - otherwise there's no point in having the bandwidth checks [01:20:40.0000] <odinho> remysharp: That's more or less implied, -- but yes, it should say that. [01:20:41.0000] <jgraham> kennyluck: I hope the actual goal is not to be obstructionist, but to be constructive [01:20:47.0000] <[tm]> Ms2ger: thanks, looking now [01:21:11.0000] <remysharp> odinho: good lord - give a vendor "implied" and we authors are fucked - if it can't be removed, it must be explicit about that. [01:21:11.0000] <jgraham> remysharp: Right, the browser would process the whole image tag at once and load the one correct resource [01:21:20.0000] <ShaneHudson> remysharp: haha! [01:21:42.0000] <jgraham> It's not implied, I assume [01:21:45.0000] <odinho> remysharp: Yes I know yes I know. :P I work in Opera. But it's still a draft under discussion :P [01:21:50.0000] <ShaneHudson> Yes I think we have all seen why the spec needs to be a tightly written as any contract [01:21:58.0000] <jgraham> Without reading the spec, I imagine it specifies which image resource should be displayed [01:22:01.0000] <kennyluck> jgraham, I am being constructive by giving ShaneHudson an answer he is looking for. [01:22:23.0000] <jgraham> Of course a browser could chose to load other resources if it wanted. But a browser *could* do anything [01:23:00.0000] <jgraham> kennyluck: I think it is more constructive to take the question a little less literally. [01:23:20.0000] <odinho> remysharp: It's specified. [01:23:38.0000] <odinho> remysharp: http://www.whatwg.org/specs/web-apps/current-work/multipage/embedded-content-1.html#update-the-image-data [01:24:26.0000] <remysharp> odinho: uses the word "updates" - should that be "load"? [01:24:26.0000] <remysharp> odinho: uses the word "updates" - should that be "load"? [01:24:40.0000] <remysharp> it seems to suggest it's resetting state (though I've not finished reading) [01:24:52.0000] <odinho> remysharp: load uses that algorithm. [01:25:10.0000] <jgraham> remysharp: first load seems like a special case of update [01:25:33.0000] <jgraham> ShaneHudson: FWIW I am *very* skeptical that anything involving media queries is the right solution [01:26:07.0000] <othermaciej> some browser engines will actually initiate the load before the "update the image data" steps would be triggered by they will presumably choose in the same way if they want to do a good job [01:26:33.0000] <jgraham> Becauseit should only be possible to vary on 3 properties: viewport width, viewport height and display density [01:26:55.0000] <othermaciej> (the spec doesn't specify anything about prefetching) [01:27:04.0000] <jgraham> Reusing a general syntax but neutering it so that most things people expect to work don't work seems like a terrible idea [01:27:06.0000] <odinho> othermaciej: It can probably use "process the image candidates" algorithm anyway: http://www.whatwg.org/specs/web-apps/current-work/multipage/embedded-content-1.html#update-the-image-data [01:27:22.0000] <jgraham> Right, prefetching is an optimisation [01:27:38.0000] <odinho> Hmmz. Why did I get a wrong URL there, meant http://www.whatwg.org/specs/web-apps/current-work/multipage/embedded-content-1.html#processing-the-image-candidates [01:27:47.0000] <annevk> remysharp: example of a browser vendor messing up that badly? [01:27:54.0000] <remysharp> video [01:28:06.0000] <remysharp> video referrers were missing from implementations [01:28:19.0000] <remysharp> it's not huge, but it was serious enough to nuke a few servers during hotlinking [01:28:56.0000] <remysharp> Specs are hard for "authors" to read - see appcache - and any ambiguity just makes everything harder for all involved [01:29:13.0000] <annevk> sure [01:29:23.0000] <jgraham> Right, that's why specs are hard to read [01:29:35.0000] <jgraham> (not just for authors, but they spend less time doing it) [01:29:42.0000] <ShaneHudson> Ok I will admit... hixie's email makes srcset look a lot better than it originally did. But I still do not like it [01:29:50.0000] <annevk> I was just curious as doing more network requests than required is something all browsers try very hard to avoid [01:30:16.0000] <jgraham> Well, not entirely [01:30:19.0000] <odinho> ShaneHudson: You want to write very clearly what you want to do. Not getting stuck on *how* you want to do it. [01:30:56.0000] <odinho> ShaneHudson: That is the best and easiest way to reply and give feedback. [01:31:03.0000] <jgraham> I think many implementations of speculation can cause network requests that are later not used [01:31:23.0000] <jgraham> Or starting loading DOM-created elements before they are inserted [01:31:49.0000] <ShaneHudson> odinho: hmm yes, I will go through it properly after I have been to the gym. But as a "front-end developer" I tend to think more of syntax than how it works technically, so need to balance the two [01:31:50.0000] <ShaneHudson> odinho: hmm yes, I will go through it properly after I have been to the gym. But as a "front-end developer" I tend to think more of syntax than how it works technically, so need to balance the two [01:33:05.0000] <odinho> ShaneHudson: Hm, you shouldn't have to think about other stuff than the frontend. But not the syntax, -- what you are getting as end-result, what your users are seeing. Etc etc. You don't need to speak in browser implementation requirements, most people don't - that's mostly for the last step. [01:34:12.0000] <odinho> E.g. what the outcome is in different scenarios for the same page :-) [01:34:38.0000] <annevk> [tm]: I think you need to modify Overview.src.html quite a bit for that [01:34:40.0000] <jgraham> To be fair, that's only really if you disagree with the requirements presented so far [01:34:52.0000] <annevk> [tm]: e.g. it needs some kind of license switch I guess [01:35:09.0000] <annevk> [tm]: a style sheet switch [01:36:04.0000] <jgraham> It seems to me that that if there is agreement that viewport dimensions / pixel density are the right axes to vary along then most of the rest of the questions are about syntax [01:36:51.0000] <jgraham> If they're not then presenting use cases - things you want to achieve - that require extra considerations is the most important first step, yes [01:37:33.0000] <odinho> :-) [01:40:01.0000] <ShaneHudson> Was the idea of bandwidth as a use case turned down completely? I think that although it is not easily possible at the moment, we should be future proofing too [01:41:03.0000] <odinho> ShaneHudson: Nope, I was maybe thinking about replying to that. [01:41:11.0000] <jgraham> I don't see how bandwidth will ever be possible to solve in this way [01:41:21.0000] <[tm]> annevk: yeah, no worries [01:41:26.0000] <[tm]> working on it now [01:41:29.0000] <jgraham> And hopefully will become less of a problem over time [01:41:37.0000] <annevk> [tm]: you can probably base it on DOM or something [01:41:57.0000] <odinho> jgraham: In the case where you just want the smallest size, it's very easy. E.g. I pay for data traffic over 1GB on my phone. [01:42:07.0000] <annevk> ShaneHudson: Hixie's email explains why bandwidth cannot be exposed as a meaningful metric [01:42:08.0000] <annevk> ShaneHudson: Hixie's email explains why bandwidth cannot be exposed as a meaningful metric [01:42:26.0000] <[tm]> need to get it done before plh goes on vacation, so we can his OK for FPWD transition [01:42:37.0000] <[tm]> annevk: I think I got it done for now [01:42:39.0000] <jgraham> But Opera have a bit of experience here; we try to use heuristics to suggest when you might want to switch to use Turbo mode which will conserve bandwidth [01:42:42.0000] <annevk> [tm]: he already gave his OK [01:42:42.0000] <odinho> annevk: It doesn't have to be exposed. As it will be an optimization, the browser can choose. [01:42:51.0000] <ShaneHudson> annevk: Does it? I know he says he cannot think how to do it, but something is likely to change or be invented in a few years [01:42:54.0000] <odinho> jgraham: Yeah, and we all know how bad that is :P [01:42:54.0000] <jgraham> This turns out to be *very* hard to get right [01:43:06.0000] <odinho> jgraham: But it's easier on phones. [01:43:12.0000] <ShaneHudson> Right I must be going, see you all later [01:43:13.0000] <[tm]> annevk: yeah, that was provisional on me actually getting it ready [01:43:21.0000] <jgraham> odinho: If you are on a data constrained plan, I suggest Opera mini :p [01:43:25.0000] <annevk> [tm]: oh lol, politics [01:43:35.0000] <odinho> jgraham: Because you can trust the OS knows if you're on GPRS, EDGE, 3G, Wifi. [01:43:44.0000] <annevk> [tm]: you'd think he would have heard of pubrules [01:43:45.0000] <jgraham> So? [01:44:16.0000] <jgraham> How to map that to "I want different assets" is decidedly non-obvious [01:44:52.0000] <jgraham> (on desktop you can also be on various different connection types, including some or all of those) [01:44:54.0000] <odinho> jgraham: ...? How? If I'm on GPRS I really want to download the smallest image, even though I'm using my KDE Spark tablet with a huge screen. [01:45:14.0000] <odinho> jgraham: Same with my main browser on my normal computer. NetworkManager also expose this information to programs running. [01:45:15.0000] <annevk> odinho: it seems you want the smallest image if you're on 3G as well because you're data-constrained [01:45:32.0000] <annevk> odinho: whereas someone with an iPad on 3G with no data constraints might not want that at all [01:45:33.0000] <jgraham> I have been on non-constrained 3G [01:45:38.0000] <jgraham> Right [01:45:42.0000] <annevk> odinho: so that doesn't seem like a meaningful metric [01:45:43.0000] <odinho> annevk: I'm not up to 1GB. [01:45:53.0000] <annevk> odinho: 1GB is constrained [01:46:01.0000] <odinho> Have any of you ever tried using GPRS? [01:46:02.0000] <jgraham> So a user would have to manually map connection type to desired assets [01:46:06.0000] <jgraham> It would be insane [01:46:07.0000] <odinho> It's very slow. [01:46:19.0000] <annevk> odinho: I know [01:46:22.0000] <jgraham> We can't have a feature that requires UI that no one would implement [01:46:28.0000] <odinho> My computer knows when I'm on GPRS. [01:46:49.0000] <odinho> I don't see why it couldn't just take the lightest assets when I am. [01:47:38.0000] <annevk> odinho: how many assets do you think the page is going to provide though? [01:47:51.0000] <jgraham> In that case I would probably turn on turbo/use mini [01:48:02.0000] <jgraham> And not rely on athors to get it right [01:48:09.0000] <annevk> yeah [01:49:52.0000] <odinho> annevk: Well, many sites are CMS driven, mine are, and I guess something like 280w, 576w, 1024w. [01:50:34.0000] <odinho> Yeah, Opera mini/turbo is a better fix, -- but it does rely on proxy servers. [01:51:18.0000] <jgraham> Sure [01:51:28.0000] <jgraham> Anyway, there was a solution to this problem in HTML [01:51:31.0000] <jgraham> lowsrc [01:51:32.0000] <odinho> But thing is, it wouldn't have to really expose anything extra. Although filesize would help, however I can't see that getting any actual use. So this is purely a potential optimization for a browser to do. [01:51:45.0000] <jgraham> It didn't really go anywhere [01:51:45.0000] <odinho> So there needs to be no spec change. [01:52:28.0000] <odinho> I'm merely saying, and meaning, that it *is* possible to use bandwidth and/or data usage as a metric and potential use case here. [01:53:51.0000] <jgraham> So your proposal is that, if it knows it is in a bandwidth constrained situation, the UA could choose the "wrong" asset to get one likely to be smaller [01:56:23.0000] <odinho> jgraham: Yep. Being a agent to the user. Or actually even, -- (although progressive images are better here) if you are in fact on a very slow network (not high latency (not saying how you find out that :P)), downloading the small file first to show in-place before doing the big one. [01:57:22.0000] <annevk> that's already allowed [01:57:24.0000] <annevk> "This allows a user agent to override the default algorithm (as described in subsequent steps) in case the user agent has a reason to do so. For example, it would allow the user agent in highly bandwidth-constrained conditions to intentionally opt to use an image intended for a smaller screen size, on the assumption that it'll probably be good enough. " [01:57:41.0000] <annevk> rtfs? [01:58:41.0000] <odinho> annevk: I have. I'm not discussing with the spec, I'm discussing with jgraham, which is saying something else :-) [01:59:14.0000] <jgraham> I'm not saying something else [01:59:27.0000] <jgraham> I'm saying having a feature for it in the spec is silly [02:00:39.0000] <odinho> jgraham: It doesn't really need a feature in the spec, -- because the current information should be enough. [02:00:51.0000] <zcorpan> Hixie: " [02:00:51.0000] <zcorpan> This seems like something that's currently relatively easily handled using [02:00:51.0000] <zcorpan> hidden="" or CSS, with some JS (or more CSS) to decide when to show what." - hidden="" and CSS don't stop the image from loading [02:00:51.0000] <zcorpan> This seems like something that's currently relatively easily handled using [02:00:52.0000] <zcorpan> hidden="" or CSS, with some JS (or more CSS) to decide when to show what." - hidden="" and CSS don't stop the image from loading [02:03:09.0000] <[tm]> annevk: can I add you as a co-editor on the URL spec? [02:07:49.0000] <annevk> [tm]: if arv is going to edit I don't think it's needed for me to be listed there [02:12:29.0000] <[tm]> OK [02:13:46.0000] <[tm]> annevk: http://dvcs.w3.org/hg/url/raw-file/tip/Overview.html [02:14:00.0000] <[tm]> I made a new make target [02:14:06.0000] <[tm]> WD target [02:14:33.0000] <[tm]> so we just flip it back to the "publish" target after WD publication [02:15:21.0000] <[tm]> anyway, break time here [02:15:26.0000] <annevk> [tm]: what I do these days is generate a TR.html copy so the editor's draft is not affected [02:15:26.0000] <[tm]> thank God almighty [02:15:30.0000] <annevk> heh [02:16:06.0000] <[tm]> annevk: OK, I can switch it to that later today [03:13:53.0000] <annevk> hmm [03:14:18.0000] <annevk> people on the WHATWG list don't seem to understand how pixel density affects images in practice [03:14:26.0000] <annevk> oh well [03:19:10.0000] <othermaciej> people seem not to get the fact that the pixel density you provide affects how the image renders [03:19:33.0000] <othermaciej> I'm not sure why this is hard to grasp (apparently) [03:20:11.0000] <othermaciej> people doing native iOS development seem to understand how 2x images fit into the picture [03:20:53.0000] <othermaciej> I think people may not realize that high pixel density means, ultimately, that CSS pixels != device pixels and most images are shown scaled up relative to the native device resolution [03:22:21.0000] <annevk> or I give feedback on how constraining URLs is not a good idea [03:22:33.0000] <annevk> "I've seen no objections about that aspect in the Community Group thread, where a number of authors have given feedback." [03:22:37.0000] <annevk> well I just gave you some [03:23:41.0000] <jgraham> Yeah, the implication that the source of the feedback somehow trumps its actual merit is distressing [03:34:45.0000] <annevk> /me stops with http://xkcd.com/386/ and goes to do something else [03:39:14.0000] <othermaciej> heh [03:48:35.0000] <annevk> so should we add new Range()? [03:52:07.0000] <[tm]> annevk: fyi, I reverted http://dvcs.w3.org/hg/url/raw-file/tip/Overview.html to being unmolested [03:52:22.0000] <[tm]> and made a TR.html for the TR version [03:52:28.0000] <[tm]> http://dvcs.w3.org/hg/url/raw-file/tip/TR.html [03:52:54.0000] <annevk> ta [03:53:38.0000] <annevk> [tm]: you should probably change the <title> and "Living Draft" in the <h2> does not work either for a TR publication [03:55:48.0000] <[tm]> hai [04:02:24.0000] <annevk> AryehGregor: how is the detach() experiment going? [04:03:45.0000] <annevk> Ms2ger: should we add baseLang? [04:03:59.0000] <annevk> we should have a better name ideally... [04:04:45.0000] <Ms2ger> What's baseLang? [04:05:21.0000] <annevk> language of the node/element [04:06:01.0000] <annevk> as determining it through script is non-trivial [04:06:26.0000] <Ms2ger> Hmm [04:06:49.0000] <annevk> can baseURI still be null btw? [04:07:01.0000] <annevk> DOMString? looks like a bug if everything starts with about:blank [04:10:44.0000] <[tm]> kinuko has landed a first draft of the Quota API spec: http://dvcs.w3.org/hg/quota/raw-file/default/Overview.html [04:12:24.0000] <annevk> respec :/ [04:14:12.0000] <[tm]> yeah [04:16:03.0000] <annevk> emailed some feedback to public-webapps [04:16:15.0000] <annevk> I assumed he's subscribed since I didn't find an email address [04:38:31.0000] <annevk> language barrier ftl https://www.w3.org/Bugs/Public/show_bug.cgi?id=17042 [04:47:03.0000] <annevk> zcorpan: hey yt? [04:47:06.0000] <annevk> zcorpan: you reported https://www.w3.org/Bugs/Public/show_bug.cgi?id=16712 [04:47:18.0000] <annevk> zcorpan: I'm trying to figure out how you could ever hit that case 2 [04:47:27.0000] <annevk> zcorpan: because it seems case 1 covers it aleady [04:48:32.0000] <zcorpan> annevk: the element doesn't have a prefix [04:53:06.0000] <annevk> zcorpan: exactly, so its namespace prefix equals prefix [04:53:10.0000] <annevk> zcorpan: they're both null [04:53:59.0000] <zcorpan> annevk: /prefix/ is "bar" [04:55:52.0000] <annevk> zcorpan: right [04:56:05.0000] <annevk> zcorpan: I guess what I'm saying is if I remove ', or whose namespace prefix is null and local name is "xmlns"' does something fall apart? [04:56:24.0000] <annevk> zcorpan: hmm I guess so [04:56:29.0000] <annevk> aah namespaces [04:56:39.0000] <zcorpan> :) [04:58:10.0000] <zcorpan> "or whose namespace prefix is null and local name is "xmlns" if prefix is null:" [04:58:42.0000] <zcorpan> (possibly also check the namespace of the "xmlns" attribute) [04:58:42.0000] <zcorpan> (possibly also check the namespace of the "xmlns" attribute) [04:58:45.0000] <annevk> I think I'll add some words for clarity [05:06:58.0000] <annevk> man this is some long sentence :( [05:14:01.0000] <[tm]> annevk: kinuko is a woman, btw [05:15:14.0000] <zcorpan> annevk: looks good [05:15:17.0000] <[tm]> and she is subscribed to public-webppas [05:17:37.0000] <annevk> cool cool [05:17:37.0000] <annevk> cool cool [05:18:09.0000] <annevk> is it just the -o or -ko that indicates the name of a woman in Japan? [05:22:21.0000] <annevk> Ms2ger: garbage collection in DOM [05:23:01.0000] <annevk> Ms2ger: where should we put that [05:23:07.0000] <annevk> yeah leave the channel alright [05:23:09.0000] <annevk> chicken [05:23:09.0000] <annevk> chicken [05:23:59.0000] <[tm]> annevk: -ko [05:24:23.0000] <charlvn> there are japanese men who also have names ending in -o or -ko afaik [05:24:32.0000] <[tm]> yeah [05:24:37.0000] <charlvn> although i think -ko is more common with female names [05:26:27.0000] <[tm]> right. there aren't a lot of men's names that end in -ko and you can almost always tell from the name if it's a men's name or a women's regardless [05:26:56.0000] <annevk> ah, the chicken returns :) [05:27:08.0000] <[tm]> I think the men's names are usually -hiko [05:27:22.0000] <charlvn> [tm]: sounds right [05:28:20.0000] <Ms2ger> Hah [05:28:44.0000] <Ms2ger> annevk, in the closet? [05:29:19.0000] <annevk> ideally, but smaug____ wants it [05:29:30.0000] <annevk> HTML has http://www.whatwg.org/specs/web-apps/current-work/#garbage-collection [05:29:31.0000] <smaug____> what do I want? [05:29:42.0000] <annevk> garbage collection notes in DOM [05:30:35.0000] <smaug____> not really gc, but definition of ownership [05:30:50.0000] <smaug____> or definition in which cases certain objects won't be deleted [05:30:54.0000] <Ms2ger> Everything owns everything :) [05:31:24.0000] <smaug____> whether gc is used internally, is implementation detail [05:33:03.0000] <annevk> not really sure we have to say much then [05:33:09.0000] <annevk> other than that sentence of HTML [05:33:23.0000] <annevk> https://www.w3.org/Bugs/Public/show_bug.cgi?id=16638 suggests saying something about weak references, but that's more of an impl detail [05:36:15.0000] <smaug____> "Unless disconnect() is called, MutationObserver observes node as long as the node exists" [05:36:20.0000] <smaug____> something like that [05:37:25.0000] <annevk> that doesn't mean much [05:39:41.0000] <smaug____> it doesn't ? [05:39:57.0000] <smaug____> it means that the MutationObserver doesn't die even if you don't keep a reference to it [06:14:24.0000] <[tm]> btw, http://lists.w3.org/Archives/Public/public-whatwg-archive/ is not available [06:14:33.0000] <[tm]> *now available [06:16:05.0000] <annevk> yeah noticed that yesterday [06:16:07.0000] <annevk> very cool [06:16:52.0000] <[tm]> now we should figure out how to add an Archived-at header to whatwg⊙wo messages [06:29:26.0000] <zewt> used to be you could just search for message-ids, but gmail makes it a pain to get it, which is annoying [06:41:41.0000] <zcorpan> [tm]: archived-at would be nice [06:45:56.0000] <dbaron> anybody know of a description of the advantages of WebVTT over TTML (or some subset thereof)? [06:48:35.0000] <annevk> dbaron: http://lists.w3.org/Archives/Public/public-html/2010May/0160.html [06:49:12.0000] <dbaron> ok, so now suppose that the TTML folks are willing to redefine it to be on top of HTML+CSS instead of XSL-FO? [06:50:27.0000] <dbaron> (may or may not actually be the case, but under discussion) [06:50:29.0000] <annevk> I guess that leaves the enormous amount of namespaces and complexity of having a lot of markup where something simpler does fine [06:51:26.0000] <annevk> and draconian error handling for a text format which would be a first as far as browser technology goes [06:51:57.0000] <zewt> annevk: i don't know anything about it, but that description makes it sound like they have no experience with web formats [06:52:47.0000] <zewt> brb reboot [06:52:49.0000] <annevk> TTML is badly designed [06:53:01.0000] <annevk> that's why we went with WebVTT [06:53:15.0000] <zewt> actually no reboot [06:55:01.0000] <zewt> dbaron: fwiw since VTT seems like it'll work fine, I don't think any amount of "what if we do this" will make web people interested in a different captioning format (ttml or otherwise), there's no problem solved by that [06:55:13.0000] <dbaron> so what's going on is: [06:55:25.0000] <dbaron> (a) w3c is looking at chartering a WG to do a new version of TTML: http://lists.w3.org/Archives/Public/public-new-work/2012Apr/0005.html [06:55:46.0000] <dbaron> (b) there's apparently a US government regulation that's going to go into effect in September that may actually require captioning in TTML [06:55:59.0000] <zewt> that sounds BS [06:56:05.0000] <zewt> (in the "not true" sense, not "it's stupid" sense) [06:56:48.0000] <zewt> (not that it wouldn't be the latter too :) [06:57:13.0000] <annevk> I heard about (b) too [06:57:13.0000] <dbaron> something from http://www.fcc.gov/encyclopedia/video-programming-accessibility-advisory-committee-vpaac [06:57:29.0000] <annevk> (a) does not matter much; W3C charters groups to work on silly stuff all the time [06:57:40.0000] <zewt> i heard something about how a specific format was given as an example of a format that could be used, and people read that and went "so now that's required!" [06:58:49.0000] <zewt> http://lists.w3.org/Archives/Public/public-texttracks/2012Apr/0001.html [06:58:58.0000] <espadrine> if the TTML group is ready to redefine it to be on top of HTML+CSS, surely the government cannot push a requirement to use something whose spec is changing [06:59:17.0000] <dbaron> apparently it would be VPAAC WG1 [06:59:30.0000] <Ms2ger> espadrine, Ha. Ha. Ha. [07:00:03.0000] <dbaron> apparently the bigger problem is that everybody implements a different subset of TTML [07:01:04.0000] <annevk> isn't the bigger problem that it's a terrible format? [07:01:27.0000] <dbaron> meant the bigger problem with the requirement [07:01:28.0000] <annevk> thanks zewt [07:01:29.0000] <dbaron> but yes, that too [07:01:54.0000] <zewt> there's nothing else leading to the "something something federal requirement" noise, right? [07:02:47.0000] <annevk> during the F2F glenn mentioned it and a couple of people were like "yup that's correct; tough luck WebVTT guys" [07:03:11.0000] <odinho> Hehe, yeah, fear struck. [07:03:25.0000] <zewt> "that" glenn? heh [07:03:32.0000] <annevk> other glenn! [07:03:34.0000] <dbaron> SMPTE-TT is a profile (with additions) of TTML, no? [07:03:54.0000] <dbaron> not a container for it as that message implies? [07:04:06.0000] <zewt> captioning isn't rocket science; if you need *profiles* of a captioning format, something seems badly amiss [07:04:28.0000] <jgraham> Well clearly something *is* badly amiss [07:04:33.0000] <zewt> something usually is [07:06:44.0000] <annevk> dbaron: looks like a super-subset, yes [07:07:19.0000] <annevk> as in, it supersets a subset of TTML [07:07:52.0000] <jgraham> Isn't that what we call a "different format"? [07:08:05.0000] <annevk> it uses namespaces so it's cool [07:08:09.0000] <zewt> heh [07:08:11.0000] <[tm]> dbaron, zewt: as far as that supposed US government requirement for TTML, see http://fjallfoss.fcc.gov/edocs_public/attachmatch/FCC-12-9A1.pdf [07:08:15.0000] <zcorpan> does it have 18 new namespaces? [07:08:28.0000] <[tm]> and look for "apparatus" [07:08:34.0000] <zewt> annevk: someone on the webgl list said he wants to push for using the URL-namespacing-gimmick to identify webgl extensions [07:08:39.0000] <zewt> raaaaaaaaage [07:09:06.0000] <zewt> re: xml tried that. it sucked. stop it [07:09:10.0000] <dbaron> is that pdf the same as https://www.federalregister.gov/articles/2012/03/30/2012-7247/closed-captioning-of-internet-protocol-delivered-video-programming-implementation-of-the ? [07:09:38.0000] <annevk> zewt: "feature" in object does not work anymore? [07:09:52.0000] <annevk> zewt: but yeah, ... [07:09:53.0000] <zewt> oh everything works fine [07:09:59.0000] <zewt> he just wanted to change to urls because *crickets* [07:10:42.0000] <odinho> [tm]: They sure say a lot of apparatus all over the place! [07:10:51.0000] <dbaron> /me heads back indoors [07:11:20.0000] <zewt> fortunately he didn't go on to push it, but it's astonishing that nonsense keeps cropping up [07:12:16.0000] <jgraham> Those damn crickets [07:12:41.0000] <annevk> http://cdn.memegenerator.net/instances/400x/20446042.jpg [07:14:30.0000] <[tm]> odinho: the section "A. Apparatus Subject to Section 203 of the Act [07:17:14.0000] <odinho> [tm]: Page 65 seems interesting. But nothing about TTML. [07:17:54.0000] <[tm]> odinho: "If a [07:17:55.0000] <[tm]> video programming owner provides captions to a video programming distributor or provider using the [07:17:57.0000] <[tm]> Society of Motion Picture and Television Engineers Timed Text format (SMPTE ST 2052-1:2010: [07:18:00.0000] <[tm]> “Timed Text Format (SMPTE-TT)” 2010) (incorporated by reference, see § 79.100), then the VPO has [07:18:03.0000] <[tm]> fulfilled its obligation to deliver captions to the video programming distributor or provider in an [07:18:06.0000] <[tm]> acceptable format. [07:18:55.0000] <[tm]> and "The VPAAC proposed that the Commission require a single standard interchange format [07:18:59.0000] <[tm]> so that video programming does not need to be re-captioned to comply with different standards. [07:19:02.0000] <[tm]> 508 [07:19:04.0000] <[tm]> The [07:19:07.0000] <[tm]> VPAAC proposed SMPTE-TT as the standard interchange format [07:19:17.0000] <odinho> [tm]: Ah, -- but the general requirements are that you can scale the text and do all sorts of other things, so it doesn't really exclude WebVVT. [07:19:37.0000] <odinho> But that last thing is not written in-doc as requirement, I guess? They only propose that as single standard? [07:20:10.0000] <[tm]> odinho: they are careful to not say it is "mandatory" [07:20:17.0000] <[tm]> but instead "safe harbor" [07:20:31.0000] <[tm]> I don't know that that particular term of art means [07:20:59.0000] <[tm]> e.g., "unlike adopting SMPTE-TT as the mandatory interchange or delivery format, commenters explain that a [07:21:08.0000] <[tm]> safe harbor approach would balance goals of efficiency, certainty, and consumer access with needed [07:22:47.0000] <odinho> e will also provide in our rules that devices that implement SMPTE-TT [07:22:48.0000] <odinho> will be deemed in compliance with our rules, while simultaneously allowing devices to achieve the same [07:22:50.0000] <odinho> functionality without implementing that standard.5 [07:23:38.0000] <odinho> "We intend to monitor the marketplace and, to the extent that additional open standards from recognized industry standard-setting organizations appear appropriate, we will consider incorporating those standards into our rules as additional safe harbors. [07:23:43.0000] <odinho> So it seems it can be done ;-) [07:54:11.0000] <zcorpan> Hixie: which characters are supposed to be paragraph separators in WebVTT for the purpose of bidi? [07:54:54.0000] <zcorpan> just Bidi_Class=Paragraph_Separator ? [07:55:32.0000] <zcorpan> tr9 isn't really explicit about which characters it wants to consider paragraph separators [07:57:16.0000] <zcorpan> e.g. U+2028 isn't class B, but i can't tell if it's an "appropriate Newline Function" or not [08:00:45.0000] <odinho> I surely don't understand why some people are so agressively against srcset. And noone can say why. [08:00:53.0000] <odinho> (... reading twitter) [08:02:12.0000] <zcorpan> oh, i think i found the definition of "newline function" now [08:02:50.0000] <[tm]> odinho: you can save yourself the confusion by not reading twitter :) [08:03:24.0000] <zcorpan> so newline function is platform-dependent, but none of the platforms listed as examples use any non-B-class character as newline function [08:03:35.0000] <zcorpan> so it's just B class [08:14:01.0000] <TabAtkins_> odinho: Because http://w3cmemes.tumblr.com/post/22831818920/the-css-wgs-primary-failure-mode , that's why. [08:17:23.0000] <annevk> "If you don't set your image's resolution appropriately, you'll get unexpected sizing effects." is not actually true [08:17:29.0000] <annevk> browsers ignore the DPI set in the image [08:18:09.0000] <TabAtkins_> annevk: What I meant is that if you ship, say, a 140dpi image but declare it to be 2x, the auto-size wont' be what you may have expected. [08:18:20.0000] <TabAtkins_> At least, it won't be the same as a 96dpi image at 1x. [08:18:27.0000] <odinho> TabAtkins_: Yeah, thinking that. But saying it alound feels wrong... :] [08:19:05.0000] <annevk> TabAtkins: it's way easier to think about it in terms of pixel density rather than resolution [08:19:27.0000] <TabAtkins_> odinho: It's not completely fair. A good bit of it is people not understanding the weaknesses of MQ for some of the use-cases we want to solve, and thus reacting against what they think is just a weird syntax for no reason. [08:19:49.0000] <annevk> TabAtkins: e.g. everything from 90-130 is typically treated as 1x [08:19:58.0000] <TabAtkins_> annevk: Do image programs talk about pixel density? I thought they explicitly talked about "dpi" when saving an image. [08:20:02.0000] <annevk> TabAtkins: whereas 260-330ppi would be 2x [08:20:26.0000] <annevk> TabAtkins_: in an image program you would just make an image with twice the amount of pixels in each direction [08:20:44.0000] <annevk> TabAtkins_: you don't use the dpi setting because it is already ignored by browsers (and has to for compatibility) [08:20:54.0000] <annevk> I sort of thought everyone involved in the discussion knew this... [08:20:54.0000] <TabAtkins_> annevk: I don't understand. Yes, that's how you map screens to pixel density. But that doesn't have much to do with how the Nx argument works in srcset. [08:20:55.0000] <annevk> I sort of thought everyone involved in the discussion knew this... [08:20:55.0000] <TabAtkins_> annevk: I don't understand. Yes, that's how you map screens to pixel density. But that doesn't have much to do with how the Nx argument works in srcset. [08:21:13.0000] <jgraham> TabAtkins_: So someone should propose <srcset><img src="something.png"><source src="somethingelse.png" width=200 height=100 density=2></srcset> or something [08:21:25.0000] <annevk> you have a 10x10 image and a 20x20 image [08:21:28.0000] <TabAtkins_> annevk: Uh, yes, I know. [08:21:29.0000] <TabAtkins_> annevk: Uh, yes, I know. [08:21:30.0000] <annevk> the latter you set as 2x [08:21:58.0000] <annevk> you don't have to care about dpi [08:22:07.0000] <jgraham> I think that is a kind of horrible syntax compared to to srcset, but maybe that's why I'm a browser QA not a web dev. [08:22:07.0000] <annevk> which is nice because it various all over [08:22:26.0000] <TabAtkins_> annevk: But that's exactly equivalent to giving an image of the same size and double resolution (because browsers only care about the pixels anyway). [08:22:46.0000] <TabAtkins_> And when you talk about resolution, people will want to do that. [08:23:24.0000] <jgraham> When people talk about "size" they mean "in pixels", no? [08:23:26.0000] <annevk> if it's double resolution it's width and height will be larger [08:23:49.0000] <TabAtkins_> jgraham: Sometimes, sometimes not. When I start GIMP, I can create an image with a size in inches and a resolution in dpi. [08:23:55.0000] <odinho> TabAtkins_: <img src=10x10 srcset="20x20.jpg 2x"> is exactly equivalent to <img src=20x20.jpg width=10 height=10> [08:24:04.0000] <TabAtkins_> odinho: Yes, I know. [08:24:22.0000] <annevk> you cannot create an image of 10x10 with more detail and have it rendered as 10x10 [08:24:27.0000] <TabAtkins_> annevk: Yes, I know. [08:24:28.0000] <annevk> there's no such thing [08:24:39.0000] <annevk> well then I've no idea what you're talking about [08:24:55.0000] <odinho> :S [08:25:19.0000] <jgraham> annevk: TabAtkins_ seems to (sometimes) be using "size" to mean something that can be measured with a ruler [08:25:30.0000] <TabAtkins_> If you create a "5 inch wide" image at 192dpi, browsers will display it as 10 inches wide. [08:25:45.0000] <jgraham> It seems clearer to me to always use "size" to mean something in pixels [08:26:48.0000] <jgraham> Because, y'know, working out what real world dimensions you get isn't that easy or helpful [08:27:13.0000] <annevk> TabAtkins_: aren't all images stored as pixels in the end? [08:27:19.0000] <TabAtkins_> annevk: Yes? [08:27:30.0000] <annevk> so then I'm not sure why we have to talk about inches [08:27:30.0000] <TabAtkins_> The point is author expectations. [08:27:39.0000] <annevk> I think authors deal in pixels too [08:27:43.0000] <TabAtkins_> I'm not *sure* if authors think that way, but I suspect that some do. I dunno. [08:27:45.0000] <necolas> 6 months ago, when we (developers) started exploring avenues for responsive images, hixie said: "<img> parsing is pretty much a lost cause" and warned us off the exact kind of syntax that is now being considered. any idea why there has been a change of heart? [08:27:49.0000] <annevk> I certainly do when creating images [08:27:50.0000] <odinho> Ah, they're doomed anyway if they think about stuff like that. [08:28:00.0000] <annevk> necolas: <img> parsing is a lost cause [08:28:01.0000] <annevk> necolas: <img> parsing is a lost cause [08:28:04.0000] <odinho> Some print guys do, and they're always way off. Have to explain. [08:28:05.0000] <TabAtkins_> necolas: Changing the parsing of the <img> element is a lost cause. [08:28:09.0000] <annevk> necolas: adding attributes does not affect parsing [08:28:20.0000] <TabAtkins_> Dammit, annevk, why are you 2 seconds faster than me? [08:28:27.0000] <necolas> adding attributes doesn't help with polyfilling either [08:28:29.0000] <odinho> TabAtkins_: In europe? [08:28:33.0000] <annevk> TabAtkins_: Europe :p [08:28:37.0000] <odinho> FTW! [08:28:40.0000] <jgraham> necolas: I answered that once already today [08:29:54.0000] <jgraham> Adding attributes really doesn't seem that bad for polyfilling, and provides the simplest-to-implement solution that will therefore get deployed the soonest [08:29:56.0000] <necolas> i guess we're all still trying to understand why developers have been cast aside on this matter, despite it being something that developers have invested a lot of time in discussing etc [08:30:16.0000] <jgraham> If you are trying to understand that, I see the problem [08:30:24.0000] <jgraham> You are working from a false premise [08:30:30.0000] <necolas> jgraham: you don't think having 2 http requests is a problem for polyfilling? [08:30:39.0000] <necolas> jgraham: and what premise would that be? [08:30:48.0000] <jgraham> "developers have been cast aside" [08:31:13.0000] <TabAtkins_> necolas: If your goal is something that doesn't render at all in legacy browsers (but will be made to work via JS), just omit @src and only use @srcset. Done. [08:31:28.0000] <necolas> that isnt the goal [08:31:48.0000] <TabAtkins_> necolas: Oh. Then I don't understand what you mean by "2 http requests are a problem for polyfilling". [08:31:52.0000] <necolas> jgraham: [the work of] developers [08:32:10.0000] <odinho> necolas: <img srcset="cat.jpg, cat⊙2 2x"><noscript><img src=cat.jpg></noscript> <---- you can polyfill that without 2 HTTP requests. [08:32:25.0000] <necolas> TabAtkins_: the browser is going to request the image in `src` before any JS can get to work [08:32:38.0000] <TabAtkins_> necolas: The work done by people developing standards (which, in this case, includes all the developers participating in Adaptive/Response Images thing) is worthless. Never, *ever* try to evaluate a solution by which had the most work put into it. [08:32:44.0000] <necolas> odinho: yeah we've discussed that but dont think it is optimal [08:33:14.0000] <TabAtkins_> necolas: ...that's why I just suggested using only @srcset and omitting @src. [08:33:49.0000] <necolas> TabAtkins_: that's not what we're doing. but basically there has been little to no engagement with the devs who were already interested in this problem [08:34:07.0000] <jgraham> necolas: The fact that people have brought up technical shortcomings does not mean that "the work has been cast aside" [08:34:29.0000] <jgraham> Wht do you think this is if not engagement? [08:34:37.0000] <TabAtkins_> necolas: I'm not sure what a thread that explicitly thanks people for all the useful work they've done, and explicitly addresses their concerns and criticisms while crafting a solution based on their work, is, other than engagement. [08:35:02.0000] <jgraham> Seriously, if there are technical problems with the srcset proposal, point them out [08:35:32.0000] <jgraham> It's a way more useful use of everyone's time than complaining about lack of engagement [08:35:40.0000] <necolas> Aside from discussions of the technical merit, I'm saying you that the developer community doesn't feel like how you described, Tab. [08:36:18.0000] <TabAtkins_> necolas: That's fine. Most people don't like it when their personal preferred solutions aren't used. That's part of standards, though. [08:36:43.0000] <TabAtkins_> necolas: Since I'm sure a lot of these devs are having their first brush with actual standards development, that mismatch is to be expected. [08:36:44.0000] <necolas> I just wish there had been better communication [08:36:55.0000] <Philip`> /me thinks there ought to be a way of standardising the technically best solution without also alienating people who have proposed other solutions [08:36:57.0000] <necolas> TabAtkins_: this has nothing to do with whos solution is getting used [08:37:24.0000] <TabAtkins_> necolas: Do you think the discussion is completely over and this is a fait accompli? That's incorrect. [08:37:24.0000] <dglazkov> good morning, Whatwg! [08:37:45.0000] <necolas> I really don't think devs are wedded to "their ideas" but are simply looking to be included in discussions and have their interests heard. [08:37:54.0000] <adactio> What necolas said. [08:37:57.0000] <odinho> Philip`: Yes, sure. But it's not super easy if people feel hurt. [08:38:02.0000] <necolas> because, initially, responsive images was not of interest to you guys, we set up a CG [08:38:07.0000] <jgraham> Philip`: I guess we could drug all the people that back the wrong solution, so they are unaware that their solution didn't get picked [08:38:08.0000] <TabAtkins_> necolas: Again, I'm not sure how what has been happening is anything if not "hearing dev's interest". [08:38:24.0000] <necolas> and Wilto put in a LOT of time. and if he feels upset, then you should look at how we can all avoid this kind of thing in the future [08:38:47.0000] <TabAtkins_> I mean, responsive images are in the spec now, when before they weren't, and it's *entirely* due to the work of devs, mostly in the CG. [08:39:05.0000] <TabAtkins_> How is this anything other than "you win, your arguments were right, here's a solution for your problem". [08:39:21.0000] <necolas> jgraham: that doesnt sound like an unpleasant experience :) [08:39:38.0000] <necolas> TabAtkins_: "thanks, we'll take over from here", right? [08:39:50.0000] <TabAtkins_> We didn't get together and politely ask everyoen which solution they liked the most and hold a vote. That's because HTML uses the benevolent-dictator model. [08:39:59.0000] <jgraham> The problem with being upset in technical discussions is that it doesn't scale very well. [08:40:05.0000] <TabAtkins_> (Also because voting is usually a bad way to make technical decisions.) [08:40:08.0000] <jgraham> You would end up being upset a lot of the time [08:40:16.0000] <odinho> Noone has really given any feedback of why they don't like the new thing though. - There is also lots of intermixing of different problems that are solving different things. [08:40:23.0000] <dglazkov> oh cool, browserfolk vs. webdevfolk [08:40:39.0000] <TabAtkins_> necolas: You are more than welcome to provide feedback on the solution that Hixie proposed, though. That's what I'm doing, for example. [08:40:41.0000] <zcorpan> fight! [08:40:42.0000] <dglazkov> /me is both. Should I be arguing with myself? [08:40:44.0000] <jgraham> Certianly *I* would if I was upset every time my initially-preferred solution didn't make the final spec [08:40:49.0000] <necolas> imo, this isnt purely about the technical details. we're not asking for votes (even though that seems to sometimes get used for CSS specs). [08:40:50.0000] <necolas> imo, this isnt purely about the technical details. we're not asking for votes (even though that seems to sometimes get used for CSS specs). [08:40:55.0000] <zcorpan> dglazkov: clearly you should hit yourself in your face [08:41:05.0000] <dglazkov> zcorpan: and enjoy it [08:41:12.0000] <hober> dglazkov: count me in! /me hits himself in the face too :) [08:41:14.0000] <jgraham> /me still thnks we should have done MathML by making browsers understand something like LaTeX [08:41:36.0000] <necolas> jgraham: well, as devs, we're used to our "awesome solution" being trashed and reimagined by someone else in the OSS community. so i guess we don't care so much :) [08:41:40.0000] <jgraham> Oh, right, Tuesday night is S&M night [08:41:48.0000] <odinho> I think many of us "browserfolk" actually are former/present webdevs as well. I'm still more dev than browserfolk still after being in for so little. [08:41:50.0000] <zcorpan> jgraham: that was on the table before we put mathml in to html, wasn't it? [08:42:10.0000] <jgraham> zcorpan: Well I tried to convivce Hixie it was a good idea then [08:42:18.0000] <jgraham> I guess it would have been a lot of work [08:42:36.0000] <jgraham> But afaict it is still the most popular way of including MathML in HTML [08:42:43.0000] <jgraham> But implemented in Javascript [08:42:57.0000] <annevk> he tried to think of a scheme that implied a lot of the markup [08:43:02.0000] <annevk> but that didn't work out [08:43:30.0000] <zcorpan> annevk: iirc real LaTeX was also considered [08:43:32.0000] <dglazkov> whoa, how did we just switch from srcset to mathml? [08:43:47.0000] <dglazkov> is this the equivalent of Godwin's law in the standards discussions? [08:44:03.0000] <dglazkov> once LaTeX is mentioned... [08:44:03.0000] <necolas> odinho: oh of course. there are many aspects of being a web developer [08:44:03.0000] <miketaylr> probably all the punching in the face [08:44:08.0000] <annevk> necolas: so Hixie did reply in length to the various emails on the subject so saying the CGs input was not even considered seems wrong [08:44:09.0000] <scottjehl> hello! [08:44:18.0000] <necolas> annevk: i didn't say that [08:44:37.0000] <TabAtkins_> odinho: I don't think that "most" browser people were webdevs. Some. But a significant fraction of the CSSWG, frex, has never written a non-trivial site. ^_^ [08:45:18.0000] <odinho> TabAtkins_: Hehe, okay. In my limited experience in Opera though, most people were webdevs :-) [08:45:20.0000] <hober> TabAtkins_: that's why you, me, and some other people get to try to make sure that wg keeps it real. :) [08:45:33.0000] <TabAtkins_> odinho: Opera might be different, I dunno. [08:46:26.0000] <TabAtkins_> necolas: The bottom line is, if you don't like the solution, say so. Complaining about process is *very rarely* productive, however. [08:46:36.0000] <adactio> hober: I'm curious. In your email to the WHATWG on May 10th when you proposed srcset, why didn't you mention the Responsive Images Community Group? Were you genuinely unaware of its existence? [08:46:51.0000] <TabAtkins_> necolas: (But, hopefully, don't rehash criticisms that have already been answered, unless you have new information.) [08:47:03.0000] <necolas> even with web devs in the whatwg, i hope you still believe it is worthwhile to have input from people outside the whatwg who are building large and complex sites with these technologies [08:47:07.0000] <zcorpan> just saying "this is not what i want!" is also not very useful, though. [08:47:20.0000] <necolas> that is not what i was saying [08:47:21.0000] <TabAtkins_> necolas: Once again, that is *exactly* what happened in this case. [08:47:40.0000] <necolas> i think it is fair to discuss the way that things get done [08:47:47.0000] <zcorpan> necolas: sorry, i didn't mean you, i just saw some bugs filed along those lines [08:47:54.0000] <TabAtkins_> necolas: A few webdevs suggested something, it got shot down, more came together and made a cogent case, it was accepted. That's exactly how things should work. [08:49:19.0000] <TabAtkins_> adactio: From what I understand, hober had written that email a few weeks ago (around the time he wrote the email proposing image-set() for CSS), he just hadn't gotten around to sending it. Hixie talking about adding it into the spec made him hurry up. ^_^ [08:49:39.0000] <TabAtkins_> adactio: I believe the timing meant that at the time of writing, the CG was either not started yet, or just starting. [08:49:43.0000] <necolas> TabAtkins_: I dont know how else to explain to you that your vision of events isn't how many people in the CG feel. so either way, there has been a communication breakdown which is worth accepting as not-ideal [08:50:41.0000] <hober> adactio: sorry, about to get off the bus. be back in <30 min; for now, http://krijnhoetmer.nl/irc-logs/whatwg/20120511#l-202 [08:50:45.0000] <TabAtkins_> necolas: I'm quite certain there are people who dont' feel it went down that way. Unfortunately, that is *always* the case. I think I'm *extraordinarly* receptive to the input of other webdevs (because I was one, and still am in a limited capacity), but I still get people angry when I don't take their solution. [08:51:14.0000] <TabAtkins_> necolas: I don't like to characterize it as sour grapes, but that's what it feels like to me. [08:51:40.0000] <necolas> well you'd be wrong [08:51:59.0000] <necolas> and i don't think you should dismiss this like that [08:52:13.0000] <TabAtkins_> necolas: I know that, for example, Wilcox is unhappy that his specific solution wasn't used. It has some merits beyond @srcset, but also some strong weaknesses (which I brought up in the thread that Wilto started). [08:52:31.0000] <necolas> most of us have pretty much nothing invested in the ideas themselves - they all came in some form from earlier mailing list ideas. [08:52:42.0000] <necolas> TabAtkins_: hah, but wilcox is always unhappy :) [08:52:47.0000] <adactio> TabAtkins_: It's not about an alternate solution being *rejected* so much as the existence of an alternate solution (in development for months) being acknowledged at all. [08:52:50.0000] <TabAtkins_> Overall I find it an unacceptable solution without changes, and even then, the fact that it relies on brand-new functionality like URL rewriting and MQ variables makes it a harder sell than @srcset. [08:53:08.0000] <necolas> definitely [08:53:40.0000] <TabAtkins_> necolas, adactio: If it's about feeling bad because you don't think you were adequately credited, then I think that's an ego problem. ^_^ [08:53:58.0000] <jgraham> That particuolar solution would probably require a huge amount of spec work to turn into something usavle, and then be hideously complex to implement so it was either not implemneted for ages, or implemented in a very buggy way [08:54:10.0000] <necolas> TabAtkins_: [facepalm] it's not about ego either [08:54:18.0000] <jgraham> *usable [08:54:59.0000] <TabAtkins_> necolas: Quote from Hixie's email " thank you to everyone for making very good points, both here on the list and on numerous blog posts and documents on the Web, referenced from these threads". [08:55:41.0000] <necolas> People put a lot of work and time into that group. They just dont want to feel like it was wasted. Feeling invested in the technologies we use and feeling in partnership with the people taking time to make the specs...is a good thing. [08:56:16.0000] <necolas> TabAtkins_: have that kind of thing filter down to the CG itself would be a great start [08:56:40.0000] <TabAtkins_> necolas: It's there own choice to feel like it was wasted. Given the fact that we went from (1) idea was rejected, (2) CG was formed, people made good arguments, (3) idea was accepted, it's clear that the time *was* useful. [08:56:55.0000] <TabAtkins_> necolas: Feel free to forward the email to the CG, then! [08:57:17.0000] <TabAtkins_> bbiab [08:57:20.0000] <scottjehl> just catching up here, but I see there are criticisms of syntax, and criticisms of srcset's applicability today. The latter is more critical. I posted some concerns on that today. Are these not valid? (perhaps they were already mentioned above..?) https://gist.github.com/2701939#gistcomment-319724 [08:57:39.0000] <adactio> TabAtkins_: Are you being wilfully obstreperous? A solution was proposed (srcset) without knowledge of the existing proposal (picture). Once the existence was acknowledged, instead of evaluating both on their merits, one was chosen simply because it came from inside the WHATWG — *not* on merit. [08:57:51.0000] <Wilto> So. [08:58:05.0000] <divya> adactio: fwiw there are some concerns raised about picture element [08:58:08.0000] <annevk> adactio: actually hober knew about <picture> [08:58:10.0000] <divya> as there are w.r.t srcset. [08:58:16.0000] <annevk> adactio: and used that in his evaluation [08:58:17.0000] <zewt> (hard to take people seriously when they use words like "obstreperous") [08:58:18.0000] <necolas> Tab, I'm only trying to highlight something that you don't seem to have been aware of, and instead of thinking that - whatever the reason - developer dissatisfaction is worthy of concern, you don't. [08:58:21.0000] <odinho> scottjehl: If giving feedback, it has to be on the mailing list for it to be seen and replied to. [08:58:24.0000] <divya> (lol zewt) [08:58:29.0000] <annevk> adactio: and <picture> and variants has come up since 2007 or so [08:58:43.0000] <annevk> adactio: it's not a recent invention [08:58:48.0000] <zewt> odinho: heh my thought was "we're hiding discussions away on github now?" [08:58:58.0000] <Wilto> My concern is that I was asked to furnish a list of use cases, potential polyfilling decisions, and _provide citation of developer sentiment_ before our proposal would be so much as considered. [08:59:00.0000] <annevk> adactio: popped up shortly after I proposed <video> I think [08:59:04.0000] <adactio> zewt: Apologies. I'll try not to use words of more than three syllables from now on. *sheesh* [08:59:12.0000] <zewt> (sarcasm is worse) [08:59:21.0000] <Wilto> Meanwhile, the only missive from the WHATWG on `img set` seems to be this: http://junkyard.damowmow.com/507 [08:59:34.0000] <necolas> annevk: re:2007, that's exactly why the CG people don't have any ego invested in the proposals - we didn't invent anything fundamentally new [08:59:40.0000] <Wilto> So I’m left to wonder if those were just tasks to get me our of everyone’s hair, or if I was simply being humored. [08:59:52.0000] <Wilto> And moreover, to wonder whether any `process` exists here in the first place. [09:00:17.0000] <Wilto> Or if merit is simply based on the fact that things were—or were not—“invented here.” [09:00:20.0000] <annevk> Wilto: the way the WHATWG works is that people email and the editor takes those emails and produces a proposal [09:00:38.0000] <annevk> Wilto: usually that proposal goes directly in the spec, sometimes he puts up something on damwmow.com or other [09:00:49.0000] <Wilto> annevk: The bottom line is that the `img set` pattern is largely indefensible—and no effort at defending it with citations and documentation have even been made. [09:01:00.0000] <Wilto> This was a decision made in a vacuum. Let’s not pretend there was any process involved. [09:01:18.0000] <Wilto> The only repeated defense has been “this is easier for implementors.” [09:01:21.0000] <annevk> Wilto: people emailed, Hixie replied with a proposal... [09:01:38.0000] <Wilto> And Hixie proceeded despite the response. [09:01:41.0000] <annevk> Wilto: I guess you haven't read the email then [09:02:00.0000] <necolas> annevk: so what are the CG designed for, just so we know in the future? [09:02:42.0000] <Wilto> I don’t care how many times I’m told “we considered feedback.” [09:02:51.0000] <annevk> necolas: the WHATWG has a WHATCG for patent reasons [09:02:52.0000] <Wilto> If none of that feedback is incorporated, we’re just being humored further. [09:03:16.0000] <annevk> necolas: I think in general they're used as a discussion group [09:03:17.0000] <scottjehl> The developer support for picture is completely overwhelming, and very very public. A List Apart articles, w3c groups, loads of high profile tweets. How could this possibly be missed by anyone in this group? [09:03:32.0000] <necolas> annevk: so, people discuss in a CG, then email hixie? [09:03:43.0000] <odinho> Wilto: Have you read the feedback? [09:03:54.0000] <Philip`> I like how the spec says "(Steps in synchronous sections are marked with .)" in my font [09:03:59.0000] <annevk> necolas: there's never been a CG for any part of HTML before really; I'm not sure why people made one now [09:04:05.0000] <Philip`> (Looks like there's meant to be some visible character there instead) [09:04:05.0000] <Wilto> Let me be frank: [09:04:11.0000] <zewt> (what does "developer support" mean? popularity is a fairly low factor in spec decisions, i think) [09:04:13.0000] <annevk> necolas: making a CG when there's already a group for HTML seems like the mistake that was made here [09:04:14.0000] <odinho> As tab wrote: Given the fact that we went from (1) idea was rejected, (2) CG was formed, people made good arguments, (3) idea was [09:04:17.0000] <odinho> accepted, it's clear that the time *was* useful. [09:04:17.0000] <Wilto> I do not care about Hixie’s opinion, any more than anyone should care about mine. [09:04:39.0000] <zewt> (bad ideas can become popular) [09:04:58.0000] <Wilto> This is all, frankly, a little sad. [09:05:08.0000] <Wilto> I’ve seen seen no defense of `img set` on technical merit. [09:05:13.0000] <Wilto> Where are the published use cases? [09:05:13.0000] <necolas> annevk: wow. the w3c set up the CG model - there are many. we discussed in the responsive images one. there is one for mobile stuff that includes facebook devs. are we all wasting our time in the community groups? [09:05:28.0000] <odinho> Wilto: ...? [09:05:31.0000] <necolas> if yes, then shut them all down [09:05:38.0000] <annevk> necolas: we're not the W3C [09:05:49.0000] <necolas> yeah, i was just talking aloud [09:05:56.0000] <Wilto> odinho: You know: that thing I was asked to do, before `picture` could ever be taken seriously? http://wiki.whatwg.org/wiki/Adaptive_images [09:06:06.0000] <odinho> Wilto: Have you read the full threads on WHATWG, and are you honestly meaning that? [09:06:09.0000] <necolas> but if people are being sent to CG's, and they have no worth, then that isn't great [09:06:19.0000] <annevk> necolas: and most CGs are not about changing HTML [09:06:28.0000] <Wilto> Where have you guys landed on the technical viability of polyfilling the `img set` pattern? Have you decided that isn’t a major consideration? [09:06:43.0000] <annevk> necolas: if you want to change HTML you should not set up a CG, for anything else they're probably great [09:06:44.0000] <jgraham> necolas: You are wasting your time if you're trying to get browsers to change something and don't have browser vendors on board [09:06:47.0000] <necolas> so HTML CG's are worthless, CSS CG's are ?, other CG's are ? [09:06:51.0000] <odinho> Wilto: It's easy to polyfill. Tell us why it's not. [09:07:09.0000] <annevk> necolas: I don't think there are CSS CGs either [09:07:11.0000] <odinho> 17:33 < odinho> necolas: <img srcset="cat.jpg, cat⊙2 2x"><noscript><img src=cat.jpg></noscript> <---- you can polyfill that without 2 HTTP requests. [09:07:20.0000] <scottjehl> zewt: it works without overhead or accessibility drawbacks in existing browsers, with fallbacks. Developers have merely been waiting on a spec or browser implementation to validate something they already can use. [09:07:23.0000] <necolas> jgraham: does srcset have browser vendors on board? [09:07:31.0000] <Wilto> odinho: That will result in two requests in browsers with JS on. [09:07:39.0000] <necolas> i bet that more people working for browser vendors were aware of our CG than this proposal [09:07:48.0000] <odinho> Wilto: No, it will not. Take a new look. No src attribute. [09:07:51.0000] <adactio> So much for the priority of constituencies. [09:07:51.0000] <jgraham> necolas: Browser vendors pay attention to the WHATWG at least [09:07:52.0000] <adactio> So much for the priority of constituencies. [09:07:52.0000] <jgraham> necolas: Browser vendors pay attention to the WHATWG at least [09:08:02.0000] <scottjehl> odhino src is required via the HTML spec [09:08:07.0000] <Wilto> odinho: So, in non-supporting browsers without JavaScript: no image. [09:08:08.0000] <odinho> scottjehl: Easy to change. [09:08:09.0000] <jgraham> adactio: What nonsense [09:08:16.0000] <Wilto> odinho: With JavaScript, rather. [09:08:23.0000] <scottjehl> not in existing browsers [09:08:33.0000] <gsnedders> adactio: If impls won't support something, then a change doesn't help users. [09:08:38.0000] <odinho> Wilto: Oh, you have to actually write the polyfill of course. :-) But that's easy. [09:08:38.0000] <odinho> Wilto: Oh, you have to actually write the polyfill of course. :-) But that's easy. [09:08:39.0000] <Wilto> If we’re being honest, the WHATWG is completely entrenched. [09:08:51.0000] <scottjehl> my gist linked above explores as much, but I'm happy to post it wherever it needs to be so it'll be read [09:08:51.0000] <Wilto> Arguing here won’t change anything. [09:08:53.0000] <zewt> (is wilto just trolling?) [09:08:58.0000] <zewt> (sure smells like it) [09:09:06.0000] <jgraham> zewt: I hope not [09:09:06.0000] <odinho> scottjehl: They don't throw any exceptions or anything if you omit the src. [09:09:32.0000] <odinho> scottjehl: It's perfectly doable. [09:09:42.0000] <Wilto> odinho, zewt, jgraham: I’ve posted scottjehl’s gist covering why polyfills may not be possible. [09:09:57.0000] <adactio> jgraham: From what you're saying, whatever solution browser makers want is what ends up getting specced. Even if it's not what developers want. That's the very opposite of the priority of constituencies. [09:09:57.0000] <odinho> I've read it. [09:10:05.0000] <Wilto> Can you prove otherwise, odinho? [09:10:09.0000] <Wilto> I’d love to see some code. [09:10:16.0000] <Philip`> scottjehl: Existing browsers don't care what the HTML spec says is a valid document; there's a totally separate set of instructions for how they must process documents regardless of validity (which in the case of missing or empty src attribute, says they don't download anytthing) [09:10:39.0000] <jgraham> adactio: What I said is that if you are discussing somethjing where there are no browser makers involved, it is very unlikely that they will implement it [09:10:44.0000] <Philip`> scottjehl: (Also, that's somewhat separate from what current browsers actually do in practice - they might not implement what the spec says they must) [09:10:49.0000] <krijnh> Wilto: digital beers received, thanks! :) [09:10:51.0000] <jgraham> That's just common sense. [09:11:03.0000] <gsnedders> adactio: Spec'ing what developers want if browsers won't impl it doesn't help the developers. [09:11:09.0000] <jgraham> But the priority of constituencies doesn't give authors magical powers [09:11:30.0000] <Philip`> scottjehl: (I don't know if anywhere reliably documents what current browsers do actually do in this case) [09:11:44.0000] <jgraham> If browser vendors agree that a different spec will work better for end users they will adopt the better spec [09:12:18.0000] <Wilto> So if `picture` is provably better in terms of use cases, polyfills, and developer sentiment: what exactly—aside from “this is easier for —was the reason `img set` was codified instead? [09:12:32.0000] <Wilto> easier for implementors* [09:12:34.0000] <zewt> developers often want things they can't have, like more synchronous APIs in the UI thread [09:12:53.0000] <jgraham> Wilto: If it is provably better, please prove it [09:12:55.0000] <Wilto> zewt: This isn’t about doing what’s best for some petulant developers. [09:12:56.0000] <scottjehl> alt text is shown in some non-JS environments for one, followed by the noscript img fallback. It's odd behavior for a recommended approach. I guess I'd expect these things would be tested before pushing out a recommendation. A lot of work has gone into ensureing picture is bulletproof today. I'm assuming some browsers will show a broken img icon without a src, but given that this spec came out wit [09:12:56.0000] <scottjehl> h so little precedence, we've yet to test how src-less images behave in existing browsers [09:12:56.0000] <scottjehl> alt text is shown in some non-JS environments for one, followed by the noscript img fallback. It's odd behavior for a recommended approach. I guess I'd expect these things would be tested before pushing out a recommendation. A lot of work has gone into ensureing picture is bulletproof today. I'm assuming some browsers will show a broken img icon without a src, but given that this spec came out wit [09:12:56.0000] <scottjehl> h so little precedence, we've yet to test how src-less images behave in existing browsers [09:13:05.0000] <annevk> Wilto: afaik <picture> does not address the pixel density use case [09:13:06.0000] <Wilto> With citations, documentation, etc.? I have, zewt. [09:13:13.0000] <zewt> what? [09:13:17.0000] <jgraham> I really haven't seen that many emails making technical objections to @srcset or arguments in favour of <picture> [09:13:22.0000] <Wilto> Y’know, nevermind. [09:13:26.0000] <adactio> jgraham: But that's my whole point: this hasn't been about what's better. It's been about where proposals are made (WHATWG) and who makes them (hober), *not* on merit. [09:13:38.0000] <jgraham> Just lots of assertions taht developers prefer <picture> [09:13:41.0000] <necolas> jgraham: but where is the proff srcset is better? [09:13:47.0000] <necolas> s/proff/proof [09:13:49.0000] <annevk> adactio: you don't think what Hixie wrote illustrates why srcset was chosen? [09:13:55.0000] <Wilto> jgraham: http://www.w3.org/community/respimg/ [09:13:55.0000] <necolas> it's a strange point to make [09:13:57.0000] <jgraham> necolas: Various arguments in its favour have been made [09:14:07.0000] <Wilto> The “lots of assertions” are from developers. [09:14:32.0000] <jgraham> It doesn't reuse an different feature in a mostly incompatible way [09:14:40.0000] <jgraham> It is relatively concise to author [09:14:48.0000] <jgraham> It is much easier for UAs to process [09:15:05.0000] <Wilto> Authors have stated that they don’t prefer the concise syntax, so let’s not pretend that’s still a factor here. [09:15:07.0000] <jgraham> (and therefore less likely to be buggy) [09:15:14.0000] <Wilto> Else, we’re just in denial. [09:15:35.0000] <Wilto> jgraham: Can you prove that “mostly incompatible” statement? [09:15:40.0000] <jgraham> To be honest, most authors have never been asked, and no one seems to have actually tried both [09:15:49.0000] <Wilto> I stress that the WHATWG wiki is publicly editable; I’d love to see a reasoned response to use cases. [09:16:03.0000] <Wilto> jgraham: Those we asked preferred picture. I’m sure you’re not saying “we didn’t ask everyone.” [09:16:09.0000] <zewt> i'm an author and i sure prefer concise syntax [09:16:10.0000] <zewt> i'm an author and i sure prefer concise syntax [09:16:21.0000] <annevk> /me too [09:16:28.0000] <Wilto> zewt: That’s a factor, then, among the many voices we’ve heard from. [09:16:32.0000] <jgraham> Wilto: Sure. Almost all of media queries isn't appropriate for this feature and so it is confusing to make it look like it is reusing media queries [09:16:33.0000] <Wilto> Two more for `img src`. [09:16:40.0000] <divya> zewt: most authors prefer what is easiest to understand [09:16:43.0000] <divya> not what is most concise. [09:17:16.0000] <annevk> Wilto: so that wiki page; what it says for high resolution displays does not actually work [09:17:17.0000] <Wilto> zewt, annevk: Perhaps you could join the other developers commenting on http://www.w3.org/community/respimg/2012/05/11/respimg-proposal/ [09:17:32.0000] <jgraham> The problem with this kind of "preference" thing is that it is rather hard to tell much from first impressions [09:17:44.0000] <scottjehl> jgraham: why does it make sense for <video> but not <picture>? [09:17:47.0000] <necolas> that goes both ways [09:18:23.0000] <annevk> scottjehl: it's not exactly clear media="" makes sense for <video> [09:18:24.0000] <jgraham> scottjehl: I don't think it does make sense for <video>; only Opera implement it and we suggested dropping it [09:18:24.0000] <jgraham> scottjehl: I don't think it does make sense for <video>; only Opera implement it and we suggested dropping it [09:18:30.0000] <zewt> Wilto: that is not where whatwg discussions take place; if you want people to see a conversation, have it on the list where it belongs [09:18:30.0000] <Wilto> jgraham: I assume the first impression of a few hundred developers is likely worth as much as the opinions of a few “key decision makers.” [09:18:40.0000] <annevk> scottjehl: it's even implemented by all browsers; <video> has it mostly for type="" [09:18:55.0000] <Wilto> zewt: So you’re stating that—because of the forum—those opinions have been disregarded by the WHATWG. [09:18:58.0000] <annevk> scottjehl: and for <track> [09:19:02.0000] <jgraham> Wilto: I don't know how to solve problems in that algebra [09:19:18.0000] <zewt> uh, i'm saying that discussions should take place in a place where people will see them, not hidden away in blog comments [09:19:21.0000] <Wilto> jgraham: That much is obvious. [09:19:37.0000] <Wilto> zewt: It was posted to the list twice. [09:19:38.0000] <jgraham> Wilto: It would be more convincing if someone actually implemented both and tried them out [09:19:54.0000] <zewt> (tip: you're going to find it hard to convince people of anything when you make everyone squint through a thick pane of annoyance and snarkiness; it's tiring and most of us have other places to spend our energy) [09:20:10.0000] <necolas> jgraham: scottjehl already made a <picture> polyfill [09:20:21.0000] <Wilto> zewt: Would you prefer I encouraged all those commenters to post their +1s on the mailing list? [09:20:34.0000] <necolas> it would be easy to make one for srcset and then see how people get on with using them [09:20:43.0000] <Wilto> zewt: Happy to do so, if that’s the only way their votes will be factored in. I assumed that would be less than ideal. [09:20:57.0000] <jgraham> necolas: So where is the srcset one for comparison? [09:21:08.0000] <jgraham> Ah you said that [09:21:09.0000] <necolas> jgraham: no one has made it yet [09:21:10.0000] <jgraham> apologies [09:21:10.0000] <Wilto> jgraham: Good question. [09:21:22.0000] <zewt> "+1s" aren't considered much of a factor at all [09:21:27.0000] <scottjehl> jgraham: 1) picturefill already works today, and we've tested it exhaustively in jQuery Mobile's device lab - loads of existing devices, no overhead drawbacks. https://github.com/scottjehl/picturefill/ [09:21:42.0000] <Wilto> zewt: So, developers cannot simply vote on their own preferences. [09:21:50.0000] <Wilto> zewt: I would _love_ to quote you on that. [09:21:56.0000] <jreading> <picture> is too much markup FWIW [09:22:04.0000] <zewt> apis should be based on technical merit, not petitions [09:22:22.0000] <Wilto> Absolutely, zewt. Developer sentiment is a factor only as much as implementor sentiment. [09:22:27.0000] <Wilto> Technical merit is key. [09:22:36.0000] <gsnedders> The only non-technical thing that matters is usability, and that's more than just first-look opinion. [09:22:45.0000] <scottjehl> jgraham: 2) some concerns around making a polyfill for imgset are here: https://gist.github.com/2701939#gistcomment-319735 (I was told to post to the wg list and I'm happy to) [09:22:59.0000] <miketaylr> (re: +1, http://wiki.whatwg.org/wiki/FAQ#.2B1) [09:22:59.0000] <Wilto> jgraham: I posted it to the list easlier today, as well. [09:23:28.0000] <hober> /me catches up again [09:23:35.0000] <Wilto> If you’re in the mood for prose, more of those technical challenges are detailed at http://www.alistapart.com/articles/responsive-images-how-they-almost-worked-and-what-we-need/ and http://www.netmagazine.com/features/state-responsive-images [09:23:45.0000] <Wilto> Again: posted to the list multiple times. [09:23:50.0000] <hober> adactio: "A solution was proposed (srcset) without knowledge of the existing proposal (picture)" is untrue; i was aware of the CG's work [09:24:09.0000] <jgraham> scottjehl: Those technical challenges hardly seem insurmountable [09:24:09.0000] <jgraham> scottjehl: Those technical challenges hardly seem insurmountable [09:24:18.0000] <hober> adactio: when I wrote the srcset draft (which was a long time before i hit send, as TabAtkins mentioned), I had been working on two drafts [09:24:42.0000] <hober> adactio: one was the srcset proposal, and one was a "why i don't think <picture> is a good idea" post [09:25:04.0000] <hober> actually, iirc they started out as one email but the second part was getting really long and unweildy and needed editing [09:25:10.0000] <hober> so i split it up [09:25:36.0000] <hober> and didn't actually send any of it because of, err, i have no idea that was a long time ago [09:25:52.0000] <necolas> jreading: im sure that could be overcome. we shouldnt turn this into @srcset vs <picture> - there may be positives from both that could lead to something beter [09:25:53.0000] <scottjehl> jgraham: perhaps, but they're there. and there are none for picture. It seems like a proposal should follow this sort of experimentation to verify if it's useful today before any spec is drafted [09:26:06.0000] <hober> anyway, while at the css f2f there was some irc chatter about this, so i thought hold on a sec, i have something drafted [09:26:13.0000] <hober> pulled up and sent the srcset email [09:26:23.0000] <TabAtkins_> Wilto: I'm pretty sure the technical challenges of polyfilling are insurmountable. You have only two choices: (1) Do something that involves having an <img src> in the document (and thus produces two requests in polyfilled browsers), or (2) Do something that *doesn't* involve having an <img src> in the document (and thus fails entirely in legacy browsers unless the polyfill runs). [09:26:33.0000] <hober> which iirc says that i'll tackle <picture> in another email [09:26:33.0000] <TabAtkins_> There's no syntax that can get around that. [09:26:55.0000] <hober> i just haven't sent that other email yet, because i haven't finished writing it [09:28:12.0000] <adactio> hober: Okay. It's a shame that the excising of your email gave the impression that you were proposing in isolation. [09:28:32.0000] <scottjehl> tabatkins: "and thus fails entirely in legacy browsers unless the polyfill runs)." That can be easily avoided with a noscript fallback. noted in picturefull repo, and noscript would be necessary for any polyfill of srcset anyway [09:29:20.0000] <jgraham> scottjehl: <picture> does seem to have the same disadvantages though. Encouraging people not to add <img src> except via script seems like a problem [09:29:46.0000] <jgraham> Since they are quite likely to just forget [09:29:48.0000] <scottjehl> jgraham: sorry, can you clarify? [09:30:19.0000] <jreading> necolas: perhaps there a way to stay DRY with <picture>, but it seems like so much cruft [09:30:43.0000] <scottjehl> hmm. plenty of standard features require careful syntax to bulletproof: @font-face is a good example of success in that [09:30:44.0000] <scottjehl> hmm. plenty of standard features require careful syntax to bulletproof: @font-face is a good example of success in that [09:31:15.0000] <jgraham> Preumably you either add <img> in normal (not <noscript>) markup and get an extra HTTP request in the polyfill case, or you make graceful fallback difficult, increasing the chance that someone will forget to do it [09:32:00.0000] <scottjehl> jgraham: have you looked at how picturefill works? [09:32:08.0000] <TabAtkins_> scottjehl: Yup, that's an interesting hack around the problem. Requires duplication, but it works in all situations. What's not to like? [09:32:16.0000] <scottjehl> I think it addresses your concerns [09:33:08.0000] <adamdbradley> they both are handling two different things, lets combine both of them. Use the srcset format for resolution variants, and source elements for setting breakpoints inline [09:33:17.0000] <hober> adactio: priority of constituencies is users over authors over etc.; i think <img srcset> is better for *users* than <picture> *for the use case <img srcset is designed to solve* [09:33:18.0000] <hober> adactio: priority of constituencies is users over authors over etc.; i think <img srcset> is better for *users* than <picture> *for the use case <img srcset is designed to solve* [09:33:51.0000] <adactio> hober: What I don't understand is why you weren't told "Provide use cases! What about backwards-compatibility?" (which is what other people would've been told). Instead your proposal was added to the spec just like that *snaps fingers*. [09:33:53.0000] <adactio> hober: This all seems far less about merit and far more about who's making the proposal (exactly what TabAtkins_ said shouldn't be happening). [09:34:26.0000] <TabAtkins_> adamdbradley: That's more verbose, though. It (or some variant) might be worthwhile if it can be shown that there are good use-cases for using more than just min/max-width MQs. [09:34:37.0000] <hober> adactio: i think you perceive a causal relationship between me sending that email and hixie making his edits [09:34:40.0000] <TabAtkins_> adactio: The CG took care of use-cases. Backwards compat was obvious. [09:34:43.0000] <hober> adactio: that simply isn't there [09:34:50.0000] <scottjehl> tabatkins: true of picture, yes. as for srcset, that's not verified, and apparently has never been tested. How do current mobile browsers render img elements without a src attribute? Can they be polyfilled? SRC been required in the spec since img was created - wouldn't that requirement factor into some browser implementations? Anyone tested this? [09:35:16.0000] <jreading> srcset is horrible syntax anyway, picture is crufty, why don't we leverage existing, validate attributes, much in the way media queries were implemented… [09:35:19.0000] <jreading> http://hellowurld.heroku.com/blog/2012/05/15/another-image-proposal-for-responsive-design/ [09:35:30.0000] <TabAtkins_> scottjehl: Requirements like "must provide a @src" are for authors, not implementors. Browsers treat a missing source like any other missing attribute. [09:35:31.0000] <adactio> hober: Okay. I will try to avoid apopheniac interpretations. [09:36:03.0000] <TabAtkins_> jreading: I and others have explained multiple times why MQs are inadequate for solving the multiple-resolution use-case. [09:36:18.0000] <jreading> lowsrc (now obsolete) would prefetch if it came first in the attrs list... [09:38:22.0000] <scottjehl> tabatkins: my brief tests show they treat it very differently [09:38:47.0000] <TabAtkins_> Details? [09:39:32.0000] <bjankord> Aynone have a recent recap, trying to figure out where the discussion is at? [09:40:12.0000] <scottjehl> opera renders the image alt text as if the img was not found. I plan to test more, and perhaps there are no other issues, but I find it surprising this exploration wasn't done up front... maybe IE shows a broken image icon. Nobody knows is my point. This is surprising to me. [09:40:35.0000] <scottjehl> (if no alt text, it spells out "IMAGE") [09:40:56.0000] <hober> adactio: i agree that the impression of work-in-isolation is unfortunate [09:41:05.0000] <TabAtkins_> Actually, that's exactly normal behavior. If you omit an attribute, it's value is the empty string. For url-valued attributes, that means "the current page". Which will never be a valid image in HTML. [09:41:10.0000] <hober> [whee, only 13 minutes behind :)] [09:41:14.0000] <TabAtkins_> So you'll get the "invalid image" fallback, which varies per browser. [09:41:52.0000] <TabAtkins_> adactio: If it makes anyone feel better, I can state definitely that the work of the CG was very useful and was important in designing what is currently in the spec. [09:42:21.0000] <hober> adamdbradley: i agree that they are addressing different problems (resolution variation of bitmaps v. affordance for arbitrary design breakpoints) [09:42:28.0000] <TabAtkins_> Hixie said as much in his draft, but people might have skipped past that sentence, or been miffed that the words "Responsive Images CG" didn't explicitly appear in the acks. [09:42:32.0000] <TabAtkins_> On that point... [09:42:47.0000] <TabAtkins_> Hixie: It might be nice to explicitly include the Responsive Images CG (as a unit) in the acks. [09:43:10.0000] <hober> adactio: for all i know, i *will* be told that when hixie gets around to processing my email. :) [09:43:15.0000] <adamdbradley> An image is content, but variants of the same image is presentation. Baking presentation in to HTML should be an option, but overall the preferred way is to solve this with CSS [09:43:37.0000] <necolas> here we go again [09:44:15.0000] <scottjehl> adamdbradley: this is about content images. design is separate. it's about delivering assets per screen size and density. in a fluid layout, that's unrelated to design breakpoints [09:44:36.0000] <TabAtkins_> adamdbradley: When the 'content' property gains the ability to make proper replaced images, and browsers let it apply to arbitrary elements like the spec says, it'll be doable in CSS. You have everything you need (once I add image-set() to CSS Images 4). It's just more verbose. [09:45:26.0000] <necolas> adamdbradley: I already wrote about how this might be possible in CSS, but still has significant drawbacks - http://nicolasgallagher.com/responsive-images-using-css3/ [09:45:59.0000] <Wilto> adamdbradley: The CSS approach—even with proposed specs—will still result in a redundant request for clients that shouldn’t receive the original src. [09:46:11.0000] <bjankord> scottjehl: I test img tag without src attr and alt text and alt text renders in IE6-IE9, FF, and Opera the same more or less, displaying the alt text [09:46:14.0000] <scottjehl> same issue as imgset actually [09:46:31.0000] <Wilto> bjankord: I don’t think we can say for certain that it won’t introduce issues in older browsers. [09:46:58.0000] <Wilto> bjankord: Or any number of mobile browsers, since those tend to exhibit a great deal of variance in the way they handle markup errors. [09:47:45.0000] <webben> both <picture> and srcset will result in a redundant request unless the <img> element is wrapped in <noscript>. [09:48:06.0000] <bjankord> Wilto: Agreed [09:49:05.0000] <necolas> TabAtkins_: "If it makes anyone feel better"... I'm disappointed that it seems no one here thinks something went wrong with the whatwg-developer-CG communication. Instead, it seems to be getting dismissed as ego. [09:49:10.0000] <bjankord> It does seem like a hack that the only way to polyfill srcset is to remove the src attribute from the img tag and us JS to add it back in. [09:49:53.0000] <scottjehl> +1 [09:49:53.0000] <odinho> bjankord: Polyfills *are* hacks. [09:49:55.0000] <Wilto> bjankord: Absolutely. [09:50:01.0000] <hober> necolas: fwiw, i do think that it's clear that some kind of communication breakdown has happened, and that that's unfortunate. [09:50:03.0000] <odinho> That's the whole reason for them existing. [09:50:14.0000] <bjankord> odinho: Agreed [09:50:15.0000] <TabAtkins_> necolas: Sorry, that's all I'm seeing. Everyone was explicitly thanked for their contributions, and specific people were responded to when appropriate. If people still don't fill like they were adequately thanked, then shrug. [09:50:42.0000] <hober> necolas: that said, most attempts to characterize that breakdown seem to me to be overly simple, us-v-them sorts of stories that don't help [09:50:52.0000] <scottjehl> is it possible that imgset could still be reconsidered at this time? Is it possible for web developers' feedback to change it? I'm unsure whether web developers have a say, and if so, where we should be focusing our efforts/disapproval. Mailing list? [09:50:53.0000] <scottjehl> is it possible that imgset could still be reconsidered at this time? Is it possible for web developers' feedback to change it? I'm unsure whether web developers have a say, and if so, where we should be focusing our efforts/disapproval. Mailing list? [09:50:54.0000] <odinho> bjankord: So I don't think you can hold that against srcset(!) [09:50:58.0000] <hober> necolas: we should all try to learn from this in order to do better in the future [09:51:11.0000] <Wilto> I second scottjehl’s question. [09:51:21.0000] <necolas> TabAtkins_: as I keep saying, it's not about people not feeling "thanked" [09:51:24.0000] <odinho> scottjehl: Yes. But you have to talk in use cases that are not being met. [09:51:31.0000] <necolas> you're basically not listening, which is part of the problem [09:51:38.0000] <odinho> scottjehl: Or ground it on technical merit. [09:51:44.0000] <hober> necolas: who is "you" in that? [09:51:53.0000] <necolas> tab [09:51:58.0000] <hober> necolas: it seems clear to me that many people are listeningto many other people :) [09:52:11.0000] <odinho> hober: English is such a inprecise language :P [09:52:19.0000] <odinho> an* [09:52:23.0000] <odinho> im* [09:52:24.0000] <odinho> lol [09:52:39.0000] <odinho> I kinda blew that, didn't I! :D [09:52:39.0000] <hober> :) [09:52:41.0000] <necolas> it's actually pretty patronising to just dismiss what several developers are saying as a problem with their egos, when it's clearly not [09:52:46.0000] <Wilto> odinho: Can we quote you as saying that one should simply remove the `src` when creating a polyfill? [09:52:51.0000] <bjankord> necolas: +1 [09:52:52.0000] <bjankord> necolas: +1 [09:53:19.0000] <odinho> Wilto: I have written the polyfill. It works :-) I only need to test it a bit more. Don't have Windows handy right now, working on it. [09:53:37.0000] <TabAtkins_> necolas: I'm sorry, but as far as I understand it, the problem is "I'm unhappy that the CG wasn't explicitly mentioned by name in Hixie's email." Am I wrong? If so, can you explain it to me better? I've been trying all morning to understand. ^_^ [09:53:53.0000] <bjankord> odinho: I can test in IE6-IE9 if you need it [09:54:02.0000] <necolas> TabAtkins_: yeah you're wrong. and I think i've already explained that at least 3 times. [09:54:04.0000] <Wilto> odinho: Cool. We’re agreed that we can’t guarantee backwards compatibility with the removal of the src, right? [09:54:18.0000] <Wilto> If that’s the official stance coming out of the WHATWG, I’m happy to pass that information around. [09:54:28.0000] <TabAtkins_> necolas: And all three times I've felt like I've gotten closer, but still am not right. I'll take that blame on myself, but still, I don't know what's wrong. [09:54:40.0000] <TabAtkins_> Wilto: Wait wait wait. Clarification. [09:54:49.0000] <odinho> Wilto: Guarantee is a very very strong word. -- But we can certainly test it many places, and see if it's a good enough solution. I very well think it may be. [09:55:00.0000] <bjankord> Shit, can't read this fast [09:55:14.0000] <TabAtkins_> Wilto: If you omit the @src from an <img>, in legacy browsers you will get the default "broken image" behavior. This varies per browser. [09:55:33.0000] <hober> Wilto: what's "an official stance coming out of the WHATWG"? that's not how things work around here... [09:55:47.0000] <TabAtkins_> Wilto: At least until the polyfill comes along and fixes it. [09:56:01.0000] <odinho> Wilto, bjankord: Bear in mind that it's proof of concept more than anything. So haven't implemented the algo etc :P Someone else has to do the real work, but that should be doable. [09:56:06.0000] <TabAtkins_> Wilto: Alternately, you can do something like use a data: url for a 0x0 image. [09:56:22.0000] <TabAtkins_> Wilto: But that seems like more work than necessary. [09:56:26.0000] <bjankord> TabAtkins_: How does the polyfill handle users without JS [09:56:43.0000] <TabAtkins_> bjankord: polyfills *don't* handle users without JS. That's the point. [09:56:44.0000] <bjankord> This is something scottjehl's picturefill is capable of [09:57:03.0000] <TabAtkins_> bjankord: If you want to support users without JS, use @src. But then you have two requests in legacy browsers with JS. [09:57:08.0000] <TabAtkins_> There are unavoidable tradeoffs. [09:57:13.0000] <Wilto> TabAtkins_: That’s hardly a _rule_. Better polyfills certainly do. [09:57:13.0000] <webben> or use noscript [09:57:13.0000] <Wilto> TabAtkins_: That’s hardly a _rule_. Better polyfills certainly do. [09:57:13.0000] <webben> or use noscript [09:57:20.0000] <adactio> TabAtkins_: Here's the problem. You are getting feedback on imgset (a fugly "solution" that authors won't grok) and you're dismissing those concerns with "Aw, you're just unhappy because your egos are bruised." It's patronising. [09:57:26.0000] <mdelcx> is the prevailing opinion still that the <picture> tag is optimal? [09:57:33.0000] <scottjehl> picture/picturefill avoids that [09:57:37.0000] <scottjehl> and it's a huge concern [09:57:40.0000] <Wilto> mdelcx: From the developer community, yes. [09:57:47.0000] <odinho> How do you turn off js in chrome? [09:57:50.0000] <bjankord> adactio: +1 +1 +1 [09:57:57.0000] <mdelcx> imset is a non-starter for me... poor spec [09:57:57.0000] <Wilto> Agreed with adactio. [09:57:59.0000] <webben> scottjehl: Avoids which? [09:57:59.0000] <mdelcx> IMO, of course [09:58:05.0000] <jgraham> adactio: You are getting requests to make technical comments and I haven't seen you make any. Others have made some, which is good. [09:58:09.0000] <bjankord> odinho: Google it [09:58:16.0000] <TabAtkins_> adactio: No. I simply am *unable* to tell what necolas is saying is wrong. It's not technical, it appears to be personal. [09:58:31.0000] <odinho> bjankord: Actually did that ,found a nice one now though, -- it was so bad the first one. Thought there had to be better. [09:58:39.0000] <TabAtkins_> scottjehl: No, <picture> doesn't really avoid it. [09:58:46.0000] <scottjehl> I have a test page [09:58:59.0000] <bjankord> TabAtkins_: Can you elaborate? [09:59:13.0000] <TabAtkins_> scottjehl: With enough extra markup crap thrown at it, <picture> can be *as good* as <img src="data:0x0 image goes here" srcset="stuff">. [09:59:53.0000] <scottjehl> but without any http overhead [09:59:54.0000] <scottjehl> http://scottjehl.github.com/picturefill/ [09:59:58.0000] <adactio> jgraham: I was responding to TabAtkins_'s request to necolas: "I'm sorry, but as far as I understand it, the problem is "I'm unhappy that the CG wasn't explicitly mentioned by name in Hixie's email." Am I wrong? If so, can you explain it to me better? I've been trying all morning to understand." [10:00:16.0000] <TabAtkins_> adactio: Let me repeat the last three sentences of that. [10:00:30.0000] <necolas> TabAtkins_: No. It's not personal. It's about the lack of communication and the form that that communication takes when it does happen [10:00:33.0000] <TabAtkins_> Am I wrong? If so, can you explain it to me better? I've been trying all mornign to understand. [10:00:50.0000] <bjankord> Has any discussion been made here on meta media variables. https://gist.github.com/2702067 [10:00:53.0000] <scottjehl> btw that markup is purposely verbose to illustrate different queries. more than a real scenario in my experience [10:00:54.0000] <webben> scottjehl: That just uses <noscript>, which could also be used with @srcset _if_ avoiding a download in browsers that don't support srcset and trying to imitate srcset using media queries is a goal. [10:01:09.0000] <bjankord> It seems one of Hixie's issues with <picture> is how verbose it can get [10:01:21.0000] <scottjehl> of course that's a goal. that's every browser that exists today. [10:01:31.0000] <TabAtkins_> necolas: So what, precisely, of Hixie's email wasnt' sufficient? [10:01:39.0000] <Wilto> TabAtkins_: So help me, you can’t possibly be struggling to understand our line here. [10:01:54.0000] <bjankord> Wilto: +1 seriously [10:02:01.0000] <Wilto> TabAtkins_: Obviously this isn’t about ego. I can’t tell if this is misdirection or just condescention. [10:02:12.0000] <webben> scottjehl: None of those browser versions will be significant in 5 years time. [10:02:17.0000] <TabAtkins_> Wilto: Either the solution is technically inadequate, in which case PROVIDE FEEDBACK so it can be changed, or it was just bad communication, in which case who gives a crap. [10:02:19.0000] <Wilto> For our part: it’s about technical merit, and has been. [10:02:27.0000] <scottjehl> webben: this is a feature we need yesterday [10:02:27.0000] <Wilto> TabAtkins_: My feedback was ignored. [10:02:27.0000] <scottjehl> webben: this is a feature we need yesterday [10:02:37.0000] <Wilto> TabAtkins_: Or shut down with “please prove it.” [10:02:39.0000] <Wilto> Which I have. [10:02:41.0000] <TabAtkins_> Wilto: omigod please get stories straight. Your line cannot be squared with necolas'. [10:02:44.0000] <webben> scottjehl: And srcset can't be fully imitated (because media queries don't tell you what the UA needs) [10:02:45.0000] <Wilto> Meanwhile, a solution made it into the spec. [10:02:50.0000] <hober> who ignored whose feedback? [10:02:59.0000] <Wilto> This is a waste of everyone’s time. [10:03:04.0000] <webben> scottjehl: But you _can_ imitate it (poorly) if you want to, using <picture> or @srcset. [10:03:05.0000] <hober> the passive voice is not helping with clarity :) [10:03:07.0000] <tkadlec> TabAtkins_: The problem doesn't lie in one email—it's the entire way that the communication has taken place. A series of blunders, not one. [10:03:08.0000] <TabAtkins_> Agreed, unfortuantely. ;_; [10:03:19.0000] <Wilto> You all, as representatives of the WHATWG, are obviously digging in your heels. [10:03:28.0000] <TabAtkins_> tkadlec: Again, was it technical, or was it communication? If it's communication, *nobody will care* in a year. [10:03:34.0000] <scottjehl> Sigh. [10:03:38.0000] <scottjehl> As someone who had never participated very much in a standards evangelism/planning process, I'm really disheartened to see how this went through. I'm just so surprised by the lack of cooperation with the community group. Even if picture wasn't chosen, we were all completely blindsided by this spec. We would have had loads of concerns to discuss - all of which you're seeing now, after the fact. It [10:03:38.0000] <scottjehl> 's not how I imagined this process worked. [10:03:38.0000] <scottjehl> As someone who had never participated very much in a standards evangelism/planning process, I'm really disheartened to see how this went through. I'm just so surprised by the lack of cooperation with the community group. Even if picture wasn't chosen, we were all completely blindsided by this spec. We would have had loads of concerns to discuss - all of which you're seeing now, after the fact. It [10:03:39.0000] <scottjehl> 's not how I imagined this process worked. [10:03:41.0000] <Wilto> So I suppose all that’s left is to publicize your reasonings, to the best we can sort them out. [10:03:45.0000] <TabAtkins_> AGHIO3IPO. [10:03:51.0000] <Philip`> TabAtkins_: "You have only two choices" - if I understand correctly (unlikely), I'd guess you could do something like <noscript><img src=foo.png srcset=...></noscript><script>polyfillSrcset()</script> which runs a script that parses document.body.lastChild.previousSibling.textContent and inserts the appropriate img element, so you don't need any content duplication [10:04:00.0000] <webben> scottjehl: "after the fact": after what fact? [10:04:00.0000] <necolas> TabAtkins_: there are 2 different things here, which can be squared. [10:04:16.0000] <webben> scottjehl: The WHATWG spec follows a commit-then-review process. [10:04:19.0000] <necolas> 1. The fact that people are not happy with how things have been handled or how they have been spoken to throughout [10:04:37.0000] <TabAtkins_> I cannot express how frustrating it is to be told "The problem is with communication" and then told "it has nothing to do with personal issues". [10:04:41.0000] <Philip`> TabAtkins_: "... the empty string. For url-valued attributes, that means "the current page"" - the spec says for img that missing or empty src means the image shouldn't be loaded (instead of resolving as relative to the base URL) [10:04:42.0000] <webben> scottjehl: Just because @srcset is there today does not mean it's there tomorrow. [10:04:43.0000] <zewt> (You're right, I'm not too happy with how Wilto talks to all of us) [10:04:57.0000] <TabAtkins_> Either I have *no idea* what definitions you are using for those words, or someone else is sending mixed messages. [10:05:07.0000] <webben> scottjehl: You wouldn't believe the amount of material that was one time included in the spec and has later been jettisoned... [10:05:14.0000] <Wilto> zewt: I assumed condescension was the established tone here, based on how I’ve been received. [10:05:20.0000] <TabAtkins_> Philip`: Ah, kk. Well, same effect in the end here. [10:05:22.0000] <necolas> TabAtkins_: being personal is different to adequate communication [10:05:24.0000] <gsnedders> scottjehl: The typical process is that based on the initial feedback, some initial proposal is put in the spec. If you have further points to raise against the proposal in the spec, do so. [10:05:40.0000] <tkadlec> TabAtkins_: I have my issues with the srcset attribute, much like most of the developers that have seen it. The bigger issue to me has been exactly what scottjehl just said—there has been a lack of communication and cooperation here. The discussion didn't feel resolved in anyway on the mailing list. To see Hixie's email stating it was added to the draft was stunning. [10:05:41.0000] <hober> what gsnedders said. [10:05:42.0000] <gsnedders> scottjehl: That fact there's now something in the spec changes little. [10:05:45.0000] <zewt> you're the only one that's been condescending and, well, frankly rather obnoxious; it's pretty uncommon here [10:06:03.0000] <beverloo> I think a primary cause of the issue is that the Responsive Images CG proposed <picture> (among other ideas) to WHATWG which ended up unresolved. They then started a community group to discuss it, several months ago, and now a solution has been specced in a very short time (4 days!) without really reaching out to them. [10:06:10.0000] <Philip`> TabAtkins_: (...but I think some browsers do still resolve the empty string, at least for an explicit src="") [10:06:14.0000] <beverloo> Being a proponent of <picture> myself, I can see that being frustrating. [10:06:22.0000] <necolas> 2. The other issue is related to the details of the technical proposals. Different issues. [10:06:37.0000] <TabAtkins_> beverloo: But omigod the solution was *based* on the month+ of feedback on the issue. [10:06:39.0000] <TabAtkins_> al;skjdf;als [10:06:46.0000] <beverloo> TabAtkins, I get it, I get it [10:06:52.0000] <bjankord> beverloo: exactly [10:06:58.0000] <beverloo> just emphasizing [10:07:01.0000] <TabAtkins_> beverloo: I don't. ;_ [10:07:05.0000] <beverloo> flip tables. [10:07:08.0000] <Wilto> beverloo: It’s not even that no one reached out to us. [10:07:14.0000] <gsnedders> beverloo: That seems an inevitable given any sort of model that works on specing something based on use-cases and then refining the spec. [10:07:15.0000] <Wilto> I reached out to the WHATWG, representing the CG. [10:07:41.0000] <bjankord> So much time and effort was put into the CG [10:07:49.0000] <Wilto> And was asked to furnish proof of `picture`s merit, which was summarily dismissed. [10:07:53.0000] <Wilto> Over the course of four days. [10:08:20.0000] <bjankord> How can it not be clear our frustration TabAtkins? [10:08:25.0000] <TabAtkins_> Wilto: It. Was. Not. Dismissed. That proposal was found wanting for reasons which were stated in the email. [10:08:56.0000] <TabAtkins_> Namely, that <picture> was overly verbose and, being MQ based, didn't address the "MQs dont' work for multi-res negotiation" problem. [10:09:26.0000] <jreading> Tab: is there a summary of the "MQs dont' work for multi-res negotiation"? [10:09:27.0000] <bjankord> multi-res negotiation, you mean min-device-pixel-ration? [10:09:30.0000] <TabAtkins_> Having your suggestion rejected is *not* the same as having your feedback dismissed. [10:09:31.0000] <TabAtkins_> Having your suggestion rejected is *not* the same as having your feedback dismissed. [10:09:48.0000] <TabAtkins_> jreading: yes. I can illustrate it best with a real-world example. [10:09:58.0000] <TabAtkins_> Assume for a moment that there was a bandwidth MQ. [10:10:07.0000] <annevk> bjankord: yes, you can query the ratio, but you do not set it for the resource [10:10:08.0000] <bjankord> There are device-pixel-ratio media queries to handle various resolution displays [10:10:25.0000] <webben> Wilto: Not sure why you think four days is an unreasonable time for evaluating a technical proposal. [10:10:27.0000] <annevk> bjankord: and you need to set it, because otherwise your image will be displayed four times as large [10:10:34.0000] <TabAtkins_> You, a good author who wants to do well by your users, use this MQ to serve high-bandwiidth people the high-res image, and low-bandwidth people the low-res iamge. [10:10:39.0000] <bjankord> The issue of verbosity can be handled by adding meta media variables as an option to the picture element and including URI templates [10:10:54.0000] <TabAtkins_> I have a phone, and visit your site. I'm initially on 4G, which triggers the high-bandwidth MQ, and get served the good images. [10:10:55.0000] <Wilto> webben: Because it makes it clear that all this discussion was going on _while_ it was specced. [10:10:57.0000] <Wilto> Not before. [10:10:58.0000] <Wilto> Not before. [10:11:16.0000] <annevk> this discussion has been going on forever [10:11:20.0000] <annevk> since 2007 [10:11:23.0000] <TabAtkins_> I then go somewhere with only 2G service, flipping into the low-bandwidth MQ. The browser will then *throw away* the already-downloaded high-res images and *re-download* the low-res ones. [10:11:24.0000] <necolas> and it's no less depressing [10:11:24.0000] <TabAtkins_> I then go somewhere with only 2G service, flipping into the low-bandwidth MQ. The browser will then *throw away* the already-downloaded high-res images and *re-download* the low-res ones. [10:11:34.0000] <TabAtkins_> This is, obviously, a bad situation. [10:12:11.0000] <TabAtkins_> And it's unavoidable as long as you try to do multi-res negotiation by a mechanism that's simple and stateless like MQ. [10:12:19.0000] <bjankord> So because bandwidth media queries are a bad situation for serving images, the picture element was disregarded? [10:12:20.0000] <bjankord> So because bandwidth media queries are a bad situation for serving images, the picture element was disregarded? [10:12:31.0000] <Wilto> bjankord: I did get that impression several times. [10:12:41.0000] <bjankord> You've got to be shitting me [10:12:42.0000] <gsnedders> Wilto: That's surely only an issue if what is in the spec is final? If it isn't, then there's no issue the two happen concurrently. [10:12:43.0000] <TabAtkins_> bjankord: Well, because of that, any solution that relied solely on MQ was disregarded. Because it's a bad solution.. [10:12:52.0000] <jreading> Tab: I get that, but I would argue that bandwidth MQ's a re bad spec as well [10:12:52.0000] <odinho> bjankord: Because media queries is not the way to solve it. [10:13:02.0000] <jreading> esp. considering latency is more of an issue than bandwdith [10:13:05.0000] <Wilto> bjankord: No one commented on the fact that picture would be much simpler to manipulate via the DOM, however, which may make `navigator.connection` a more viable option. [10:13:20.0000] <webben> Wilto: Why is it a problem if people produce alternate proposals while evaluating a proposal? [10:13:26.0000] <Wilto> bjankord: Which I posted to the WHATWG wiki, and linked in the mailing list twice. [10:13:27.0000] <TabAtkins_> jreading: You can argue that, but you'd be wrong. ^_^ "Fixing" bandwidth MQ is *very* complicated, and can't be done without making MQs stateful, which is completely inconsisten with the rest of the model. [10:13:32.0000] <scottjehl> Tab, why would it have to do that? It seems presumptive to suggest bandwidth queries couldn't be implemented more smartly, especially for a feature that doesn't exist [10:13:39.0000] <miketaylr> what about zooming in on a page, from your desktop? the low-res/mobile image MQ will be triggered. [10:13:39.0000] <miketaylr> what about zooming in on a page, from your desktop? the low-res/mobile image MQ will be triggered. [10:13:41.0000] <jreading> so because of bandwidth MQ, MQ are bad… makes no sense [10:13:49.0000] <odinho> Wilto: Then it's another proposal. Right now we're taking about the flexible image resolutions. [10:13:51.0000] <Wilto> miketaylr: Not when using ems. [10:13:55.0000] <scottjehl> or doesn't yet work anyway... [10:14:01.0000] <miketaylr> Wilto: ORLY [10:14:02.0000] <miketaylr> Wilto: ORLY [10:14:10.0000] <TabAtkins_> jreading: No. Because bandwidth MQ are bad, but negotiating res based on bandwidth is an important use-case, anything which uses *only* MQ is bad. [10:14:18.0000] <Wilto> It’s worth noting the the specced solution seems to be fully pixel dependent. Is that correct? [10:14:22.0000] <TabAtkins_> Because that means, by definition, that it's not solving an important use-case. [10:14:29.0000] <Wilto> As no detailed use cases have been published, we don’t know. [10:14:33.0000] <bjankord> The network information API is the closest thing we have to detecting a users bandwidth connection. http://www.w3.org/TR/netinfo-api/ [10:14:38.0000] <miketaylr> can everyone start using ems? that's one of my biggest complaints about "responsive" sites. zooming in gets me a mobile layout [10:14:47.0000] <hober> miketaylr: :) [10:14:58.0000] <beverloo> TabAtkins_, the spec could define a fallback model. If a 3G version is available in cache/memory, use that, otherwise fall back to the 2G version. [10:15:16.0000] <beverloo> there's other issues with that, though [10:15:21.0000] <gsnedders> Wilto: What use-cases does it fail to fulfil on the discussion on the mailing list? [10:15:21.0000] <TabAtkins_> beverloo: It could. But then your MQ is stateful. What about the *rest* of the properties that you're applying via the bandwidth MQ? [10:15:21.0000] <gsnedders> Wilto: What use-cases does it fail to fulfil on the discussion on the mailing list? [10:15:22.0000] <TabAtkins_> beverloo: It could. But then your MQ is stateful. What about the *rest* of the properties that you're applying via the bandwidth MQ? [10:15:28.0000] <Wilto> miketaylr, hober: <3 ems. [10:15:37.0000] <beverloo> I'm not sure whether I agree with a bandwidth MQ at all, actually [10:15:44.0000] <beverloo> it'd be based on estimates [10:16:00.0000] <Wilto> gsnedders: The one I just mentioned. [10:16:03.0000] <beverloo> "wifi" can be significantly slower than "3g" [10:16:10.0000] <TabAtkins_> beverloo: I agree with you. ^_^ That's the point. Bandwidth MQs are a bad idea for multiple reasons. So you can't rely on MQs to do res negotiation. [10:16:18.0000] <zewt> seems to me that connection-based selection is so fiddly, imprecise and heuristic, that the only sane solutions are ones that give the browser metadata and say "do what you think is best" [10:16:32.0000] <Wilto> I’m not sure we’re all on the original topic anymore. [10:16:35.0000] <TabAtkins_> Yes, the Network Info API is *really* flawed. [10:16:40.0000] <beverloo> So we're excluding a feature such as <pictures> on a use-case that relies on a hypothetical, not yet specified feature (being bandwidth MQs)? [10:16:51.0000] <odinho> beverloo: it will never be right, in Opera, we've tried to at least make it useful as an heuristic, -- but that also falls flat on its bottom. Although it sounds like it can be done, it's extremely hard to even get an approximation. [10:16:52.0000] <Wilto> beverloo: It seems so. [10:17:05.0000] <TabAtkins_> beverloo: No, we're excluding it based on a ues-case (res negotiation) which requires an ability that MQ can't do. [10:17:15.0000] <beverloo> odinho, yes, I'm aware of that (I'm on the Chrome on Android team) [10:17:40.0000] <Wilto> beverloo: Android Chrome! You do yeoman’s work, beverloo. [10:18:11.0000] <bjankord> TabAtkins_: If the Network Information API is flawed, why have Chrome and Firefox added support for it? [10:18:14.0000] <TabAtkins_> Wilto: Chat with me privately. I need to find out what pieces of your knowledge are missing so I can fill them in properly. [10:18:17.0000] <beverloo> TabAtkins, as odinho says, getting bandwidth negotiation right is extremely hard to begin with. It's likely that we won't ever get past heuristics here [10:18:19.0000] <TabAtkins_> bjankord: Because it's in a spec. [10:18:20.0000] <TabAtkins_> bjankord: Because it's in a spec. [10:18:20.0000] <odinho> beverloo: You obviously knew something about it, just wanted to back you up by extra data point :-) [10:18:31.0000] <ShaneHudson> Bandwidth media queries would be really nice but they are not accurate at the moment, perhaps never will be. That is absolutely no reason to deminish <picture> or the other solutions the community group as proposed. [10:18:36.0000] <TabAtkins_> beverloo: heuristics are fine. the *real* problem is statefulness, which MQ can't provide. [10:18:47.0000] <Wilto> TabAtkins_: I’d prefer to see more information on the specced solution posted publicly. [10:18:55.0000] <beverloo> TabAtkins, changing a mobile device's orientation will influence the width [10:19:00.0000] <Wilto> You don’t have to convince me; the developer community will need citations. [10:19:01.0000] <bjankord> ShaneHudson: +1 [10:19:02.0000] <beverloo> potentially causing the same problem [10:19:17.0000] <TabAtkins_> Wilto: I've responded on multiple threads about why MQ-based solutions are flawed. Did you read those posts? If not, I can fill you in quickly. [10:19:35.0000] <Wilto> TabAtkins_: Sure. Sum up for me, real quick. [10:19:43.0000] <Wilto> The community will want an easy-to-digest reasoning. [10:19:45.0000] <annevk> beverloo: so does resizing the browser window on desktop [10:19:52.0000] <TabAtkins_> I explained *literally* a screen or so ago, to jreading. [10:19:56.0000] <odinho> beverloo: A browser has much better control over herusticts, and can do special things based on its special knowledge. Say, the phone knows it's roaming, -- or that the user has set a bandwidth limitation, -- those it can use in a useful way to always take the smallest pictures to save bandwidth. [10:19:56.0000] <beverloo> annevk, TabAtkins, so neither solution is really stateful [10:19:57.0000] <beverloo> annevk, TabAtkins, so neither solution is really stateful [10:20:03.0000] <odinho> beverloo: MediaQueries can't do that. [10:20:21.0000] <TabAtkins_> beverloo: Yeah, I think that Hixie's proposal is somewhat flawed because of that. [10:20:38.0000] <Wilto> TabAtkins_: It seemed like you were discussing bandwidth MQ. [10:20:38.0000] <TabAtkins_> beverloo: I think it should take one length and use the smaller of the window dimensions to check it, precisely for that reason. [10:20:38.0000] <Wilto> TabAtkins_: It seemed like you were discussing bandwidth MQ. [10:20:38.0000] <TabAtkins_> beverloo: I think it should take one length and use the smaller of the window dimensions to check it, precisely for that reason. [10:20:40.0000] <beverloo> odinho, absolutely, the heuristics would be much more accurate than what a user can do [10:20:45.0000] <annevk> beverloo: one solution describes the resource, the other queries the device and selects a resource based on that [10:20:56.0000] <annevk> beverloo: you need the former for e.g. pixel density stuff [10:21:01.0000] <TabAtkins_> Wilto: I was explainign why you can't do bandwidth MQ, yes. And that's why MQ-based solutions are bad. [10:21:19.0000] <bjankord> Tab: What solution do you suggest? [10:21:27.0000] <Wilto> So, neither of these solutions will use hard-coded bandwidth detection, is that correct? [10:21:30.0000] <odinho> TabAtkins_: I think that too. A bounding box. Did I write that on mailing list, or only here? [10:21:33.0000] <beverloo> TabAtkins, issue there is that websites may present different versions based on the orientation (again due to MQs) [10:21:33.0000] <Wilto> TabAtkins_: ^ [10:21:36.0000] <TabAtkins_> bjankord: For what? The general problem? [10:21:39.0000] <Wilto> Nothing inline, in the markup, I mean. [10:21:43.0000] <TabAtkins_> odinho: I think it was you suggesting that, here in the chatroom. [10:21:46.0000] <jreading> bandwidth MQ are bad [10:21:48.0000] <beverloo> TabAtkins, for which you may want a different picture too. [10:21:58.0000] <jreading> none of this makes sense… MQ are still part of the style layer in the presentation layer, there's navigator.connection and UA ISP sniffing, server solutions to work with bandwidth. Latency is more of an issue than bandwidth with wireless carriers anyway, bandwidth means nothing with 600ms latency. Why is this a concern of the presentation layer? Are all mobile web solutions through the letterbox of markup and css? [10:22:00.0000] <ShaneHudson> <picture> could work the same way as srcset if need be, without the media queries. The biggest problem with srcset is that it has the potential to be an extremely large attribute which would be unreadable and horrid. It also has very litttle potential for expansion [10:22:03.0000] <TabAtkins_> beverloo: Eh, true. Welp, whatever. [10:22:24.0000] <Wilto> TabAtkins_: If neither is using an attribute that detects bandwidth, I’m not sure why we’re still talking about it. [10:22:33.0000] <Wilto> I agree that it’s a flawed idea. [10:22:35.0000] <bjankord> Agreed [10:22:37.0000] <Wilto> On both markup patterns. [10:22:40.0000] <Wilto> Let’s move on. [10:22:46.0000] <beverloo> annevk, it's really a subset, MQs can poll for the DPI as well. Due to it being a subset however, most of the questions we're discussing now are out of scope (let alone parsing issues) [10:22:51.0000] <Wilto> Apart from _theoretical_ media queries, that is. [10:22:54.0000] <TabAtkins_> Wilto: You, um, tell the browser what your image's DPI is. It then decides based on its own metrics (such as bandwidth history) which one to download. [10:23:03.0000] <annevk> beverloo: polling for the DPI is not enough [10:23:07.0000] <annevk> beverloo: you need to set it [10:23:12.0000] <Wilto> TabAtkins_: And the specced pattern does not, then? [10:23:17.0000] <annevk> beverloo: because otherwise the image will be displayed four times the size [10:23:23.0000] <Wilto> It incorporates bandwidth in a meaningful way? [10:23:27.0000] <TabAtkins_> Wilto: Wait, have you read the specced proposal? [10:23:30.0000] <beverloo> annevk, got you [10:23:30.0000] <webben> Wilto: Providing information about the resource allows browsers to use any information (including bandwidth information, power information, viewport size, orientation etc) to select the optimal image. [10:23:35.0000] <beverloo> annevk, thanks :) [10:23:40.0000] <zewt> TabAtkins: well, you also need to tell it the image's size (even if it's only a hint) [10:23:48.0000] <TabAtkins_> Wilto: I'm not sure if you're asking questions that I should answer, or if you're being rhetorical, if you're trying to lead to a point that I suspectis inaccurate. [10:23:55.0000] <annevk> beverloo: I'm somewhat surprised everyone thinks MQ magically solve that... [10:24:01.0000] <zewt> file size, I mean [10:24:31.0000] <TabAtkins_> zewt: Filesize *should* correspond to the Nx multiplier you give. [10:24:36.0000] <webben> zewt: Yeah. I'm inclined to suggest srcset should include HxW (always) and would benefit form size. [10:24:50.0000] <webben> TabAtkins_: Doesn't that depend on e.g. jpeg quality? [10:24:50.0000] <beverloo> annevk, let's make a new image format containing the target DPI! [10:24:51.0000] <zewt> TabAtkins: you can also have multiple images at the same resolution with eg. different jpeg compression ratios [10:24:51.0000] <beverloo> annevk, let's make a new image format containing the target DPI! [10:24:52.0000] <beverloo> /me runs [10:24:52.0000] <webben> TabAtkins_: Doesn't that depend on e.g. jpeg quality? [10:24:52.0000] <zewt> TabAtkins: you can also have multiple images at the same resolution with eg. different jpeg compression ratios [10:24:52.0000] <beverloo> /me runs [10:24:53.0000] <TabAtkins_> zewt: Unless you're being perverse and providing different-sized images for different resolutions. [10:25:26.0000] <TabAtkins_> Yeah, compression ratio is somewhat harder to deal with, particularly since it's so format-specific. [10:25:36.0000] <hober> /me mumbles something about jpeg2000 [10:25:37.0000] <hober> /me mumbles something about jpeg2000 [10:25:43.0000] <TabAtkins_> I think by the time you're dealing with that, you should just go res-independent. [10:25:44.0000] <bjankord> We could sit here all day discussing pros and cons of each solution, yet still get no where... [10:25:53.0000] <odinho> beverloo: They already have it. But let's rather make a "Real_Image_DPI_I_Really_Mean_It" attribute for JPG et al! *runs* [10:25:54.0000] <annevk> beverloo: you mean one that browsers are not forced to ignore? [10:26:39.0000] <zewt> TabAtkins: also, file size doesn't change linearly with resolution [10:26:50.0000] <zewt> not always, anyway [10:26:53.0000] <beverloo> annevk, odinho, yup. [10:26:54.0000] <grigs> Wow, lot's of conversation. Took forever to catch up. A couple of quick notes. [10:27:27.0000] <bjankord> grigs: yes, there is a lot to digest here [10:27:38.0000] <tantek> TabAtkins - re: "just go res-independent", let me know when you know of cameras that output SVG photographs :p [10:27:53.0000] <grigs> 1. Was really pleased to hear more background from hober. Good to know that he looked at a bunch of info before making his rec. [10:28:03.0000] <annevk> beverloo: good luck with that :) [10:28:19.0000] <beverloo> annevk, :D [10:28:24.0000] <TabAtkins_> zewt: Yeah, it's not a perfect fit. But I suspect it's probably a good enough approximation. [10:28:30.0000] <grigs> 2. A couple of people asked earlier why we went to a community group to discuss this. It was because we were asked to take the conversation off the whatwg list. [10:28:52.0000] <annevk> grigs: pointer for that? I asked before but never got it [10:28:54.0000] <annevk> grigs: pointer for that? I asked before but never got it [10:29:29.0000] <ShaneHudson> It should be in a community group anyway, it is the kind of thing the community should have input on [10:30:02.0000] <grigs> @annevk I'll dig it up. [10:30:19.0000] <zewt> TabAtkins: at least on impression (without having looked at it too hard), having different JPEG qualities seems useful; if I'm looking at pictures, I'd rather have a high-resolution image go from q12 to q10 than cut the resolution in half [10:31:03.0000] <annevk> ShaneHudson: WHATWG is a community developing HTML [10:31:13.0000] <annevk> ShaneHudson: no need for subgroups in the past eight years [10:31:18.0000] <zewt> i suppose that if browsers are given too many axes to choose from, the chances of making wrong decisions rises, which could be a barrier to using the feature [10:31:32.0000] <grigs> annevk: original objection to the conversation on the list http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2012-February/034775.html [10:31:48.0000] <TabAtkins_> zewt: I think that, in general, you won't request the double-res image unless you're on a double-res device anyway, regardless of bandwidth. [10:31:57.0000] <TabAtkins_> (Or, as people have said, you're printing, or saving, etc.) [10:32:04.0000] <grigs> annevk: my response and concern http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2012-February/034790.html (which in retrospect seems pretty spot on ;-) ) [10:32:28.0000] <annevk> grigs: interesting, I don't know who that is [10:32:37.0000] <TabAtkins_> zewt: Within a particular resolution level, jpeg quality might be useful, but I think it's a much smaller concern. [10:32:46.0000] <grigs> annevk: still looking for the specific suggestion that we take it to a community group. [10:33:01.0000] <TabAtkins_> Huh. I have also never heard of this Ronjek guy. [10:33:04.0000] <zewt> TabAtkins: i mean, for example, if I'm on a 1920x1200 display, and a site wants to show a 1920x1200 image, selecting an appropriate quality for my connection--it can make sense to choose a 1920x1200 q10 image if you want to save bandwidth, rather than a 1280x720 q12 one [10:33:30.0000] <annevk> grigs: the list has 1500 people so if one person tells you something that doesn't mean it's definitive [10:33:31.0000] <zewt> at least for that (isolated, off-the-top-of-my-head) case, it's almost always the better choice [10:33:53.0000] <annevk> grigs: thanks [10:33:55.0000] <tkadlec> grigs: I believe here? http://lists.w3.org/Archives/Public/public-whatwg-archive/2012Feb/0194.html [10:33:58.0000] <annevk> i have to go now [10:33:59.0000] <grigs> annevk: I get that. To me, that’s the frustrating part. [10:34:09.0000] <divya> annevk: you are just being defensive here. ultimately who do we consider as the 'voice'? [10:34:13.0000] <zewt> but then it's also much more tied into connection heuristics, where resolution selection at least isn't, so it's much more likely to be useful in practice I guess [10:34:13.0000] <divya> if we mention to hixie [10:34:15.0000] <divya> thats not enough [10:34:17.0000] <divya> if we mention here [10:34:19.0000] <divya> thats not enough [10:34:24.0000] <divya> if we mention on mailing list its not enough [10:34:27.0000] <divya> when is it enough? [10:34:34.0000] <annevk> divya: not sure what you're saying [10:34:40.0000] <annevk> but I have to go watch a movie [10:34:45.0000] <grigs> tkadlec: thanks, that's the link [10:34:53.0000] <grigs> annevk: enjoy! [10:35:03.0000] <webben> http://lists.w3.org/Archives/Public/public-whatwg-archive/2012Feb/0195.html seems relevant. [10:35:31.0000] <grigs> webben: yep. [10:36:13.0000] <webben> divya: http://wiki.whatwg.org/wiki/FAQ#How_does_the_WHATWG_work.3F [10:36:48.0000] <TabAtkins_> zewt: Sure, it's the better choice. I personally don't think it'd be worth the effort of speccing compression levels in a cross-format way (or handling format-specific stuff). [10:37:07.0000] <zewt> TabAtkins: i'd just include the file size by itself [10:37:22.0000] <TabAtkins_> Ah, yeah, that's valid. [10:37:34.0000] <TabAtkins_> Hixie mentions that in a response on the latest thread. [10:37:52.0000] <zewt> that would make the assumption that all available options are in the same or comparable compression types, which is probably reasonable [10:38:01.0000] <divya> webben: all of it was done. [10:38:20.0000] <divya> hober: it would have been nice when you proposed the solution why you decided it over picture [10:38:35.0000] <divya> hober: it seems like the whole work was summarily dismissed even though now it looks like it wasnt [10:38:45.0000] <webben> divya: I'm linking you to that to explain why it doesn't make sense to ask for the voice of the WHATWG. [10:39:14.0000] <webben> divya: The WHATWG as an organization speaks only for administrative reasons basically. [10:40:16.0000] <jgraham> Hmm, I am now behind on the discussion, obviously, but it occured to me that one problem with an element-based solution is that people will be tempted to polyfill it before there are any browser implementations. This will make it very hard to modify in the face of implementor feedback without breaking existing content. By contrast with an attribute-based solution one could use data-srcset or similar from script until there are shipping native implem [10:40:52.0000] <divya> jgraham: i actually am not attached to any one solution [10:40:58.0000] <divya> i suspect none of the people are [10:41:11.0000] <divya> jgraham: only concern is the unreadable syntax that srcset proposes [10:41:22.0000] <grigs> divya: +1 [10:41:22.0000] <divya> and the other concerns that scottjehl raises [10:41:34.0000] <divya> jgraham: we know people will go wrong [10:41:35.0000] <divya> jgraham: we know people will go wrong [10:41:50.0000] <divya> jgraham: i think having an easy to understand solution would make it less difficult to go wrong [10:42:24.0000] <grigs> i'll also add that while hober's original spec was clear, despite multiple readings of the expanded version, i still don't fully grok how it will work. [10:42:54.0000] <grigs> not trying to be difficult here, but will readily acknowledge I may be dense. ;-) [10:43:20.0000] <divya> just because the GCD is stupid [10:43:35.0000] <jgraham> divya: hmm i don't think I said you were. Just sharing some "thinking" from the bus ride home [10:43:36.0000] <grigs> divya: GCD? [10:43:45.0000] <divya> Greatest common denominator :))) [10:43:56.0000] <Philip`> zewt: Your example of a 1280x720 image on a 1920x1200 display sounds like it's effectively using downsample+upsample as a lossy compression algorithm, but a pretty dumb one, in which case it makes sense that any non-terrible lossy compression algorithm (e.g. JPEG) ought to be able to do a better job at the same bitrate [10:44:12.0000] <divya> doesnt mean we should just do whatever is the most easy for implementors. [10:44:32.0000] <divya> jgraham: ya ya i am just ranting [10:44:35.0000] <paul_irish> jgraham: every feature will be polyfilled. polyfill all the things. if there is a way polyfill solutions can better coexist with iterative browser implementations, then that'd be good too. [10:45:06.0000] <jgraham> paul_irish: There is: don't use new elements before they are implemented; for things with no implementation use data- attributes [10:45:09.0000] <divya> jgraham: i am yet to come across a polyfill problem [10:45:21.0000] <zewt> Philip`: sure, the point of the example was why you might want to select among different jpeg quality versions for the same resolution--why file size might be useful in addition to resolution [10:46:04.0000] <divya> jgraham: is there one? [10:46:43.0000] <jgraham> divya: My concern is that people add special processing for <picture> based entirely on polyfill, then a browser implements it, we realise that the spec doesn't work for some reason that the polyfill doesn't care about (e.g. during dynamic changes) and changing the spec breaks existing content [10:46:44.0000] <paul_irish> <x-picture> is the equivalent. it could work effectively the same from the polyfill side. [10:47:13.0000] <jgraham> Yeah, but inventing elements is non-conforming and highly controversial [10:47:38.0000] <divya> jgraham: like we havent been doing that for eons :P [10:47:53.0000] <jgraham> divya: How many things *that no browser implements*? [10:47:56.0000] <scottjehl> jgraham: the polyfill would of course be updated to match implementation [10:48:01.0000] <Philip`> zewt: Yeah, I'm not trying to argue any case - just musing about how lowering resolution is simply a poor lossy compression algorithm [10:48:02.0000] <Philip`> zewt: Yeah, I'm not trying to argue any case - just musing about how lowering resolution is simply a poor lossy compression algorithm [10:48:08.0000] <divya> jgraham: severalll [10:48:09.0000] <jgraham> scottjehl: But existing sites would break [10:48:11.0000] <divya> HOW ABOUT HGROUP [10:48:22.0000] <jgraham> /me -> food [10:48:36.0000] <divya> kk bai [10:50:22.0000] <jreading> there.must.be.a.better.way [10:50:43.0000] <jreading> http://davidbcalhoun.com/present/mobile-performance-amazon/#navigator.connection this should be in headers and we can kill that MQ bandwidth nonsense [10:52:11.0000] <jreading> variablize the MQ's in <picture> (like my blog post) and be more DRY [10:52:12.0000] <jreading> variablize the MQ's in <picture> (like my blog post) and be more DRY [10:54:51.0000] <paul_irish> jgraham: speaking of changing implementations, that spec changed. :p [10:54:59.0000] <paul_irish> oops. jreading, rather. [10:55:07.0000] <odinho> I've written a polyfill for srcset now. [10:55:14.0000] <zewt> hey, i've been using gmail for years and I finally figured out how to extend a damned quote block [10:55:41.0000] <odinho> To see it could be done. But, yes, it does have some problems at edge cases. [10:55:44.0000] <paul_irish> jreading: agreed we need http://dvcs.w3.org/hg/dap/raw-file/tip/network-api/index.html in MQs and some sort of relation to <picture>/srcset [10:56:20.0000] <scottjehl> odhino: link? [10:56:30.0000] <odinho> scottjehl: Yeah, I need someone to test it in IE :P [10:57:46.0000] <scottjehl> ...and loads of mobile devices [10:58:14.0000] <scottjehl> so what's next for picture element supporters? is there anywhere worthwhile to direct our existing and ongoing efforts? [10:58:41.0000] <divya> scottjehl: i am not sure if we should be so affiliated to pic element [10:59:30.0000] <ShaneHudson> I am not 100% attached to picture element, but it is far surpior to srcset and to be honest, I prefer most suggested solutions over srcset [10:59:46.0000] <scottjehl> divya, what do you mean? [11:00:37.0000] <grigs> scottjehl: i'm still trying to definitely figure out if srcset addresses one of the core use cases. [11:00:52.0000] <divya> scottjehl: i am saying we shouldnt be attached to a syntax [11:01:09.0000] <divya> its okay as long as we find a solution that works for implementors and devs alike [11:02:00.0000] <scottjehl> sure, primary concern is something that solves the problem for users. That said, many are attached to picture's syntax because we already know it does solve that problem today. [11:02:35.0000] <grigs> scottjehl: to my mind there are three questions about srcset. 1) does it do what we need it to do? 2) can the syntax be improved to provide more clarity? 3) does the real or perceived concerns about the syntax mean that srcset will have a harder time being adopted by developers? [11:02:43.0000] <kevinSuttle> At this point, I don't think this is our problem to solve. I'm trying to bring a bit more scope to the discussion. http://www.w3.org/community/respimg/2012/05/11/respimg-proposal/#comment-780 (Going to keep posting this until it gets in front of decision makers/influencers) [11:02:45.0000] <tantek> In following this discussion, I'm a bit surprised to see so little reasoning from actual use-cases, that is, what seems to be documented here: http://www.w3.org/community/respimg/2012/04/16/summary-of-use-cases-and-requirements/ and here: http://www.w3.org/community/respimg/wiki/Main_Page and perhaps originating here: https://etherpad.mozilla.org/responsive-assets [11:03:07.0000] <Philip`> scottjehl: If you define yourself as being in the pro-foo camp, you're implicitly defining everyone else as anti-foo, which makes it emotionally harder for them to ever support foo, so you're shooting yourself in the foot :-) [11:03:23.0000] <grigs> tantek: exactly [11:03:42.0000] <tantek> /me agrees with Philip - why jumping straight to pro/anti any specific syntax? [11:03:59.0000] <tantek> /me has yet to see any consensus on which use-cases matter and why. [11:03:59.0000] <tantek> /me has yet to see any consensus on which use-cases matter and why. [11:04:25.0000] <tantek> if you can't agree on use-cases and priority, then of course you're going to get different solutions. it's pointless to argue about syntax when you can't agree on purpose. [11:04:35.0000] <grigs> tantek: amen! [11:04:56.0000] <tantek> grigs - I'm looking at the folks proposing <picture> as well [11:05:04.0000] <grigs> tantek: OH I KNOW [11:05:07.0000] <grigs> :-) [11:05:18.0000] <tantek> I've yet to see the clear math-style proof of, starting from these use-cases, with this priority, here's how we end up designing the <picture> element. [11:05:57.0000] <tantek> until you show your work, I'm not buying the proof (proposal) [11:06:09.0000] <grigs> same true for srcset as well then, eh? [11:06:15.0000] <tantek> grigs - of course [11:06:59.0000] <odinho> I think the camps disagree about the use cases. -- Because many use cases can't be done with srcset, - and many can't be done with <picture>. -- I align much more with srcset's use cases, and think they're more important to tackle. However, -- the other use cases should be tackled as well, maybe later. Maybe the best solution for those is a new <picture> element. [11:06:59.0000] <odinho> I think the camps disagree about the use cases. -- Because many use cases can't be done with srcset, - and many can't be done with <picture>. -- I align much more with srcset's use cases, and think they're more important to tackle. However, -- the other use cases should be tackled as well, maybe later. Maybe the best solution for those is a new <picture> element. [11:07:16.0000] <tantek> odinho - that's bass-ackwards [11:07:28.0000] <tantek> "align much more with srcset's use cases" [11:07:47.0000] <grigs> i do think there's a lot of that going on though. [11:07:50.0000] <tantek> are you seriously framing a set of use-cases by one particular syntax? that's so limiting in thinking I don't even know where to begin. [11:07:52.0000] <jgraham> tantek: Not if you are charitable [11:08:23.0000] <tantek> those use-case documents I linked to above are reasonable starting points. the only nit I would pick is to put them on a *generally* accessible/usable wiki (e.g. w3.org/wiki/ or wiki.whatwg.org ) [11:08:31.0000] <odinho> tantek: No :-) I'm not wedded to any syntax at all! But I want certain things to be able to do in the browser. [11:08:50.0000] <adactio> tantek: And yet one of the proposed solutions is in the spec while the other is not. Both, as you rightly point out, should first be tested through use-cases, backward-compatibility, etc. There is an unequal weighing of proposals. [11:08:54.0000] <tantek> group/community specific wikis are a bit of a silo-iziation/tribalization and not useful for broadening consensus (actually harmful) [11:08:55.0000] <jgraham> tantek: You could assume he meant "I think that the use cases that informed the design of srcset are more important than those that influenced <picture>" [11:09:23.0000] <grigs> jgraham: that's the way I read it. [11:10:00.0000] <tantek> as long as you frame the use-cases by specific syntaxes, you're going to suffer more arguments/friction than if you simply give the use-cases their own (perhaps fragment) permalinks and discuss them individually. [11:10:19.0000] <odinho> http://www.w3.org/community/respimg/2012/04/16/summary-of-use-cases-and-requirements/ <- in fact these are actually very nice. [11:10:30.0000] <tantek> odinho - except it's not editable [11:10:53.0000] <tantek> hence: "put them on a *generally* accessible/usable wiki (e.g. w3.org/wiki/ or wiki.whatwg.org )" [11:10:59.0000] <tantek> give them their own sections so you can link to them [11:11:05.0000] <ShaneHudson> So how do we go about settling on the use-cases? We have had hundreds of posts on the CG, the mailing list and in here about the different use cases. What is the recommented way to compile them, the wiki? [11:11:16.0000] <tantek> example: http://wiki.whatwg.org/wiki/Time_element [11:11:23.0000] <odinho> tantek: I agree with all of them. -- But the media queries doesn't solve them as nicely as the srcset proposal does. [11:11:44.0000] <tantek> ShaneHudson - yes, wiki please [11:11:48.0000] <grigs> one lesson coming out of this experience should be some better guidance on the hoops developers should jump through when they really want something in the spec. [11:11:52.0000] <tantek> that way we can iterate on them in one place [11:12:00.0000] <scottjehl> grigs +1 [11:12:04.0000] <tantek> rather than go round-round in the support forums known as "mailing lists" [11:12:06.0000] <tkadlec> grigs: +1 [11:12:33.0000] <jgraham> tantek: Characterising the whatwg list as a "support forum" is just inaccurate [11:12:39.0000] <tantek> grigs - what part of "document use-cases on the wiki" was not communicated?: [11:12:56.0000] <ShaneHudson> http://wiki.whatwg.org/wiki/Adaptive_images Has everyone seen that page? [11:12:56.0000] <jgraham> Although I agree that using the wiki in this case might be useful [11:13:02.0000] <grigs> we've had people say we shouldn't discuss on the whatwg list, we should use a community group, we should build use cases, we should have use cases on publicly editable wiki (why isn't that part of the community group), math proofs, etc. [11:13:04.0000] <grigs> we've had people say we shouldn't discuss on the whatwg list, we should use a community group, we should build use cases, we should have use cases on publicly editable wiki (why isn't that part of the community group), math proofs, etc. [11:13:13.0000] <Wilto> tantek: Just tuning back in, but you saw this, yeah? http://wiki.whatwg.org/wiki/Adaptive_images [11:13:19.0000] <Wilto> I posted it to the list a couple of times. [11:13:42.0000] <jgraham> Wilto: That still has proposed solutions listed under "use cases" [11:13:46.0000] <ShaneHudson> grigs: I agree. For new members at least, it is pretty caotic! [11:13:51.0000] <tantek> Wilto - problem with that wiki page is that it is already assuming <picture> as syntax [11:13:55.0000] <odinho> grigs: That was one person of a list of 1400+, and at least I haven't seen him much. But it's good to start from use cases. [11:13:56.0000] <tantek> so you're not going to get people to read past that [11:14:13.0000] <tantek> this style of documenting use-cases is better: http://www.w3.org/community/respimg/2012/04/16/summary-of-use-cases-and-requirements/ [11:14:19.0000] <grigs> odinho: how the hell were we supposed to know that? :-) [11:14:30.0000] <ShaneHudson> Okay, lets make a new page. [11:14:44.0000] <grigs> At various points we were told that modifying the img tag was a non-starter so we coached others not to do so. [11:14:51.0000] <tantek> jgraham - the one person that told people to not discuss responsive images on the whatwg list proves my point about it being a support forum. [11:14:58.0000] <Wilto> That’s helpful, tantek. We can put up a page more along those lines, if that’ll help. [11:15:00.0000] <tantek> all email lists eventually devolve into support forums [11:15:10.0000] <TabAtkins_> grigs: That was misinterpreted, unfortunately. Modifying <img> to have *children* is a non-starter. Adding attrs is always okay. [11:15:16.0000] <odinho> grigs: That's also a technicality, -- you were told you couldn't change the parser for img element, that's a non-starter. Adding attrs is ofc. [11:15:17.0000] <jgraham> grigs: That was a misunderstanding about what "modify" meant [11:15:30.0000] <tantek> actual design happens in smaller / quicker fora like IRC, with anything of any substance being documented on the web (wiki pages) [11:15:46.0000] <tantek> the whatwg list *appears* to work only because Hixie treats it as his inbox. [11:16:05.0000] <jgraham> tantek: I don't think that is true and I think you are assuming your hypothesis and then picking and choosing data to support it [11:16:16.0000] <grigs> Back to my original point, my biggest frustration with the process feeling like there was a lot of wasted time despite asking a lot how to go about making something happen. [11:16:29.0000] <tantek> no, I'm just concluding from the vast majority of email that gets to sent to the whatwg list [11:16:33.0000] <Philip`> tantek: Not even maths has maths-style proofs of the suitability of its syntax - it just has people like Newton and Leibniz coming up with something arbitrary that works for them, until one of them becomes most widely adopted and only weirdos like physicists use the other [11:16:37.0000] <tantek> sure reads like a support forum [11:16:55.0000] <jgraham> You must be reading a very different list to me [11:17:04.0000] <ShaneHudson> It is a shame we cannot modify the <img> to have children.. Bruce Lawon and Christian Heilmann explained it to me on Twitter, but it would make things so much easier if we could [11:17:11.0000] <odinho> grigs: It was not wasted. Read what hober said about his proposal. And also, how easy everyone now agrees that the use cases are sane and good. [11:17:20.0000] <odinho> grigs: It was not like that last time. [11:17:30.0000] <ShaneHudson> How do we register for this wiki? Says I have to be an admin to register... [11:17:45.0000] <Philip`> ShaneHudson: Many things would be much easier if we could ignore reality :-) [11:17:46.0000] <grigs> odinho: Actually, no, we don’t seem to agree on use cases (which was tantek's earlier point). [11:17:47.0000] <tantek> ShaneHudson - any W3C account can sign into w3.org/wiki [11:18:14.0000] <odinho> grigs: I agree with the use cases you've put up. -- And putting media queries up to those points, I find it lacking. [11:18:23.0000] <ShaneHudson> tantek: Ah, so not wiki.whatwg? Ok [11:18:30.0000] <jgraham> ShaneHudson: The whatwg wiki? [11:18:44.0000] <jgraham> Uh Hixie is an admin, maybe annevk, AryehGregor perhaps [11:18:44.0000] <grigs> odinho: happy you agree, not clear everyone else does (again tantek's point) [11:18:44.0000] <jgraham> Uh Hixie is an admin, maybe annevk, AryehGregor perhaps [11:18:45.0000] <grigs> odinho: happy you agree, not clear everyone else does (again tantek's point) [11:18:46.0000] <odinho> grigs: media queries in <picture> do _more_ stuff though, -- and I thought there was more use cases people wanted, -- but I didn't see them on respimg. [11:19:18.0000] <Wilto> The official instructions for creating an account on the WHATWG wiki are to either ask someone in here to create it, or contact Hixie directly about one. [11:19:20.0000] <tantek> anyway, now that there's WHATCG, it probably makes reasonable sense to just use w3.org/wiki - unless there's major WHATWG objection to doing so. [11:19:46.0000] <odinho> So I thought people wanted the extra power of media queries as well (and had made some use cases that wanted that), but it seems not. :-) [11:20:15.0000] <tantek> odinho - indeed, it's easy to draw incorrect conclusions when we don't have specific use-cases to refer to by URL. [11:20:25.0000] <jreading> i have a hard time seeing a dozen or so <picture> elements on a page, each with 2 or 3 MQ's sources as a good solution… [11:20:39.0000] <tantek> jreading - a good solution to *what use-cases*? [11:20:41.0000] <grigs> odinho: i'm pleased hober found the threads useful. i was thankful to see his comments this morning. still hard not to see time wasted writing a spec for picture if that wasn't really a good next step. [11:20:43.0000] <ShaneHudson> Right I am going to grab dinner. Does anyone disagree with the wiki page being http://www.w3.org/wiki/Images ? [11:21:05.0000] <tantek> ShaneHudson - go ahead and start it. we can rename/move later if needed. [11:21:12.0000] <jreading> is their a github of use cases that I can submit pull requests? [11:21:33.0000] <tantek> jreading - wiki editing is easier / more usable/accessible that git for documentation/text [11:21:35.0000] <zewt> let's make discussing a feature as complicated as possible [11:21:48.0000] <ShaneHudson> jgraham: Use the wiki I just linked to.. I will be creating the page after dinner [11:22:27.0000] <tantek> just created it as a stub ;) [11:23:04.0000] <jreading> tantek: I was thinking markup, but I'd be happy to contribute some examples to the wiki that might illustrate my pain [11:23:06.0000] <jgraham> That doesn't follow the general layout of the W3C wiki [11:23:29.0000] <jgraham> But I don't really care [11:23:39.0000] <Wilto> jreading: I started a repo here a while back: https://github.com/Wilto/respimg [11:23:52.0000] <jgraham> (I think the WhatWG wiki would be better since it doesn't require W3C Member access) [11:23:56.0000] <Wilto> jreading: The WHATWG wiki page is more thorough, though. [11:24:30.0000] <TabAtkins_> grigs: Time wasted writing specs is, fortunately, never a consideration when deciding on which solution to use. If we prioritized "time spent by an editor", we'd make *much* worse decisions in general. [11:24:55.0000] <grigs> TabAtkins_: not my point. seriously. [11:24:56.0000] <tantek> TabAtkins indeed. [11:24:56.0000] <TabAtkins_> grigs: It definitely doesn't *feel* good to have something you've put time into not be used, but shrug. Good design must be egoless. [11:24:57.0000] <TabAtkins_> grigs: It definitely doesn't *feel* good to have something you've put time into not be used, but shrug. Good design must be egoless. [11:24:57.0000] <tantek> TabAtkins indeed. [11:25:22.0000] <TabAtkins_> grigs: "still hard not to see time wasted writing a spec for picture if that wasn't reallly a good next step." Sorry if I misinterpreted this. [11:25:27.0000] <tantek> good design must be based on referenceable real world use cases [11:25:30.0000] <odinho> TabAtkins_: And sunken cost-less! [11:25:46.0000] <TabAtkins_> odinho: Yeah. [11:27:11.0000] <tantek> ShaneHudson - I added the URLs I mentioned below to help kickstart the /Images page - please feel free to rewrite the page as you see fit - I gave it a bit of skeleton structure to get started: http://www.w3.org/wiki/Images [11:27:19.0000] <tantek> s/below/above [11:27:25.0000] <grigs> TabAtkins_: that was in response to someone saying there wasn't wasted time spent in the community group, not in response to a question about the solution selected. [11:28:11.0000] <tantek> grigs - time is usually most wasted on email lists / support forums. [11:28:12.0000] <tantek> grigs - time is usually most wasted on email lists / support forums. [11:29:39.0000] <ShaneHudson> tantek: Thank you :) I have not written anything quite like this before so anyone feel free to help and suggest improvements [11:29:54.0000] <ShaneHudson> Will get started on it in about 30 mins, not cooked dinner yet [11:31:23.0000] <tantek> ShaneHudson - we had similar problems with trying to expand the time element, and only by writing down discrete use-cases on the wiki were we able to discuss them individually, find out which were actually important (and which were not) and then *lastly* decide on syntax to support the rough consensus of useful/important use cases. [11:31:31.0000] <grigs> my point is that for people outside the standards process who would like to see something change in html, the process for contributing was confusing. so we asked a lot of questions. it was unclear who could provide definitive answers. a lot of time was spent with incorrect assumptions (no modifications to img tag) and jumping through hoops that may or may not have made sense (writing a spec when maybe we should have been working on use cases). [11:31:44.0000] <tantek> point being, not all the use-cases were satisfied, that's OK. [11:31:45.0000] <tantek> point being, not all the use-cases were satisfied, that's OK. [11:32:06.0000] <grigs> anyways, for next time, it would be nice to have something a little clearer for outsiders. which was my original point. :-) [11:32:09.0000] <tantek> " we asked a lot of questions. it was unclear who could provide definitive answers. a lot of time was spent with incorrect assumptions " ==> sign of a support forum [11:32:12.0000] <TabAtkins_> grigs: Ah, sure. Yeah, standards is hard. [11:32:13.0000] <grigs> any disagreement with that? [11:32:51.0000] <TabAtkins_> The takeaway is: present persuasive arguments, and figure out who will actually be making the decision, so you can decide what sort of presentation will work best for them. [11:33:00.0000] <TabAtkins_> Everything else is details. [11:33:16.0000] <tantek> grigs - yes, framing "outsiders" is imprecise. at the end of the day, Hixie is the editor of HTML, everyone else contributes their input/opinions. [11:33:53.0000] <TabAtkins_> For Hixie, for example, you just need persuasive use-cases. Examples of solutions can be useful, but are not required. However, they can be used to point out specific corners of the problem-space that need to be looked at. [11:37:10.0000] <tantek> adactio (late follow-up to your message) - Hixie experiments with adding things to the spec, and then sometimes removing them. If you're ok with (a believer in) "HTML the living standard" - then you're ok with that level of spec frothiness. The current alternative is to wait for things to stabilize / percolate into a W3C snapshot on TR. [11:37:34.0000] <grigs> tantek: would replacing "outsiders" with "people new to the whatwg standards process" be sufficient? i think we're just talking semantics. [11:37:53.0000] <tantek> outsiders just means anyone not editing the spec ;) [11:38:14.0000] <adactio> grigs: What do you mean exactly by "semantics?" ;-) [11:38:19.0000] <tantek> adactio LOL [11:38:31.0000] <grigs> adactio: LOL [11:39:05.0000] <grigs> https://twitter.com/#!/adactio/status/197343204683694080 [11:39:18.0000] <grigs> I’ve got that tweet favorited. [11:39:33.0000] <zewt> so many quotostrophes [11:42:53.0000] <grigs> tantek: a support forum would be nice. my fear in starting the community group is that we would persist in silo-iziation/tribalization and wouldn’t get the feedback we needed from people with more experience to really test our assumptions and conclusions. [11:43:33.0000] <zewt> the biggest problem I've seen with CGs is that for some bizarre reason, they all feel the need to set up little isolated mailing lists instead of using the larger lists that people are actually on [11:43:51.0000] <tantek> grigs, sometimes it's useful to get a smaller group of people together to agree on some set of clear principles, use-cases etc. in order to write them down as such for review/feedback by a broader audience. [11:44:00.0000] <grigs> tantek: perhaps some specific place to get that guidance would have helped. it could have made it clear that persuasive use cases was where to focus energy [11:44:26.0000] <zewt> tantek: i disagree; better off having the largest (relevant) audience possible, so you don't waste time with false starts that could be avoided with the right input (which you miss due to isolation) [11:44:31.0000] <tantek> zewt - the CG mechanics setup those lists by default. It's just bad UX, or rather, email-list-silo based UX (which is a traditional assumption of most standards orgs) [11:44:36.0000] <grigs> zewt: blame the tool makers, not the CG users. [11:44:56.0000] <tantek> grigs - from my understanding the recommendation to document persuasive use-cases is quite prevalent [11:44:56.0000] <tantek> \ [11:45:25.0000] <tantek> e.g. http://wiki.whatwg.org/wiki/FAQ#Is_there_a_process_for_adding_new_features_to_a_specification.3F [11:45:37.0000] <tantek> steps 1 and 2 [11:45:49.0000] <tantek> 1. Research the use cases and requirements by discussing the issue with authors and implementors. [11:45:49.0000] <tantek> 2. Come up with a clear description of the problem that needs to be solved [11:46:29.0000] <scottjehl> tantek did the CG not follow that process in your opinion? [11:47:03.0000] <grigs> tantek: that seems fair. perhaps we got a lot of bad advice during the process. hard to figure out where we went astray. [11:47:22.0000] <tantek> scottjehl - no - based on my request to see a simple wiki page of use-cases and there not being one. [11:47:31.0000] <tantek> the blog post was an excellent start [11:47:37.0000] <tantek> but blog posts are snapshots [11:47:40.0000] <tantek> they're not "living" [11:47:58.0000] <tantek> grigs - "bad advice during the process" - yes that's what happens on mailing lists / support forums. [11:48:30.0000] <grigs> tantek: and the alternative is? [11:49:02.0000] <grigs> tantek: i thought you were suggesting support forums. i think i missed your earlier point. [11:49:04.0000] <tantek> 1. document your research on an open wiki page, 2. post URLs on IRC etc. to gather feedback [11:49:14.0000] <tantek> posting actual "content" on email lists is pretty much a waste of time [11:49:18.0000] <jgraham> To be clear tantek's views on how to do things are personal. [11:49:27.0000] <Wilto> tantek: This sounds a lot like what I did with http://wiki.whatwg.org/wiki/Adaptive_images? [11:49:30.0000] <tantek> jgraham - no, the data fits my generazliations. [11:49:38.0000] <tantek> generalizations even. [11:49:50.0000] <jgraham> So you claim [11:49:56.0000] <tantek> the bad advice that the respimg folks got on the whatwg list is just the latest example [11:50:00.0000] <tantek> no, not claim - have URLs [11:50:04.0000] <tantek> people already posted them above [11:50:26.0000] <jgraham> Anyway the point that you should collect use cases first and present use cases before solutions is sound [11:50:42.0000] <scottjehl> jgraham we certainly did that [11:50:52.0000] <jgraham> scottjehl: That isn't clear [11:51:07.0000] <tantek> scottjehl - I have to agree with jgraham on that - that part wasn't clear. [11:51:13.0000] <jgraham> At least the wiki page I have seen that claims to have use cases mostly has solutions [11:51:19.0000] <tantek> posting use-cases on a wiki page helps make it *more* clear [11:51:20.0000] <othermaciej> tantek: what bad advice do you think the respimg guys got on the whatwg mailing list? [11:51:37.0000] <othermaciej> (interested in hearing because it would be good to ensure future contributors get good advice) [11:51:37.0000] <tantek> jgraham - agreed. intermingling a particular syntax among use-cases is not helpful. [11:51:38.0000] <Wilto> No one responded to my post linking http://wiki.whatwg.org/wiki/Adaptive_images to tell me that it wasn’t the correct way to handle use cases. [11:51:52.0000] <Wilto> It just got ignored, I assume because it is “done wrong.” [11:51:53.0000] <tantek> Wilto - what post? to a mailing list? [11:51:58.0000] <Wilto> Yes. [11:52:00.0000] <tantek> Wilto - most email is ignored [11:52:02.0000] <grigs> wilto: yeah, tantek did. he said it includes picture element. [11:52:06.0000] <tantek> it's an inherent problem with the medium [11:52:08.0000] <tantek> especially lists [11:52:12.0000] <tantek> "someone else will handle it" [11:52:14.0000] <Wilto> Eesh. [11:52:21.0000] <jgraham> Wilto: At the time I tried to suggest that you focused more on use cases. Apparently I wasn't cler nough nd didn't follow thorugh. Sorry. [11:52:29.0000] <grigs> Wilto: i missed his comment when it went past the first time, but it is in the irc logs. [11:52:54.0000] <Wilto> I mean, that is fair: this was mentioned, and you guys did try to help. I think I just misinterpreted the feedback somewhat. [11:53:00.0000] <grigs> Wilto: nevermind. i misunderstood. [11:53:15.0000] <Wilto> But, yeah—nothing on the mailing list. [11:53:37.0000] <scottjehl> the CG is where we were told to post and plan it, though. @wilto constantly asked around to ensure we were following the procedures that would get us into the conversation. If that was an inadequate location for planning the feature, we would have loved to have been told that before a completely new solution was invented. [11:54:01.0000] <grigs> Yeah. :-( [11:54:11.0000] <Wilto> Dead-on as always, scottjehl. [11:55:09.0000] <zewt> mail to the whatwg list always gets a reply ... but hixie is overloaded, so sometimes it takes a *long* time [11:55:14.0000] <jgraham> I think the big thing lacking from the CG was implementors. You are unlikely to get people to implement something that they don't know about and haven't given feedback on [11:55:43.0000] <jgraham> The place where implementors (except Microsoft) typically work on HTML is WHATWG [11:55:57.0000] <grigs> jgraham: amen. that is what I asked about specifically BEFORE the CG was created: http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2012-February/034790.html [11:57:03.0000] <grigs> If there is anything that bums me out about this, it is the fact it was perfectly predicable and despite our efforts, seemingly unavoidable. [11:57:14.0000] <zewt> the mail you're replying to was wrong; whatwg is definitely the place for that discussion, IMO [11:57:19.0000] <jgraham> FWIW the CG might have been a good place to gather *requirements* [11:57:32.0000] <scottjehl> this is all very hard to hear now, naturally [11:57:47.0000] <smaug____> Web Audio API is horrible. Hard to even start reviewing because everything is so under-defined [11:58:36.0000] <odinho> scottjehl: http://odin.s0.no/web/srcset/polyfill.htm [11:58:42.0000] <jgraham> grigs: So we will do better next time :) [11:58:49.0000] <odinho> scottjehl: So, come with all the bugs. :P [11:59:19.0000] <zewt> (don't know who "ronjec viktor" is, don't recall seeing his name before) [11:59:29.0000] <tantek> jgraham - yes, CGs are plenty fine places to gather and even prioritize requirements and use-cases [11:59:37.0000] <Wilto> I’m not comfortable saying “oh well, this is a forgone conclusion because the process is broken; better luck next time.” [11:59:40.0000] <tantek> but I think bikeshedding a solution is too tempting for any group to avoid [12:00:17.0000] <jgraham> Wilto: There is no "conclusion" yet [12:00:37.0000] <tantek> grigs, Wilto, scottjehl, more on the problems of doing most work on email lists (rather than doing most work on wikis instead) http://microformats.org/wiki/wiki-better-than-email [12:00:39.0000] <Wilto> Fair. [12:00:52.0000] <jgraham> Not much has happened really [12:01:05.0000] <tantek> /me agrees with jgraham [12:01:13.0000] <jgraham> Hixie has been convinced that a set of requirements exist that are worth solving [12:01:21.0000] <zewt> lots of very useful discussions like this happen on email; it's when people go to a wiki that I cringe at the distracting waste of time [12:01:30.0000] <tantek> there's still quite the opportunity to define prioritized requirements and use-cases [12:01:34.0000] <Wilto> Oh no—by no means to I consider this a done deal. [12:01:34.0000] <jgraham> Whether or not these are the same requirements that you have is still unclear [12:01:37.0000] <Wilto> do I. [12:01:53.0000] <jgraham> There is no implementation and so no lock-in [12:01:56.0000] <tantek> zewt - I mostly see discussions either repeated, or based in theory in email [12:02:24.0000] <grigs> Slightly different topic: I have a question about a use case that I’ll document on the stub page later today. [12:03:38.0000] <grigs> One of the use cases that I see is for the image changing at different sizes. I used a Nokia Browser page as an example on the mailing list: http://browser.nokia.com/smartphones.html [12:03:43.0000] <zewt> tantek: a wiki doesn't stop email; it just duplicates it [12:04:18.0000] <grigs> The browser for Meego image changes depending on the viewport width. [12:04:52.0000] <scottjehl> jgraham: why isn't the feature driven by web developers' well-known documented needs? I think the web development community has been very clear about the requirements we've needed for this, placing them in the place we were told would be noticed by implementors and spec writers. Why didn't these use cases drive the spec? Is the problem truly that our planning was in the wrong location on the web? [12:05:02.0000] <grigs> I see this as being something that a lot of sites would do with header graphics on home pages. There aren't many examples in the wild because you can't do it with img yet. [12:05:56.0000] <jgraham> scottjehl: I think you are overestimating the clarity of the requirements [12:06:07.0000] <grigs> I'm pretty sure to support this use case, the change in image has to be aware of breakpoints because not only does the image change, but also things like whether or not the text next it to floats also changes. [12:07:03.0000] <scottjehl> I don't think anyone involved in the CG was unclear about them [12:07:18.0000] <grigs> For example, it seems likely that Apple would need something like this if they ever implemented as responsive design as their header images contain text and are <img> tags. [12:08:20.0000] <odinho> scottjehl: The use cases I saw on the blog seems to be covered by Hixie in his email. -- There's maybe one of cropping pictures (doing a landscape and portrait) that I'm not really sure I agree with, -- but it's in the current srcset spec right now. So that use case is also working. [12:08:32.0000] <grigs> Anyways, the questions I have are 1) how to document use cases when you don’t have existing sites doing it, but seems likely 2) does the language I’m using make sense (e.g., header images). I fear I don’t have the right terminology to explain what I mean. [12:09:30.0000] <scottjehl> grigs: perhaps, but often even a large image, when compressed well, is small enough in filesize that negotiation isn't necessary. I see this more useful for photography - say an article lead feature image. Dramatically different file sizes at different dimensions. [12:09:45.0000] <tantek> zewt - actually it does, because it makes it trivial for anyone to reply with a URL and say - already discussed, see this. [12:09:54.0000] <tantek> email duplicates email [12:10:26.0000] <scottjehl> anyway, both use cases are kind of the same [12:10:44.0000] <grigs> scottjehl: ha ha. your comment makes it clear my fear that i'm not explaining myself clearly is well-founded. :-) [12:11:01.0000] <tantek> scottjehl " web development community has been very clear about the requirements we've needed for this" - if anything the lack of such clarity should be obvious from this IRC conversation! [12:11:21.0000] <grigs> /me going to lunch. back later. [12:11:35.0000] <zewt> tantek: that's already trivial; a link to list archives [12:11:35.0000] <scottjehl> tantek rehashing things we've already agreed upon in the wrong channel is all. These are the same use case [12:11:52.0000] <odinho> Wilto: Could you also quote what you are replying to, and answer at the bottom of that? I know it's a common concern, but when reading so much email on my phone as I'm doing now, it's nicer and easier to read/follow when people write it in that style on WHATWG list. [12:11:54.0000] <tantek> "Why didn't these use cases drive the spec?" - were the use cases linked to with permalinks in all discussions of them? (I think not) [12:12:26.0000] <Wilto> Sure thing, odinho. My mistake. [12:12:32.0000] <jgraham> grigs: That use case is described as something like "on a large viewport width, a large wide image is loaded. On a smaller viewport a different image (e.g. a crop of the original) is loaded that is narrower allowing a different layout more sutiable for the screen size" or something [12:12:48.0000] <tantek> "Is the problem truly that our planning was in the wrong location on the web?" - to some extent yes. Some places on the web are more findable than others. E.g. wiki pages are FAR more findable than archived emails (which may be just a fraction of a message in a longer thread which takes too long to re-read and parse) [12:13:26.0000] <tantek> it's pretty simple actually. stop putting substantial content into email lists assuming people will see it and/or find it. they won't. this has been repeatedly shown/understood. [12:13:48.0000] <tantek> put substantial content on the web where it can be trivially found by search engines. wikis tend to be very good for that. [12:14:23.0000] <zewt> i find mailing list conversations constantly; stuff on wikis is more hidden [12:14:35.0000] <tantek> if someone asks you to use an email list, use it only to communicate very short summaries and permalinks to the aforementioned wiki pages [12:14:41.0000] <zewt> gross [12:14:47.0000] <odinho> Wilto: Thanks. :-) [12:14:52.0000] <scottjehl> tantek: re: permalinks: Fair enough, but we had no idea that any of our documentation was inadequate for the working group. We would have been thrilled if whatwg asked us to clarify portions to aid in their planning. No contact at all. [12:14:57.0000] <tantek> google (bing etc.) find thing things on wikis MUCH more easily than mailing lists [12:15:00.0000] <zewt> wikis are pretty much useless for any kind of discussion [12:15:07.0000] <tantek> so is email [12:15:11.0000] <tantek> irc is a bit better [12:15:12.0000] <zewt> uh, no it isn't? [12:15:56.0000] <zewt> i've had countless useful discussions on email, so the claim just doesn't make any sense [12:16:19.0000] <scottjehl> zewt sorry, my claim? [12:16:39.0000] <tantek> scottjehl - there is no "whatwg" to have "asked" you to "clarify portions" - that's an illusion perpetrated by some individuals [12:16:48.0000] <tantek> there is the editor of the spec, Hixie [12:17:01.0000] <zewt> (your what?) [12:17:10.0000] <scottjehl> zewt disregard sorry [12:17:12.0000] <tantek> and there is the support forum known as the whatwg email lists, where people who claim to be a part of or not of it claim to speak on its behalf [12:17:30.0000] <odinho> Email is nice for discussions, -- irc is nice for more rapid-fire and other types of discussion (even less permanent than email, and shorter bursts, no as thought out). -- Wiki's are the place to always update while you're having the discussion, filling out and having a nice permanent page for all the important points and conclusions being done whilst discussing in another forum. [12:18:16.0000] <jgraham> You know this whole doublespeak thing of saying "support forum" instead of "mailing list" gets old fast [12:18:17.0000] <tantek> etherpad is a nice realtime improvement upon wikis also - works better for some discussions [12:18:17.0000] <zewt> http://krijnhoetmer.nl/irc-logs/whatwg/20120515#l-2154 sort of permanent [12:18:18.0000] <tantek> etherpad is a nice realtime improvement upon wikis also - works better for some discussions [12:18:38.0000] <tantek> jgraham - I don't think you understand what "doublespeak" means. [12:18:46.0000] <scottjehl> yeah, picture planning originated in etherpad [12:18:47.0000] <tantek> it's more like, acts like a duck, probably is a duck [12:18:59.0000] <tantek> email lists that act like support forums may as well be considered support forums [12:19:16.0000] <odinho> zewt: The hilighting makes things easier to follow. But still, it's really horrendous to read IRC logs when you want info. It's much better, but still very hard to read email discussions when you want to find out of something. Wiki's (or normal authored web pages) are better for that. [12:19:35.0000] <zewt> (i didn't say any of that, heh) [12:19:40.0000] <odinho> zewt: Think if the HTML spec was the entire email archive, -- you just had to read through it (most often some of the last emails) to find out what happened :] [12:19:41.0000] <tantek> hmm krijnhoetmer.nl appears to be slow to respond for me [12:19:42.0000] <tantek> hmm krijnhoetmer.nl appears to be slow to respond for me [12:19:47.0000] <bjankord> CG was a great place to read info, far better than IRC logs or email lists [12:19:58.0000] <scottjehl> +1 [12:20:14.0000] <bjankord> CG was also open to everyone [12:20:23.0000] <bjankord> The best ideas rise to the top this way [12:20:38.0000] <zewt> odinho: you seem to be responding to things i didn't say :) [12:22:45.0000] <odinho> 21:19 < zewt> http://krijnhoetmer.nl/irc-logs/whatwg/20120515#l-2154 sort of permanent [12:22:50.0000] <scottjehl> ...that the formatting of it was insufficient was never communicated to us, but we would have gladly made any changes that would have helped. [12:22:58.0000] <odinho> zewt: You said that irc was sort of permanent. [12:23:01.0000] <necolas> if the mailing list is the place to communicate with the whatwg / hixie, then perhaps it should be acknowledged that it is up to whatwg / hixie to communicate with other parties via the place they prefer to be involved (e.g. CG or w/e) [12:23:03.0000] <scottjehl> Loads of developers actually thought they were taking the right steps to help plan this feature. [12:23:12.0000] <zewt> odinho: ... none of what you said has anything to do with that :) [12:23:33.0000] <zewt> i never said anything about reading IRC logs [12:24:02.0000] <zewt> only responding to "less permanent" [12:24:16.0000] <bjankord> entering lurk mode [12:25:39.0000] <scottjehl> http://www.alistapart.com/articles/responsive-images-and-web-standards-at-the-turning-point/ [12:25:59.0000] <odinho> zewt: I talked about it, -- and your comment seemed to be pointed at me. [12:26:34.0000] <MarcDrummond> The thing which I find disappointing is that there seems to be a whole lot of discussion about process going on here, and not a lot of discussion of the spec change. [12:26:35.0000] <MarcDrummond> The thing which I find disappointing is that there seems to be a whole lot of discussion about process going on here, and not a lot of discussion of the spec change. [12:26:47.0000] <zewt> you said irc isn't permanent, which sounds like you don't know about the public logs, so i showed them to you; nothing more [12:27:09.0000] <MarcDrummond> Obviously, process is important, but what we really care about is the implementation for responsive images. [12:27:17.0000] <zewt> MarcDrummond: that belongs on the list, most of the time [12:28:17.0000] <MarcDrummond> And so the wheel continues to spin. [12:28:20.0000] <zewt> ... [12:28:33.0000] <jgraham> MarcDrummond: If you have points that require real-time interaction feel free to discuss them here [12:28:56.0000] <jgraham> If you have points that you want documented, use the wiki or the mailing list [12:29:48.0000] <jmather> scottjehl: thanks for the link [12:30:28.0000] <odinho> zewt: Hehe, I know very well about the krijnhoetmer.nl public logs. :-) I ever wrote a irc log indexer internally for Opera. [12:31:15.0000] <tantek> thanks scottjehl for the ALA link - I added it to: http://www.w3.org/wiki/Images#see_also [12:31:16.0000] <tantek> thanks scottjehl for the ALA link - I added it to: http://www.w3.org/wiki/Images#see_also [12:31:24.0000] <tantek> (which you should feel free to edit directly as well) [12:32:21.0000] <MarcDrummond> One concern that I have, as an author, with srcset, is locking dimensions to pixels. I may want to define an image in ems, so that it can flex with text size changes, since my layout is defined with ems. Even though the image would degrade, it would still stay in proportion to the design. [12:33:02.0000] <odinho> MarcDrummond: It has nothing to do with how the picture is presented. None whatsoever. [12:33:12.0000] <jmather> MarcDrummond: I think srcset only sets rules on when to use which image, you can size it with css width/hight as normal [12:33:20.0000] <odinho> MarcDrummond: You can write 100000000w for a picture that is 234px wide. [12:33:20.0000] <odinho> MarcDrummond: You can write 100000000w for a picture that is 234px wide. [12:33:23.0000] <scottjehl> sure thing [12:33:25.0000] <tantek> MarcDrummond, when there's a breakdown (as seems obvious in this area), it makes sense to unwind to the point of where the breakdown in order to fix it and make progress again towards a solution for responsive images. It's pretty clear that the breakdown occurred at a difference of use-cases etc., so we're trying to figure out how to best resolve/communicate that and reach a rough consensus on use-cases (and their priorities). [12:33:39.0000] <tantek> A solution that fails to solve the "important" use-cases is useless. [12:34:05.0000] <tantek> s/solution/implementation etc. [12:34:35.0000] <zewt> (discussing here is fine, of course, but discussion *is* going on on the list, and if you want to make points that the wider audience--rather than whoever happens to be here right now--will see, that's the place to do it) [12:34:43.0000] <jmather> scottjehl: this is totally OT, but i've been following your trip on twitter -- absolutely fantastic. It must be incredible. [12:34:44.0000] <jmather> scottjehl: this is totally OT, but i've been following your trip on twitter -- absolutely fantastic. It must be incredible. [12:35:24.0000] <jgraham> Right, the Aw Bh syntax is *somewhat* like a max-width:Apx max-height:Bpx media query [12:35:28.0000] <tantek> zewt - apparently the list has a wider audience that includes sufficient misinformation (the "go away" email on whatwg) to actually *harm* discussion. [12:35:29.0000] <tantek> zewt - apparently the list has a wider audience that includes sufficient misinformation (the "go away" email on whatwg) to actually *harm* discussion. [12:35:52.0000] <jgraham> Uh, +viewport somewhere [12:36:09.0000] <zewt> tantek: that's uncommon, and present in all media [12:36:22.0000] <MarcDrummond> I thought somebody said above that nobody reads email, or at least not in a timely manner. Was that in relation to the listserv or emails sent directly to a person? [12:36:22.0000] <jgraham> It basically means "don't display this image unless the viewport is at least A x B" [12:36:32.0000] <tantek> zewt - nope. such miscommunication is quickly corrected in places like IRC. but not apparently email. [12:36:38.0000] <scottjehl> jmather ah! Kind of you to say, thanks :) [12:36:39.0000] <tantek> this is a failing of email, and especially heavy lists. [12:36:52.0000] <tantek> too much crap goes by that no one bothers to correct, so it gets propagated. [12:37:01.0000] <ShaneHudson> Talking of the list.. I posted to it earlier, I got a reply from Matt Wilcox but it is not in the archive... I did reply to him, so did I not send it properly? I had the list as a cc [12:37:04.0000] <jgraham> MarcDrummond: tantek has an aversion to mailing lists which he spreads with evangelical fervour [12:37:15.0000] <jmather> scottjehl: not kind really… I'm just jealous. Hah. [12:37:22.0000] <tantek> jgraham - you can stop the ad hominem anytime you like. [12:37:35.0000] <jgraham> tantek: Which part isn't true? [12:37:37.0000] <tantek> I'm simply communicating based on actual experiences (with citations) [12:37:43.0000] <othermaciej> tantek dislikes email and mailing lists more than most, Hixie (person who actually does much of the editing) likes email and mailing lists more than most [12:37:55.0000] <tantek> you're on the otherhand simply being defensive about emailing lists [12:38:08.0000] <odinho> othermaciej: Nice summary. [12:38:12.0000] <jgraham> tantek: "Citations" being more or less anecdotes [12:38:32.0000] <tantek> othermaciej - right, whatwg list works as inbox for Hixie, beyond that, it's effectively a support forum. [12:38:48.0000] <jgraham> I'm saying that overall the WHATWG list has been super-effective [12:38:52.0000] <othermaciej> in my observation, the whatwg list is reasonably functional, and to the extent it has failure modes, "support forum" does not strike me as a fair characterization thereof [12:39:36.0000] <tantek> jgraham - to URLs of actual occurrences yes. on the otherhand all you present is knee-jerk defensive reactions in IRC. anecdotes vs. IRC defensiveness, anecdotes are more persuasive. [12:39:39.0000] <jgraham> And wikis have all sorts of problems that you don't mention e.g. people unwilling to trample other people's edits, not knowing which stuff is significant, edit/revert wars, etc. [12:40:00.0000] <jgraham> But I am not really interested in discussing the merits of these things [12:40:06.0000] <jmather> scottjehl: I was a little disheartened when Jen pointed me to your picture implementation though. I thought I had something novel… :D [12:40:09.0000] <tantek> then why do you keep bringing it up? [12:40:16.0000] <othermaciej> maybe I don't understand what is meant by "support forum" but I imagine it means questions like "how do I use <video> on my website to play Flash videos?" or whatever, which seem rare [12:40:38.0000] <tantek> othermaciej - it's a support forum for standards development for folks who don't realize they need a support forum. [12:40:39.0000] <jgraham> The fact is that in order to engage with WHATWG/Hixie the mailing list is the primary veichle [12:40:53.0000] <bjankord> Agreed [12:41:03.0000] <tantek> nah, the whatwg list is a nice convenient public inbox for Hixie, nothing more really. [12:41:06.0000] <jgraham> But you keep trying to discourage its use [12:41:08.0000] <scottjehl> dropping off. [12:41:08.0000] <othermaciej> whatwg is culturally friendly to use of a wiki in my experience, but if you avoid use of the mailing list entirely, you're gonna have a bad time [12:41:09.0000] <scottjehl> dropping off. [12:41:09.0000] <othermaciej> whatwg is culturally friendly to use of a wiki in my experience, but if you avoid use of the mailing list entirely, you're gonna have a bad time [12:41:30.0000] <jgraham> and are trying to coin some phrase to keep people away from it (the support forum thing) [12:41:41.0000] <tantek> othermaciej, sure, because Hixie prefers it as his inbox, so it makes sense to send emails to whatwg accordingly. [12:42:10.0000] <tantek> jgraham, I'm not trying to discourage you from contributing to a support forum, please go on doing so. [12:42:19.0000] <jgraham> See there we go again [12:42:22.0000] <jgraham> It is tiresome [12:42:25.0000] <jgraham> Plese stop [12:42:30.0000] <jgraham> We are not children [12:42:34.0000] <othermaciej> tantek: it is kind of trollish to refer to a communication medium repeatedly using a term that its participants would not self-apply [12:43:04.0000] <tantek> primarily Hixie's inbox? [12:43:49.0000] <othermaciej> "support forum" [12:44:27.0000] <othermaciej> the whatwg.org site identifies it as "A discussion list for feedback on the specs", and says "The WHAT Working Group discusses issues and new proposals on an open and public mailing list, whatwg⊙wo Discussion from interested parties is welcome." [12:44:41.0000] <othermaciej> in my experience, many active participants on the list see it that way as well [12:44:50.0000] <tantek> othermaciej, perhaps those that tend to realize it's more of a support forum then go ahead and spend their energies elsewhere, leaving only those who haven't yet realized it. so it's not necessarily unreasonable that current active participants would not self-apply the description. [12:45:18.0000] <tantek> btw - I would say the same about public-html as well FWIW. [12:45:18.0000] <othermaciej> you may well have a point that discussion lists have intrinsic dysfunctions by nature [12:45:38.0000] <tantek> sure, there are intrinsic dysfunctions, that's perhaps a broader statement. [12:45:44.0000] <othermaciej> but referring to the discussion list with a dismissive term is somewhat rude and not a very effective way to make that point [12:45:57.0000] <tantek> I mean more that discussion lists tend to become support forums, and have seen this occur with pretty much every standards list. [12:46:00.0000] <othermaciej> doing so is likely to generate more heat than light [12:46:01.0000] <jgraham> I am unaware of any communication medium that doesn't have some dysfunction [12:46:53.0000] <tantek> othermaciej - if it helps a few more folks come to the support forum realizations and then spend their energies more effectively elsewhere (or focus their list posts accordingly), then that provides a net benefit. [12:47:18.0000] <jgraham> And for any reasonable definition of "support forum" the whatwg list isn't. People almost never ask how-to type questions there. [12:47:22.0000] <tantek> you're right that it will/does cause some to react defensively and simply dig-in to support a tradition. [12:47:31.0000] <jgraham> Anyway like I say, this is dull [12:47:53.0000] <MarcDrummond> Wilto has does an excellent job writing up a description of the current conundrum, and why picture would likely work far better than srcset: http://www.alistapart.com/articles/responsive-images-and-web-standards-at-the-turning-point/ [12:48:16.0000] <othermaciej> tantek: mentioning a few times that "mailing lists tend to turn into support forums, in my experience" or something like that seems like a fine and effective way to make the point [12:48:16.0000] <tantek> right, it's not an explicit support forum, it's an implicit support forum. people don't ask "how to" questions, people ask for features etc. which are implicit "how do I do this" type questions. [12:48:44.0000] <othermaciej> tantek: thereafter always using the term "support forum" to refer to the list is likely to cause more disruption and meta-discussion than effective persuasion [12:48:46.0000] <ShaneHudson> I was surprised to see he was able to ublish to ALA so quickly,I would expect to wait for ages! [12:48:51.0000] <jmather> tantek: I just joined the chat here, but i wanted to tell you, you're coming off as someone who seems to have an axe to grind, and nobody suitable to grind it with. I don't know what you're expecting to get out of the exchange you're engaging in but I hope it's worth it. [12:48:52.0000] <tantek> MarcDrummond, why limit yourself to two possible solutions? [12:48:52.0000] <jmather> tantek: I just joined the chat here, but i wanted to tell you, you're coming off as someone who seems to have an axe to grind, and nobody suitable to grind it with. I don't know what you're expecting to get out of the exchange you're engaging in but I hope it's worth it. [12:48:53.0000] <tantek> MarcDrummond, why limit yourself to two possible solutions? [12:49:07.0000] <tantek> when the use-cases aren't even broadly understood, prioritized, and agreed upon? [12:49:23.0000] <othermaciej> particularly since the core of a community is likely most invested in its traditions, yet also ultimately needs to be influenced if there's a desire to change the process' [12:49:58.0000] <tantek> othermaciej - I'm not sure about that, I think if Hixie is influenced, that's sufficient for change (in the spec or elsewhere) [12:49:58.0000] <tantek> othermaciej - I'm not sure about that, I think if Hixie is influenced, that's sufficient for change (in the spec or elsewhere) [12:50:12.0000] <tantek> the "community" doesn't really have much power in that way [12:51:56.0000] <tantek> jmather, as a counter to your "axe to grind", I'd offer only the fact that it's jgraham who has used terms like "doublespeak" and "evangelical" in a directed personal attack manner, not I. Such ad hominem behavior is usually an indication of failure to engage in more substantial discussion. [12:52:10.0000] <othermaciej> well, if you want to point out to Hixie that his mailing list is a support forum and encourage him to stop using it (as much), it might be less disruptive to do that elsewhere [12:52:11.0000] <ShaneHudson> So is Hixie in charge of the entire spec? Still trying to get my head around the pecking order [12:52:12.0000] <necolas> tantek: agreed. it shouldn't be about one early proposal vs another. [12:52:31.0000] <tantek> ShaneHudson, yes, Hixie is the editor of HTML. [12:53:07.0000] <othermaciej> Hixie derives his influence in large part from being responsive to other influential people [12:53:12.0000] <jmather> tantek: Like I said, I just entered the room here recently, so I don't have historical context, but i'd hope if I were coming off poorly someone would let me know, so I was just trying to do you a favor. Not trying to be rude, just maybe suggesting to grab a breather and reassess of what you're doing right now is worth the effort. :) [12:53:16.0000] <annevk> tantek: is there a wiki page for your suggested process or do I have to extract it from the IRC logs somehow? [12:53:27.0000] <annevk> /me can probably read a little backlog [12:53:27.0000] <tantek> othermaciej, I've discussed this with Hixie, and for him, using the list as a personal public inbox works for him, so I'm not going to argue with that. we all have our own personal ways to get our work done. [12:53:38.0000] <ShaneHudson> tantek: Thanks.. I didn't realise before there was one person in charge [12:53:41.0000] <tantek> annevk - I was citing whatwg FAQ earlier. [12:53:44.0000] <necolas> i'd just like to reiterate the point that "forward hixie's email to the CG" doesn't really constitute what i would consider an acceptable level of outreach to the wider community [12:54:02.0000] <othermaciej> that being said, he <3s email as a communication channel, and it's unlikely he could be convinced to use it much less without very good reson [12:54:18.0000] <tantek> jmather - you're right, this discussion is likely beyond the point of diminishing returns. [12:54:22.0000] <annevk> tantek: I thought your point was that the mailing list didn't work; I was wondering what you think would be better [12:54:22.0000] <annevk> tantek: I thought your point was that the mailing list didn't work; I was wondering what you think would be better [12:54:38.0000] <MarcDrummond> tantek: In theory, there are an infinite number of solutions. In practice, numerous web developers have coalesced around the picture solution on one hand, while WHATWG has published srcset as the proposed solution, with very little discussion (or use cases). [12:54:39.0000] <tantek> annevk - I've continuously reiterated the request for well documented use-cases. [12:54:53.0000] <ShaneHudson> tantek: I am working on the use-cases now :) [12:54:55.0000] <tantek> which from my understanding is what WHATWG tends to prefer ;) [12:55:04.0000] <annevk> tantek: there's http://wiki.whatwg.org/wiki/Adaptive_images [12:55:17.0000] <MarcDrummond> tantek: But I'm quessing your question was rhetorical, since you know all that. [12:55:20.0000] <tantek> annevk - that page is problematic because it presumes <picture> too much [12:55:24.0000] <annevk> tantek: and they have been provided via email too, I think Hixie mentioned them upfront [12:55:31.0000] <annevk> tantek: easy enough to read around that [12:55:41.0000] <jgraham> annevk: Beter to edit around it [12:55:48.0000] <jgraham> i.e. rewrite the page [12:55:48.0000] <othermaciej> annevk: do you understand how the "DOM Scripting and Bandwidth" item on that paye is a use case? [12:55:50.0000] <tantek> annevk - better is: http://www.w3.org/community/respimg/2012/04/16/summary-of-use-cases-and-requirements/ [12:55:54.0000] <jgraham> But some people wanted accounts [12:55:57.0000] <tantek> but that's not on a wiki page [12:56:01.0000] <jgraham> Do you have admin access? [12:56:04.0000] <tantek> so it can't be edited/ prioritized etc. [12:56:10.0000] <ShaneHudson> annevk: We came to the decision that use-cases are all over the places, so we are putting together a page on the wiki to define them once and for all [12:56:14.0000] <annevk> othermaciej: I don't agree with all the use cases :) [12:56:21.0000] <ShaneHudson> http://www.w3.org/wiki/Images [12:56:24.0000] <annevk> jgraham: I do [12:56:32.0000] <jgraham> Someone also started a page on the W3C wiki, but you need a W3C ccount to edit that [12:56:38.0000] <annevk> haha [12:56:42.0000] <annevk> four places already [12:56:46.0000] <tantek> annevk - exactly, we need this on a wiki so we can all contribute to the use-cases [12:56:46.0000] <annevk> way to go internet :) [12:56:47.0000] <tantek> annevk - exactly, we need this on a wiki so we can all contribute to the use-cases [12:56:47.0000] <ShaneHudson> (I am writing it at the moment, but that is the page it will be) [12:57:04.0000] <tantek> thanks ShaneHudson [12:57:04.0000] <tantek> thanks ShaneHudson [12:57:14.0000] <jgraham> tantek: Which of the muliple wikis it is on did you have in mind [12:57:19.0000] <othermaciej> <http://www.w3.org/community/respimg/2012/04/16/summary-of-use-cases-and-requirements/> is cool, would be nice to consolidate that with one of the wiki-based efforts [12:57:28.0000] <jgraham> (all with non-overlapping sets of authorised users) [12:57:33.0000] <jmather> othermaciej: I think someone is working on that... [12:57:34.0000] <tantek> othermaciej - my conclusion exactly [12:57:39.0000] <annevk> tantek: anyway I was talking about your meta point, the mailing list being a "support forum" [12:57:43.0000] <ShaneHudson> In fact, if everyone wants to post or email (shane⊙sn) what use-cases you think there are then I will compile them with the others [12:57:47.0000] <jmather> but i could have misunderstood what's going on here. :D [12:57:52.0000] <Philip`> Someone should set up a Tumblr with a list of all the wikis [12:57:53.0000] <annevk> tantek: you mentioned that irl before, but I didn't quite what you meant or what you want to replace it with [12:58:17.0000] <tantek> jgraham - re: which wiki, I asked, and people didn't seem to have a preference. so given that WHATCG exists right now, I suggested w3.org/wiki (not the respimg community wiki) so that anyone with a w3c account could edit/contribute [12:59:08.0000] <jgraham> That seems to exclude most developers [12:59:10.0000] <tantek> annevk - it's more of a transition than a replacement outright. the more content can be published in places that can be edited/updated/linked-to, the better. email for sending around links to such content is fine. [12:59:31.0000] <jgraham> Also, http://www.w3.org/community/respimg/2012/04/16/summary-of-use-cases-and-requirements/ is indeed very nice [12:59:37.0000] <jmather> jgraham: which is kind of what is causing the current uproar, which might not want to be perpetuated... [12:59:37.0000] <tantek> annevk - here's some more reading on it: http://microformats.org/wiki/wiki-better-than-email [13:00:07.0000] <tantek> jgraham - yes, even just moving that blog post to a wiki for iteration would be a huge help [13:00:09.0000] <ShaneHudson> jgraham: I will add those points in that article to the wiki. Please stop moaning, instead helping would be appreciated [13:00:25.0000] <tantek> I think we just need to give ShaneHudson some time to update w3.org/wiki/Images :) [13:00:37.0000] <jmather> I don't think Anselm would be upset if someone copy/pasted it. [13:01:03.0000] <anatolbroder> /me just wants to tell thank you to othermaciej and other nice guys who brought the srcset attribute [13:01:32.0000] <othermaciej> very little of the credit goes to me! [13:02:11.0000] <othermaciej> I did suggest the name "srcset" instead of "set", but otherwise most of the design and spec credit goes to hober and Hixie (as well as to people who provided use cases) [13:05:05.0000] <tantek> annevk - since you asked, here's some more documentation on the matter: http://www.wikinomics.com/blog/uploads/wiki_collaboration2.jpg from http://www.wikinomics.com/blog/index.php/2008/03/26/wiki-collaboration-leads-to-happiness/ [13:06:17.0000] <jmather> Can I ask a question and have everyone just assume I'm not trolling and give me an honest answer? Because I'm honestly curious why srcset doesn't reuse media query "powers." [13:06:32.0000] <jgraham> ShaneHudson: I am not moaning. I think I am suggesting that the W3C wiki is a bad choice of venue. But I don't want to stop you doing the work [13:07:25.0000] <annevk> tantek: mkay, I think we use wikis pretty often actually; they just haven't been used in this instance [13:07:25.0000] <annevk> tantek: mkay, I think we use wikis pretty often actually; they just haven't been used in this instance [13:07:28.0000] <jmather> jgraham: can developers not get w3c logins? Is that an issue? or is it just that developers aren't likely to already have w3c logins and having it be a barrier to participate? [13:07:37.0000] <tantek> jgraham - if you could provide a link to your criticisms of the w3c wiki, I'd like to understand that better. thanks. [13:07:49.0000] <jgraham> tantek: How does one get an account? [13:07:57.0000] <tantek> annevk - agreed. that was part of my point. [13:07:57.0000] <annevk> tantek: see e.g. timed tracks, the recent canvas additions [13:07:58.0000] <tantek> annevk - agreed. that was part of my point. [13:08:08.0000] <ShaneHudson> jmather: it is easier to access the w3c wiki than it is the whatwg, which is why we decided on it [13:08:35.0000] <ShaneHudson> General question (perhaps tantek could answer).. I presume everyone is fine with me writing the wiki in British English? [13:08:44.0000] <tantek> jgraham, right at the top of the home page: http://www.w3.org/wiki/Main_Page [13:08:52.0000] <tantek> Request a Public W3C Account to get started. [13:08:53.0000] <tantek> Request a Public W3C Account to get started. [13:09:03.0000] <tantek> where Public W3C Account links to: [13:09:04.0000] <tantek> http://www.w3.org/Help/Account/Request/Public [13:09:09.0000] <MarcDrummond> Wish somebody would answer jmather's very reasonable question. [13:09:20.0000] <tantek> since it's right at the top of the home page, I'm not sure how that could be made more obvious [13:09:22.0000] <import-logic> ShaneHudson: I don't see why it would matter, if anyone had a problem with that they would be way too picky :) [13:09:26.0000] <jmather> ShaneHudson: ok, seems fair enough to me then. I was just curious if it was a "restricted membership" or something. I think it's fair to say if developers are stumped by a signup form they probably shouldn't participate. :D [13:09:28.0000] <tantek> (i.e. not sure that adding it to an FAQ would help any) [13:09:48.0000] <jgraham> tantek: It sounds a lot like it is for people who are becoming invited experts [13:10:00.0000] <import-logic> jmather: lol'd [13:10:08.0000] <jgraham> Anyway if it works for people I don't care [13:10:43.0000] <tantek> jgraham really? the prose seems to make it clear there are many reasons: "Public accounts are necessary for a number of interactions with W3C, including ..." [13:11:11.0000] <tantek> if you'd like to suggest different prose, I'm sure we could ask W3C to improve the content. [13:11:11.0000] <tantek> if you'd like to suggest different prose, I'm sure we could ask W3C to improve the content. [13:11:14.0000] <annevk> afaik they can be created by anyone [13:11:33.0000] <annevk> if people want an account on the WHATWG wiki btw I can create them [13:11:36.0000] <jgraham> I'm reading the bit that says "W3C Public Accounts are for individuals who are not Member employees and who require access to the W3C Web site to register for W3C events and as part of the Invited Expert process." [13:11:38.0000] <tantek> jmather, exactly, perhaps consider the sign-up form to be a light-weight captcha ;) [13:11:44.0000] <annevk> we disabled account creation because of the amount of spam :( [13:11:50.0000] <jmather> jgraham: it seems the CG account and the w3c account are one in the same (or somewhere along the line long ago i signed up with my new password -- which would be really weird.) [13:11:50.0000] <jgraham> Neither of which is the case here [13:11:55.0000] <jreading> marcDrummond & jmather: srcset doesn't do MQ because there is some concern over the "statefulness" of MQ over new spec'd solution. why it is considered a bug and not a feature I'm not sure [13:12:29.0000] <annevk> jreading: MQs don't handle the pixel density case [13:12:30.0000] <jgraham> jmather: If all the people who are interested in contributing have accounts there, that's fine [13:12:36.0000] <tantek> jgraham - that text, e.g. "not Member employees" is nowhere on the page I linked to: http://www.w3.org/Help/Account/Request/Public so I'm not sure where you're getting that from [13:12:37.0000] <tantek> jgraham - that text, e.g. "not Member employees" is nowhere on the page I linked to: http://www.w3.org/Help/Account/Request/Public so I'm not sure where you're getting that from [13:12:41.0000] <jmather> annevk: I thought they did? [13:12:56.0000] <annevk> jreading: they can query it, but you need to set the pixel density [13:13:28.0000] <jgraham> http://www.w3.org/wiki/index.php?title=Special:UserLogin&returnto=Main_Page then "Account Request Form" [13:13:51.0000] <tantek> a-ha! thanks [13:14:24.0000] <tantek> that page, http://www.w3.org/Help/Account/ , is quite confusing :/ [13:14:37.0000] <jmather> jreading: hrm… i haven't seen anything about statefulness… link? [13:15:00.0000] <jmather> annevk: which pixel density case? link? [13:15:33.0000] <annevk> jmather: you cannot do <img src=lala srcset="tralla 2x"> in a way that works properly using the <picture> proposal [13:16:05.0000] <MarcDrummond> annevk: Why not? [13:16:07.0000] <jmather> annevk: device-pixel-ratio: 2 in the media query should do it, no? [13:16:22.0000] <annevk> jmather: how does that downscale "trallla"? [13:16:55.0000] <annevk> "lala" is 10x10; "tralla" is obviously 20x20 [13:17:01.0000] <jmather> the width/height defined to the pixel element forms the size from what I understand (if I get your question right) [13:17:05.0000] <annevk> if you just display tralla it would be way bigger [13:17:23.0000] <ShaneHudson> Is resolution and DPI technically synomynous? [13:17:26.0000] <annevk> jmather: so each source element would always have a width/height attribute? [13:17:29.0000] <annevk> ShaneHudson: sure [13:17:41.0000] <annevk> jmather: because all the <picture> examples got this wrong [13:17:43.0000] <jreading> width/height attrs [13:17:45.0000] <MarcDrummond> annevk: You would provide the 2x version only if 2x is needed and another version for the standard. You can do that with MQs and picture just fine. [13:17:46.0000] <MarcDrummond> annevk: You would provide the 2x version only if 2x is needed and another version for the standard. You can do that with MQs and picture just fine. [13:17:47.0000] <ShaneHudson> annevk: thanks [13:17:51.0000] <annevk> jmather: so nobody seems to be knowing what is going on [13:17:54.0000] <jmather> annevk: from my understanding, sizing would work the same with srcset and picture [13:17:59.0000] <annevk> jmather: and including them all over seems bloat [13:18:07.0000] <jreading> I may need to read through the email threads but, i don't know why the idea of tokenizing the MQ state in a global attr is not being throw around… both <picture> and srcset rely on too much repetition [13:18:14.0000] <scottjehl> img: max-width: 100%; is commonly used in responsive layouts. It'd address that [13:18:18.0000] <annevk> jmather: again, how does <picture> downscale it if you don't set height/width? [13:18:35.0000] <annevk> scottjehl: that doesn't work for icons [13:18:35.0000] <jreading> and why not work to improve MQs instead of creating a whole new spec outside of that... [13:18:44.0000] <scottjehl> icons? [13:18:58.0000] <MarcDrummond> annevk: why prioritize avoiding code bloat over clarity for authors and extensibility? [13:19:12.0000] <annevk> jreading: MQ is about reading device info, it's not about sizing images [13:19:25.0000] <jmather> annevk: I'm not the expert but i'd assume if we knew the ratio was 2 and had no other indicator, the browser could assume to display as 50% to producee a properly dimensioned image [13:19:31.0000] <annevk> MarcDrummond: dunno, you think including height/width attributes all over is better? [13:19:34.0000] <tantek> ok, I'm going to step back from this discussion until ShaneHudson has said he's got a draft of consolidated use-cases on w3.org/wiki/Images [13:19:46.0000] <annevk> jmather: we don't know the ratio [13:19:48.0000] <jmather> annevk: no other width/height would be included than with srcset [13:19:50.0000] <jmather> annevk: no other width/height would be included than with srcset [13:19:53.0000] <jmather> annevk: but we do [13:19:57.0000] <annevk> jmather: how? [13:20:08.0000] <scottjehl> icons are the use case the makes picture impractical? [13:20:11.0000] <jmather> device-pixel-ratio knows the density [13:20:15.0000] <scottjehl> the/that [13:20:19.0000] <jmather> and then downloading the picture gets the dimensions [13:20:20.0000] <annevk> scottjehl: no [13:20:33.0000] <annevk> jmather: what if the ratio of the device is greater than 2? [13:20:44.0000] <annevk> jmather: you need to know the ratio of the image, not the device [13:20:49.0000] <jmather> on a 2x image, 1/2 the width/height would be the natural display size [13:21:01.0000] <annevk> but the MQ applies to the device, not the image... [13:21:01.0000] <annevk> but the MQ applies to the device, not the image... [13:21:02.0000] <jmather> annevkhwo often do you include images with no width/height? [13:21:17.0000] <jmather> I've never done it, at least recently [13:21:22.0000] <annevk> jmather: I do it all the time, but I'm told I'm not a normal author [13:21:28.0000] <jmather> always at least width [13:21:47.0000] <annevk> but you are shifting the argument now [13:21:48.0000] <annevk> but you are shifting the argument now [13:21:49.0000] <jmather> but then my main job is maintaining a CMS, so I have a different angle than a lot of people :) [13:21:53.0000] <annevk> and again [13:21:56.0000] <scottjehl> very common in responsive design to omit width and height attrs [13:22:01.0000] <annevk> all the <picture> examples thus far got this wrong [13:22:05.0000] <jmather> Not trying to shift the argument, just answering your questions [13:22:27.0000] <scottjehl> annevk - I'm unclear what "this" is that you're referring to [13:22:28.0000] <MarcDrummond> annevk: I really don't understand the concern of what to do if device is not 2x. Authors are going to include the standard res version AND a HD version. Or Super HD. Or whatever. I don't understand what the issue is with that? [13:22:28.0000] <annevk> scottjehl: it's not just icons, you wouldn't want to stretch a big photo all the time either, it would not look great [13:22:28.0000] <MarcDrummond> annevk: I really don't understand the concern of what to do if device is not 2x. Authors are going to include the standard res version AND a HD version. Or Super HD. Or whatever. I don't understand what the issue is with that? [13:22:33.0000] <jmather> scottjehl: but even then you would still have width: 100% or whatnot in the css applied to it somewhere, right? [13:22:55.0000] <scottjehl> jmather yes max-width: 100% is the usual fluid images CSS [13:23:00.0000] <jmather> right [13:23:16.0000] <jmather> annevk's argument is what happens when there is /no/ width/height guideline [13:23:18.0000] <tantek> ShaneHudson: re: "everyone is fine with me writing the wiki in British English?" - I think that's fine for the W3C wiki. W3C specs use US English, and as does the microformats wiki. http://microformats.org/wiki/en-us and http://www.w3.org/2001/06/manual/#Spelling [13:23:25.0000] <annevk> MarcDrummond: <img src=x srcset="y 2x"> x is 10x10; y is 20x20 [13:23:32.0000] <annevk> MarcDrummond: tell me how to do that with <picture> [13:23:36.0000] <jreading> <picture> <source media="device-pixel-densiity:2" src="2xbig.jpg" width="50%"> [13:23:52.0000] <jreading> how's that? [13:24:09.0000] <annevk> that's not the same [13:24:14.0000] <jreading> .picture {width:100%} [13:24:18.0000] <scottjehl> no no. width: 100% would stretch an image, sure. But max-width: 100% never goes beyond the dimensions of the image itself [13:24:26.0000] <ShaneHudson> annevk: I am not paying much attention here at the moment, please let me know if you have found something all the other articles (of which I am using to compile the wiki) have gotten wrong [13:24:43.0000] <ShaneHudson> tantek: Thanks, I don't think I am ready to write the actual spec quite yet :p [13:24:48.0000] <annevk> scottjehl: it would be displayed four times the size if you just use max-width [13:24:50.0000] <ShaneHudson> Hard enough writing a wiki haha [13:24:56.0000] <MarcDrummond> annevk: <picture alt=""> <source src="x.jpg" /> <source src="y.jpg" media="min-device-pixel-ratio: 2" /> <img src="x.jpg" /> </picture> [13:25:04.0000] <annevk> MarcDrummond: yeah that fails [13:25:13.0000] <MarcDrummond> annevk: Why? [13:25:20.0000] <annevk> MarcDrummond: if y is selected it would be displayed four times as big as x [13:25:27.0000] <jmather> annevk: why? [13:25:33.0000] <jgraham> FWIW I think this is reenforcing my belief that media queries are only a superficially good match for this use case [13:25:46.0000] <annevk> the reason is this [13:25:46.0000] <MarcDrummond> annevk: I don't see why that would be the case. [13:25:47.0000] <annevk> the reason is this [13:25:47.0000] <MarcDrummond> annevk: I don't see why that would be the case. [13:26:01.0000] <annevk> MQ asks the device if it's pixel ratio is at least 2; device says yes [13:26:06.0000] <scottjehl> annevk: it'll still fit to a parent container's width at most, given that rule. This is trivial to work with. We have tested picture in this way and it worked as expected [13:26:10.0000] <annevk> browser thus selects y [13:26:14.0000] <tantek> ShaneHudson - just figured I'd give you a heads-up and provide citations accordingly ;) [13:26:15.0000] <annevk> browser starts laying out y [13:26:20.0000] <annevk> its 20x20 [13:26:42.0000] <annevk> there's no information about DPI associated with the image thus the browser uses 1 CSS pixel for each image pixel [13:26:57.0000] <MarcDrummond> annevk: That association goes in the CSS. [13:27:02.0000] <annevk> and you get something that's four times as big as x [13:27:06.0000] <jreading> <picture alt="" width="100%">    <source src="x.jpg" />     <source src="y.jpg" media="min-device-pixel-ratio: 2" width="50%" />    <img src="x.jpg" /> </picture> [13:27:12.0000] <annevk> MarcDrummond: are you saying your example was incomplete? [13:27:16.0000] <jreading> pave the freakin' cowpaths [13:27:16.0000] <jmather> annevk: I see where you're coming from, but I don't think it actually works out as a problem in actual practice. [13:27:19.0000] <annevk> jreading: that's not the same as my example [13:27:35.0000] <annevk> jmather: it's a problem in all the <picture> examples [13:27:42.0000] <scottjehl> jmather +1 [13:28:01.0000] <jmather> annevk: in theory, yes, but scottjehl has done quite a bit of testing on this thus far and found it to not be the case [13:28:06.0000] <jmather> logically, I totally see your point [13:28:17.0000] <jmather> but scott says he hasn't seen this particular issue actually occur in the wild [13:28:27.0000] <jgraham> Isn't that going to break if the pixel ratio is not exactly 2 [13:28:30.0000] <jmather> and he's done quite a bit of leg-work regarding picture [13:28:30.0000] <annevk> well he's wrong [13:28:30.0000] <jmather> and he's done quite a bit of leg-work regarding picture [13:29:03.0000] <jmather> annevk: umm, not sure how actual results from testing an implementation directly are … wrong. [13:29:09.0000] <MarcDrummond> annevk: Why do you say he's wrong. Have you tested this? He has. [13:30:03.0000] <jgraham> They can be wrong if they only test a subset of all devices (e.g. only phones avaliable today) and don't account for future devices or other formats [13:30:03.0000] <annevk> I can tell he's wrong because I know how media queries and images work [13:30:12.0000] <TabAtkins> He has seriously attempted to view an image that is authored as, say, 5 inches wide and 192dpi, and seen it lay out as 5 inches on the screen? [13:30:43.0000] <scottjehl> no the testing we did was more... deliver a higher density image to a retina iphone [13:31:06.0000] <TabAtkins> If you're using explicit sizes on the <img>, it'll "work", because the browser is downscaling. [13:31:11.0000] <MarcDrummond> annevk: Inmost cases, a picture is going to be defined with a width that is a certain percentage of its parent. That definition is going to go in the CSS. The picture will fill that width, rather than just simply expanding out to any old size. [13:31:14.0000] <TabAtkins> And in a retina environment, the downscaling will work well. [13:31:17.0000] <jmather> I don't think we have any 3x displays to test on yet annevk [13:31:33.0000] <annevk> jmather: not sure how that's an argument [13:31:39.0000] <MarcDrummond> annevk: Yes, if you don't have a width defined, it's going to go all over the place. But that defeats the entire point of responsive images anyhow. [13:31:41.0000] <annevk> scottjehl: I'd be interested in a pointer to the test [13:32:25.0000] <jmather> annevk: basically just that there's time to solve it [13:32:40.0000] <annevk> jmather: it doesn't work now either [13:32:54.0000] <annevk> jmather: unless you specify width/height all the time which is rather insane [13:32:59.0000] <scottjehl> Example: a picture element sitting inside a 500px wide column. CSS could be... picture { max-width: 100% }. Regardless of the source in play, the image won't expand beyond its container. An HD media query would make a denser image. This is the sort of testing we did. Is this different than what we're talking about? [13:32:59.0000] <annevk> if you ask me anyway [13:32:59.0000] <scottjehl> Example: a picture element sitting inside a 500px wide column. CSS could be... picture { max-width: 100% }. Regardless of the source in play, the image won't expand beyond its container. An HD media query would make a denser image. This is the sort of testing we did. Is this different than what we're talking about? [13:33:04.0000] <jmather> so, specify width/height? :) [13:33:21.0000] <jmather> I dunno. [13:33:35.0000] <jreading> scottjehl: that's what I'm trying to figure out wtf they are saying... [13:33:36.0000] <jmather> What does srcset do with a 2x image on a 3x display? [13:33:54.0000] <TabAtkins> Whatever the UA decides is appropriate. [13:34:31.0000] <scottjehl> did my example make sense? [13:34:48.0000] <TabAtkins> Yeah, that example works. [13:35:01.0000] <annevk> scottjehl: if you have a 500px width and you create a 500 and 1000 wide image then sure it'll work if you define the width somewhere [13:35:03.0000] <scottjehl> that's what I imagine to be the 95% use case [13:35:07.0000] <TabAtkins> If you set a size, the "original" size doesn't matter. [13:35:31.0000] <jmather> TabAtkins: but what is spec'd to happen? [13:35:37.0000] <annevk> it doesn't seem at all clean to me to not have that semantic embedded [13:35:39.0000] <TabAtkins> jmather: Literally what I just said. [13:35:56.0000] <scottjehl> right. this fact can be used to our advantage. How is this a failing of picture? [13:36:08.0000] <jmather> TabAtkins: that sounds like IE all over again. Not trying to start aright, just saying, that sounds highly … open to … alternative implementations. [13:36:22.0000] <annevk> scottjehl: it requires a bunch of extra attributes [13:36:30.0000] <TabAtkins> scottjehl: The only issue is that deciding when to send the "high-dpi version" (that is, the 1000px-wide image) may be best made by data that you don't have easy access to. [13:36:34.0000] <scottjehl> what attributes? [13:36:52.0000] <annevk> width/height [13:37:00.0000] <annevk> if CSS is disabled the image will be way large [13:37:05.0000] <TabAtkins> Here, I just now wrote a blog post about it so I can stop explaining why it's best to do resolution negotation by just telling the browser about it: http://www.xanthir.com/blog/b4Hv0 [13:37:13.0000] <annevk> *otherwise [13:37:17.0000] <scottjehl> tabatkins that's a very different subject, no? [13:37:28.0000] <scottjehl> annevk no width or height attrs are used here [13:37:29.0000] <TabAtkins> scottjehl: Not really, no. [13:37:46.0000] <annevk> scottjehl: exactly, but you need to [13:37:47.0000] <MarcDrummond> annevk: The point of responsive images is to maintain hierarchy of images to content at various container sizes. By default, that means defining the relationship of images to their container. Again, usually as a percentage of the width of the parent. That is done in the CSS for all images (and can be segmented out by classing, etc.). So yes, there will be widths, and this will work. [13:37:55.0000] <TabAtkins> If you're doing *anything* with high-res images, you want the browser to be the one deciding when to request them. [13:37:55.0000] <annevk> scottjehl: because otherwise with CSS disabled it'll turn ugly [13:38:02.0000] <jmather> TabAtkins: your first paragraph after tl;dr is … i believe highly inccorect [13:38:28.0000] <jmather> that's the central starting point for any responsive image implantation, i think [13:38:35.0000] <jmather> that or ipad3… either way [13:38:48.0000] <scottjehl> annevk... disabling css on a retina ipad is the reason picture isn't practical? [13:39:34.0000] <MarcDrummond> What's the percentage of users out there with retina displays that have CSS disabled but images enabled? It has to be pretty darned low. [13:39:46.0000] <jmather> annevk: can you even disable css on an iPad in safari? [13:39:55.0000] <jmather> /me has never even thought to try [13:40:24.0000] <annevk> that was just to illustrate a point [13:40:39.0000] <annevk> that you want the semantic that the image is twice its actual size in the markup [13:40:47.0000] <MarcDrummond> What point? What use case are you highlighting where this wouldn't work? [13:40:55.0000] <annevk> so that if you do anything with that image it's known what is going on [13:41:01.0000] <scottjehl> partially kidding there, but seriously, this doesn't seem like a real problem we're facing [13:41:34.0000] <jmather> annevk: thus why with picture you instruct the element to it's size and then source is used to fill the element [13:42:35.0000] <jmather> annevk: which is why the styling/sizing applies to picture and then whatever source is selected is used to fill it [13:42:36.0000] <TabAtkins> jmather: Interesting. private message me to discuss? [13:43:07.0000] <annevk> jmather: that doesn't work if you want to vary both width and pixel density [13:43:18.0000] <jmather> TabAtkins: sure, but I'm not sure it really merits it… I just think it would be difficult to talk about responsive images without having a retina display on the forefront of the conversation [13:43:38.0000] <annevk> scottjehl: e.g. you hit that problem as soon as you draw the image on a <canvas> [13:43:49.0000] <scottjehl> I could easily update the picture demo to include a high-density source [13:43:50.0000] <annevk> scottjehl: because instead of only taking up 500px it would take up 1000px [13:44:03.0000] <TabAtkins> jmather: I meant relative to your assertion that my tl;dr paragraph is wrong. [13:44:03.0000] <jmather> the iPad/iphone retina displays are what threw so much fuel on the fire of getting something working [13:44:03.0000] <TabAtkins> jmather: I meant relative to your assertion that my tl;dr paragraph is wrong. [13:44:03.0000] <jmather> the iPad/iphone retina displays are what threw so much fuel on the fire of getting something working [13:44:10.0000] <annevk> scottjehl: whereas with srcset we know the image is only 500px because of the 2x indicator [13:44:14.0000] <scottjehl> that also seems fairly avoidable to me [13:44:14.0000] <scottjehl> that also seems fairly avoidable to me [13:44:20.0000] <jmather> TabAtkins: not your tl;dr, first para after. [13:44:44.0000] <TabAtkins> jmather: Oh, that's even more interesting if you think it's wrong. [13:44:54.0000] <annevk> jmather: yeah and only Apple managed to make a proposal that works [13:44:55.0000] <annevk> jmather: yeah and only Apple managed to make a proposal that works [13:44:59.0000] <annevk> thus far anyway [13:45:12.0000] <scottjehl> that's unfair, I think [13:45:14.0000] <krijnh> Quiet day on the internets today, what's happening? [13:45:24.0000] <jmather> TabAtkins: the ascertation that the CG didn't take iPhone display into account is the only thing I have issue with, though perhaps you just got a different take on it than I did. [13:45:53.0000] <jmather> annevk: the only time that comes in to play is when no dimensions of any kind are set on picture [13:46:03.0000] <TabAtkins> jmather: Oh, okay. Well, the proposals that the CG put forward weren't taking resolution into account, afaict. [13:46:08.0000] <jgraham> krijnh: Heh. Going for gold in your own logs? [13:46:19.0000] <annevk> jmather: actually no, that would always occur [13:46:19.0000] <annevk> jmather: actually no, that would always occur [13:46:29.0000] <jmather> i don't know how to get any sort of statistics on that, but i would have to bet it's fairly low [13:46:36.0000] <jreading> my recommendation is toss bandwidth MQs, tokenize MQ and use that in the picture element, add bandwidth to headers. It's not like images are the only bandwidth concern [13:46:38.0000] <annevk> jmather: unless there's a semantic that tells the pixel density [13:47:06.0000] <TabAtkins> jreading: Note my blog post linked above - bandwidth is *not* the only consideration you want to make, and it's not static. [13:48:02.0000] <jmather> annevk: it doesn't need pixel density info [13:48:19.0000] <jmather> it needs to know how many css pixels the image has to fill [13:48:19.0000] <scottjehl> what works seems highly subjective here. [13:48:32.0000] <scottjehl> sorry, gotta drop off again. thanks [13:48:52.0000] <jreading> Tab: seems like bandwidth MQ is the only sticking point for why MQs fail with respimgs, no? [13:48:53.0000] <jmather> TabAtkins: depends on what context you mean [13:49:25.0000] <jmather> resolution of the linked image, no. And that's an interesting piece of meta data that could lead to some fun stuff I think [13:49:57.0000] <jmather> but i don't think it's really required [13:50:04.0000] <ShaneHudson> Would there ever be any need for different file formats to be shown? Thinking if one is better for lower quality while one is better for higher? [13:50:05.0000] <MarcDrummond> annevk: It really seems you are beating an imaginary issue into the ground. [13:50:05.0000] <ShaneHudson> Would there ever be any need for different file formats to be shown? Thinking if one is better for lower quality while one is better for higher? [13:50:05.0000] <MarcDrummond> annevk: It really seems you are beating an imaginary issue into the ground. [13:50:23.0000] <jmather> at least from a content guy's perspective… I have a box, 100x100 in css pixels [13:50:27.0000] <jmather> I want it filled with an image [13:50:44.0000] <jmather> if it's a high resolution screen, i want them to use this other image, because it has 2x the info, and will look sharper [13:51:15.0000] <jmather> ShaneHudson: that's something that really interested me about picture as well [13:51:22.0000] <jmather> not so much in that i think we need it [13:51:29.0000] <jmather> but that it would give room for new file formats to grow [13:51:43.0000] <jmather> like, say, a stereographic image format, perhaps? [13:52:07.0000] <jmather> it's neither here nor there, but I liked that picture opened up the opportunity for someone else to be able to explore that. [13:52:30.0000] <annevk> MarcDrummond: hey man, I'm just telling you what's wrong with <picture> [13:52:39.0000] <annevk> MarcDrummond: at the end of the day, I don't really care what happens here [13:52:48.0000] <MarcDrummond> annevk: To me, the use case of "Author adds retina display image but can't be bothered to define a width" is not a persuasive use case. [13:52:52.0000] <annevk> feel free to ignore me [13:53:23.0000] <annevk> MarcDrummond: even if they add width, it would still blow up on <canvas> [13:53:29.0000] <MarcDrummond> annevk: I do care. As do a lot of other developers/authors. Feel free to pay attention to our concerns! [13:53:30.0000] <annevk> MarcDrummond: as I mentioned before [13:53:30.0000] <jmather> annevk: you're saying one issue that applies in only the most minimal of instances and can be worked around rather trivially, at least as far as I have been able to ascertain. I'm not trying to minimize your argument, I'm just trying to phrase it how it's coming across. [13:53:43.0000] <TabAtkins> MarcDrummond: More important is the use-case "author adds retina display image, but doesn't want to send it to normal-dpi screens, and doesn't want to send it to retina screens on low bandwidth, and..." [13:54:20.0000] <jmather> TabAtkins: picture covers for that well [13:54:31.0000] <annevk> jmather: you're coming accross as someone who doesn't want to hear about problems with <picture> [13:55:06.0000] <TabAtkins> jmather: Based on the last thing I've seen of <picture>, it doesn't, unless you hack in @srcset functionality. [13:55:21.0000] <jmather> annevk: I figured as much. I'm trying not to, but it's hard. I understand the point your'e trying to make, but i don't think it's as big of an issue as you seem to think it is, is all. [13:55:34.0000] <annevk> that ease of authoring is not a serious concern; or drawing images with a higher pixel density on <canvas> is not a concern [13:55:41.0000] <TabAtkins> In which case the difference between "<picture> with <source srcset>" and "<img srcset>" is verbosity and slight differences in how you do fallbacks. [13:55:59.0000] <MarcDrummond> TabAtkins: So, image displays ONLY for retina displays with sufficient bandwidth, otherwise no image at all? That also seems unlikely. [13:56:00.0000] <MarcDrummond> TabAtkins: So, image displays ONLY for retina displays with sufficient bandwidth, otherwise no image at all? That also seems unlikely. [13:56:07.0000] <jmather> TabAtkins: <picture><sourcr src="image⊙2" media="min-device-pixel-ratio: 2"><source src="image.jpg"></picture> I think [13:56:16.0000] <jmather> well, aside from bandwidth, granted [13:56:32.0000] <jmather> but if a bandwidth mq were added, easy enough [13:56:41.0000] <annevk> a bandwidth MQ? haha [13:57:14.0000] <gsnedders> Have we not been over why MQ don't work for bandwidth often enough yet? [13:57:16.0000] <jmather> annevk: I wasn't trying to make a joke… :D I'm not sure why that's so funny though. If the browser can be trusted to figure it out, why couldn't it expose it to an mq? [13:57:28.0000] <MarcDrummond> The reality is that MQs offer a lot more possibilities for addressing issues like bandwidth than the goofy syntax in srcset that doesn't resemble anything else in HTML. [13:57:32.0000] <annevk> jmather: expose it how? [13:57:49.0000] <TabAtkins> jmather: Once again, read my blogpost <http://www.xanthir.com/blog/b4Hv0>. To do that *well*, you need *at least* a bandwidth MQ, and bandwidth MQs have very bad behavior. [13:57:54.0000] <annevk> jmather: did you read Hixie's email explaining the problems with bandwidth and how they're shifting from size to latency etc.? [13:58:39.0000] <annevk> jmather: btw, your media queries thus far miss the required parenthesis and would therefore fail [13:58:43.0000] <gsnedders> FWIW, I don't like the srcset syntax at first glance, but that's a syntatual issue [13:58:44.0000] <gsnedders> FWIW, I don't like the srcset syntax at first glance, but that's a syntatual issue [13:58:56.0000] <jmather> I haven't seen Hixie's email, no. But TabAtkins seems to propose letting the browser make those bandwidth aware decisions, and so if the browser can make a decision, it could be exposed in mq in some form as well [13:59:04.0000] <annevk> jmather: ease of authoring is important ;) [13:59:14.0000] <kevinSuttle> Think about hotel wifi on a laptop vs an iPhone on wifi at home. What should a bandwidth API tell us in that case? [13:59:16.0000] <davatron5000> jmather: +1 [13:59:18.0000] <jmather> annevk: I'm just trying to get the general point across is all and hopefully get an idea of where you guys are coming from with srcset [13:59:38.0000] <annevk> I didn't come up with srcset [13:59:52.0000] <annevk> I don't really care for it either [13:59:59.0000] <jmather> kevinSuttle: I have no idea [14:00:01.0000] <annevk> but given the alternative... [14:00:07.0000] <TabAtkins> jmather: It can't be cleanly exposed in an MQ manner. The way it's going to be handled is with image-set(), which is basically the same as @srcset. [14:00:20.0000] <gsnedders> It's not us-v-them like many are putting forward. Plenty of us have issues with srcset… and bigger issues with the picture proposal. [14:00:21.0000] <jmather> but TabAtkins wanted the browser to account for bandwidth in it's downloading provision [14:00:40.0000] <kevinSuttle> @Jmather: Exactly. Which was part of my point here: http://www.w3.org/community/respimg/2012/05/11/respimg-proposal/#comment-780 We're solving the wrong problem. [14:01:06.0000] <jmather> I'm not trying to make it us-vs-them, just trying to understand srcset as i still don't like it, and was hoping more information could help that. [14:01:24.0000] <zewt> shouldn't you understand it before deciding you don't like it :) [14:01:38.0000] <jmather> zewt: gut instincts are there for a reason :) [14:01:44.0000] <jmather> not to say they can't be wrong [14:02:25.0000] <jreading> so i think the srcset folks agree with me that bandwidth MQ is a bad idea [14:02:38.0000] <jreading> also, seems that's the ONLY sticking point [14:02:42.0000] <jreading> so lose it [14:02:51.0000] <gsnedders> jmather: Someone saying an issue with picture (or bandwidth MQs, etc.) doesn't mean they like srcset is my point [14:03:02.0000] <jmather> gsnedders: true enough [14:03:15.0000] <zewt> (you can do bandwidth without trying to actually strictly define it, eg. as i suggested with the file-size hint) [14:03:26.0000] <kevinSuttle> No one seems to be able to answer why we're only focused on images. Is it because we have indirect control by not having to deal with audio/video codecs? [14:03:27.0000] <kevinSuttle> No one seems to be able to answer why we're only focused on images. Is it because we have indirect control by not having to deal with audio/video codecs? [14:03:29.0000] <annevk> zewt: but not in MQs [14:03:44.0000] <annevk> MQs are about device capabilities [14:03:52.0000] <TabAtkins> jmather: Once you accept that bandwidth MQs are bad, and so you accomodate resolution negotiation some other way, it boils down to a preference for more verbose but possibly more readable (<picture>) versus more compact and typeable (<img srcset>). [14:03:52.0000] <TabAtkins> jmather: Once you accept that bandwidth MQs are bad, and so you accomodate resolution negotiation some other way, it boils down to a preference for more verbose but possibly more readable (<picture>) versus more compact and typeable (<img srcset>). [14:03:55.0000] <jmather> kevinSuttle: I think it's mainly because video/audio is considered "done" with the tags [14:03:55.0000] <annevk> whereas we are concerned here with capabilities from the image [14:03:59.0000] <zewt> annevk: i'm talking about srcset, don't know much of anything about MQ [14:04:01.0000] <annevk> such as its size and aspect ratio [14:04:10.0000] <gsnedders> jmather: Nah, they aren't done. They can always have more done to them, as can img. [14:04:16.0000] <jmather> TabAtkins: at which point with html , verbose and readable usually wins [14:04:21.0000] <zewt> and file size, which is a relative representation of content quality (relative to the other options, at least) [14:04:33.0000] <jreading> and if the concern is the flipping on/off of bandwidth MQ, make them behave differently or lose it [14:04:39.0000] <TabAtkins> jmather: That's an arguably point. ^_^ [14:04:42.0000] <jmather> Unfortunately I have to take off now :) [14:04:44.0000] <annevk> using device queries (=MQs) for resource selection is the wrong solution (for <video> too imo) [14:04:45.0000] <annevk> using device queries (=MQs) for resource selection is the wrong solution (for <video> too imo) [14:04:59.0000] <kevinSuttle> @Jmather: I don't think the media elements are done until they're consistent. Why is it OK to serve one file size of video/audio to any browser, but not images? [14:05:00.0000] <kevinSuttle> @Jmather: I don't think the media elements are done until they're consistent. Why is it OK to serve one file size of video/audio to any browser, but not images? [14:05:05.0000] <jmather> TabAtkins: it does seem to go back and forth, but I do like matt's argument that picture is less error brone [14:05:23.0000] <jmather> kevinSuttle: because html video sucks. :D [14:05:35.0000] <MarcDrummond> The thing that really gets me is that all of the objections with picture seem to revolve around, you didn't document your use cases! You didn't post things in the right places! But srcset comes along, without well-documented use cases, and bam, it's in, because Hixie likes it. And since this is a dictatorship, it feels like we're a mob storming a gate, rather than being able to debate things with some hope of coming to a reason [14:05:37.0000] <annevk> jmather: I haven't seen much people make mistakes with srcset yet, plenty with <picture> though... [14:05:53.0000] <jmather> Not a good answer, but it is my own, hah. I end up using youtube/vimeo for everything video… just avoid that issue altogether. [14:06:03.0000] <jreading> so bandwidth MQ is out and <picture> is back in, right? [14:06:17.0000] <zewt> (nothing screams troll like calling someone a "dictator"; do you even know how standards work?) [14:06:27.0000] <jmather> alright, c-ya guys. Gotta run. [14:06:31.0000] <kevinSuttle> @jmather haha. no one is arguing that. My point is that those media elements are handled by a server that determines quality. Why can't we do the same with images? [14:06:32.0000] <annevk> me too [14:06:42.0000] <TabAtkins> MarcDrummond: I don't think anyone credible has actually made those objections. Use-cases were well-documented, and no one gives a fuck where it was posted. [14:06:59.0000] <TabAtkins> MarcDrummond: HTML is a benevolent dictatorship, but that seems to be a successful model for tech. [14:07:20.0000] <gsnedders> It doesn't matter whether the use-cases are documented in one place or fourty. It'd be nice for the former, but… [14:07:23.0000] <MarcDrummond> zewt: Thank you for your condescension. [14:07:25.0000] <jmather> kevinSuttle: I think it's an accessibility thing… it's harder to expect someone to install/maintain a server for basics like images than it is for video, probably because of the amount of usage of images as opposed to video -- but -- gosh, i really have to run. HAH… [14:07:25.0000] <zewt> TabAtkins: it's not, since implementors effectively have (as a unit, not individually) veto power [14:08:09.0000] <TabAtkins> zewt: Yeah, sure. But most of the time, on most issues, we let Hixie do his thing on the assumption that he makes good choices. [14:08:11.0000] <gsnedders> zewt: Well, as a unit of two or more [14:08:26.0000] <gsnedders> Not necessarily as a unit of all of them. [14:09:00.0000] <zewt> gsnedders: sure--just didn't want to claim that they *individually* have veto power (it just gets muddier when just one vendor balks) [14:09:28.0000] <gsnedders> zewt: Yeah, I was just trying to clarify more than correct *you*. I think you know well enough how things work. [14:09:58.0000] <bjankord> What happens when Hixie makes bad choices? [14:10:14.0000] <zewt> everyone tells him :) [14:10:24.0000] <bjankord> Does he listen? [14:10:33.0000] <bjankord> Does he care? [14:10:41.0000] <TabAtkins> If he doesn't, browsers do what they want anyway. [14:10:56.0000] <TabAtkins> And that means the spec doesn't match brwosers, which lowers the reputation of the spec and it's power in general. [14:11:06.0000] <TabAtkins> So yes, it's ultimately the slave of the browsers. [14:11:07.0000] <Hixie> /me pops his head in and then ducks the incoming tomatoes [14:11:36.0000] <tantek> but at least they're high-resolution tomatoes. [14:11:45.0000] <bjankord> +1 [14:11:47.0000] <Hixie> 'sup people [14:12:00.0000] <Hixie> please direct your ire at me :-) [14:12:02.0000] <adactio> Oh hai. [14:12:05.0000] <Hixie> happy to answer any questions [14:12:14.0000] <zewt> /me BEAM [14:12:15.0000] <davatron5000> Hixie: You have lots of explaining' to do [/rickyricardo] [14:12:38.0000] <bjankord> Hixie: Is there any chance the picture element would be reconsidered [14:12:42.0000] <kevinSuttle> OK guys, gotta run. @Hixie: I'd love your feedback on this comment: http://www.w3.org/community/respimg/2012/05/11/respimg-proposal/#comment-780 Will check in in a bit. [14:12:56.0000] <bjankord> I've heard if it is revised there is a chance that it will be reconsidered [14:13:31.0000] <Hixie> bjankord: if there is new information, absolutely [14:13:41.0000] <Hixie> bjankord: however generally speaking it's better to talk about use cases, not solutions [14:13:53.0000] <Hixie> bjankord: and to point out what is wrong with the existing solutions [14:14:16.0000] <Hixie> bjankord: so e.g. "srcset="" doesn't address use case X" or "srcset="" is overly complicated for use case X" [14:14:23.0000] <necolas> Hixie: by "existing solutions" do you mean @srcset? [14:14:27.0000] <adactio> Apparently, according to zewt, referring to Hixie as a dictator automatically means you're a troll. By which definition, TabAtkins is a troll for describing the WHATWG as a benevolent dictatorship. "My mixed messages: let me show you them." [14:14:32.0000] <Hixie> necolas: now, yes [14:14:36.0000] <zewt> zzz [14:14:36.0000] <zewt> zzz [14:14:52.0000] <TabAtkins> adactio: Welcome to a multitude of opinions. [14:14:55.0000] <jgraham> In other news TabAtkins and zewt are the same person [14:15:07.0000] <zewt> :| [14:15:10.0000] <Hixie> adactio: referring to me as a dictator doesn't automatically mean you're a troll, but it does mean you're talking about process and not the technical stuff, which usually isn't helpful [14:15:15.0000] <TabAtkins> adactio: That's why references to the "cabal" are so misleading. ^_^ [14:15:41.0000] <annevk> there's disagreement in the cabal? [14:15:49.0000] <annevk> quick krijn, disable public logging! [14:15:53.0000] <jgraham> annevk: Not in #secret-treehouse [14:15:55.0000] <Hixie> bjankord: (the point being that agreement that there is a problem to solve is more important than getting agreement on the solution) [14:15:58.0000] <adactio> zewt: Seriously, almost every contribution you've made here has been unconstructive and unhelpful, particularly people new to the process just trying to figure out how things are supposed to work. [14:16:01.0000] <jgraham> Just for show [14:16:05.0000] <annevk> jgraham: ssssh [14:16:13.0000] <bjankord> Hixie: Would examples of use cases of picture element be helpful, or you specifically looking for use cases of srcset? [14:16:14.0000] <tantek> adactio, I thought the accepted term was BDFL per http://en.wikipedia.org/wiki/Benevolent_Dictator_for_Life [14:16:32.0000] <zewt> (I don't feel like dignifying that with a reply, beyond this one) [14:16:48.0000] <Hixie> bjankord: use cases don't mention solutions [14:16:48.0000] <necolas> Hixie: we can consider it progress that it is now agreed that there is a problem to solve [14:16:49.0000] <Hixie> bjankord: use cases don't mention solutions [14:16:49.0000] <necolas> Hixie: we can consider it progress that it is now agreed that there is a problem to solve [14:16:57.0000] <Hixie> bjankord: srcset and <picture> are solutions [14:17:04.0000] <MarcDrummond> Hixie: One of my concerns with srcset is how radically different its syntax is from anything else in HTML. Whereas picture fits the markup pattern established from audio and video. Using similar markup, even if it is slightly more verbose, would help to make this important new feature easier for many to understand and for it to be adopted. [14:17:08.0000] <zewt> bjankord: you're approaching this as "what arguments can I make to convince you to use picture", instead of "what problems do I want to solve that srcset does not" [14:17:36.0000] <Hixie> MarcDrummond: one of the lessons we learnt from <video> is that hte <source> pattern is a bad one, unfortunately [14:18:18.0000] <davatron5000> The problem with the <source> pattern is multiple codecs across various browsers. not @media, imo. [14:18:23.0000] <MarcDrummond> Hixie: Just curious, in what ways is that markup pattern considered bad? I hadn't heard that. [14:18:24.0000] <Hixie> MarcDrummond: it leads to all kinds of problem e.g. with the parser needing to notify the rendering logic so that all the sources can be considered; the problem with orphan sources being grafted in random places; the problem with dealing with inter-element content; the problems with verbosity; etc. [14:18:33.0000] <bjankord> I feel like both can be used to solve the same problem, one is just more verbose then the other. But there is a solution that makes <picture> less verbose then srcset - https://gist.github.com/2702067 [14:18:40.0000] <zewt> if the use cases lead to picture, that's fine, but it's not great when the goal is to use a particular solution and to go looking for arguments for it [14:18:45.0000] <Hixie> MarcDrummond: also we really haven't had much luck with media="" on <Source> so far for <video> [14:18:50.0000] <ShaneHudson> Hixie: Hey Hixie, did you read that we are going to be focusing on defining use-cases to focus on since everybody has their own opinions? I am writing a base for it at the moment at http://www.w3.org/wiki/index.php?title=Images currently trying to compile all the different use-cases from everywhere [14:19:07.0000] <zewt> *blink* [14:19:07.0000] <Hixie> bjankord: you're still talking about solutions not problems :-) [14:19:07.0000] <zewt> *blink* [14:19:07.0000] <Hixie> bjankord: you're still talking about solutions not problems :-) [14:19:26.0000] <MarcDrummond> Hixie: Thanks. Helpful to understand those problems. [14:19:27.0000] <MarcDrummond> Hixie: Thanks. Helpful to understand those problems. [14:19:46.0000] <Hixie> ShaneHudson: i think i listed the use cases that led to srcset="" in my big e-mail, are there others? [14:19:48.0000] <necolas> Hixie: so what is the lesson learned from the <source> problems? that sounds like part of the problem was adding something to the draft without working out the problems it might result in [14:20:09.0000] <Hixie> necolas: numerous lessons, e.g. the ones i just mentioned. [14:20:28.0000] <necolas> sure, but those are lessons about the actual details [14:20:45.0000] <bjankord> Hixie: Thanks for the feedback, good to know what direction to go from here. [14:20:49.0000] <tantek> is there a wiki.whatwg.org equivalent to http://microformats.org/wiki/irc-people where people note their IRC nickname, perhaps link to their website etc.? it would be useful to know who is a browser implementer for example. [14:20:49.0000] <TabAtkins> Do you mean "the lesson is: don't use <source> children"? [14:20:50.0000] <jgraham> necolas: The problem is that oftentimes we don't realise what the problems are until after people have worked hard at making interoperable implementations [14:20:50.0000] <tantek> is there a wiki.whatwg.org equivalent to http://microformats.org/wiki/irc-people where people note their IRC nickname, perhaps link to their website etc.? it would be useful to know who is a browser implementer for example. [14:20:58.0000] <ShaneHudson> Hixie: Well we realised that all the solutions have been focusing on different use-cases [14:21:07.0000] <ShaneHudson> tantek: good idea [14:21:12.0000] <annevk> TabAtkins: it's not equivalent to max-width [14:21:22.0000] <annevk> TabAtkins: because it would be used for a viewport of 700px [14:21:31.0000] <annevk> (re mailing list) [14:21:34.0000] <TabAtkins> annevk: Oh. So it's equivalent to min-width? [14:21:34.0000] <Hixie> TabAtkins: well it's not that simple, i'm not sure how we'd have done otherwise for <video>. but certainly "don't assume it's a good pattern". [14:21:53.0000] <TabAtkins> /me thinks the microsyntax for MQ is a bad idea. [14:21:57.0000] <Hixie> ShaneHudson: in the threads i replied to i'm not sure that was the case, but it's certainly possible. [14:21:58.0000] <Hixie> ShaneHudson: in the threads i replied to i'm not sure that was the case, but it's certainly possible. [14:22:09.0000] <annevk> TabAtkins: it's not a MQ microsyntax though [14:22:11.0000] <Hixie> microsyntax for MQ? [14:22:17.0000] <annevk> TabAtkins: e.g. 2x is not something MQ can do [14:22:18.0000] <Hixie> srcset="" isn't mq [14:22:23.0000] <TabAtkins> Hixie: The "100w 100h" part. [14:22:24.0000] <TabAtkins> Hixie: The "100w 100h" part. [14:22:30.0000] <TabAtkins> annevk: Yes, I'm only talking about the w/h part. [14:22:40.0000] <Hixie> that's just describing the environment for the image, it's not a mq-equivalent [14:22:41.0000] <jgraham> that is a really limited subset of mq [14:22:43.0000] <Hixie> it doesn't evaluate to true or false [14:22:44.0000] <necolas> the fact that we're stuck with <source> suggests that it might have been prematurely added to the draft. stuff like parser & rendering logic, orphan sources, verbosity seem like they wouldn't have needed implementation to be considered potential problems [14:23:20.0000] <Hixie> necolas: everything is "prematurely" added to the standard. we can't learn the lessons until things are implemented, at which point it's too late. [14:23:22.0000] <jgraham> necolas: Video had *lots* of discussion and changes to the design [14:23:22.0000] <TabAtkins> Hixie: I think you're misunderstanding me. That, or you've complicated the w/h thing in a way that's non-obvious beyond what MQ can do. [14:23:23.0000] <necolas> im just interested if that is considered one of the lessons learnt, rather than trying to make judgements [14:23:37.0000] <adactio> Hixie: I'm genuinely confused. You keep saying "provide use cases, not preferred solutions" (which is excellent advice IMHO) but you've gone and put a preferred solution into the HTML: The Living Standard document. Again: mixed messages. [14:23:40.0000] <Hixie> TabAtkins: i don't understand the relevance of mq here [14:23:51.0000] <zewt> jgraham: it's describing the image, and leaving what to do with it to the implementation (I believe), which is a different approach [14:24:00.0000] <annevk> adactio: based on the use cases to date [14:24:11.0000] <jgraham> zewt: Well it's not actually describing the image [14:24:19.0000] <zewt> jgraham: describing attributes about the image [14:24:23.0000] <Hixie> necolas: the lessons learnt are those i listed above, about how the multi-element pattern for url selection has technical implications that are good to avoid if possible [14:24:28.0000] <necolas> jgraham: in those discussions, were concerns about verbosity, orphan <source>, rendering logic discussed? or did it not occur at the time? [14:24:40.0000] <TabAtkins> Hixie: Adding "500w" to a src acts like either min-width or max-width (I'm not sure off the top of my head), throwing away that source if it doesn't pass the test. [14:24:43.0000] <jgraham> It's describing some things about the browser environment that will be used to pick the right image [14:24:44.0000] <TabAtkins> Hixie: Correct? [14:25:09.0000] <Hixie> necolas: the problem with <video> is we don't really have a good alternative for handling it other than <source> (and <track>) [14:25:17.0000] <Hixie> necolas: so it's not that we made a bad decision [14:25:18.0000] <Hixie> necolas: so it's not that we made a bad decision [14:25:31.0000] <Hixie> necolas: it did teach us that the decision is not as obviously good as one would have thought [14:25:40.0000] <Hixie> TabAtkins: not really [14:25:46.0000] <Hixie> TabAtkins: the browser can pick whichever image it wants [14:26:03.0000] <Hixie> TabAtkins: there's a recommended algorithm that picks the image based on some priorities [14:26:09.0000] <necolas> Hixie: i'm not suggesting that. i'm simply curious if the experience also had an impact on the criteria you consider before adding new things (in general) to the draft now [14:26:22.0000] <Hixie> TabAtkins: but even that algorithm doesn't knock things out necessarily [14:26:25.0000] <jgraham> Hixie: BTW "return a random image" in the spec is wrong [14:26:29.0000] <necolas> you live and learn [14:26:36.0000] <Hixie> TabAtkins: e.g. if you only have one image and it has 100w, it doesn't matter if the width is 50 or 200, it'll be used [14:26:50.0000] <jgraham> "return an image according to an algorithm of the UA's choosing" would be right [14:27:04.0000] <Hixie> necolas: yes, all the lessons we learn with everything we do impact how we make future decisions [14:27:14.0000] <hober> jgraham: agreed [14:27:21.0000] <Hixie> jgraham: did i really say "random" in the spec text? [14:28:02.0000] <jgraham> "Optionally, return the URL of a random entry in candidates, and that entry's associated pixel density, and then abort these steps." [14:28:04.0000] <necolas> Hixie: so do you have any concerns that something like @srcset might start to get implemented, and be problematic, before further discussion and exploring of the problem-space can occur? [14:28:08.0000] <Hixie> jgraham: oh man. let me fix that. [14:28:11.0000] <TabAtkins> Hixie: Yeah, the "if nothing matches, choose X" is fine. But aside from that, it acts like a strict filter, right? [14:28:13.0000] <jgraham> Heh [14:28:52.0000] <Hixie> necolas: of course. that happens with everything we do. [14:28:52.0000] <Hixie> necolas: of course. that happens with everything we do. [14:28:57.0000] <jgraham> TabAtkins: http://www.whatwg.org/specs/web-apps/current-work/multipage/embedded-content-1.html#processing-the-image-candidates [14:29:05.0000] <Hixie> necolas: at the end of the day though it's better to have something mediocre than nothing at all. [14:29:23.0000] <ShaneHudson> Right I have got to the point where I will be writing gobblygoop. Have updated http://www.w3.org/wiki/Images but it is certainly not up to Hixie's standard! tantek and anyone else please feel free to add/edit/destroy as you wish. I will carry on with it tomorrow if I am wanted to [14:29:24.0000] <Hixie> TabAtkins: well, modulo the way the UA can do whatever it wants, sure [14:29:27.0000] <necolas> Hixie: in which case, what was compelling enough to add it to the draft before further discussion could take place? [14:29:44.0000] <zewt> TabAtkins: step 16 means "the browser can come up with its own decision/heuristics", from what I understand [14:29:50.0000] <zewt> (the "random" line they're talking about) [14:30:04.0000] <MarcDrummond> Hixie: I think a big concern with srcset is that it locks down the vectors for what is considered to simply width and height and device resolution. Media queries seem to offer more flexibility for the future for other ways to vary image source beyond just width, height and device resolution. [14:30:35.0000] <zewt> TabAtkins: eg. you can make more complex decisions (like "we're a small device, but the viewport is zoomable and we have lots of bandwidth, so let's download the high-res one anyway") [14:30:51.0000] <Hixie> necolas: there was a clear (imho) statement of a problem that needed resolving, there had already been a broad investigation of the solution space, and the discussion was no longer progressing. That's usually the point at which I try to go through all the e-mails and distill the discussion into a decision, which we then see if the browser vendors are ok with implementing. [14:30:52.0000] <Hixie> necolas: there was a clear (imho) statement of a problem that needed resolving, there had already been a broad investigation of the solution space, and the discussion was no longer progressing. That's usually the point at which I try to go through all the e-mails and distill the discussion into a decision, which we then see if the browser vendors are ok with implementing. [14:31:09.0000] <jgraham> MarcDrummond: FWIW I consider that an advantage unless there are other axes along which we have a clear need to vary stuff [14:31:12.0000] <Hixie> MarcDrummond: the format is intentionally extensible [14:31:27.0000] <Hixie> MarcDrummond: but in general unless there's something specific you have in mind, it's good to not be open-ended. [14:31:50.0000] <jgraham> Using a syntax that suggests lots of things *ought* to work but finding out that they don't, or that they have bad perf. characteristics is not good [14:31:57.0000] <annevk> wow http://lists.w3.org/Archives/Public/public-whatwg-archive/2012Feb/0195.html is an instant classic [14:32:03.0000] <zewt> TabAtkins: (oh. it says as much in the note right below it. :) [14:32:26.0000] <Hixie> jgraham: fixed [14:32:48.0000] <annevk> so this whole "create a CG" thing traces back to someone who to my knowledge doesn't contribute a whole lot and a suggestion that person got from dom [14:32:55.0000] <MarcDrummond> Hixie: For example, media queries offer the ability to provide a different version specifically for print. The srcset syntax does not seem as capable of handling such a use case. [14:33:55.0000] <jgraham> Hixie: Thanks [14:34:19.0000] <TabAtkins> Okay, so yes, the "XXXw" syntax is equivalent to a "min-width" MQ, and the same for h and min-height, except that it has an escape clause for when nothing matches. [14:34:32.0000] <zewt> annevk: i'm surprised nobody (if nobody actually did) responded to that "not the place" mail with a correction, since it seems pretty much opposite to the attitude of the list [14:34:36.0000] <Hixie> TabAtkins: certainly they're in a similar space, sure [14:34:53.0000] <annevk> zewt: yeah [14:34:54.0000] <annevk> zewt: yeah [14:35:15.0000] <Hixie> MarcDrummond: i do not believe that use case came up in the whatwg discussions, though i may have missed it. but as it happens, srcset="" does handle that case, i even gave an (indirect) example of it in the spec. [14:35:18.0000] <TabAtkins> zewt: I think I just ignored the email. The responsive images stuff generated a bunch of noise, so I only skimmed things. [14:35:30.0000] <annevk> zewt: doesn't seem like anyone got through in time :( [14:35:35.0000] <Hixie> MarcDrummond: (assuming the only thing you need for print images is an even higher res) [14:35:51.0000] <zewt> (it's also pretty odd for someone apparently new to a list to come on and tell people what they can talk about) [14:35:58.0000] <necolas> Hixie: given that the problem has been discussed and explored for quite some time, i think it could have been wise to invite a few more days/weeks of discussion. it would have allowed a bit more time for the developers with interests to be included in the discussion of the more recent proposals. [14:36:15.0000] <zewt> i wasn't following those threads; too much traffic vs. not enough personal interest in the subject [14:36:20.0000] <annevk> zewt: yep, guess I'll be doing more careful reading of threads I'm not too interested in going forward [14:36:46.0000] <tantek> ShaneHudson, ok, since you said good idea, here's a stub. I added a few folks that I recognized from here in the channel recently and in the Recent Changes on the wiki - please feel free to add yourself(ves) and others: http://wiki.whatwg.org/wiki/Irc-people [14:36:48.0000] <Hixie> necolas: discussion is always welcome, and can continue even with a proposal in the spec. it's not like browser vendors will immediately implement what we put in! [14:36:48.0000] <annevk> at least to make sure nobody is pointing people away from where they should be... [14:36:49.0000] <Hixie> necolas: discussion is always welcome, and can continue even with a proposal in the spec. it's not like browser vendors will immediately implement what we put in! [14:36:49.0000] <annevk> at least to make sure nobody is pointing people away from where they should be... [14:37:10.0000] <Hixie> necolas: if there is information that should lead to a different solution being selected, whether it's given before or after the first draft is specced doesn't matter [14:37:22.0000] <MarcDrummond> Hixie: Another option for the print use case would be providing a grayscale image. I don't think the srcset solution would make it easy to do that. [14:37:35.0000] <Hixie> MarcDrummond: are you aware of anyone trying to do that? [14:37:38.0000] <ShaneHudson> Hixie: Could I have an account on the whatwg wiki please? [14:37:39.0000] <zewt> MarcDrummond: that'd be an easy addition: add a greyscale flag [14:37:44.0000] <Hixie> ShaneHudson: e-mail? [14:37:51.0000] <ShaneHudson> Hixie: shane⊙sn [14:38:10.0000] <TabAtkins> zewt: I don't think that's scalable, honestly. Translating a significant set of MQ into the srcset microsyntax is silly. [14:38:13.0000] <necolas> Hixie: that's true. [14:38:32.0000] <jgraham> TabAtkins: Is there any evidence it is a significant set? [14:38:42.0000] <Hixie> ShaneHudson: account details in the mail [14:38:58.0000] <TabAtkins> jgraham: I dunno. [14:39:08.0000] <ShaneHudson> Hixie: Thank you :) [14:39:08.0000] <ShaneHudson> Hixie: Thank you :) [14:39:20.0000] <Hixie> i didn't see anyone asking for colour. i saw an e-mail mentioning it and all it got iirc was a reply saying that it was not necessary and no disagreement. [14:39:31.0000] <Hixie> hence it not being one of the use cases i considered [14:39:40.0000] <TabAtkins> 'A [14:39:46.0000] <necolas> Hixie: in which case, do you feel that perhaps a better job needs to be done of communicating with the wider developer community about the process etc. [14:39:53.0000] <TabAtkins> Actually, looking over MQ, the only one I might consider useful is the grayscale one. [14:39:55.0000] <MarcDrummond> Hixie: No, I don't know of somebody trying to do that (though there certainly could be interest in that), but it seems that would be easier to handle with media queries than having to create a different flag like grayscale for every use case. What is the issue with media queries? [14:39:58.0000] <TabAtkins> So, shrug. [14:40:19.0000] <zewt> Hixie: fwiw (thinking about the monochrome thing--not proposing it), "file.jpg w:1000 h:800 x:1.5" might be marginally better for future uses, eg. "c:mono" seems better than something like "monoc" [14:40:26.0000] <ShaneHudson> necolas: defintely agree with you there. Until yesterday I had no idea anyone could join whatwg, thought we could only go as far as the working group. [14:40:35.0000] <Hixie> necolas: i am not convinced we could have gotten any more useful input on this. we got literally hundreds of e-mails on it. what input do you think we could have gotten that we didn't get? [14:40:51.0000] <necolas> Hixie: rather than this addition to the draft being viewed positively as evidence that the whatwg agrees that there is a problem worth solving, it has instead been interpreted as further evidence of a disconnect between the interested parties. that's unfortunate [14:41:08.0000] <divya> necolas++ [14:41:30.0000] <annevk> Hixie: maybe we should rename "Contribute" to "Join" on http://www.whatwg.org/ [14:41:31.0000] <TabAtkins> necolas: That's honestly something that could be corrected by you and others. You *know* it's untrue. (If you haven't yet been convinced, then it's hopeless.) [14:41:31.0000] <Hixie> MarcDrummond: we try to design to use cases, not to open-ended problem spaces. If it's not a problem, and the only things that matter are width, height, and pixel density, then it's better to optimise for those and have a terse syntax than to require verbose media queries for every image. [14:41:47.0000] <Hixie> zewt: it could just be "mono" [14:41:47.0000] <annevk> although it does state pretty clearly already everyone can make proposals [14:41:58.0000] <necolas> TabAtkins: again, you seem to think it is *our* job to correct your failings [14:42:04.0000] <divya> exactly [14:42:15.0000] <annevk> but "Join" is nicer and more what we mean than "Contribute" I think; it's a community after all [14:42:15.0000] <Hixie> ShaneHudson: any idea why you didn't think the whatwg was open? i mean, we're pretty radically open and say so everywhere we can... [14:42:17.0000] <necolas> Fortunately, hixie has done a better job of explaining the situation [14:43:29.0000] <jgraham> Pretty sure the only kind of community you have to join is a gated one [14:43:31.0000] <Hixie> necolas: well i am certainly eager to reduce any disconnect, but i don't know how to do it. I don't think two extra weeks of discussion would have made any difference there if the result was the same (which, unless there is new information that hadn't yet been raised, it would have been) [14:43:37.0000] <necolas> As I said for the beginning, IMO what has been going on is more an issue of miscommunication. But that doesn't mean it should just be dismissed and not considered a problem that the whatwg might be wise to want to remedy. [14:43:40.0000] <TabAtkins> necolas: Dunno why you think it's productive to point fingers. There was a communication mismatch, largely caused by Hixie operating as usual, but with a lot of people new to the process having some (unknown) expectation of how it should work that wasn't met. [14:43:42.0000] <annevk> jgraham: so leave Contribute? :) [14:43:44.0000] <TabAtkins> Shrug. [14:43:59.0000] <Hixie> annevk: done [14:44:12.0000] <jgraham> annevk: I think so :) [14:44:21.0000] <necolas> TabAtkins: to be fair, you were the one who started suggesting that the problem lay with *us* and "egos" etc. [14:44:27.0000] <ShaneHudson> Hixie: well I wanted to get involed with the spec etc about a year ago.. I went onto the w3c and everything I saw was about paying to be a business member etc. So I avoided for a few months. Eric Meyer and I spoke on twitter about it, and I then emailed the w3c support team, they told me that although you have to pay to be a member of the w3c there is the community groups. It was not until yesterday that I realised the working group was also free to c [14:44:30.0000] <TabAtkins> Hixie's not the best communicator, but he's not a troll either. ^_^ If people new to the process are misinterpreting, and you know who they are and already have a connection to them, feel free to correct them! [14:44:52.0000] <Hixie> jgraham: we tried contribute for a while, let's try join for a while and see if it helps :-) [14:45:04.0000] <necolas> I just felt that it wasnt a good situation for anyone to not have developers feel invested in these processes in some form. I didn't expect to be shooed away for saying so. [14:45:11.0000] <zewt> when i decided to look at specs, i just subscribed to webapps and whatwg and started reading, heh [14:45:13.0000] <TabAtkins> necolas: I was trying my best to understand what you were saying. ^_^ If still seems to boil down to "the words he said weren't what I was expecting, even though the content is more-or-less okay". [14:45:18.0000] <Hixie> ShaneHudson: ah, yeah, if you approached the w3c then i could see why you'd get that impression [14:45:19.0000] <zewt> (couldn't say how I stumbled across the lists in the first place) [14:45:24.0000] <Hixie> ShaneHudson: unfortunately we cannot control the w3c's messaging [14:46:00.0000] <Hixie> necolas: i'm eager to change things to be more welcoming for developers and users and anyone else who isn't currently contributing [14:46:07.0000] <necolas> TabAtkins: well people are entitled to disagree over proposals. any frustration centred around that is distinct from those of people feeling disconnected. [14:46:08.0000] <Hixie> necolas: if you have any ideas i'm all ears [14:46:15.0000] <Hixie> necolas: but i don't think just talking more would have achieved that [14:46:31.0000] <adactio> TabAtkins: If you're still trying to understand how this all appears to people outside the WHATWG, Wilto has done a good job of summarising here: http://www.alistapart.com/articles/responsive-images-and-web-standards-at-the-turning-point/ [14:46:33.0000] <zewt> if there's anything whatwg doesn't lack, it's talking more :) [14:46:33.0000] <ShaneHudson> Hixie: I think maybe a ALA article or something that everybody reads, explaining how we can get involved. Perhaps there is already and I just hadn't seen anything [14:46:35.0000] <Hixie> necolas: since we had already been talking about this for months at least (the earliest mail in the thread i just replied to with that big mail was from january) [14:46:38.0000] <divya> Hixie: completely agreed. [14:46:44.0000] <necolas> TabAtkins: and fwiw, no hard feelings. i know you have the right intentions [14:47:26.0000] <divya> Hixie: at this point my only concern is as a dev i am having a hard time understanding how learning yet another syntax would make srcset adoption better or less painful. [14:47:30.0000] <necolas> Hixie: i think talking more would actually help. but just as you expect the talking directed to the spec-work to be via the mailing list, developers expect the communication directed at them to be via channels that *they* use or can easily consume [14:48:03.0000] <TabAtkins> adactio: I know how people can misinterpret things here. I've been subject to it myself before. :/ Doesn't mean that anything's wrong, just that people can have wrong expectations sometimes, and we can't please everyone. [14:48:22.0000] <TabAtkins> divya: Heh, every single CSS property in existence is a brand new microsyntax to learn. ^_^ [14:49:03.0000] <Hixie> ShaneHudson: good idea. any idea how one of us can go about writing an article for ALA? [14:49:12.0000] <adactio> I share divya's concern. I'm not wedded to picture or srcset but I do think that "Avoid needless complexity" is a design principle that we should remember in this (and every other) case. [14:49:36.0000] <annevk> isn't sharing the syntax with image-set therefore good? [14:49:39.0000] <adactio> Hixie: write the article. Send it to ALA. I can put you in touch if you want. [14:50:03.0000] <divya> TabAtkins: at the minimum they have adaptability, or some level of consistency with other syntaxes [14:50:05.0000] <TabAtkins> annevk: It's the "NNNw NNNh" part that's new. [14:50:13.0000] <TabAtkins> annevk: And not shared with image-set(). [14:50:16.0000] <annevk> also, given how much markup <picture> requires; avoid needless complexity seems considered... [14:50:19.0000] <TabAtkins> annevk: And duplicated from MQ. ^_^ [14:50:22.0000] <Hixie> divya: well any solution in this space would involve new syntax of some sort. new elements, new attribute, something has to be new. [14:50:31.0000] <adactio> annevk: Yes, it is. It's one of the things in srcset's favour. [14:50:50.0000] <divya> Hixie: yeah i feel if it is a pattern devs can relate with. [14:51:00.0000] <divya> it would be less difficult to make mistakes with. [14:51:01.0000] <divya> it would be less difficult to make mistakes with. [14:51:05.0000] <annevk> TabAtkins: someone proposed using 10x10 instead like rel=icon; but often you don't need one or the other [14:51:16.0000] <divya> annevk: i would prefer verbose/readable to concise machine language. [14:51:16.0000] <Hixie> necolas: that's reasonable. What channels should we be using? We have forums, a blog, a wiki, a Google+ page, and a FAQ. What else should we be doing? [14:51:18.0000] <TabAtkins> annevk: I agree with that criticism. [14:51:27.0000] <miketaylr> Hixie: Wilto is the "technical curator" of ALA. Publishing wouldn't be hard. [14:51:44.0000] <Hixie> adactio: how do we "send it to ALA"? [14:51:57.0000] <TabAtkins> annevk: My preference is something like "w:200px" like Zewt suggested, or full-bore embedding of some MQs like "min-width:200px". [14:52:17.0000] <ShaneHudson> miketaylr: So THAT is how he managed to get an article out today so quickly! [14:52:25.0000] <annevk> TabAtkins: why use units? [14:52:32.0000] <annevk> TabAtkins: nothing in HTML has units [14:52:33.0000] <miketaylr> ShaneHudson: heh [14:52:37.0000] <TabAtkins> annevk: Because people are used to it. And by people, I mean "me". [14:52:41.0000] <Hixie> TabAtkins: w:200px is still new syntax, and if we're going to have new syntax, 200w seems better. [14:52:48.0000] <divya> annevk: this discussion seems to mirror the discussion around open font ligatures [14:53:00.0000] <divya> from what i gather the movement there has been from terse opentype syntax [14:53:04.0000] <TabAtkins> The big problem I have is that sometimes I want to use min-width, and sometimes it's clearer to say max-width. [14:53:05.0000] <divya> to a longer descriptive one [14:53:12.0000] <divya> but then again, i maybe wrong. [14:53:14.0000] <adactio> Hixie: I'll drop you a line in an email and introduce you to some of the ALA people. [14:53:15.0000] <annevk> divya: don't think I was involved in that discussion [14:53:15.0000] <ShaneHudson> Hixie: submit⊙ac is their official submit email address, but there are quite a few people that I am sure could speed it up [14:53:22.0000] <annevk> divya: or know much about it [14:53:24.0000] <zewt> 200w vs w:200 seems pretty much dead-on even to me (intuitively) as far as readability/typability/etc [14:53:30.0000] <necolas> Hixie: The G+ page, for example, hasn't been updated for over 3 weeks. There is no mention of any of the discussions or plans surrounding "responsive images" [14:53:31.0000] <necolas> Hixie: The G+ page, for example, hasn't been updated for over 3 weeks. There is no mention of any of the discussions or plans surrounding "responsive images" [14:54:01.0000] <annevk> necolas: you want to help maintain it? [14:54:16.0000] <necolas> If you have the channels, then it seems like the information isn't being fed into them before (or even after, in this case) the fact. [14:54:39.0000] <Hixie> necolas: if you'd like to help that would be awesome :-) the problem is i don't know what to mention -- there's over 1000 e-mails in the queue for me to deal with, i can't very well mention each one of them on g+ :-) [14:55:04.0000] <zewt> being able to parse all entries as "%c:something" and tell what it is based soley on the %c seems a plus--but a minor one (it's not likely to end up with tons of possible values), so I don't feel strongly about it [14:55:05.0000] <Hixie> necolas: i don't know what is an issue developers want to be told about and what isn't [14:55:17.0000] <dreadnaut> With css using units after the number, 200w would look a bit strange. [14:55:20.0000] <divya> Hixie: if you could like broadly group what you are looking to solve currently. [14:55:24.0000] <divya> and mention it publicly. [14:55:26.0000] <divya> it would be useful. [14:55:36.0000] <TabAtkins> dreadnaut: Spot on. That's my objection, I think. [14:55:38.0000] <divya> we can simply subscribe to threads within that set that is useful. [14:55:50.0000] <ShaneHudson> Hixie: If there is anything I can do to help, let me know.. I have been wanting to get involved for ages as I said earlier [14:55:57.0000] <Hixie> divya: i don't know what i'm going to be working on. I go through e-mail more or less on a first-come-first-served basis, so it's all over the place. [14:56:00.0000] <necolas> annevk Hixie: i'd be tentatively willing to be involved with others in helping out. i've got a full time job + open source commitments also consume most of my time. [14:56:06.0000] <Hixie> divya: http://whatwg.org/issues/ lists all the e-mails [14:56:11.0000] <Hixie> divya: "broadly categorised" [14:56:23.0000] <zewt> if any variable string values are ever added then x: is better, but I can't think of any (even theoretically) [14:56:36.0000] <annevk> ShaneHudson: necolas: we've had this "WHATWG Weekly" running on http://blog.whatwg.org/ but I've been getting worse at maintaining it [14:56:45.0000] <divya> perhaps what we need then Hixie is someone to summarize your current week's responses. :P [14:57:04.0000] <divya> its right now too much effort to follow everything you respond to. [14:57:05.0000] <divya> its right now too much effort to follow everything you respond to. [14:57:10.0000] <annevk> ShaneHudson: necolas: if you think having blog posts of standards activity will help and you can help out with that I can set you up [14:57:11.0000] <MarcDrummond> Hixie: One concern I have with the srcset syntax is that the breakpoints are defined in pixels. However, for a responsive design, it's better to set breakpoints in ems, so that the layout can shift if the text is resized. This could lead to a disconnect between the breakpoints in the layout and for the images. [14:57:12.0000] <zewt> sounds like a good way to confuse things :) [14:57:21.0000] <Hixie> divya: anne was doing that with whatwg weekly but yes, it'd be awesome to have people help him with that [14:57:30.0000] <Hixie> divya: shelley helped for a while [14:57:34.0000] <Hixie> which was cool [14:57:37.0000] <necolas> annevk Hixie, you know, it might be worth also looping some people in who are already trying to do this stuff: bruce lawson, the html5doctor people. [14:57:37.0000] <divya> yeah :) [14:57:37.0000] <necolas> annevk Hixie, you know, it might be worth also looping some people in who are already trying to do this stuff: bruce lawson, the html5doctor people. [14:57:38.0000] <divya> yeah :) [14:58:53.0000] <ShaneHudson> annevk: An 0 comments on most of the articles.. that blog needs promoting! I am surprised actually that with the amount of high profile people I have seen in the irc today that promotion has not been easy! [14:59:39.0000] <Hixie> necolas: bruce is around sometimes [15:00:12.0000] <Hixie> MarcDrummond: (i'm not familiar with the term "breakpoint" as you use it... do you have a link to a page explaining that by any chance?) [15:00:47.0000] <annevk> ShaneHudson: it has about 7000 people subscribed to it via Google Reader [15:00:53.0000] <Hixie> MarcDrummond: in practice 1rem = 16px and the font size changes are all known to the author so i'm not sure it really matters if it's pixels or ems [15:01:02.0000] <TabAtkins> Hixie: A "breakpoint" is a point where you suddenly shift your layout. [15:01:11.0000] <Hixie> ah [15:01:12.0000] <TabAtkins> In other words, where your current layout breaks and you need to rearrange. [15:01:12.0000] <Hixie> ah [15:01:33.0000] <divya> Hixie: i think what we need is what you are currently responding to :P basically a hixie-response-summary-bot [15:01:40.0000] <annevk> necolas: I talk with Bruce now and then; he's not so close to the activities though that he could summarize what is going on [15:01:41.0000] <jreading> OMG [15:01:53.0000] <ShaneHudson> annevk: Really? Ok I reckon I could help out with the blog... though I am new to this group [15:02:06.0000] <jreading> 'nuff…. [15:02:36.0000] <necolas> Hixie and annevk, what is a good avenue to continue the discussion and bring in others who are already partially involved in commenting on whatwg happenings? this feels like something that would be worth attempting. [15:02:37.0000] <Hixie> annevk: (btw, the width and height in srcset are viewport width/height, not image width/height. still about the images, but not necessarily the image dimensions per se.) [15:02:43.0000] <necolas> mailing list? :P [15:02:52.0000] <gsnedders> necolas: Yeah, the mailing list is the place [15:02:56.0000] <adactio> Hixie: you've got mail (some intros to ALA folk). [15:03:00.0000] <Hixie> adactio: thanks [15:03:15.0000] <zewt> Hixie: that's on the unintuitive side [15:03:34.0000] <Hixie> zewt: yeah, unfortunately we're kinda stuck on that. see the e-mail. [15:04:35.0000] <zewt> "viewport-width: ..." would be clearer (but not crazy about the verboseness) [15:04:50.0000] <Hixie> yeah [15:05:04.0000] <annevk> CSS has vw / vh but they mean something else [15:05:30.0000] <MarcDrummond> Hixie: TabAtkins summarized breakpoints well. file.jpg 600w 200h would have breakpoints at 600 pixels wide and at 200 pixels high. The issue is that I would define the layout shift at 37.5em. If the user increases their text size, then the layout shift and the file shift might not sync up. [15:05:34.0000] <TabAtkins> Unfortunately "max-width" is both a MQ and a property. [15:05:48.0000] <zcorpan> annevk: it's not describing the image [15:05:59.0000] <annevk> yeah I was just told [15:06:05.0000] <annevk> silly me [15:06:12.0000] <Hixie> MarcDrummond: if the user increases their font size, they also increase their pixel density, but they don't change the number of pixels [15:06:16.0000] <zcorpan> ah. i failed at reading the logs then :-) [15:06:17.0000] <MarcDrummond> Hixie: This could be really important because I might want images to fill 100% of the width of their container below 37.5 em and only 40% of the container above 37.5em. [15:06:48.0000] <Hixie> MarcDrummond: (don't change the pixel font size, i should say) [15:06:55.0000] <Hixie> MarcDrummond: (assuming they change font size using regular browser zoom) [15:07:13.0000] <TabAtkins> Hixie: Problem still occurs when they have a minimum font-size. [15:07:25.0000] <MarcDrummond> Hixie: Yes, but the issue is if they change the actual text size rather than using browser zoom. [15:07:51.0000] <Hixie> TabAtkins: does that affect teh computed font size? [15:07:56.0000] <Hixie> TabAtkins: i guess in certain cases it does [15:08:15.0000] <TabAtkins> Hixie: Yeah, that's the whole point. [15:08:20.0000] <Hixie> MarcDrummond: i don't expect browsers going forward to really support changing the font size. too many sites break if 1rem != 16px. [15:08:36.0000] <TabAtkins> I've considered enforcing a 16px minimum font size before, but I don't think Chrome lets me do that. [15:08:50.0000] <Hixie> TabAtkins: web breaks if you do that, i have in the past tried [15:08:56.0000] <zcorpan> can the syntax for 200w be more obvious about what it means, maybe? [15:08:56.0000] <zcorpan> can the syntax for 200w be more obvious about what it means, maybe? [15:09:05.0000] <MarcDrummond> Hixie: Lots and lots of browsers do support changing the actual font size right now. I haven't seen evidence that ability will be eliminated. [15:09:17.0000] <zewt> TabAtkins: pretty sure they finally added the option for that (but chrome is being stupid and timing out loading the ... options screen? so I can't check) [15:09:33.0000] <MarcDrummond> Hixie: And if there was, there would be a huge outcry from the accessibility community. The primary purpose of changing font size is for low vision users. [15:10:02.0000] <TabAtkins> zewt: Aww yeah! [15:10:12.0000] <TabAtkins> Everything's big now, woo! [15:10:19.0000] <Hixie> zcorpan: suggestions always welcome, mail the list :-) [15:10:27.0000] <zewt> yeah it's sort of hidden but it's there [15:10:38.0000] <MarcDrummond> Hixie: That's why the best practice is to use ems to determine break points rather than pixels. That allows your layout to still work if the base font size is not equal to 16 px. [15:10:45.0000] <TabAtkins> I was pretty sure that pkasting was fighting against it at some point, but whatever. [15:11:02.0000] <zewt> minimum font size is a pretty critical feature, in a world where people with better-than-20/20 vision design websites with no understanding of people with less-than-20/20 [15:11:06.0000] <Hixie> MarcDrummond: the primary purpose of zooming is for low vision users, too, but without the disadvantages of text size changing. [15:11:19.0000] <Hixie> MarcDrummond: i would expect browsers going forward to more and more drop their font size settings ui [15:11:37.0000] <zewt> (i don't want to zoom in and make *everything* needlessly big; i just want excessively small text fixed) [15:11:42.0000] <zcorpan> zewt: better-than-20/20? [15:12:04.0000] <zewt> yes, 20/20 isn't a maximum [15:12:21.0000] <zewt> don't recall how it works offhand (not that it's really the point :) [15:12:24.0000] <MarcDrummond> Hixie: Zooming works well too. But the reality is that people *can* change font sizes right now. And probably next month. And probably the next month after that. [15:12:35.0000] <zcorpan> /me looks it up [15:13:04.0000] <MarcDrummond> Hixie: This isn't an insurmountable problem. I'm just saying that if srcset is used as a syntax, it would be good to be able to define the breakpoints in units beyond pixels. [15:13:36.0000] <zcorpan> so 20/10 means that you have supervision, and 20/100 means that you have crappy vision, and 20/20 is "normal" [15:13:49.0000] <zewt> cool, we have a Mr. Development posting on the list [15:13:58.0000] <TabAtkins> zewt: A/B means that, at A feet, you see things about as well as a "normal" person does a B feet. [15:14:03.0000] <ShaneHudson> Right, added myself to the irc users wiki page :) It is times like this that I realise how boring I am! No books, no active blogs, no awards and not even working for a browser :[ [15:14:06.0000] <ShaneHudson> * :p [15:14:21.0000] <zcorpan> at least when looking at something 20 feet away [15:14:25.0000] <zewt> at least he's not posting as a company name; that's one of the more annoying things on lists [15:14:37.0000] <Hixie> MarcDrummond: in general we're designing for 10 years from now, so what they'll be doing in a few months isn't a huge concern. i am pretty sure that font size setting ui is headed out, and that it's ok to treat 1rem=16px. [15:14:38.0000] <zewt> (that is, as a company name with no real name) [15:14:52.0000] <zcorpan> usually people's screens are closer than that, so 20/20 is not really relevant [15:14:55.0000] <TabAtkins> Hixie: I'm curious why you think minimum font-szies are heading out. [15:15:06.0000] <TabAtkins> Do you think people are going to inevitably get better vision? [15:15:07.0000] <zewt> Hixie: i don't have any opinion on whatever he's arguing, but I'd object ferociously if people removed minimum font sizes [15:15:19.0000] <Hixie> TabAtkins: minimum font sizes may not be but they're not really where you see this kind of image [15:15:48.0000] <Hixie> i'm talking about the definition of rem, not the definition of 0.1rem [15:16:58.0000] <MarcDrummond> Hixie: There's a lot of websites that have been developed between 2002 and 2012. There will be even more between 2012 and 2022. The markup pattern for srcset *will break* how images work within responsive design until font size uis are removed in 2022. That is a problem. [15:17:07.0000] <TabAtkins> Hixie: The point is that if a webdev, say, sets "html { font-size: 10px; } body { font-size: 1.6rem; }" (which I've seen in tutorials), it may break the page. [15:18:09.0000] <Hixie> TabAtkins: that page will be broken long before srcset="" gets involved, though, no? [15:18:29.0000] <MarcDrummond> Hixie: To be clear, it's not that the font size will be different, and that will screw up the image. The problem is that the layout is based on the font size. So if the font size changes, the entire layout can change, the widths of containers can change. [15:18:55.0000] <TabAtkins> Hixie: Maybe, but maybe not. If the author is being very consistent about sizing in rems, though, the page will at least scale up uniformly. It would suck if the image breakpoints weren't allowed to do the same. [15:19:06.0000] <Hixie> MarcDrummond: maybe i'm not understanding what you mean. can you show me a realistic example of where one would use srcset that would be broken if someone set a minimum font size? [15:19:10.0000] <TabAtkins> MarcDrummond: Yeah, that's right. [15:19:29.0000] <MarcDrummond> MarcDrummond: This isn't an isolated use case either. The best practice for responsive design is to change layouts based on ems, not px. And responsive design is swiftly becoming one of the primary techniques for designing new sites. [15:19:48.0000] <Hixie> TabAtkins: well i don't disagree in principle, but there's already a solution: don't set a 10px font size if you want a 16px font size. [15:19:57.0000] <TabAtkins> Hixie: You use a breakpoint at a particular spot, set in ems or rems or something, that changes your layout via MQ. You set up your @srcset to match it, translating into px based on your assumptions about the user's font size. [15:20:09.0000] <Hixie> TabAtkins: no need to massively increase the complexity of srcset="" to handle something that can already be addressed, imho [15:20:17.0000] <TabAtkins> The user's minimum font-size kicks in, suddenly your conversions to pixels are inaccurate, and your images swap at weird points. [15:20:18.0000] <TabAtkins> The user's minimum font-size kicks in, suddenly your conversions to pixels are inaccurate, and your images swap at weird points. [15:22:01.0000] <zcorpan> TabAtkins: why would the conversion be inaccurate? surely the minimum font-size only affects actual text, not computed values of other things like width:10rem ? [15:23:14.0000] <MarcDrummond> Hixie: Here's a simple test page. http://www.vistasmith.com/responsive-ems/responsive-ems.html [15:23:50.0000] <TabAtkins> zcorpan: 10rem is the font-size of the root element. If the minimum font-size causes the font-size of the root to go up, well... [15:23:51.0000] <Hixie> zcorpan: no, minimum font-size affects the computed value for 'em' [15:24:03.0000] <Hixie> zcorpan: see the bottom of http://hixie.ch/specs/css/font-size-ui/font-size-ui [15:24:04.0000] <Hixie> zcorpan: see the bottom of http://hixie.ch/specs/css/font-size-ui/font-size-ui [15:24:19.0000] <Hixie> (it doesn't affect inheritance though) [15:24:43.0000] <zcorpan> hmm. ok. [15:24:58.0000] <Hixie> MarcDrummond: i see no images there [15:25:07.0000] <TabAtkins> Yeah, it's just literally "if an element's font-size is below the minimum, change it to the minimum". [15:25:12.0000] <MarcDrummond> Right. It's just a test page I already have. [15:25:42.0000] <MarcDrummond> Hixie: I'll throw some in there quick. [15:25:49.0000] <Hixie> MarcDrummond: (fundamentally, the problem is that we can't use 'em's because 'em's rely on layout and we need to be able to pick the image before layout happens, btw. so unfortunately even if you convince me that it would be nice to do this, i think we're stuck here.) [15:25:59.0000] <Hixie> MarcDrummond: (but i would still like to understand the problem.) [15:26:20.0000] <MarcDrummond> So the short summary is this. [15:27:34.0000] <MarcDrummond> With that layout, there's two columns above 30ems up to 880px. At 1em=16px, that's 480px to 880px. [15:27:49.0000] <zcorpan> TabAtkins: so if you have something like <h1 style="width:30rem">Hello!</h1>, that will be way too wide if the user increases the minimum font-size [15:28:26.0000] <ShaneHudson> After seeing what it has been like with responsive images... I am really glad I was not involved on the spec side of the original browser wars! [15:28:30.0000] <MarcDrummond> However, if the base font size is 1em = 20px, then there will be two columns between 600px and 1100px. [15:29:08.0000] <zcorpan> but maybe it makes sense for rem-based layouts to get bigger when minimum font-size is increased [15:29:09.0000] <necolas> adactio Hixie: if an ALA article happens, that might be a good time both to highlight whatwg communication channels with the wider dev community, and ask for people to help filter the key mailing list content for summary on the whatwg blog etc. [15:29:33.0000] <adactio> necolas: Agreed. [15:30:21.0000] <akamike> necolas: I've sent a call to action to the other H5Ds, so we might be able to help [15:30:26.0000] <Hixie> zcorpan: it's only a problem if the font-size of the root element is smaller than the minimum, which it shouldn't be unless the minimum is really high, in which case imho the user should just be using zoom (and the UA UI imho should just be taking care of that detail automatically; see the top of the aforementioned page) [15:30:31.0000] <TabAtkins> zcorpan: Yes. [15:30:35.0000] <MarcDrummond> That filters down to the column widths. So if you have an image set to the full width of the second column, that's going to be a different number of pixels when the font size increases. [15:30:36.0000] <necolas> akamike: nice! [15:30:40.0000] <ShaneHudson> akamike: What is a H5D? [15:30:54.0000] <Hixie> MarcDrummond: sure but when is the user going to set a minimum font size of 20px?? the entire web would fall apart. [15:31:08.0000] <akamike> ShaneHudson: http://html5doctor.com/ [15:31:25.0000] <ShaneHudson> akamike: Oh! Of course, sorry and thanks :) [15:31:26.0000] <MarcDrummond> Hixie: The point of defining the layout in ems rather than px is so that the web does not fall apart. [15:31:48.0000] <Hixie> MarcDrummond: i mean the existing web [15:32:23.0000] <ShaneHudson> What is the opinion in here of rem instead of em? [15:32:27.0000] <Hixie> MarcDrummond: not future web pages written by perfect web authors, the web pages that already exist that implicitly assume that when they say font-size: 16px it won't turn into 20px [15:33:23.0000] <MarcDrummond> Hixie: So the problem is that at 16px, the columns shift at480px, but at 20px, they shift at 600px. However, if srcset has the image coded to shift at 480px, then at 20px, the wrong image will be delivered. [15:33:41.0000] <MarcDrummond> Hixie: Unfortunately, that assumption is incorrect. [15:34:08.0000] <MarcDrummond> Hixie: Lots and lots of web developers design their sites assuming that users *can* change their font sizes. [15:34:09.0000] <MarcDrummond> Hixie: Lots and lots of web developers design their sites assuming that users *can* change their font sizes. [15:34:27.0000] <MarcDrummond> Hixie: That's the best practice that's being taught with responsive design right now. [15:34:54.0000] <Wilto> I use em-based media queries, yeah. [15:35:07.0000] <MarcDrummond> Hixie: So if the srcset breakpoints are hard-coded as px, incorrect images will be delivered. [15:35:18.0000] <Wilto> CloudFour has a great write-up on them; it’s really impressive stuff in practice. http://blog.cloudfour.com/the-ems-have-it-proportional-media-queries-ftw/ [15:35:37.0000] <zcorpan> so for em- or rem-based layouts, it seems it would be useful to be able to use em/rem in srcset as well to describe the viewport [15:35:51.0000] <Hixie> MarcDrummond: lots and lots of web developers design their sites assuming that users *can* change their font sizes, but they are utterly swamped by the number of web developers who do not. [15:35:53.0000] <MarcDrummond> zcorpan: Exactly. [15:35:54.0000] <MarcDrummond> zcorpan: Exactly. [15:36:26.0000] <TabAtkins> Hixie: It kind of sucks to tell webdevs that are doing responsible things "sucks to be you, maybe you should do what all the sucky devs are doing instead". [15:36:38.0000] <MarcDrummond> Hixie: That doesn't mean that people who are following best practices should be unable to deliver a solution for their users. [15:36:44.0000] <zcorpan> Hixie: we don't need to cripple a new feature just because most authors don't use rem-based layouts :-P [15:36:54.0000] <Hixie> TabAtkins: agreed, but i don't know what else we can do. [15:36:55.0000] <ShaneHudson> Hixie: But that does not mean the ones that do not are right for not doing it [15:36:57.0000] <Wilto> Completely agreed, on all counts. [15:37:07.0000] <TabAtkins> Hixie: Um, use zewt's idea? [15:37:11.0000] <TabAtkins> w:200px [15:37:18.0000] <TabAtkins> Or w:40em [15:37:20.0000] <TabAtkins> whatever. [15:37:20.0000] <Wilto> I mean, I have an idea, but… y’know. [15:37:26.0000] <MarcDrummond> Hixie: I'm not trying to criticize you here. I'm just saying that this is a very important use case, and that incorporating the usage of ems into the syntax of srcset is essential. [15:37:34.0000] <Hixie> MarcDrummond: my point is that people who are following best practices _can_ deliver a solution for their users anyway, because users aren't going to pick a minimum font size of great than 16px. [15:37:39.0000] <Hixie> MarcDrummond: they'll just zoom instead [15:37:45.0000] <Hixie> MarcDrummond: and zoom works fine if you use pixels [15:37:50.0000] <Hixie> MarcDrummond: since it zooms the CSS pixels [15:38:13.0000] <TabAtkins> Hixie: Bonus advantage of zewt's suggested syntax: it doesn't like a CSS "w" unit. [15:38:37.0000] <Hixie> TabAtkins: as i said earlier, we can't actually use em even if we want to, because we need to be able to do the calculations before layout happens. i'm just trying to work out if there is a problem because if there is then maybe we need another solution altogether. [15:38:40.0000] <MarcDrummond> Hixie: What evidence do you have that users *never* change their default font size? Because accessibility folks would tell you, they do. [15:38:57.0000] <TabAtkins> Hixie: Make the same assumption that MQs do about rem. [15:39:09.0000] <TabAtkins> MQs interpret em as relative to the initial font size. [15:39:15.0000] <TabAtkins> s/rem/em/ [15:39:25.0000] <TabAtkins> (rem is identical, obviously). [15:39:32.0000] <Wilto> Except Opera, I think. [15:39:37.0000] <TabAtkins> Bwuh? [15:39:41.0000] <Wilto> I _think_ it uses the body font size? [15:39:47.0000] <TabAtkins> That's nonsense. [15:39:51.0000] <zcorpan> what? [15:39:53.0000] <MarcDrummond> Hixie: I don't think you have to calculate the layout. You just have to calculate the base font size. Then both the images and the layout are calculated from that font size. [15:39:56.0000] <TabAtkins> (Opera might do it, but it's still nonsense.) [15:40:01.0000] <Hixie> TabAtkins: initial font size is always 16px, so then you're fine with pixels. [15:40:08.0000] <Wilto> Oh, right. I mean, submitted without comment. [15:40:24.0000] <TabAtkins> Hixie: Nope, initial font-size is adjustable by the user. [15:40:25.0000] <MarcDrummond> Hixie: Initial font size is not always 16px. [15:40:26.0000] <zcorpan> i don't follow, what does opera do? [15:40:30.0000] <Hixie> TabAtkins: not in practice. [15:40:36.0000] <Hixie> TabAtkins: we should just fix it to 16px and use zoom. [15:40:39.0000] <TabAtkins> Hixie: Sigh. [15:41:16.0000] <Hixie> see http://hixie.ch/specs/css/font-size-ui/font-size-ui [15:41:52.0000] <TabAtkins> So, nobody does anything like what you suggest there. [15:42:04.0000] <TabAtkins> We have font-size prefs, and zoom is always a full-page zoom. [15:42:04.0000] <Hixie> i should point out that i used to say the same kind of things as you (MarcDrummond, TabAtkins) are saying, but the truth is it just isn't compatible with the web. [15:42:12.0000] <Hixie> too many sites make bad assumptions for it to work. [15:42:23.0000] <Hixie> (i know this from personal experience of intentionally browsing with unusual font defaults) [15:42:36.0000] <TabAtkins> That's what web browsers do *right now*, so browsers apparently don't find it too incompatible. [15:43:28.0000] <MarcDrummond> Hixie: I understand how you found that lots of websites don't work at other font sizes. However, that doesn't invalidate the fact that some people do so. [15:43:51.0000] <MarcDrummond> Hixie: I have seen people in my office who have changed the font size on their browser, because of their eyesight. [15:43:58.0000] <MarcDrummond> Hixie: They are not alone. [15:45:19.0000] <Hixie> i 100% agree we should address their use case [15:45:28.0000] <Hixie> i'm just saying that what we have now does a bad job and we can do much better [15:45:34.0000] <Hixie> and when we do, the issue about using pixels goes away [15:45:56.0000] <ShaneHudson> I often zoom into pages, but haven't changed my default font-size [15:46:33.0000] <Hixie> exactly [15:47:23.0000] <zcorpan> odinho: 200w isn't a hint. read the spec :-P [15:47:34.0000] <MarcDrummond> Hixie: All I ask is that you look into the possibility of using zewt's suggested syntax of putting w: before the number rather than after, and allow for ems to be used instead of px. It really would make a difference for developers. [15:47:52.0000] <MarcDrummond> Hixie: Heading out. Thanks for listening. [15:48:36.0000] <Hixie> ttyl [15:49:01.0000] <ShaneHudson> Quick question, probably stupid.. In Zurb's Foundation Template they use the classes hide-on-phones show-on-desktop etc. Has there been any suggestion (or is it worth it) making it so only one appears in the source? Seems very unsemantic to have duplicates. Or is it recommended to just keep as much in the css as possible? [15:49:51.0000] <TabAtkins> The latter, but when necessary, use @hidden to hide the stuff that is irrelevant. [15:49:58.0000] <TabAtkins> Or just use JS to remove it. [15:50:48.0000] <ShaneHudson> Ok, fair enough [15:51:19.0000] <zcorpan> a team of editors for html5? awesome [15:57:04.0000] <Velmont> TabAtkins: Nice blog post. (odinho here btw) [15:57:10.0000] <ShaneHudson> It is getting late. It has been good to meet you all and to finally see where to contribute. Goodnight all [15:57:56.0000] <TabAtkins> Velmont: Ah, keep a consistent ircnick! [15:59:38.0000] <tantek> TabAtkins indeed, hence we now have: http://wiki.whatwg.org/wiki/Irc-people [15:59:51.0000] <tantek> add yourself and others [16:03:13.0000] <zcorpan> there's http://gavinsharp.com/irc/whatwg.html already [16:06:19.0000] <smaug____> if gavin's scripts just were updating that page regularly [16:06:38.0000] <gavin> they are? [16:07:05.0000] <smaug____> gavin: looks like that page is updated randomly [16:07:27.0000] <gavin> smaug____: there were a few cases where the machine was disconnected recently [16:07:30.0000] <smaug____> the "Statistics generated on" gives usually too optimistic date [16:07:44.0000] <gavin> but it's on a cron job to update every day, with a full rebuild of the cache every three days [16:08:00.0000] <zcorpan> gavin: would be nice if the page showed /whois in the first table [16:20:00.0000] <Hixie> so i'm going to be mostly offline for the next few days, is there anyone who would like to work on that ALA article we discussed earlier? [16:20:04.0000] <Hixie> if not i can work on it when i get back [16:20:07.0000] <Velmont> TabAtkins: Can't, odinho is taken by my work computer :P I should really just not speak when I'm reading from home. [16:20:50.0000] <Hixie> or use odinho_ :-) [16:20:56.0000] <Hixie> that's what most people do :-P [16:22:10.0000] <hober> didn't someone already write an ALA article about how to contribute at the whatwg? [16:22:15.0000] <hober> that sounds really familiar [16:22:41.0000] <Velmont> 00:48 < zcorpan> odinho: 200w isn't a hint. read the spec :-P <---- Ohwell, but you can put 5000w or something else. You can make it up, -- and it will do removal based on your device width. However, -- as far as I understand, browser is actually free to use those numbers as it want. Breaking that would at least not be really evil (as far as I can see, although I haven't really thought about it). [16:23:20.0000] <Velmont> Hixie: Heh, have my legacy nick that people in other channels than #whatwg know me by. :-/ -- But I can just as well pose as two different people. Might be less confusing overall :P [16:23:40.0000] <adactio> hober: it would be good to see an ALA article that responded directly to the concerns raised there today by Wilto: http://www.alistapart.com/articles/responsive-images-and-web-standards-at-the-turning-point/ [16:23:50.0000] <adactio> hober: would you up for writing such an article? [16:25:51.0000] <Velmont> adactio: You're speaking much truth in your answer to my mail. [16:27:13.0000] <adactio> Velmont: Thanks. Just trying to take a step back and look at the problem we're trying to solve ...and realising that actually it's problem*s*, plural. So that's why I suspect that instead of a solution, there probably needs to be solution*s*, plural. [16:27:34.0000] <Velmont> adactio: Yes. We had a talk a bit earlier about that. Hmmm... [16:30:55.0000] <Velmont> Having trouble finding it. So extremely many emails :P [16:33:56.0000] <zcorpan> Velmont: oh, right. for some reason i thought the heuristics step was only for choosing between 2x etc, not for width/height [16:34:09.0000] <Velmont> adactio: http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2012-May/035819.html like this strawman. [16:34:55.0000] <zcorpan> but is that what we want? why do we want to allow heuristics for deciding on the width/height descriptors? [16:35:06.0000] <grigs> adactio: thank you for that email. good summary of what I think is going on as well. [16:36:40.0000] <Hixie> zcorpan: we don't, generally, hence the discouragement in the spec. but there are certain cases where maybe the UA has reason to do otherwise, e.g. it wants to use the image for the biggest possible viewport width because it has reason to believe the user is about to switch to a wide viewport width (e.g. because the tab is only displayed narrowly when in the background) [16:36:44.0000] <adactio> grigs: Thanks. I think a good next step would be to consolidate the use-case examples for art-directed responsive images (e.g. Boston Globe) *separately* from the use-case examples of retina/bandwidth responsive images (e.g. Apple.com). [16:37:08.0000] <adactio> grigs: Also: your blog post does an excellent job of separating out the two use-cases. Thanks for writing it. [16:38:24.0000] <grigs> one thing i've been thinking a lot about in the last few days is that in art directed changes, it may be both a combination of the image and the layout changing. [16:38:49.0000] <zcorpan> Hixie: rendering background tabs at a different width than they would be when active seems like asking for problems [16:39:20.0000] <grigs> which is why even if the Obama example can be handled by srcset (which Hixie’s email argued it could) i don’t think it would work for art direction overall. [16:40:05.0000] <grigs> The Obama example keeps the same ratio for dimensions which won't necessarily be the case. [16:40:06.0000] <Hixie> zcorpan: yeah, it's rarely the right thing, but i don't want to make it completely non-conforming, because there are some rare cases where it makes some sense. [16:40:06.0000] <grigs> The Obama example keeps the same ratio for dimensions which won't necessarily be the case. [16:40:06.0000] <Hixie> zcorpan: yeah, it's rarely the right thing, but i don't want to make it completely non-conforming, because there are some rare cases where it makes some sense. [16:42:04.0000] <zcorpan> Hixie: my knee-jerk reaction is to not allow it [16:42:20.0000] <Hixie> yeah that was mine too, had to go back and add in that step [16:42:30.0000] <Hixie> hence why it was so poorly written originally, as jgraham pointed out :-) 2012-05-16 [17:08:33.0000] <zcorpan> TabAtkins: time to flip a table? [17:08:56.0000] <TabAtkins> zcorpan: For fucks sake, yes. [17:10:58.0000] <TabAtkins> Oh, weird. If you @reply someone who doesn't exist, it doesn't count, and shows in other people's feeds anyway. [17:11:16.0000] <TabAtkins> In other words, for the purpose of showing/hiding @-replies, you're considered to follow all non-existent people. [17:13:58.0000] <TabAtkins> Shit, I keep ending tweets with periods. Gotta strip that shit. [17:19:31.0000] <grigs> TabAtkins: I *think* I may have finally figured out how to describe where I see the disconnect on the art direction side in the email I just sent to the list. I’m curious to hear what you think given our attempts earlier today to get on the same page. [17:19:43.0000] <grigs> TabAtkins: http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2012-May/035895.html [17:22:00.0000] <tantek> Good to see the progress on http://www.w3.org/wiki/Images [17:22:38.0000] <tantek> grigs - are the use-cases you mention, "art direction" vs "retina/bandwidth" adequately described / mentioned in that page? [17:22:39.0000] <tantek> grigs - are the use-cases you mention, "art direction" vs "retina/bandwidth" adequately described / mentioned in that page? [17:22:50.0000] <tantek> and uniquely enough to refer to by URL? [17:23:41.0000] <grigs> tantek: I don't think so. I was looking at that page earlier and trying to figure out where to start editing it. [17:23:54.0000] <tantek> hmm - list items make it harder to reference by frag - going to do some heading edits to make it more granular [17:24:01.0000] <TabAtkins> grigs: Answering your email now. You're misunderstanding the syntax, which may be causing the disagreement. [17:24:49.0000] <grigs> tantek: it isn't organized the way my brain works so i'm struggling a little to figure out how to contribute. not insurmountable, just giving me a moments pause (i opened it up to edit a few minutes ago and stopped) [17:24:58.0000] <tantek> grigs - let me take a shot [17:24:59.0000] <othermaciej> right now the wiki page seems more like a list of issues than of use cases [17:25:04.0000] <grigs> TabAtkins: Cool. [17:25:05.0000] <grigs> TabAtkins: Cool. [17:25:19.0000] <grigs> /me nods in agreement with othermaciej [17:25:46.0000] <othermaciej> "There are many different screen sizes" is not a use case, "authors want to serve different content to browsers with different viewport sizes, specifically like X, Y or Z" would be a use case [17:32:16.0000] <tantek> othermaciej - agreed - I'm on it [17:33:35.0000] <grigs> TabAtkins: hmm… so the width in srcset isn't the width of the image? but the "the algorithm here could be to sort the images by width, and remove all those that are wider than the available width" [17:33:45.0000] <tantek> grigs, othermaciej - take a look: http://www.w3.org/wiki/Images [17:34:10.0000] <tantek> Hopefully I've helped make it clearer what's missing in terms of documenting the use-cases [17:34:35.0000] <grigs> looks better. i'll take a stab later tonight. gotta head out soon. [17:34:48.0000] <tantek> thanks grigs - any incremental iteration / editing / contribution is appreciated [17:34:50.0000] <zcorpan> grigs: don't read the 507 thing. read the spec [17:35:18.0000] <grigs> zcorpan: i'm reading Hixie’s email. [17:35:37.0000] <zcorpan> ah. ok [17:35:49.0000] <grigs> zcorpan: not sure what the 507 thing is. URL for what I should read just to make sure I’m in the right spot? [17:36:07.0000] <zcorpan> i guess by "images" there, he really meant "candidates" [17:36:14.0000] <tantek> grigs - I think the "DPI" use-case is attempting to cover the "retina" use-case you mention: http://www.w3.org/wiki/Images#DPI [17:36:35.0000] <tantek> and the "slow networks" use case is attempting to cover the "bandwidth" use-case you mention: http://www.w3.org/wiki/Images#Slow_Networks [17:36:53.0000] <tantek> but I don't see anything that would correspond to the "art direction" use case you mention [17:36:58.0000] <grigs> correct [17:37:04.0000] <tantek> does that have to do with portrait vs. landscape? or something else? or both? [17:37:21.0000] <grigs> no, two things [17:37:29.0000] <zcorpan> grigs: it's just an extract from his email [17:37:44.0000] <grigs> first is easiest to explain here http://blog.cloudfour.com/a-framework-for-discussing-responsive-images-solutions/ [17:37:56.0000] <grigs> The Obama example in my post. [17:39:08.0000] <tantek> different cropping for different sizes? [17:39:34.0000] <tantek> is that what you mean by art direction? or are there other modifications going on? (different color palette, contrast, etc.) [17:39:43.0000] <grigs> The problem with the Obama example is that the photo doesn’t change proportions nor do I mention that the layout may change as well. In email, I looked at Nokia’s browser site as another example. http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2012-May/035795.html [17:39:53.0000] <zcorpan> if the image is *just* cropped, why isn't CSS crop good enough? [17:40:10.0000] <tantek> zcorpan - you beat me to it [17:40:24.0000] <grigs> zcorpan: downloading unnecessary data. [17:40:42.0000] <grigs> zcorpan: also, the nokia browser example shows the image making a more substantial change. [17:40:57.0000] <zcorpan> grigs: what if you have already downloaded the bigger image, and then the viewport gets narrower? [17:41:02.0000] <grigs> part of the reason why my Obama example was ok, but sort of misleading. [17:41:36.0000] <grigs> zcorpan: i'm not sure i understand the question. [17:41:46.0000] <tantek> a-ha, so the Obama example is the simple art direction use-case [17:42:00.0000] <grigs> tantek: right. very basic. [17:42:09.0000] <tantek> and the Nokia phone changing orientation (horizontal vs vertical) and crop is the more advanced art direction use-case [17:43:44.0000] <zcorpan> grigs: you said it would download unnecessary data. but going from a wide viewport to a narrow viewport would also download unnecessary data (a cropped image when you could just crop the bigger image you already have with css) [17:44:53.0000] <grigs> zcorpan: unnecessary data on a small screen--many of which can never be resized and thus will never see the extra data. [17:45:22.0000] <tantek> grigs, could you give your <h3> tags 'id' attributes so I could link directly to those sections here? http://blog.cloudfour.com/a-framework-for-discussing-responsive-images-solutions/ [17:45:48.0000] <zcorpan> grigs: mobiles can usually be flipped which changes the viewport width :-P [17:46:23.0000] <zcorpan> anyway, i shouldn't be here [17:46:53.0000] <grigs> tantek: done [17:52:10.0000] <grigs> heading out. i'll take a crack at the use cases wiki later. [17:53:12.0000] <tantek> ah just missed him. updated: http://www.w3.org/wiki/Images [17:53:39.0000] <tantek> for those interested in the "Art Direction" use-case, I attempted to capture what Grigs was explaining and emailing here: http://www.w3.org/wiki/Images#Art_Direction [17:53:59.0000] <tantek> I don't claim to have done it justice, please feel free to rewrite/extend what I'd written there to improve the use-case description. [17:55:55.0000] <tantek> also - if anyone wants to sort the use-cases by what they see as priority/importance (and add reasoning as to such), please feel free to re-order them - they're not in any particular order right now. [18:40:09.0000] <Flakerimi> Having <menu><ul><li></li></ul></menu> and styling with: menu { border:1em solid red} works, but is it wrong? I know its more xml but it helps. Creating custom tags should be treated as divs! [18:41:35.0000] <Flakerimi> So instead of creating <div id="wrapper"> why not just <wrapper> [19:11:15.0000] <jmather> TabAtkins: just a follow up thought re: bandwidth-based mq's… what if they're only calculated once per page load? That would prevent throwing assets out if something changes while a page loads up. [19:27:59.0000] <zewt> jmather: pages stay loaded for a long time now; bandwidth changes (wifi vs. cell vs. docked) [19:28:51.0000] <jmather> And that would be a problem for people loading large assets after-the-fact [19:28:55.0000] <jmather> ok, it's clicking now [19:29:09.0000] <jmather> Reading Dylan's reply to me on the CG [19:29:59.0000] <jmather> The problem is nothing really solves the core issue, as I understand it, because what we really need, for a future proof way to move forward, is a way to describe which image to select relative to the space available for it, as opposed to viewport size/whatnot [19:30:25.0000] <jmather> any other means of describing when to use which image means lots of headaches of one type or another come redesign time [19:31:15.0000] <jmather> and we can't do that because we don't know that until css is processed, which means we can't prefetch images [19:31:33.0000] <jmather> I'm sure i'm restating the obvious to a lot of people but it's just fully dawning on me :D [19:31:50.0000] <tantek> jmather - the use cases I've seen so far refer to device viewport dimensions, rather than "space available for it" [19:32:07.0000] <tantek> which use-case are you referring to? http://www.w3.org/wiki/Images [19:33:07.0000] <jmather> tantek: i'm not looking at any in particular. I'm processing in my head… the issue Dylan raises is how horrific the syntax can get, and what an unmaintainable nightmare it would become for static html when a redesign comes up [19:34:28.0000] <tantek> yeah, I haven't seen an "elegant" syntax solution yet either, nor am I sure what/which use-cases are solved by which. [19:34:30.0000] <jmather> and i started to reply that the media queries should describe how to display the image relative to the space available for it, which clicked the whole "we have to rely on the viewport, not the image space" deal, which lead me to the fact that none of this fits my personal distilled view of the problem. [19:35:16.0000] <tantek> indeed - I've found the best way to broaden our collective distilled views of the problem is to share/document the use-cases until there's some agreement there at least [19:36:04.0000] <jmather> I'm going to try to distill my use-case into the wiki and see if I can make something coherent [19:36:50.0000] <tantek> it's different from the existing use-cases? [19:37:03.0000] <jmather> essentially, when you redesign the site 4 years down the line, you shouldn't have to alter anything, because you have already described how to display the image for a variety of space considerations…. which I think invalidates every approach talked about thus far, but I could be wrong. [19:37:31.0000] <jmather> Is that off base or does that make sense? [19:37:44.0000] <tantek> jmather - I'm not sure any solution will handle the "redesign the site 4 years down the line" scenario [19:37:49.0000] <tantek> given how fast things are changing [19:37:55.0000] <jmather> because the goal of HTML5 was to begin to make things autonomous [19:38:01.0000] <tantek> it was? [19:38:06.0000] <jmather> part of it anyway [19:38:10.0000] <jmather> with <article> in particular [19:38:27.0000] <tantek> <article> essentially just reflected what we already knew from years of experience with RSS [19:38:27.0000] <jmather> should contain self-… something content, which can be pulled out and used elsewhere [19:38:31.0000] <jmather> right [19:38:44.0000] <tantek> nothing "future proof" there [19:38:45.0000] <jmather> but if you embed site-specific requirements on image displays inside an <article> [19:38:48.0000] <tantek> more like paving the footpaths [19:39:08.0000] <jmather> then that content is no longer container independent, so to speak [19:39:54.0000] <jmather> I take a little while to get myself to where my conclusion has taken me… give me a minute to rethink/rework/whatnot… :) [19:41:42.0000] <jmather> ok, let me try and re-step-through that logic set [19:42:35.0000] <jmather> theoretically, content wrapped in an <article> tag should be able to be taken off the site and used elsewhere w/ limited issues. Am I right-ish here? [19:43:45.0000] <tantek> as much as RSS <item>s can be, I suppose. [19:46:34.0000] <jmather> so, if we're embedding rules on when to use what image, there would be a valid concern in ensuring the rules described when to pick which image relative to the image's container. Because if the rules described how to select an image based on browser viewport, that would/could cause an issue where it's loading the 800px wide image when the container is only … ok, well, i guess … now i've gone and circled myself back around to where [19:46:35.0000] <jmather> ago [19:47:34.0000] <jmather> what if picture had a default style of max-width: 100%? It still doesn't solve the bigger than intended, but it should keep it from looking absolutely horrible in almost every situation... [19:48:09.0000] <jmather> it'd be hard to apply that to img w/ backwards compatibility issues, so it wouldn't work for srcset (I don't think) [19:50:56.0000] <jmather> I still think the best idea would be to be able to describe relative to container width, but I think it's been established as not going to happen, so I don't know at this point. :D [19:52:03.0000] <tantek> jmather - I'm having trouble following your general example. if you could describe it with a specific example similar to http://blog.cloudfour.com/a-framework-for-discussing-responsive-images-solutions/#artdirection that would help [19:54:10.0000] <jmather> Yeah… let me think on it to see if I can solidify it. It's hard for me to think it through at the moment though because I am pretty sure srcset initially tried/thought to do this and it got shot down by the browsers, so the width/height then became relative to viewport instead of container. [19:59:30.0000] <TabAtkins_> jmather: The maintenance issue is unavoidable if you want to do this in HTML proper. And you *do* want that, because it allows prefetching. [20:00:17.0000] <jmather> TabAtkins_: right, except maybe it'd be worth giving up prefetching to gain a longer lifetime of usefulness. [20:03:35.0000] <jmather> Just a thought, because, well, I like thinking about crazy things. [20:06:24.0000] <jmather> Most of the page will use image-set in CSS to load the layout and such, so the only things affected are content images anyway [20:06:31.0000] <jmather> eventually anyway [20:09:47.0000] <jmather> I'm just trying to maybe throw some off-the-wall ideas out there. If these are crackpot ideas, feel free to tell me so. I just know a lot of the time people sometimes get stuck thinking inside of some constraints, and so I like to try and see if they can be re-evaluated. [20:10:25.0000] <jmather> /me has a team who is good at making him readjust outside of the box. [20:14:36.0000] <tantek> jmather - no problem with posing off-the-wall/crackpot ideas for exploration/discussion, especially when framed as such. [20:15:31.0000] <jmather> tantek: cool. I'm just thinking from my perspective… there's not enough man-power in the world at my employer to go back and upgrade 4 year old content. [20:15:49.0000] <jmather> well, there is, but not that will get paid for. [20:16:08.0000] <jmather> It'd be a miserable summer for a couple interns. [20:16:41.0000] <tantek> othermaciej - earlier http://krijnhoetmer.nl/irc-logs/whatwg/20120515#l-2032 you asked (and I missed amongst other discussion) "what bad advice do you think the respimg guys got on the whatwg mailing list?" [20:16:44.0000] <jmather> And I hate having that be the solution to anything [20:16:47.0000] <tantek> and the answer was/is: http://lists.w3.org/Archives/Public/public-whatwg-archive/2012Feb/0169.html [20:17:27.0000] <tantek> jmather, designing for "future proof" is pretty much impossible (if you figure this out, plenty of folks want to hear about it) [20:17:37.0000] <tantek> however, designing for "future friendly" may be possible [20:18:09.0000] <jmather> tantek: i'd be happy with friendly, but i don't think any of the responsive solutions even hit that target very squarely [20:18:24.0000] <jmather> unless i'm wrong about srcset's implementation, which i very well could be [20:18:44.0000] <jmather> (that the width/height refer to browser, not container) [20:19:48.0000] <tantek> jmather, I leave you with http://futurefriend.ly/ - and note that many of the folks behind that were/are involved with the respimg CG (and some have even been participating right here in the channel today) [20:20:10.0000] <jmather> Yeah, I know [20:20:18.0000] <jmather> I've been there and all [20:20:55.0000] <jmather> I'm just saying, looking at it as a whole, selecting the content image to display based on browser size, is a terrible idea. [20:21:44.0000] <jmather> Well, terrible implementation anyway [20:27:22.0000] <jmather> Is that too off the wall, or does that make sense? [20:31:27.0000] <jmather> Just realized I hadn't actually joined the CG group so I can't post my crazy idea. Maybe that's a good thing, but request pending anyway… :D [20:34:53.0000] <tantek> jmather - anyone with a W3C account can edit the W3C wiki: w3.org/wiki [20:35:00.0000] <tantek> no need to wait to join the CG or anything else [20:36:25.0000] <jmather> I figured I would use the CG to ensure I'm not a total moron before I let /everyone/ know I'm an idiot, and I don't think the wiki is the right context for a 'what if we are looking at the wrong requirements' type question [20:37:17.0000] <jmather> Maybe they've already thought of this and it's simply so un-doable that they've dismissed it [20:40:27.0000] <jmather> Alright, off for the night. Good chat, thanks for dealing with my crazy. Hopefully tomorrow will be more cohesive. [20:42:17.0000] <tantek> jmather - if you follow the example of describing a specific use-case like the URL I gave earlier, http://blog.cloudfour.com/a-framework-for-discussing-responsive-images-solutions/#artdirection then it's unlikely that you'll appear as a moron/idiot ;) [20:42:20.0000] <jmather> Brain hurts from all this today, heh [20:42:34.0000] <tantek> welcome back grigs [20:42:46.0000] <tantek> w3.org/wiki/Images edited per our discussion here earlier [20:42:58.0000] <jmather> tantek: yeah, that's the plan. It's pretty off the wall though, so I think I'll have to put on my flame retardant underwear before I post it. [20:43:12.0000] <grigs> thanks tantek. :-) [20:44:01.0000] <grigs> jmather: fwiw, i think i've been following your logic and i don't think you're off the wall. [20:44:51.0000] <jmather> grigs: I've got an even more crackpot idea at this point, unless you've been watching in chat for the last hour or so? :D [20:45:19.0000] <grigs> jmather: i just reread the chat logs before jumping back in. [20:45:30.0000] <jmather> Ah! awesome [20:45:36.0000] <grigs> little easier to catch up on the conversation tonight than it was earlier in the day. [20:45:42.0000] <jmather> Yeah, I bet [20:46:45.0000] <jmather> I'll think on it over night, but the more I think about it, the more I think the premise of our solutions in general have been wrong. The images should be selected based on the size of the content area available. I know that means it can't be selected until css has been processed, but maybe that's just something we accept as part of the solution to get the best solution. [20:50:00.0000] <jmather> I have to run for a few [20:50:03.0000] <jmather> bbl [20:50:43.0000] <grigs> tantek: the write up of art direction is good [21:00:38.0000] <tantek> thanks grigs [21:03:10.0000] <grigs> btw, this slide deck from bryan rieger was one of the first examinations i saw of this issue. http://www.slideshare.net/bryanrieger/prime-sky in particular, slide 20 shows the scaling issues. [21:03:54.0000] <grigs> was tempted to add it to the use cases wiki, but think it might confuse things because it was written before responsive design was a thing. [22:15:07.0000] <AryehGregor> annevk, subscribe to the Mozilla bug and you'll know as much about detach() as I do. If anyone complains, it will likely be filed as blocking that bug, so we'll see it then. [22:15:42.0000] <AryehGregor> The metrics I added are actually useless in practice. I should have added something that counts how many times an exception would have been thrown and now wasn't. [23:47:31.0000] <hsivonen> I should probably read all the threads before commenting, but it seems that the responsive images debate conflates two issues: [23:47:57.0000] <hsivonen> 1) Whether it's better to have <picture> and <source> or an attribute-based microsyntax [23:48:41.0000] <hsivonen> 2) Whether Media Queries are the appropriate way to declare things that pertain to the selection of the alternative image files [23:49:14.0000] <othermaciej> those are two of the differences between <picture> and <img srcset> [23:49:36.0000] <othermaciej> I believe there is at least one other important difference: [23:50:11.0000] <othermaciej> 3) Should selecting an image intended for a particular display resolution automatically rescale it, or should it be up to the author to do that with CSS? [23:50:24.0000] <othermaciej> and one other difference the importance of which has not been established: [23:50:44.0000] <othermaciej> 4) Are there useful criteria for selecting images besides width, height, and pixel density? [23:51:12.0000] <othermaciej> those are the actual non-bikeshed differences between the two options [23:51:15.0000] <hsivonen> othermaciej: about #3: do you mean it's up for debate if the box of the image should change size relative to the surroundings as part of the responsiveness function? [23:51:41.0000] <hsivonen> othermaciej: yeah, #4 is part of #2 [23:51:50.0000] <othermaciej> hsivonen: I'm not sure I understand the framing of your question, but... [23:52:09.0000] <hsivonen> othermaciej: I'm not sure what you mean by not automatically rescaling [23:52:29.0000] <othermaciej> let's say I have a 600x200 version of an image and a 1200x400 version of an image, the latter intended for devices with a 2x scale factor (e.g. an iPad with Retina display) [23:53:07.0000] <othermaciej> the following always has the same intrinsic size: <img src="foo.jpg" srcset="foo⊙2 2x"> [23:53:18.0000] <hsivonen> so if I have a device with 1.5 scale factor, I'd expect to see the 1200x400 version scaled down [23:54:14.0000] <othermaciej> the following does not always have the same intrinsic size: <picture><source src="foo⊙2" media="min-device-pixel-ratio: 2"><source src="foo.jpg"></picture> [23:54:17.0000] <hsivonen> also, if I have an 1x device and zoom in by a factor of 2, I'd expect to see the 1200x400 image as soon as it has downloaded [23:54:43.0000] <othermaciej> the <picture> version will have different intrinsic size based on which image is selected [23:54:49.0000] <hsivonen> ah [23:55:04.0000] <hsivonen> that would be bad for swapping the images when zooming [23:55:07.0000] <othermaciej> the <img> version will have consistent intrinsic size, as when the 2x version is selected, it gets downscaled by a factor of 2 [23:55:33.0000] <othermaciej> you'd have to either apply fixed size, or have a CSS rule that applies image-resolution based on a media query [23:55:42.0000] <othermaciej> with the <picture> version [23:56:00.0000] <othermaciej> otherwise, you get something twice as large, not something with twice the density [23:59:29.0000] <hsivonen> I don't particularly like responding to issues by rallying troops from one's group of colleagues or fans instead of addressing *why* something is better or worse [00:00:12.0000] <othermaciej> me neither [00:00:55.0000] <othermaciej> I think the important thing to do is identify use cases (or other pragmatic considerations) in favor of one approach or the other, rather than counting Facebook likes [00:01:05.0000] <othermaciej> (or their equivalent) [00:01:16.0000] <smaug____> hmm, how to say politely that "given its quality, this spec sure feels like an early draft" [00:01:52.0000] <othermaciej> smaug____: what formal maturity level is it? [00:02:16.0000] <smaug____> WD [00:02:22.0000] <othermaciej> for an FPWD, the polite way to say it would be to not mention it, since it goes without saying [00:02:34.0000] <othermaciej> for a PR, presumably even stronger words are called for [00:02:57.0000] <othermaciej> for LCWD, the way you said it is probably appropriately polite [00:03:02.0000] <smaug____> well, I'd want to answer to comment " This is not an early draft of a spec." [00:03:21.0000] <smaug____> ah, yeah, I could just say it is WD [00:12:10.0000] <anatolbroder> othermaciej: any feedback on srcset from devs of Wordpress, Drupal, Joomla, $anotherCMS? I’m wondering which CMS will be the first to implement it. [00:14:46.0000] <othermaciej> anatolbroder: no idea; haven't heard from any of them [00:15:13.0000] <othermaciej> I doubt it will get big-time deployment until at least one significant-share browser supports it [00:17:23.0000] <anatolbroder> othermaciej: I suppose, Webkit will support it very soon (<6 month), right? [00:18:01.0000] <othermaciej> I hope so - depends on how fancy we want to get [00:19:03.0000] <hsivonen> anatolbroder: I sure hope no CMS deploys anything on the responsive images topic before there's a browser build to test with [00:19:50.0000] <hsivonen> premature and subtly incorrect deployment in a popular CMS is a way to poison a feature for everyone [00:26:07.0000] <anatolbroder> hsivonen: sure. But I hope the guys from the popular CMSs pay attention to that topic. If $cms_1 supports responsive images before $cms_2, I would start my next cat’ face project with $cms_1. [00:33:27.0000] <jgraham> see also: premature polyfills [00:35:04.0000] <Von_Davidicus> Yes, it's me again, with questions of markup. <: ) [00:35:06.0000] <othermaciej> it's better to polyfill only once there's something to fill [00:35:49.0000] <jgraham> Right. Making the whole wall out of polyfill is likely to be structurally unsound, or something. [00:36:05.0000] <jgraham> Although I think that's not quite the right extension of the metaphor [00:36:14.0000] <Von_Davidicus> How often were webpages written in XML + XSLT? [00:36:24.0000] <jgraham> It's more like "will get in the way of building the actual wall" [00:36:54.0000] <jgraham> Von_Davidicus: I doubt anyone has statistics. Is "too often" too cynical? [00:38:04.0000] <Von_Davidicus> So I'm not the only one who was that weird? [00:38:56.0000] <jgraham> Von_Davidicus: It depends whether you were responsible for http://thedailywtf.com/Articles/Where-the-Wild-Web-Things-Are.aspx or not [00:39:02.0000] <hsivonen> Von_Davidicus: often enough that removing support for XSLT is not actively pursued but rarely enough that it doesn't make sense to invest in improving XSLT support [00:40:34.0000] <Von_Davidicus> Okay. [00:40:34.0000] <Von_Davidicus> Okay. [00:40:46.0000] <jgraham> Oh, that was quite a good answer [00:40:50.0000] <hsivonen> there's even at least one Chrome Experiment that uses XSLT. For shame. [00:41:02.0000] <jgraham> Although I think it is the same as "too often" [00:41:03.0000] <Von_Davidicus> /me did most of a website in XSLT + XML (+ Schema + DTD + WTHWIT) and another set of pages with XML + XSLT. [00:41:04.0000] <Von_Davidicus> /me did most of a website in XSLT + XML (+ Schema + DTD + WTHWIT) and another set of pages with XML + XSLT. [00:41:36.0000] <anatolbroder> What’s about Firefox and Opera, are they willing to support the srcset soon (<6 month)? [00:41:52.0000] <Von_Davidicus> But I wasn't going to count those because that would be cheating. [00:42:36.0000] <jgraham> anatolbroder: We (Opera) don't generally make promises about our roadmap. [00:42:59.0000] <hsivonen> anatolbroder: it would be unwise for anyone to promise support at this point [00:43:33.0000] <jgraham> anatolbroder: A reasonable point is that a simpler design is likely to lead to faster implementations [00:43:51.0000] <Von_Davidicus> Here's another question: What's wrong with XML + XSLT? [00:43:52.0000] <anatolbroder> jgraham, hsivonen: I understand. [00:44:06.0000] <jgraham> If the feature needs extensive development + QA work it will take longer than if one bored developer can do it in an afternoon [00:44:09.0000] <tantek> FWIW server-side XSLT is used some production deployments. [00:44:13.0000] <tantek> *in some [00:44:18.0000] <hsivonen> yeah, the parts of srcset that would fall on my plate to implement would be simpler than the parts of <picture> that would fall on my plate [00:44:31.0000] <hsivonen> server-side XSLT is OK [00:44:45.0000] <tantek> hsivonen - it's still a pain to maintain. [00:44:55.0000] <jgraham> (I am not claiming that any design is simple enough to only be an afternoon of work) [00:45:13.0000] <tantek> but once it works, it seems quite reliable. [00:45:51.0000] <Von_Davidicus> But--and here's the reason I used it--It resulted in a smaller website (pertaining to file size) [00:48:25.0000] <othermaciej> serving xml+xslt to the client has some downsides [00:48:27.0000] <Ms2ger> Wow [00:48:39.0000] <othermaciej> 1) you have to transform the whole document before you can render it [00:48:55.0000] <Ms2ger> Is there anything I should read in the million lines of Responsive Images trolling in yesterday's logs? [00:49:11.0000] <othermaciej> 2) you can't really do dynamic updates to your model very well and have them properly restyled (as with css0 [00:49:42.0000] <jgraham> Ms2ger: I think you will livwe [00:49:49.0000] <jgraham> *live [00:49:53.0000] <othermaciej> 3) if you are using generic xml or a custom xml vocabulary with your xslt, you are serving something that doesn't have useful widely understood semantics for data mining type consumers [00:50:13.0000] <othermaciej> so, less efficient to render, less dynamically adaptable, and less semantic [00:50:23.0000] <othermaciej> using xslt on the server side does not have any of these issues [00:50:32.0000] <othermaciej> assuming your output is html [00:50:41.0000] <anatolbroder> jgraham: I see the point. I really love the scrset spec. It’s so smart. This is the solution I always dreamed of. Now I can tell, the web is not print anymore. I build my current picture based project with scrset in mind. Thanks again to all you devs! [00:51:25.0000] <Ms2ger> anatolbroder, are you a webdev? [00:51:33.0000] <Von_Davidicus> Okay, so 1) means that the pages actually load slower than an HTML page? [00:51:50.0000] <hsivonen> Von_Davidicus: yes [00:52:05.0000] <anatolbroder> Ms2ger: I create web pages. [00:52:17.0000] <Von_Davidicus> Wow. [00:52:33.0000] <hsivonen> 4) XSLT doesn't benefit from speculative preloading of images, scripts and CSS [00:52:40.0000] <Von_Davidicus> Could you explain 2) a bit more clearly? [00:52:51.0000] <Ms2ger> That can't be true, several people came in here yesterday and made it clear that ALL web developers HATE srcset [00:54:10.0000] <othermaciej> Von_Davidicus: 2) is relevant only if you have client-side scripts or ever plan to add any [00:55:24.0000] <Von_Davidicus> Oh, okay. None of my XML+XSLT pages had any scripts--certainly none that fiddled with the DOM. [00:56:09.0000] <anatolbroder> Ms2ger: It was very funny, indeed. Most of that “web developers” are control freaks, caught in the print world. They just don’t want to understand, what is web for. [00:56:25.0000] <Ms2ger> anatolbroder, I'm not going to comment on that :) [00:58:02.0000] <Von_Davidicus> Assuming the DOM is what you were referring to when you said "model" and stuff like "document.createElement()" was what you meant by doing "dynamic updates". :) [00:58:02.0000] <Von_Davidicus> Assuming the DOM is what you were referring to when you said "model" and stuff like "document.createElement()" was what you meant by doing "dynamic updates". :) [01:06:19.0000] <Ms2ger> tantek, note, regardless of whether you want to call Hixie a benevolent dictator, you know he's not appointed for life. [01:06:20.0000] <Ms2ger> tantek, note, regardless of whether you want to call Hixie a benevolent dictator, you know he's not appointed for life. [01:09:07.0000] <tantek> Ms2ger - I don't think any of the BFDLs referenced in the Wikipedia article are "appointed for life" - from my understanding it is a deliberate term of exaggeration. [01:09:59.0000] <tantek> sorry, BDFLs [01:18:03.0000] <Ms2ger> Also, winter/summer on http://wiki.whatwg.org/wiki/Irc-people ... Northern or Southern hemisphere? [01:20:31.0000] <charlvn_> very good question, but is daylight savings going to make so much of a difference in any case? [01:20:38.0000] <charlvn_> we're talking about an hour sooner or later [01:20:51.0000] <charlvn_> most people's irc times are pretty variable in any case [01:22:06.0000] <tantek> charlvn_ on the contrary, the data shows most people's irc times are pretty consistent within certain blocks of hours - see the preponderance of particular color blocks for people in general here: http://gavinsharp.com/irc/whatwg.html [01:22:21.0000] <Gargoyle> If you were going to specify a timezone, what is wrong with the "Europe/London" syntax? [01:22:35.0000] <tantek> Gargoyle - because that's political, you can't trust it over time. [01:22:45.0000] <zcorpan> TabAtkins: for the square and rectangle example, if you wanted to use the rectangle as the fallback image, the syntax doesn't provide for that, right? [01:23:55.0000] <tantek> Ms2ger - by the preponderance of people listed there it is easy to conclude Northern hemisphere. If you have suggestions for clarifying, please feel free to edit the wiki accordingly. [01:24:08.0000] <charlvn> tantek: i guess you might be right [01:24:24.0000] <charlvn> perhaps it's just my own times that are quite variable :) [01:24:39.0000] <tantek> charlvn - it's not about me, I'm simply communicating what the data appears to show. feel free to provide alternate interpretations. [01:24:44.0000] <Gargoyle> tantek: Isn't that a benefit, so you don't need to wory about DST? [01:24:48.0000] <Ms2ger> tantek, how so? Europe/London seems more reliable than +0000 [01:24:51.0000] <othermaciej> what is the purpose of the people/timezone list? knowing when to try to contact someone? [01:25:19.0000] <Ms2ger> I guess one is more likely to move than to have their time zone changed under them... [01:25:26.0000] <tantek> othermaciej - I don't remember why it was introduced to the list of irc-people on microformats.org - but given it's presence there I simply copied the template rather than re-bikeshed it. [01:25:33.0000] <tantek> its presence even [01:25:55.0000] <tantek> feel free to edit the template if you have ideas for improving it [01:26:42.0000] <tantek> Ms2ger - named timezones are the parlance of politicians and tend to be less stable than numerical offsets from Z which is more the parlance of scientists. [01:27:00.0000] <tantek> borders change, when summer times start/end change etc. [01:27:21.0000] <hsivonen> tantek: looking at the recent i18n comments on HTML, it seems the political instability of named time zones is considered a feature [01:27:31.0000] <Ms2ger> tantek, well, I guess Europe/London wouldn't make much sense if you were in Lisbon, but it seems sensible enough for someone who actually lives in London [01:27:45.0000] <tantek> hsivonen, that's sad to hear. [01:27:57.0000] <othermaciej> tantek: what is the intended purpose of the irc-people list in general? would be glad to improve it if I knew what it is meant to be used for [01:28:08.0000] <othermaciej> hsivonen: oh? [01:28:12.0000] <tantek> help people get to know each other better [01:28:13.0000] <tantek> help people get to know each other better [01:28:17.0000] <hsivonen> tantek: it seems they want to be able to specify named time zones so that if the politicians move the named zone, the times anchored to the named zone change accordingly [01:28:41.0000] <Gargoyle> tantek: Is'nt that what whois is for? [01:29:02.0000] <tantek> whois doesn't do irc nickname lookups AFAIK [01:29:14.0000] <hsivonen> (if that's not what they want, asking for named zones makes no sense) [01:29:33.0000] <tantek> plus a wiki makes it possible for anyone to curate such information about anyone [01:30:27.0000] <tantek> othermaciej - with so many folks in the channel, it helps to be able to lookup who is a person, what have they worked on, are they a browser implementer etc. - especially should a person choose to document that on their User page. [01:30:45.0000] <hsivonen> I suppose the use case is that if you agree that a meeting takes place at noon on a given day next year, if the politicians change the UTC offset, the meeting still takes place at noon according to the new offset [01:30:52.0000] <tantek> the breadth of participants has far exceeded a small # of folks who mostly know each other. [01:30:52.0000] <tantek> the breadth of participants has far exceeded a small # of folks who mostly know each other. [01:31:00.0000] <othermaciej> ok, so not just list of nicks but ideally some sort of brief information about their context-relevant background and interests [01:31:20.0000] <tantek> othermaciej - the "brief information about their context-relevant background and interests" can go on their User page - it's already there for that purpose [01:31:28.0000] <tantek> better to just have the list link to user pages [01:31:32.0000] <tantek> rather than duplicate that content inline [01:31:33.0000] <tantek> rather than duplicate that content inline [01:31:45.0000] <othermaciej> is it considered bad form to edit someone else's user page? [01:32:00.0000] <tantek> othermaciej - nah, people are assume to be benevolent on a wiki until proven otherwise [01:32:10.0000] <tantek> it's part of the wiki-way (as far as I understand it) [01:32:11.0000] <tantek> it's part of the wiki-way (as far as I understand it) [01:32:12.0000] <hsivonen> (I think politicians who think it's cool to have daylight saving time and non-integral hour time zones should be required to implement calendar software with some punishment for each bug) [01:32:15.0000] <Ms2ger> Depends on the wiki, I've been told off for it long ago [01:32:22.0000] <tantek> hsivonen +1 [01:32:23.0000] <tantek> hsivonen +1 [01:32:34.0000] <othermaciej> I've fixed enough security vulnerabilities to consider everyone malevolent by default, myself included :-) [01:32:44.0000] <Ms2ger> othermaciej++ [01:32:45.0000] <tantek> Ms2ger - yeah? sorry to hear that. haven't had that problem with any wiki I've edited. [01:33:34.0000] <othermaciej> I had the impression that user pages, unlike general wiki content, were in some way "owned" by the relevant user [01:33:42.0000] <othermaciej> but I am no expert on wiki etiquette [01:33:46.0000] <charlvn> the irc regulars list sounds like an interesting idea, but while we are on the topic of the wiki, it seems like i am both unable to perform password recovery and unable to register a new account [01:33:59.0000] <tantek> othermaciej - mostly it's summed up by be nice and forgiving. [01:34:14.0000] <tantek> that tends to handle the vast majority of cases. [01:34:28.0000] <tantek> beyond that, wikipedia has various escalations for conflict resolution. [01:35:02.0000] <tantek> charlvn - new whatwg wiki accounts are registered thru Hixie. [01:35:06.0000] <othermaciej> I'm totally down with being nice to and forgiving of wiki edits made by others, just not sure what kinds of edits it is nice to make [01:35:26.0000] <odinho> 09:36 < othermaciej> it's better to polyfill only once there's something to fill <-- BTW totally agreed, -- I made my proof of concept (very shortcutty) polyfill merely to demonstrate that "it can't be polyfilled!11" argument was incorrect. [01:35:27.0000] <annevk> charlvn: if you give me your email address I can get you an account [01:35:27.0000] <othermaciej> I will take your word for it that editing other people's user pages is provisionally ok unless the person in question says otherwise [01:35:33.0000] <charlvn> tantek: thanks, just saw the notice on the main page as well, i'll speak to him next time i see him online or just send him a mail [01:35:36.0000] <tantek> othermaciej - perhaps the kinds of edits you wouldn't mind if someone else made for you? [01:36:46.0000] <tantek> btw anyone who'd like to discuss political timezone arcana should ping dbaron when he's around - he's got the most thorough database of historical timezone changes (nearly weekly?) that I've seen. [01:38:00.0000] <othermaciej> wikipedia has a lot of policy related to their use of user pages http://en.wikipedia.org/wiki/User_page [01:38:42.0000] <AryehGregor> Wikipedia convention is definitely that you should generally avoid touching other people's user pages unless you have very good reason. [01:39:01.0000] <AryehGregor> I've corrected typos and formatting errors and gotten away with it, and I had people do things like substitute templates that were about to be deleted. [01:39:03.0000] <othermaciej> man, wikipedia has more policies than the html wg [01:39:10.0000] <AryehGregor> And of course, if you violate policies . . . [01:39:17.0000] <AryehGregor> othermaciej, Wikipedia is a lot bigger than the HTMLWG. [01:39:38.0000] <AryehGregor> Also, its "policies" are whatever someone happens to have decided to write, and half the time they're ignored. [01:40:01.0000] <tantek> othermaciej - their policies show a history of the abuses they've had to deal with [01:40:09.0000] <AryehGregor> Real and perceived. [01:40:10.0000] <tantek> not unlike other processes I know of. ahe. [01:40:17.0000] <tantek> *ahem [01:40:28.0000] <charlvn> yeah considering the number of people who edit the whatwg wiki is fairly small, i don't see this as a big problem in the short-mid term [01:40:41.0000] <charlvn> i'm sure if there are some specific cases of abuse/unhappiness we can get them resolved fairly easily [01:40:45.0000] <tantek> oh look - I totally missed that the Astrology company withdrew their copyright suit over the Olson timezone database: http://www.computerworld.com/s/article/9224557/Astrolabe_withdraws_copyright_suit_over_Internet_time_zone_database [01:40:56.0000] <tantek> well that was hilarious while it lasted. [01:41:14.0000] <othermaciej> html wg policy has grown cumulatively in response to real or perceived problems with events that actually occurred [01:41:55.0000] <hsivonen> othermaciej: unfortunately, the policy hasn't been very good at adapting to deal with problems arising from the policy [01:41:58.0000] <othermaciej> I think it could be potentially improved by having less of it [01:42:39.0000] <tantek> othermaciej - I wonder if there is a way to grow processes in a more usable/accessible/positive manner. that is, instead of so much focus on avoiding negative events/behavior, can we craft process to presume positive / cooperative behaviors, and then separately list "exception handling" so that it doesn't interfere with understanding the main process flow of control. [01:42:41.0000] <tantek> othermaciej - I wonder if there is a way to grow processes in a more usable/accessible/positive manner. that is, instead of so much focus on avoiding negative events/behavior, can we craft process to presume positive / cooperative behaviors, and then separately list "exception handling" so that it doesn't interfere with understanding the main process flow of control. [01:43:02.0000] <othermaciej> it is hard to design a system that cannot be gamed (as a general rule) [01:43:11.0000] <hsivonen> I think the HTML WG policy optimizes for avoidance of accusation of the chairs being unfair but I think it should optimize for good technical outcomes and efficient use of the participants' time [01:43:21.0000] <tantek> indeed [01:43:25.0000] <othermaciej> tantek: I originally envisioned the issue process in general, and the survey part in particular, to be "exception handling" [01:43:45.0000] <othermaciej> but that created an incentive for some people to purposefully fall into the exception case [01:43:47.0000] <tantek> in practice it appears to have descended into a primary use [01:43:57.0000] <tantek> I'm not sure there is incentive, as much as inaction [01:44:05.0000] <othermaciej> because there is no incentive not to [01:44:05.0000] <othermaciej> because there is no incentive not to [01:44:13.0000] <tantek> it seems if people disagree, and do nothing, surveys happen [01:44:21.0000] <othermaciej> at least, if you care a lot about a narrow set of issues [01:44:22.0000] <othermaciej> at least, if you care a lot about a narrow set of issues [01:44:50.0000] <tantek> but I don't have a specific proposed fix, so I can't really criticize particulars [01:45:00.0000] <tantek> hence my more general musing above [01:45:20.0000] <othermaciej> I prefer systems that give people positive incentives to do the right thing [01:46:00.0000] <othermaciej> sometimes negative incentives to discourage undesirable behavior are also required but ideally should not be the focus [01:46:00.0000] <othermaciej> sometimes negative incentives to discourage undesirable behavior are also required but ideally should not be the focus [01:52:29.0000] <hsivonen> so typical. I'm trying to fix one document.open() regression and notice that the code doesn't appear to quite match the spec in other cases, either [01:52:59.0000] <hsivonen> sigh. instead of just adding a couple of boolean checks, it seems I now have to write test cases [01:54:02.0000] <zcorpan> hsivonen: the real fun doesn't start until you find that the spec is wrong :-P [02:06:18.0000] <AryehGregor> hsivonen, is writing test cases a bad thing? [02:06:21.0000] <AryehGregor> :) [02:38:16.0000] <odinho> Oh my. http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2012-May/035920.html [02:42:22.0000] <annevk> classic [03:04:04.0000] <hsivonen> it's rather hard to get a good idea of how srcset works conceptually by reading the algorithms in the spec. :-( [03:05:33.0000] <annevk> there's a note that explains it except it has minimum/maximum reversed [03:07:27.0000] <hsivonen> oh ok. so srcset doesn't do what I thought it was doing. [03:07:43.0000] <hsivonen> seems to just reinvent viewport querying [03:07:58.0000] <hsivonen> instead of declaring the characteristics of the image [03:08:22.0000] <hsivonen> now I don't understand the point of doing srcset instead of <source media> [03:08:34.0000] <annevk> the 2x bit is not querying [03:08:55.0000] <annevk> well, Xx bit [03:09:53.0000] <hsivonen> did Hixie post to the list with design rationale? [03:10:49.0000] <annevk> hsivonen: http://lists.w3.org/Archives/Public/public-whatwg-archive/2012May/0247.html [03:11:50.0000] <hsivonen> annevk: thanks [03:13:18.0000] <jgraham> hsivonen: testharness.js testcases + W3C submission please? :) [03:14:23.0000] <hsivonen> jgraham: is testharness.js still hard to use? [03:14:38.0000] <jgraham> hsivonen: It really isn't that hard to use [03:14:47.0000] <jgraham> If you mean "is it Mochitest" then no [03:15:11.0000] <hsivonen> I don't want mochitest. I want something that I can run in all browsers and then mochitestify [03:16:20.0000] <jgraham> Well testharness.js tests run in all browsers [03:16:29.0000] <jgraham> For a sutiable definition of "all" [03:17:20.0000] <smaug____> mochitests should run on all browsers if you don't use browser specific features [03:17:40.0000] <smaug____> the problem is to set proxies correctly in the browsers [03:25:53.0000] <odinho> testharness.js is nice. I like it alot. Being grown up with testharness I see lots of weaknesses in other systems when I have to use them. [03:48:35.0000] <jgraham> So, can anyone suggest what the simplest thing I can do is to cause a task to be queued? Preferably not involving resource loading [03:49:03.0000] <annevk> setTimeout [03:49:54.0000] <jgraham> Not good enough since that has a delay [03:50:13.0000] <annevk> jgraham: if you specify 0 it shouldn't [03:50:13.0000] <jgraham> (if setTimeout 0 was sure to not have a delay it would be) [03:50:40.0000] <jgraham> Well I can try that, but the spec doesn't actually say it has to work [03:50:50.0000] <odinho> It has no delay first time called? Isn't there an optimization like that? -- Yeah, spec... [03:50:56.0000] <jgraham> So if there is a better option, I would prefer it [03:51:15.0000] <jgraham> odinho: "Optionally wait a User Agent specified amount of time" or something [03:51:33.0000] <odinho> jgraham: IndexedDB transaction? :P [03:52:03.0000] <jgraham> odinho: No :p [03:55:40.0000] <zcorpan> jgraham: postMessage [03:56:42.0000] <hsivonen> soo... does "completely loaded" mean something other than readyState == "complete"? http://www.whatwg.org/specs/web-apps/current-work/#completely-loaded [03:57:24.0000] <jgraham> zcorpan: Wrong task source [03:57:26.0000] <hsivonen> apparently so [03:57:38.0000] <hsivonen> gotta love terminology like that [03:57:39.0000] <annevk> hmm, so even the ALA article from yesterday on <picture> gets the media query syntax wrong [03:58:15.0000] <hsivonen> I wonder how Gecko tracks the "completely loaded" concept [03:58:17.0000] <wilhelm> annevk: Is there a spec somewhere? I haven't paid attention. [03:58:36.0000] <annevk> wilhelm: there's a spec for media queries [03:58:36.0000] <odinho> wilhelm: For what? srcset? [03:59:00.0000] <wilhelm> odinho: Yes. Or <picture>. Or whatever it's called now. (c: [03:59:13.0000] <odinho> wilhelm: http://www.whatwg.org/specs/web-apps/current-work/multipage/embedded-content-1.html#attr-img-srcset [04:00:03.0000] <annevk> beautiful, all the examples on https://github.com/Wilto/respimg are also invalid [04:00:17.0000] <annevk> you'd think if you propose using media queries you'd at least know how to use them [04:00:41.0000] <hsivonen> Hixie: is it intentional that per spec document.open() doesn't return document when the steps are terminated due to the user refusing to allow the document to be unloaded? [04:01:55.0000] <zcorpan> jgraham: which task source do you need? [04:02:10.0000] <wilhelm> odinho: Thanks. [04:02:59.0000] <jgraham> zcorpan: The one that is used for the task related to the load event [04:03:04.0000] <jgraham> DOM manipulation [04:03:52.0000] <zcorpan> annevk: i think being grumpy doesn't help [04:04:32.0000] <jgraham> There is so much grumpy going round that not even dglazkov can balance it out [04:04:33.0000] <zcorpan> jgraham: window load? [04:04:39.0000] <jgraham> Yeah [04:04:58.0000] <zcorpan> yeah there is, and clearly it doesn't help [04:05:52.0000] <zcorpan> /me opens "The End" [04:06:06.0000] <hsivonen> ha. at least Gecko isn't randomly non-compliant. It's consistent with IE: http://hsivonen.iki.fi/test/moz/document-open-during-network-parse.html [04:06:34.0000] <hsivonen> Gecko and Trident don't return document from document.open() contrary to the spec [04:06:44.0000] <hsivonen> WebKit and Presto return document per spec [04:09:27.0000] <zcorpan> jgraham: have a syntax error in a dedicated worker and listen for worker.onerror [04:09:46.0000] <zcorpan> though that loads a resource too eh [04:10:39.0000] <zcorpan> jgraham: <input autofocus> [04:11:02.0000] <hsivonen> hmm. the spec seems bogus in this case: http://hsivonen.iki.fi/test/moz/document-open-during-script-created-parse.html [04:12:01.0000] <zcorpan> jgraham: or the storage event [04:12:23.0000] <hsivonen> never trust the spec [04:12:28.0000] <hsivonen> always write test cases anyway [04:12:29.0000] <hsivonen> sigh [04:13:23.0000] <zcorpan> jgraham: or use <marquee> [04:13:34.0000] <jgraham> marquee?! [04:13:43.0000] <zcorpan> yep [04:14:07.0000] <zcorpan> the start event [04:14:18.0000] <zcorpan> and finish [04:15:18.0000] <jgraham> Wow [04:15:27.0000] <jgraham> Every day's an adventure in HTML land [04:16:08.0000] <jgraham> That looks perfect, and delightfully bizzare. Thank you! [04:16:22.0000] <zcorpan> np [04:46:00.0000] <zcorpan> odinho: isn't this the wrong way around? [04:46:01.0000] <zcorpan> <img src="fallback..." srcset="logo-150px.jpg, logo-50px.jpg 600w"> [04:46:40.0000] <odinho> zcorpan: img without anything will get Infinite w and Infinite h. [04:46:43.0000] <zcorpan> logo-50px.jpg will be used when the viewport is 600px or *bigger* [04:47:35.0000] <zcorpan> also you need at least one descriptor, or the comma will be part of the url [04:49:00.0000] <zcorpan> so it should be <img src="fallback..." srcset="logo-150px.jpg 600w, logo-50px.jpg 1x"> or <img src="logo-50px.jpg" srcset="logo-150px.jpg 600w"> [04:49:45.0000] <odinho> zcorpan: Ah, very true. [04:50:21.0000] <zcorpan> (the syntax doesn't support the case where you want to use the 150 image as the fallback, afaict) [04:50:53.0000] <zcorpan> maybe there should be an attribute to give descriptors for the src attribute [04:51:40.0000] <zcorpan> or have a max-width descriptor [04:51:47.0000] <odinho> zcorpan: I did rock around that one a few times also, wondering if I should write something correct or just small and easy and only to the point. But yes, it'd be all wrong. Anyway, that is written wrong, but it can be done like I say anyway. Just look past the error :P [04:52:08.0000] <odinho> zcorpan: I don't understand. [04:52:15.0000] <odinho> zcorpan: I think it can support anything as fallback. [04:52:49.0000] <zcorpan> the fallback is a candidate, but you can't change its descriptors [04:53:08.0000] <zcorpan> let's say you want to have 600w for the fallback [04:54:01.0000] <zcorpan> having two candidates with the same descriptors is invalid [04:54:48.0000] <zcorpan> (so <img src="fallback..." srcset="logo-150px.jpg 600w, logo-50px.jpg 1x"> is invalid actually) [04:55:04.0000] <zcorpan> (since fallback... and logo-50px.jpg have the same descriptors) [04:55:30.0000] <odinho> Hmm. Either you or I am wrong here. srcset="logo-150px.jpg 600w, logo-50px.jpg 1x", I would've written srcset="logo-150px.jpg 1x, logo-50px.jpg 600w". [04:55:39.0000] <odinho> zcorpan: I don't see why that is valuable. [04:57:24.0000] <odinho> The attribute essentially takes a comma-separated list of URLs each with one or more descriptors giving the maximum viewport dimensions and pixel density allowed to use the image. [04:57:33.0000] <odinho> </quote> from spec. [04:57:52.0000] <odinho> ... To specify an image, give first a URL, then one or more descriptors of the form 100w, 100h, or 2x, where "100w" means "maximum viewport width of 100 CSS pixels", [04:58:50.0000] <zcorpan> "If there are any entries in candidates that have an associated width that is greater than max width, then remove them," [04:59:01.0000] <zcorpan> so if the viewport is 400px, the 600w candidate is removed [04:59:12.0000] <odinho> That's conflicting. [04:59:20.0000] <zcorpan> odinho: what you quoted might be wrong :-) [04:59:36.0000] <odinho> Yes, or the steps in the algorithm might be. :P [05:00:00.0000] <zcorpan> Hixie: ^ [05:00:01.0000] <hsivonen> maybe we should have the <picture><source> thing for the "art directed" axis and srcset="foo.jpg 2x" for the different pixel samplings of the same image [05:00:15.0000] <odinho> hsivonen: Or punt, art directed. As I want. [05:00:33.0000] <odinho> hsivonen: Have a way to say "don't do prefetch load" maybe. -- So that it can be done in javascript. [05:00:47.0000] <zcorpan> what's clear with the current syntax is that even people who read the spec get it wrong [05:00:50.0000] <zcorpan> that's bad [05:01:09.0000] <zcorpan> at least the width/height descriptors [05:01:28.0000] <hsivonen> zcorpan: yeah. it can't stay in the spec as written [05:01:58.0000] <hsivonen> It's terribly confusing that Nw and Mh don't refer to the image file's characteristics [05:02:07.0000] <zcorpan> yes [05:02:27.0000] <zcorpan> and even if you know it refers to the viewport, it's not clear if it's max-width or min-width [05:03:09.0000] <hsivonen> that, too [05:05:39.0000] <odinho> zcorpan: Hmmmm. Yeah, now I wonder more, what is supposed to be correct - the detailed spec, or the overview... [05:05:43.0000] <odinho> zcorpan: I'll go back to the source. [05:07:45.0000] <odinho> <img alt="Obama spoke at the factory." src="factory.jpeg" srcset="obama-factory-face.jpeg 500w"> [05:08:21.0000] <odinho> If viewport 1000px. factory Infinity is greater than 1000. So that one is removed. [05:08:41.0000] <odinho> The spec algorithm steps are reversed/wrong then. Based on Hixie's example in his email. [05:09:17.0000] <zcorpan> i'll fire off an email [05:10:22.0000] <odinho> This one: If there are any entries in candidates that have an associated width that is greater than max width, then remove them, unless that would remove all the entries, in which case remove only the entries whose associated width is greater than the smallest such width. [05:10:54.0000] <odinho> Should probably say "associated width that is less than max width". (or less than or the same, or similar). [05:11:11.0000] <odinho> As far as I've understood it. [05:13:49.0000] <jgraham> It has to be min width [05:14:17.0000] <jgraham> Uh, have I confused myself [05:15:14.0000] <jgraham> Well I guess it could work either way, depending on what you want the behaviour of src to be [05:15:40.0000] <jgraham> The important point is that it should be one or the other though [05:15:52.0000] <jgraham> So you can't create gaps [05:16:19.0000] <jgraham> So yeah, maybe a slightly more verbose syntax would help [05:17:00.0000] <odinho> zcorpan: I strongly believe it should be min-width, and that was the intended thing, that only the spec steps got it wrong. [05:17:25.0000] <jgraham> odinho: I thought the opposite. Why do you strongly believe that? [05:18:05.0000] <odinho> jgraham: Because of all the code hixie has been giving makes sense when it is that, all code I've written assumes that as well. [05:18:16.0000] <zcorpan> odinho: you mean max-width? :) [05:18:44.0000] <jgraham> It's probably clearer not to talk in terms of media query properties [05:18:48.0000] <odinho> zcorpan: I mean, steps should say "associated width that is less than max width". [05:18:58.0000] <odinho> zcorpan: I don't know if that is min- og max- tbh :P [05:19:16.0000] <odinho> yeah probably max when you say it like that. [05:19:20.0000] <odinho> Oh man [05:19:26.0000] <jgraham> odinho: Do you think that an image with 600w shoud load if the viewport is wider than 600px or narrower, given sutiable alaternatives? [05:20:02.0000] <odinho> jgraham: 600w, it picks that picture if the viewport is 600 or less. [05:20:23.0000] <jgraham> And you think that is the right behaviour? [05:20:40.0000] <jgraham> (I agree that is what the spec says, and I agree it makes sense) [05:21:05.0000] <jgraham> Because it means that the @src image is always the biggest one [05:21:23.0000] <jgraham> Which seems like the right fallback behaviour [05:21:46.0000] <zcorpan> (except it's not what the spec says in the algorithm) [05:22:21.0000] <odinho> And that's our confusion ;-) [05:23:50.0000] <jgraham> It's not? [05:24:23.0000] <odinho> jgraham: Nope. Spec is contradicting itself in the overview-text, and the actual steps. As far as me an zcorpan has been able to see. [05:24:48.0000] <zcorpan> http://krijnhoetmer.nl/irc-logs/whatwg/20120516#l-1182 [05:24:48.0000] <annevk> there's a bug on the note already [05:24:56.0000] <jgraham> If the viewport is narrower than the width descriptor, you remove the image from the list of candidates [05:25:05.0000] <jgraham> /me isn't looking at the note [05:25:14.0000] <jgraham> it's in green -> irrelevant :) [05:25:29.0000] <odinho> annevk: Shouldn't be a bug on the note, should be in the steps. IMHO. [05:25:39.0000] <odinho> annevk: But the bug goes for the same anyway ;-) [05:26:10.0000] <odinho> annevk: Where is it? [05:26:41.0000] <jgraham> ...which doesn't match what you expect [05:26:42.0000] <jgraham> OK [05:34:29.0000] <annevk> odinho: https://www.w3.org/Bugs/Public/show_bug.cgi?id=17057 [05:35:33.0000] <smaug____> ++foolip_ [05:39:08.0000] <annevk> smaug____: context? [05:41:27.0000] <smaug____> annevk: reviewing a badly defined spec [05:41:34.0000] <smaug____> (WebAudio) [05:50:28.0000] <adactio> I'm glad it's not just me who was confused by what's in the spec. [05:54:15.0000] <adactio> Also, this text in the spec is very presumptuous: " This attribute allows authors to provide alternative images for environments with smaller screens..." It could equally well be used to provide alternative images for environments with *larger* screens (in fact, using the "Mobile First" approach, it should be). [05:55:55.0000] <odinho> adactio: Heh, don't get too worked up over that. ;-) [05:56:26.0000] <adactio> odinho: I'm not getting worked up. I'm just pointing out an unwarranted assumption in the spec text. [05:57:16.0000] <Kolombiken> Better? " This attribute allows authors to provide alternative images for different screens in different environments…" [05:57:17.0000] <Kolombiken> No assumptions [05:57:21.0000] <odinho> adactio: It's only explanation text (non-normative), it's not "real" part of the spec. [05:57:25.0000] <odinho> Kolombiken: Yeah. [05:57:35.0000] <foolip_> smaug____, you rang? [05:57:51.0000] <adactio> I thought that non-normative text was in green. [05:58:08.0000] <adactio> Kolumbiken: Your rewording sounds good. [05:58:10.0000] <odinho> adactio: I didn't actually check where it was, just sounded super non-normative. [05:58:39.0000] <adactio> odinho: That was why I brought it up: I would expect to see use-case sentences like that in the non-normative text. [05:58:40.0000] <Kolombiken> adactio: Thanks :) [05:59:17.0000] <odinho> adactio: Ah, ohwell. It's easy to fix. [05:59:31.0000] <slikts> is this the what what in the butt group? [05:59:58.0000] <odinho> Kolombiken: It doesn't speak about different screen sizes, or different pixel densities though. [06:01:10.0000] <Philip`> adactio: Anything that isn't phrased with "must"/"should"/etc, and that isn't a definition used by such phrases, is non-normative regardless of typography [06:01:29.0000] <Kolombiken> odinho: Which one doesn't? [06:01:31.0000] <adactio> Philip: Okay. Got it. [06:01:49.0000] <jgraham> The sentence seems right per spec. The img @src will always be used on infinitley big screens and the srcset allows you to provide images for smaller screens [06:02:05.0000] <odinho> jgraham: Ah, that's very true. [06:02:10.0000] <Philip`> adactio: (according to http://ln.hixie.ch/?start=1140242962&count=1) [06:02:15.0000] <jgraham> There is some confusion about whether this is what is intended, but it is a statement of fact [06:02:35.0000] <jgraham> i.e. neither normative nor non-normative [06:02:52.0000] <jgraham> well strictly non-normative [06:02:57.0000] <jgraham> But different to a note [06:06:37.0000] <adactio> So the example in the spec is wrong, no? srcset="banner-phone.jpeg 100w" will be served up to every device with a viewport width greater than 100 pixels, over-riding the image in the src attribute (where the intention was to only serve it up to small screens). So src="banner-phone.jpeg" srcset="banner.jpg 100w" would actually be correct. [06:07:14.0000] <odinho> adactio: Nope, opposite. [06:07:25.0000] <odinho> adactio: Although you've hit the bug we've been talking about. :-) [06:07:27.0000] <jgraham> adactio: The spec is buggy at the moment [06:07:32.0000] <adactio> odinho: That is *extremely* worrying. [06:07:33.0000] <odinho> adactio: Because the spec says two things about that right now. [06:07:45.0000] <jgraham> There is no point in discussing this until the spec bug is fixed [06:08:02.0000] <adactio> So is the plan to effectively make it impossible to mobile-first development? [06:08:22.0000] <odinho> adactio: Nope. [06:09:20.0000] <odinho> <img src="mobile-first.jpg" srcset="mobile-first.jpg 400w, normal.jpg 1000w, supersizemeomg 4069w"> [06:09:24.0000] <Ms2ger> /me yawns [06:09:52.0000] <odinho> adactio: granted, if you have a bigger screen than 4069px now, you'll get the mobile image :P -- But you can set it very high if you want. [06:10:11.0000] <adactio> That. Sucks. [06:10:21.0000] <adactio> Seriously. That's fucking awful. [06:10:24.0000] <Ms2ger> Yes [06:10:36.0000] <Ms2ger> Did you miss the part where there is a spec bug? [06:10:47.0000] <Ms2ger> Did you also miss the idea that bugs can get fixed? [06:11:10.0000] <odinho> Well, I want that behaviour I said now. [06:11:54.0000] <adactio> Ms2ger: But this isn't about the spec bug, this about the desired behaviour. [06:12:33.0000] <Ms2ger> Go and send a polite comment to the whatwg list, then [06:12:38.0000] <adactio> Ms2ger: The spec bug will get resolved to make it clearer what the intended use is. And the intended use is desktop-first. [06:12:49.0000] <adactio> Ms2ger: Oh, I plan to. [06:14:18.0000] <Lachy> gsnedders, Ms2ger, in Anolis, how do I get the --dump-xrefs argument to work? It seems to just hang when I try running: $ anolis --dump-xrefs=data/xrefs/output.json Overview.html [06:14:26.0000] <jgraham> I don't understand foo-first [06:14:31.0000] <jgraham> It's src-first [06:15:50.0000] <wingo> greets [06:16:00.0000] <adactio> jgraham: It's a question of whether you assume a large-screen environment by default (in the src attribute) that gets over-ridden for small-screen environments (in the srcset attribute) vs. assuming a small-screen environment by default (in the src attribute) which gets over-ridden for large-screen environments (in the srcset attribute). [06:16:16.0000] <wingo> are there html5lib folks here? where do I send a patch? [06:16:36.0000] <adactio> It sounds like the plan is to go with an assumption of large-screen. That is very, very troubling. [06:17:58.0000] <jgraham> adactio: Rather than talking about "the plan" being "very very troubling" why not send technical feedback on how the existing solution has a problem meeting your usecase, phrased in a polite way that doesn't assume a conspiracy to screw you over? [06:18:38.0000] <adactio> jgraham: I'm writing an email right now. And I didn't say anything about conspiracies. [06:18:48.0000] <adactio> jgraham: Physician, heal thyself. [06:19:01.0000] <jgraham> wingo: Yes. Attach it to an issue, perhaps? [06:19:33.0000] <adactio> jgraham: And what wording should I use rather than "the plan"? "desired behaviour"? Seriously, I'd like to know. [06:20:08.0000] <Ms2ger> Lachy, does data/xrefs/output.json exist? [06:21:22.0000] <jgraham> adactio: "The srcset draft" [06:21:23.0000] <Lachy> no, should I create it first? [06:21:36.0000] <Ms2ger> Yes [06:21:44.0000] <Ms2ger> On which revision are you? [06:21:44.0000] <Lachy> empty file or with something in it? [06:21:46.0000] <adactio> jgraham: That describes the thing, not the intention. [06:21:57.0000] <Lachy> anolis --version says 1.2pre [06:21:58.0000] <jgraham> adactio: Yes. [06:22:01.0000] <Ms2ger> { "url": "foo#" } [06:22:17.0000] <Lachy> and I just updated html5lib to the most recent version too 0.95 [06:22:26.0000] <wingo> jgraham: ah, hum; just mailed it to the list [06:22:32.0000] <wingo> issue is probably better [06:22:33.0000] <Ms2ger> Lachy, well, it's the HG changeset I'm looking for :) [06:22:48.0000] <wingo> modulo the obligatory grumbling about google accounts :) [06:22:49.0000] <adactio> jgraham: I need a phrase to describe the approach that "the srcset draft" is taking. The approach. The plan. The desired behaviour. [06:22:59.0000] <jgraham> adactio: No you don't [06:23:08.0000] <bjankord> Yeah he does [06:23:09.0000] <adactio> jgraham: Yes, I do. [06:23:13.0000] <bjankord> We all do [06:23:16.0000] <jgraham> You need to explain how the current draft fails to meet your use case [06:23:29.0000] <adactio> jgraham: Okay. Finally. [06:24:24.0000] <bjankord> With srcset, if the viewport width changes will the img src change? [06:25:12.0000] <Lachy> Ms2ger, ok, well, I should have the most recent changes. I downloaded and installed a fresh copy from the repo yesterday. [06:25:20.0000] <Ms2ger> Ah, hmm [06:25:35.0000] <Ms2ger> Because that version should have a useful error message... [06:25:58.0000] <Lachy> anyway, even after creating the json file, I get the same result. It just sits there doing nothing until I kill it a few minutes later with Ctrl-C [06:26:36.0000] <wingo> jgraham: http://code.google.com/p/html5lib/issues/detail?id=205 [06:27:35.0000] <jgraham> wingo: Oh, nice, thanks! [06:27:41.0000] <Ms2ger> Oh [06:28:51.0000] <wingo> you're welcome :) [06:28:53.0000] <Lachy> it also seems to make no difference regardless of which input file I use, whether I use Overview.src.html or Overview.html. I even tried using a copy of the whatwg spec. [06:28:58.0000] <Ms2ger> Lachy, you need something like anolis --dump-xrefs=data/xrefs/output.json Overview.src.html /tmp/foo [06:29:13.0000] <Lachy> what's the /tmp/foo path for? [06:29:53.0000] <bjankord> ... With srcset, if the viewport width changes, will the img src update? [06:30:03.0000] <bjankord> For example, with the following markup: <img alt="The Breakfast Combo" [06:30:03.0000] <bjankord> src="photo-mobile-portrait.jpeg" [06:30:03.0000] <bjankord> srcset="photo-mobile-portrait.jpeg 320w, photo-mobile-landscape.jpeg 480w"> [06:30:20.0000] <Lachy> oh, I see. It needs an output file. [06:30:38.0000] <Ms2ger> Yeah [06:30:53.0000] <bjankord> Would the img element update to use the photo-mobile-landscape on orientation change from portrait to landscape? [06:31:08.0000] <Ms2ger> I guess I can probably make it work without one, but I'm not sure how much refactoring it would need [06:31:25.0000] <samuelfine> bjankord: That's a good question. Not sure of the answer, but it makes me wonder how the browser will handle loading or pre-loading those srcset images. [06:32:00.0000] <samuelfine> bjankord: If the src does update, will there be a FOUC-esque load of the new image? Etc. [06:32:30.0000] <Lachy> thank you. [06:32:35.0000] <bjankord> I believe preloading/prefetching is a valid concern, one that sould be considered in any solution. [06:33:08.0000] <Ms2ger> I've made it throw an exception for now [06:33:26.0000] <bjankord> As far as FOUC, there are some JS based solutions that swap out images in the src attribute that do not seem to have noticeable FOUC [06:34:04.0000] <odinho> bjankord: If you read the spec, you'll find the section where the algorithm for updating is. [06:34:24.0000] <odinho> bjankord: It is specced, but browsers doesn't *have* to do it when the environment changes. But they are free to do so. [06:35:07.0000] <odinho> bjankord: So a desktop browser will most probably do it. -- Whilst a phone might opt to save the bandwidth required to do a new fetch. [06:35:22.0000] <bjankord> Ah I see [06:35:30.0000] <bjankord> Yeah I'm reading through - http://www.whatwg.org/specs/web-apps/current-work/multipage/embedded-content-1.html#attr-img-srcset [06:36:50.0000] <odinho> bjankord: The algorithms are a bit further down. http://www.whatwg.org/specs/web-apps/current-work/multipage/embedded-content-1.html#processing-the-image-candidates and then scroll down the full first algorithm, down to: "The user agent may at any time run the following algorithm to update an img element's image in order to react to changes in the environment. (User agents are not required to ever run this algorithm.)" [06:37:38.0000] <bjankord> odinho: Thanks! [06:39:36.0000] <samuelfine> Yes, thanks odinho. Looks like FOUC shouldn't be an issue, images aren't replaced until new images are fully downloaded. [06:40:45.0000] <odinho> samuelfine: But it might flash over to the new image. If it doesn't fade-between or something. Guess that's QoI (Quality of Implementation), something user agents can compete on to make look nice. [06:42:35.0000] <samuelfine> odinho: I would hope browsers would drop in new images over old ones before removing, but you're right. Up to the implementers to decide. [06:43:28.0000] <bjankord> samuelfine: That does sounds like the best way to handle the image swap, if/when it happens. [06:43:51.0000] <odinho> samuelfine: Oh, I think noone would leave a gap inbetween! It's possible yes, but why would you? :P I was thinking about just doing a fast swap, or a swap with a 0.2s dissolve between the images. [06:45:41.0000] <samuelfine> I try to avoid assuming any browser will default to the Right Way. ;) [06:46:18.0000] <zewt> dear world: please stop starting your own threads when we have 8 already for the same topic [06:48:01.0000] <wingo> jgraham: http://code.google.com/p/html5lib/issues/detail?id=206 as well [06:48:21.0000] <odinho> zewt: Haha. Yeah. [06:48:28.0000] <wingo> i was going to make planet venus add preload=none to all videos, but it seems not so straightforward [07:01:17.0000] <Ms2ger> Oh look, Gecko supports <source media> now [07:02:43.0000] <odinho> Ms2ger: Hehe. [07:03:38.0000] <Ms2ger> I bet Opera did it first [07:04:15.0000] <odinho> Ms2ger: ^^ How much? [07:04:33.0000] <odinho> :P [07:04:42.0000] <Ms2ger> Hmm? [07:06:16.0000] <odinho> Ms2ger: Choke up the money, lad ;] [07:06:44.0000] <Ms2ger> Oh, heh [07:06:55.0000] <Ms2ger> 10000 dollars? ;) [07:07:08.0000] <odinho> Okay! :D I'll go in on that oone. [07:07:10.0000] <Ms2ger> /me is secretly running for US president [07:11:12.0000] <zcorpan> opera indeed did do it first :-) [07:11:16.0000] <zcorpan> but now i want to drop it :-P [07:13:16.0000] <Ms2ger> Too late ;) [07:13:42.0000] <odinho> WAT! [07:13:49.0000] <odinho> I didn't know we had done it. [07:14:09.0000] <odinho> Ms2ger: Good thing it was Monopoly dollars! Now I just need your address to send it to you. [07:14:17.0000] <odinho> :P [07:14:24.0000] <Ms2ger> Hah [07:17:55.0000] <zcorpan> odinho: we've supported it from the start (10.50) [07:18:51.0000] <odinho> zcorpan: Yeah, saw that now. Way before my time. [07:19:22.0000] <Ms2ger> Newb ;) [07:19:25.0000] <odinho> :( [07:20:02.0000] <odinho> :D [07:20:14.0000] <odinho> (I just couldn't let that negative hang there...) [07:20:43.0000] <Ms2ger> http://xkcd.com/859/ [07:21:25.0000] <odinho> Ms2ger: Yeah, it does feel bad. [07:33:59.0000] <zcorpan> odinho: gotta love when someone top-post replies to your email asking not to top-post [07:36:19.0000] <odinho> zcorpan: lol, haven't seen it yet. [07:36:35.0000] <odinho> zcorpan: It may be like a "I really hate you" [07:37:51.0000] <zcorpan> you replied to it, so you must have seen it :-P [07:39:37.0000] <odinho> zcorpan: Oh, I forgot that one. I asked for a new one now. And Matt did inline reply then. And Mike Taylor answered him with the guidelines, -- and then Matt answered back with a top-post. Which was kinda funny. :P [07:40:16.0000] <odinho> zcorpan: Oh, -- yes, I also asked for non-top-posting in the email *before* that. [07:40:50.0000] <zcorpan> yeah [07:52:50.0000] <zcorpan> step 6 "If url is empty, then jump to the step labeled descriptor parser." - if url is empty, doesn't that mean that you have reached the end of the attribute value? [07:57:30.0000] <annevk> zcorpan: just means you hit some space characters [07:57:43.0000] <annevk> well one [08:05:46.0000] <zcorpan> julian's concern with the comma seems valid. what can we use instead? newline? [08:06:02.0000] <zcorpan> the examples seem to use newline anyway for readability [08:06:38.0000] <hober> newline is way worse than comma [08:06:51.0000] <zcorpan> why? [08:07:14.0000] <annevk> what's wrong with comma? [08:07:32.0000] <jgraham> Newline is awful [08:07:47.0000] <zcorpan> annevk: urls can contain comma [08:08:00.0000] <ShaneHudson> newline is fine for presentation but remember it will often be minimised [08:08:06.0000] <annevk> zcorpan: sure [08:08:29.0000] <ShaneHudson> Use comma but add a new line as syntax style rather than syntax itself [08:15:41.0000] <hober> the comma comes after the scale factor, which isn't a url, and before the next url. if you think your urls might start with a comma, use whitespace. [08:17:26.0000] <zcorpan> hober: i've seen one example where the descriptors were all omitted. that's invalid, sure, but it doesn't mean that the current syntax is golden [08:20:12.0000] <jgraham> Should have used JSON in the attribute value. srcset = "[{'url':'foo.jpg', 'height':200, width:20, resolution:2}]" everyone loves JSON [08:20:35.0000] <jgraham> Note: this was not a serious suggestion [08:20:52.0000] <hober> zcorpan: indeed. i believe i listed "new microsyntax" as a drawback of the design in my proposal :) [08:20:57.0000] <smaug____> annevk: so, if a new event interface doesn't have init*Event, should one be able to create such event using createEvent() ? [08:21:24.0000] <annevk> as per spec yes [08:21:31.0000] <zcorpan> hober: the newness isn't a problem in itself :-) [08:21:43.0000] <annevk> smaug____: I'm open to changing that either way [08:21:44.0000] <smaug____> annevk: why ? [08:21:47.0000] <smaug____> annevk: yeah [08:22:03.0000] <annevk> smaug____: well the spec is generic and doesn't list any interfaces [08:22:09.0000] <smaug____> annevk: if the event doesn't have init*, createEvent is useless [08:22:28.0000] <annevk> it will still have initEvent() [08:22:35.0000] <annevk> but sure [08:22:36.0000] <smaug____> well, sure [08:23:23.0000] <Ms2ger> jgraham, duh, should have used XML [08:25:02.0000] <odinho> <img src=ohman.jpg srcset="<srcset><img src='foo.jpg' height='1000' width='1000' resolution='2'></srcset>"> [08:25:24.0000] <odinho> OH MY GOD I FORGOT TO SELFCLOSE /> [08:26:06.0000] <Philip`> Got to use &lt; instead of < to make sure it's polyglottal [08:27:58.0000] <odinho> Philip`: Oh man, totally forgot - good you reminded me. It's going to be awsum. [08:28:10.0000] <jgraham> Oooh, there's a meme for this situation [08:28:25.0000] <odinho> <img src="ohman.jpg" srcset="&lt;srcset>&lt;img src='foo.jpg' height='1000' width='1000' resolution='2' />&lt;/srcset>" /> [08:28:35.0000] <odinho> There! what a beautify! spec it up [08:30:10.0000] <zcorpan> aha! <source> in srcset=""! [08:30:16.0000] <zcorpan> everybody wins! [08:30:40.0000] <odinho> why didn't I think of that :-/ [08:31:42.0000] <annevk> needs more media query [08:32:19.0000] <odinho> MEDIA QUERY ALL THE THINGS! [08:32:23.0000] <odinho> o/ [08:33:25.0000] <samuelfine> Refresh my memory as to why this isn't a job for media queries? That seems like the most obvious solution to me. I'm probably missing something though. [08:33:55.0000] <Philip`> odinho: Make srcset conditional on a media query, like <img srcset-min-width-800px="&lt;srcset>..." srcset-min-width-400px="..."> [08:34:59.0000] <zcorpan> samuelfine: http://www.xanthir.com/blog/b4Hv0 [08:37:05.0000] <samuelfine> None of the proposals gracefully manage bandwidth concerns, though. [08:38:05.0000] <zcorpan> bandwidth concerns is up to the UA to deal with in the srcset proposal [08:38:37.0000] <zcorpan> there was a suggestion to add byte size information, but no suggestion on what to do with that information [08:38:53.0000] <samuelfine> That's certainly a good reason to avoid using them *now*, but perhaps we should consider improving media queries instead of adding brand new attributes? [08:39:33.0000] <samuelfine> I don't know how this process works, I'm just a lowly developer. It just feels like an overwrought solution to a problem that's already mostly solved. [08:39:37.0000] <dglazkov> good morning, Whatwg! [08:39:59.0000] <odinho> samuelfine: media has conflict with how img is loaded. [08:41:01.0000] <samuelfine> odinho: I'm not sure what you mean. [08:43:19.0000] <odinho> samuelfine: Maybe start with http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2012-May/035886.html -- and also all the other emails in that thread. There's stuff scattered around in that list. [08:44:25.0000] <samuelfine> odinho: Thanks. I'm still new to this mailing list thing, and trying to grok the site gives me a clot in my brain. [08:45:22.0000] <odinho> samuelfine: It's essential to have a good mail reader. -- I started using Opera M2 again half a year ago, and am very happy I've done that now. It's really nice to have such a good overview over complex discussions. [08:46:14.0000] <samuelfine> odinho: @opera.com recommending Opera Mail? I'm shocked! (Seriously though, I'll check it out. Mail for OS X is the worst.) [08:46:55.0000] <odinho> samuelfine: lol, -- I don't recommend stuff I don't use. I don't use Opera IRC, I use irssi. [08:47:40.0000] <ShaneHudson> I use Sparrow Mail, it is very good [08:48:16.0000] <ShaneHudson> Biggest problem is that I signed up with my email address that also goes to my phone.. not a good idea [08:48:43.0000] <odinho> People in Opera choose their own mail reader, so people use all sorts of different readers. Except Outlook which is banned. [08:49:01.0000] <ShaneHudson> Haha [09:09:18.0000] <odinho> foolip_: Funny email :-) [09:09:32.0000] <foolip_> odinho, which one? [09:09:46.0000] <odinho> foolip_: WHATWG one. <source media> [09:09:50.0000] <foolip_> oh :) [09:10:06.0000] <odinho> I should just stop writing emails in those threads. They just keep growing when I do it :-/ [09:10:34.0000] <foolip_> You can never shrink a thread by replying to it, that is true... [09:10:44.0000] <odinho> So bad, I hate that. [09:10:49.0000] <annevk> I think part of the reason we had <source media> is because the idea was that media queries would be used to determine whether or not the user would want to see subtitles and such by default [09:11:02.0000] <annevk> I think since then we realized that would be way complex and invented something more clever [09:12:37.0000] <foolip> annevk, I didn't search the email archives around that time, do you remember where it was discussed? [09:14:54.0000] <odinho> ShaneHudson: lol, go write that book. [09:15:28.0000] <annevk> foolip: now we have http://lists.w3.org/Archives/Public/public-whatwg-archive/ you can search in both lists at the same time [09:15:42.0000] <annevk> foolip: David Singer brought this up a few times [09:16:25.0000] <annevk> foolip: but given the date of revision 724 this should be easy enough to figure out [09:16:45.0000] <ShaneHudson> odinho: haha I want to write a book somewhen tbh [09:20:29.0000] <foolip> annevk, http://lists.w3.org/Archives/Public/public-whatwg-archive/2007Jun/0126.html is the oldest I can find, but that's after the commit [09:22:28.0000] <annevk> foolip: cannot find the follow up to http://lists.w3.org/Archives/Public/public-whatwg-archive/2007Apr/0059.html [09:23:23.0000] <foolip> I guess we could just ask singer and hixie, if we really wanted to know. [09:24:20.0000] <annevk> did find some youthful optimism http://lists.w3.org/Archives/Public/public-whatwg-archive/2007May/0088.html [09:25:01.0000] <foolip> :) [09:25:26.0000] <foolip> I also found a mail where Lachlan argued something slightly weird about Theora. [09:27:43.0000] <annevk> http://lists.w3.org/Archives/Public/public-whatwg-archive/2007Mar/0663.html mentions media queries [09:33:07.0000] <tantek> annevk - you're like a whatwg list human-search-engine [09:34:02.0000] <Ms2ger> foolip, eww, git :) [09:43:46.0000] <ShaneHudson> adactio: Very good article by the way! (http://adactio.com/journal/5474/) [09:44:32.0000] <adactio> ShaneHudson: thanks. [09:44:49.0000] <tantek> jgraham, FYI, the link you found on that W3C wiki login page has been fixed to link to something more understandable (same as the link on the W3C wiki home page) - for signing up for new accounts: http://www.w3.org/wiki/index.php?title=Special:UserLogin&returnto=Main_Page - thanks to brett in irc://irc.w3.org:6665/sysreq [09:54:37.0000] <foolip> adactio, <picture><source> doesn't mimic the behavior of video at all, the video resource selection algorithm uses the first source that passes all the tests and then never changes [09:55:32.0000] <foolip> but it's not surprising that many people make this mistake, given the complexity of the resource selection algorithm. [09:55:33.0000] <adactio> foolip: A moot point at this juncture, wouldn't you say? [09:56:48.0000] <adactio> foolip: Also, I was talking about how <picture> mimicked the *syntax* of <video>, not necessarily the behaviour. But point taken. [09:57:26.0000] <foolip> adactio, just thought you might want to know :) [09:58:46.0000] <foolip> It's also probably not true that "using media queries on the sources of videos has proven to be very tricky for implementors" but I haven't read all the email so I don't know who claimed that. [10:06:11.0000] <jmather> tantek: wrote up my idea on the CG. Maybe I'm crazy. :D [10:07:44.0000] <annevk> foolip: that's probably in part because for <video> it goes only one way, and when evaluation of the media query changes, the resource does not change [10:08:15.0000] <annevk> foolip: which is quite different from how people expect it to work for <picture> (although nobody has quite articulated how it would work I think) [10:08:25.0000] <foolip> Yes, but it's easy to implement, it just isn't useful. [10:08:41.0000] <foolip> at least it was easy in Opera, or I wouldn't have done it. [10:09:01.0000] <foolip> /me counts to 6 lines of code to support it [10:09:23.0000] <annevk> yeah at this point I have a hard time seeing why it was added too [10:09:31.0000] <ShaneHudson> Talking about implementing, is there any ooks or indepth articles for getting started with browser development? Or is it just, clone the source and jump in? [10:10:08.0000] <foolip> ShaneHudson, unfortunately you can't clone Opera, but I'd presume Gecko and WebKit start in that way :) [10:11:31.0000] <jmather> The more I think about it the more I realize MQ's aren't the right way, even if the picture style syntax is/might be/whatever. As has been said, re-evaluation is/would be problematic, plus, the media query selector that's really needed doesn't exist. [10:12:33.0000] <ShaneHudson> foolip: Yes, I wouldn't expect to be able to jump into a paid position! :) [10:20:40.0000] <tantek> ShaneHudson, reporting bugs to various browser's bug databases is one way of getting noticed. [10:26:30.0000] <othermaciej> ShaneHudson: WebKit is pretty approachable [10:26:37.0000] <othermaciej> http://www.webkit.org/building/checkout.html [10:26:41.0000] <othermaciej> http://www.webkit.org/building/build.html [10:29:01.0000] <tantek> well if we're dropping URLs ;) [10:29:10.0000] <tantek> http://www.mozilla.org/contribute [10:29:39.0000] <tantek> https://developer.mozilla.org/en/Introduction [10:29:56.0000] <othermaciej> ShaneHudson: finding and filing bugs is also a great way to get involved, as mentioned by tantek [10:30:35.0000] <annevk> dropping URLs? http://opera.jobs/ [10:38:24.0000] <tantek> ShaneHudson, summary for your convenience: http://wiki.whatwg.org/wiki/Getting_started_with_browser_development [10:39:07.0000] <tantek> /me wasn't sure if that was ontopic for the WHATWG FAQ or not so didn't add it - if someone more familiar with the FAQ feels it is ontopic enough, feel free to add it. [11:03:36.0000] <MikeSmith> jgraham: a couple weeks there was something you said I should add the platform.html5.org, but I have since forgotten what it was [11:08:02.0000] <annevk> MikeSmith: should prolly drop From-Origin at this point [11:08:12.0000] <MikeSmith> hai [11:08:44.0000] <MikeSmith> done [11:09:23.0000] <othermaciej> annevk: is From-Origin abandoned? [11:12:40.0000] <Ms2ger> 'parently [11:12:48.0000] <othermaciej> should DNT be added to platform.html5.org maybe? [11:12:50.0000] <othermaciej> or too soon? [11:13:22.0000] <jwalden> /me wonders if there's some sort of conference happening that caused #whatwg to be abnormally active yesterday [11:13:56.0000] <othermaciej> jwalden: the ongoing responsive image discussion [11:14:47.0000] <Ms2ger> jwalden, you don't want to read it [11:14:54.0000] <Ms2ger> It's trolling all over [11:15:00.0000] <MikeSmith> othermaciej: I will add DNT. It seems clear it's going to become part of the platform in some form [11:15:03.0000] <jwalden> yeah, I skimmed backlogs super-heavy [11:15:23.0000] <jwalden> Ms2ger: I hear that Ms2ger troll is especially worth not reading [11:15:38.0000] <Ms2ger> I wasn't around, fortunately [11:15:42.0000] <jwalden> haha [11:15:52.0000] <jmather> Ms2ger: i think a lot of it is just more mis-understanding the sides honestly. At first I didn't entirely get where the srcset proposal was coming from [11:16:25.0000] <jmather> i still don't particularly get it or agree with it, but I get you guys have your reasons… I'm now firmly in the camp that everyone is wrong and solving the problem the wrong way, so maybe i'm jus the odd one out. :D [11:17:10.0000] <othermaciej> jmather: do you have a specific better proposal in mind, or better yet, an important use case that is not well addressed by other proposals? [11:17:35.0000] <hober> jmather: i think it's fair to say that there is no, single "the problem" [11:17:42.0000] <jmather> http://www.w3.org/community/respimg/2012/05/16/shouldnt-we-be-defining-content-not-context/ I tried to work in a use-case type scenario but the comment below has a more susinct use case [11:17:59.0000] <jmather> hober: granted. It's a multi-faceted issue, which makes it harder to solve [11:18:23.0000] <jmather> and there is a lot of politics involved with what happened in yesterdays uproar [11:18:37.0000] <annevk> othermaciej: afaik nobody is implementing it [11:18:47.0000] <hober> annevk: DNT? [11:18:51.0000] <jmather> /me gets enough politics at work, tries to avoid it elsewhere, hah. [11:18:57.0000] <annevk> hober: from-origin [11:19:15.0000] <hober> annevk: oh, yeah. that makes me sad. i really like from-origin [11:19:37.0000] <annevk> me too :) [11:20:13.0000] <Kasey> jwalden I'm not sure about anybody else, but this is why I just joined the channel: http://www.alistapart.com/articles/responsive-images-and-web-standards-at-the-turning-point/. Really interested to see how this turns out. [11:20:16.0000] <hober> /me is still somewhat bitter that we didn't agree to use it for @font-face [11:20:51.0000] <Kasey> the article was just published yesterday [11:21:22.0000] <jwalden> /me likes the "I was also there, for some reason." line in that article :-) [11:21:37.0000] <othermaciej> annevk: we'd be down for implementing it in WebKit but it's probably somewhat less useful if not applicable to fonts [11:21:55.0000] <annevk> othermaciej: as currently defined it applies to all resources [11:22:18.0000] <jwalden> jmather: "you must be new here" :-) [11:22:22.0000] <othermaciej> so what's the issue with it and @font-face? [11:22:36.0000] <Ms2ger> The cabal, I suppose [11:22:39.0000] <jwalden> web standards is politics to a decent extent :-) [11:22:39.0000] <jmather> jwalden: Yep. Sure am. That's why I have the crazy ideas. :) [11:22:50.0000] <annevk> @font-face requires CORS now I think [11:22:51.0000] <othermaciej> (mind you, I also don't like the name From-Origin very much but I'm not sure of a better name) [11:23:18.0000] <jwalden> that said, generally we get pretty good ideas out of stuff even with politics in the undertones [11:23:52.0000] <othermaciej> the CSS3 Fonts WD doesn't seem to hard-require CORS, or at lest it leaves the From-Origin possibility open [11:23:59.0000] <jwalden> /me has been lurking here four years or so, has contributed directly on the list in the past, mostly has been busy in recent years tho with other stuff [11:24:02.0000] <othermaciej> http://www.w3.org/TR/css3-fonts/#same-origin-restriction [11:24:28.0000] <othermaciej> I guess that is deleted from the ED [11:24:40.0000] <Ms2ger> jwalden, just a decent extent? :) [11:25:04.0000] <annevk> othermaciej: I haven't followed up on that much [11:25:53.0000] <jmather> I just had some ideas one day about who to work towards responsive images and tried to follow along with the CG… now a lot of them are talking about meta variables in the head and that's the complete opposite solution i see as viable… making the content dependent on head variables and such… yikes. :D [11:26:30.0000] <jmather> but that's what made me realize that the approach where browser size determines the image to pull is … problematic. [11:26:36.0000] <jwalden> Ms2ger: eh, it depends; some of the stuff around window.postMessage back in the day (I am showing my age here, and get off my lawn) had a little bit of politics to it, some of it fair-ishly characterizable as coming from Mozilla, and we still got something pretty decent out in the first cut there (that everyone originally implemented, before various improvements to it got spec'd) [11:26:53.0000] <jwalden> probably depends a bit on the issue [11:27:27.0000] <jwalden> /me remembers how late that landed in the 3.0 cycle with Mozilla, is glad that argument about changing things so late in a cycle is ancient history now [11:27:33.0000] <jwalden> trains ftw [11:29:46.0000] <jwalden> so having skimmed that article, I think it's safe to say I'm not close enough to web development to know how well either satisfies a reasonable set of the needs people are trying to solve [11:30:02.0000] <jwalden> new elements like <picture> tend to be a hard sell, tho [11:30:24.0000] <jwalden> but shoving everything into a microsyntax in an attribute is unpleasant enough, certainly [11:30:32.0000] <jwalden> meh [11:30:59.0000] <annevk> best go shopping? [11:31:34.0000] <jwalden> assuming you meant "let's", sure :-) [11:31:48.0000] <jmather> at this point i'm for either implementation if we can make it relate to the space for content [11:31:50.0000] <othermaciej> the srcset micro syntax was more readable before the width/height additions [11:32:00.0000] <othermaciej> maybe if you use only width or only resolution it's not so bad [11:32:23.0000] <jmather> though i think i like the longer picture syntax w/ srcset style attributes [11:32:41.0000] <tantek> jwalden, are the use cases you care about covered here? http://www.w3.org/wiki/Images (if so, feel free to add details, if not, add more use cases) [11:32:43.0000] <jmather> the only problem is that polyfill's for srcset implementation will have the double-hit download [11:33:22.0000] <annevk> jwalden: relevant: http://plasmasturm.org/log/495/ [11:33:38.0000] <annevk> jwalden: includes a working(!) pointer to Markup Barbie as a bonus [11:33:47.0000] <jwalden> tantek: the things I can think of that I'd want seem covered by either proposal; as I said, I'm not close enough to web development these days to really experience firsthand the concerns people have, or to super-productively contribute on the point [11:34:25.0000] <tantek> that's interesting because the two proposals appear to focused on solving different problems/use-cases. [11:34:31.0000] <jwalden> /me says a prayer for diveintomark.org, bless its soul [11:34:40.0000] <jwalden> tantek: also I didn't read either super-closely [11:34:48.0000] <jwalden> tantek: really I'm probably best ignored on this topic :-) [11:34:58.0000] <tantek> jwalden - hence I'm asking a question from the perspective of needs rather than syntax/proposals [11:35:16.0000] <jmather> tantek: that's where I think the authors and browser guys are butting heads… i think we're trying to solve our own issues and not looking at what the other side is trying to solve [11:35:30.0000] <jwalden> I will claim to be worth listening to when I think there's a chance that's actually the case, and I don't think I would add anything here if I read more deeply :-) [11:35:32.0000] <jmather> using we collectively there :) [11:35:36.0000] <othermaciej> it looks like nearly every thread on whatwg@ is now on responsive images [11:35:44.0000] <Ms2ger> Correct [11:35:49.0000] <hober> yeah, i'm over 100 mails behind :( [11:36:59.0000] <jmather> From what I understand, browser side, the concern is knowing enough about the image to try and make better decisions about what to display and when [11:38:22.0000] <jmather> Am I missing something there or is that a fair summary of srcset's origins? [11:39:03.0000] <annevk> jmather: its origins are clearly explained here http://lists.w3.org/Archives/Public/public-whatwg-archive/2012May/0247.html [11:40:22.0000] <jmather> Ok, I've raid that. I really wish I could just scream size of the viewport is irrelevant from the mountain tops, but alas… there are no mountain tops here. :D [11:40:35.0000] <jmather> sorry, read, not raid, hah [11:44:38.0000] <tantek> welcome ChrisWilson [11:45:41.0000] <ChrisWilson> :) [11:58:02.0000] <Ms2ger> WebKit guys, has https://bugs.webkit.org/show_bug.cgi?id=32695 been fixed? [11:59:12.0000] <ShaneHudson> python [11:59:17.0000] <ShaneHudson> oops, wrong window! [11:59:39.0000] <Ms2ger> perl! [11:59:43.0000] <othermaciej> Ms2ger: working ok in my build [12:07:46.0000] <ChrisWilson> ms2ger: and in Chrome Stable [12:09:23.0000] <Ms2ger> Eek, it's ChrisWilson [12:09:31.0000] <Ms2ger> And working for Google [12:09:54.0000] <Ms2ger> Still feels weird, I got to say [12:10:19.0000] <ChrisWilson> heh. [12:10:25.0000] <ChrisWilson> Not to me. [12:10:46.0000] <ChrisWilson> Freeing, you might say. [12:19:48.0000] <jgraham> /me reads that as "ChrisWilson has 'no underwear' in his Google contract", sells it to an online gossip rag [12:20:48.0000] <ChrisWilson> jgraham: that would not be the worst thing said about me by a very long short. [12:20:55.0000] <ChrisWilson> er, shot. [12:20:58.0000] <ChrisWilson> :) [12:25:20.0000] <annevk> ChrisWilson: go on :p [12:26:15.0000] <Ms2ger> "@Tab - yes I do remember, sorry. I'm being a bloody idiot." [12:26:23.0000] <Ms2ger> If that's your entire email, yes. [12:28:42.0000] <ChrisWilson> annevk: Pretty sure a few things said in this very forum make the cut. :) [12:33:43.0000] <dglazkov> I wonder why the idiot is bloody? [12:33:53.0000] <dglazkov> perhaps there was a terrible accident [12:36:51.0000] <jgraham> dglazkov: I always imagined it more like the black knight from monty python [12:38:58.0000] <dglazkov> oh wow! Monty Python is free on Amazon Prime. Yay, time to watch it all again! [13:32:27.0000] <jgraham> http://pretty-rfc.herokuapp.com/ [13:39:01.0000] <othermaciej> hmm, Media Queries is not a very good spec [13:43:05.0000] <annevk> othermaciej: when I fixed most of its major problems (parsing was undefined and such) it was a bit too late to rewrite the whole thing [13:43:23.0000] <othermaciej> annevk: parsing now seems to be defined, but as far as I can tell, evaluation is undefined [13:43:40.0000] <othermaciej> I can't find a normative statement that would define when a media query property expression is true [13:43:41.0000] <annevk> othermaciej: example? [13:44:02.0000] <annevk> othermaciej: it doesn't use RFC 2119 [13:44:09.0000] <othermaciej> (it also uses "must" in the non-RFC2119 sense a bunch) [13:44:29.0000] <othermaciej> but even overlooking the terminology [13:44:43.0000] <othermaciej> what defines when a media query expression evaluates to true, and when it evaluates to false? [13:45:20.0000] <annevk> I think that would be [13:45:30.0000] <annevk> 'A media query is a logical expression that is either true or false. A media query is true if the media type of the media query matches the media type of the device where the user agent is running (as defined in the "Applies to" line), and all expressions in the media query are true' [13:45:43.0000] <othermaciej> and what defines when an expression evaluates to true? [13:46:06.0000] <annevk> that [13:46:28.0000] <annevk> oh sorry [13:47:30.0000] <othermaciej> that statement is defined in terms of expressions in the media query all being true, and I have a hard time finding an explicit statement of when expressions are true [13:49:03.0000] <annevk> third bullet point in section 4 [13:49:27.0000] <annevk> actually no, that's only for shorthands [13:49:30.0000] <othermaciej> there are some specific defined conditions for when expressions are false [13:49:48.0000] <annevk> yeah so this would have been caught if I had rewritten it [13:50:10.0000] <annevk> anyway, I just patched known problems rather than taking a fresh look at the problem [13:50:30.0000] <othermaciej> it also does not define how to evaluate relative units other than maybe in an example [13:50:50.0000] <annevk> does section 6 not define that? [13:52:41.0000] <othermaciej> ah, mea culpa [13:56:29.0000] <othermaciej> hmm, I have a hard time understanding what the initial value of font-size is (based on CSS 2.1 at least) [13:57:34.0000] <Ms2ger> It's CSS, what'd you expect? :) [13:58:04.0000] <annevk> othermaciej: undefined [13:58:10.0000] <annevk> othermaciej: up the UA [13:58:15.0000] <annevk> othermaciej: typically 16px [13:58:28.0000] <othermaciej> is 'medium' supposed to be a fixed value per UA, or is it supposed to match the value of the current configured font size? [13:58:46.0000] <Ms2ger> I think the latter [14:01:22.0000] <othermaciej> I was trying to figure out if media queries based on 'em' units are actually useful, and my conclusion is that I can't tell [14:18:47.0000] <smaug____> dglazkov: Salty liquorice ice cream is way more effective than beer ;) [14:19:01.0000] <Ms2ger> /me keeps that in mind [14:28:09.0000] <annevk> othermaciej: they react to minimum font-size settings but if you change font-size on the root and base your calculations on that it'll go wrong [14:29:58.0000] <TabAtkins_> othermaciej: Heh, I can't believe we didn't actually define when MQs are true. ^_^ [14:36:00.0000] <dtharp> TabAtkins_: yt? [14:36:47.0000] <othermaciej> TabAtkins_: I can't believe it got to PR that way [14:38:09.0000] <dglazkov> smaug____: http://www.icecreamsource.com/Black-Licorice--3-Gallons_p_361.html? [14:38:41.0000] <TabAtkins_> dtharp: pong [14:39:01.0000] <dtharp> TabAtkins_: Hi, I've been implementing MS ietestcenter tests in webkit, and recently found that webkit's 'computeLength()' is not handling vw, vh and vmin. I'm working on a patch for this. The issue I have a question with is that MS seems to have aliased vmin to 'vm' in their test htm. Apparently this alias has been in effect since IE6. The question is, should we alias vmin as 'vm' in webkit as well for compatibility? 'vm' is, of course, not [14:39:02.0000] <dtharp> mentioned in the spec. [14:39:55.0000] <TabAtkins_> 'vm' was the original name used by the spec, but we changed it because it was impossible to tell whether it meant "min" or "max". [14:40:19.0000] <TabAtkins_> I don't think there's a compat problem. [14:40:49.0000] <dtharp> TabAtkins_: Ah, that makes sense. So you think it's fine to add the alias? [14:40:53.0000] <smaug____> dglazkov: :p [14:42:36.0000] <dglazkov> I like licorice. Never had licorice ice cream [14:43:49.0000] <ShaneHudson> I have never even heard of licroice ice cream! [14:43:58.0000] <ShaneHudson> Mint ice cream is good though [14:45:55.0000] <tantek> Mint chocolate chip even. [14:46:00.0000] <ShaneHudson> mmmmm [14:46:34.0000] <ShaneHudson> I have none :( but I am about to grab some gingernut biscuits from the cupboard :) [14:49:17.0000] <smaug____> https://en.wikipedia.org/wiki/File:TotallyBlack.jpg is good [14:52:20.0000] <dtharp> TabAtkins_: sorry for being dense here... when you say "I dont think there's a compat problem", do you mean that if we implment the 'vm' alias for vmin, then at that point we do not have a compatibility issue? or are you saying we *currently* don't have a compat problem, and therefore should NOT implement the 'vm' alias? [14:54:36.0000] <annevk> per spec there is no vm unit [14:54:53.0000] <annevk> I think it would be wrong for a browser to support it [14:55:07.0000] <dtharp> IE10 supports it [14:55:27.0000] <ShaneHudson> Does anyone know when IE10 is being released yet? [14:55:52.0000] <annevk> dtharp: just saying; if it's somehow required then it should be in the spec... [14:56:11.0000] <annevk> dtharp: cannot have every browser have to figure out on their own what to implement [14:57:22.0000] <dtharp> annevk: i agree. I have seen concessions for "common in the wild" stuff before, so was just trying to get clarification. It sounds like the right path here is to get MS to update thier test [15:08:18.0000] <TabAtkins_> dtharp: Sorry for the delay. I think that there is no need to implement a vm unit, so we should leav eit alone. [15:09:02.0000] <dtharp> TabAtkins_: sounds good. I'll work with MS to have them update their test per spec. Thanks! [15:09:38.0000] <TabAtkins_> I didn't even realize we had a computeLength() function, by the way. [15:11:11.0000] <dtharp> TabAtkins_: its actually CSSPrimitiveValue::computeLengthDouble() in CSSPrimitiveValue.cpp [15:11:39.0000] <TabAtkins_> Oh, one of our internal functions. Gotcha. [15:11:41.0000] <dtharp> TabAtkins_: its the target fror a bunch of templates [15:12:59.0000] <ShaneHudson> Are any of you from Google? [15:14:40.0000] <TabAtkins_> ShaneHudson: I am. I don't think dtharp is. [15:15:48.0000] <TabAtkins_> (At least, his name doesn't have any matches when I search on our employee database.) [15:15:58.0000] <ShaneHudson> Excellent, is it alright to pm you? [15:16:26.0000] <TabAtkins_> ShaneHudson: Go for it. [15:18:21.0000] <TabAtkins_> annevk: I doubt there's a compat need to support vm. IE was the only browser to support it, and use of the viewport-relative values wasn't common afaik. [15:21:46.0000] <annevk> yeah [15:21:52.0000] <annevk> should have a negative test for vm really [15:36:41.0000] <zewt> annevk: https://www.khronos.org/webgl/public-mailing-list/archives/1205/msg00135.html :| [15:40:03.0000] <TabAtkins_> zewt: Holy mother of god no. [15:40:16.0000] <zewt> was: re: :| [15:40:18.0000] <TabAtkins_> Why? Why won't people learn? We made mistakes! [15:40:24.0000] <zewt> things not to take inspiration from: [15:40:24.0000] <zewt> java [15:40:25.0000] <zewt> xml [15:40:33.0000] <zewt> note: list not exhaustive. [15:41:03.0000] <zewt> (several times i've seen webgl people go "we can't do it that way. java did it and it sucked!" because clearly if java can't do something well, it can't be done, right?) [15:41:11.0000] <Philip`> /me recently found that OpenGL and OpenGL ES have incompatible extensions with exactly the same name, which is fun, especially since some desktop GL drivers implement both GL and GLES extensions [15:41:14.0000] <TabAtkins_> I mean, for chrissakes, the XML people themselves finally admit that namespaces were a mistake. [15:42:53.0000] <zewt> call me cynical, but it smells a lot more like "i want my name in the webgl spec, implement this random thing i thought of for a bunch of contrived reasons" [15:43:09.0000] <zewt> maybe i've just been doing the open source thing too long, heh [15:43:23.0000] <gsnedders> WebGL is a clusterfuck. Is this not obvious? [15:43:31.0000] <gsnedders> (News at 10, move on.) [15:43:38.0000] <zewt> it's not XML-level clusterfuck [15:43:49.0000] <zewt> that's more like clusterbomb-fuck [15:43:52.0000] <zewt> (is that a thing?) [15:44:40.0000] <gsnedders> clusterbomb-clusterfuck. [15:45:52.0000] <gavinc> at least XML got unicode and utf-8 by default right? ... than it added letting you set your own encoding which sort of lets it down a bit... [15:46:07.0000] <zewt> we could go in circles for hours about what xml got wrong :) [15:47:38.0000] <gsnedders> gavinc: Look at the EDBIC stuff in it, like dfn of space characters in 1.1 (there again, nobody gives a shit about 1.1) [15:48:06.0000] <gavinc> Speaking of which... if anyone here would like to point out everything we got wrong http://dvcs.w3.org/hg/rdf/raw-file/default/rdf-turtle/index.html# BEFORE going to last call. Other than it's RDF and is killing the web. [15:48:19.0000] <gavinc> Last call next week and all that [15:49:18.0000] <zewt> does "i have no clue what this is for after reading the introduction" count? [15:49:26.0000] <gavinc> Yes. [15:50:04.0000] <zewt> (i have no idea what RDF is; a paragraph summarizing this in isolation, assuming no knowledge of whatever that is, might be useful--if that's possible) [15:50:18.0000] <gsnedders> gavinc: Have a N-Triple example in the Intro too. [15:50:31.0000] <gsnedders> (Ideally semantically identical to the Turtle one) [15:50:52.0000] <zewt> ("turtle" makes me think it has something to do with Logo. heh) [15:51:11.0000] <gavinc> gsnedders: good point [15:51:44.0000] <gsnedders> gavinc: Don't recommend CDATA sections for embedding in XHTML (as then people will fail to handle the case where ]]> appears in a string) [15:52:05.0000] <gavinc> gsnedders: if you can show any other way of embedding in XHTML I'm all for it [15:52:07.0000] <gsnedders> "THe HTML lang attribute" has an obvious typo [15:52:19.0000] <gsnedders> gavinc: Escape everything, as painful as it is. [15:52:28.0000] <gsnedders> gavinc: But otherwise people are unlikely to handle ]]> correctly. [15:52:43.0000] <gavinc> gsnedders: mmm... does that work in polyglut documents? [15:52:47.0000] <TabAtkins_> I agree with gsnedders, since ]]> is a vital part of ascii fishes. [15:52:59.0000] <gsnedders> At least note "]]>" cannot be represented within a CDATA section. [15:53:16.0000] <gsnedders> gavinc: CDATA doesn't, escaping everything does. [15:53:22.0000] <gavinc> Perfect. [15:53:40.0000] <gavinc> also, don't use XHTML... but hey ;) [15:53:56.0000] <gsnedders> gavinc: 12.3 uses what BNF form? [15:54:14.0000] <gavinc> ah, yes W3C EBNF... [15:54:24.0000] <gsnedders> gavinc: That seems to be noted in 5.4 but not 12.3 [15:54:29.0000] <gavinc> /me nods [15:54:37.0000] <gsnedders> gavinc: Why does it cite XML 1.0 3rd Edition and not 4th? [15:54:46.0000] <gsnedders> (Or, perhaps contravercially, 5th?) [15:54:56.0000] <gavinc> should be 5th... [15:55:21.0000] <gavinc> the link is to 5th... [15:55:33.0000] <gsnedders> http://www.w3.org/TR/REC-xml/#sec-notation is the URL given, and it explicitly states an edition, which is obviously nonsense [15:55:52.0000] <gavinc> Yes. [15:56:01.0000] <gavinc> remove words Thrid Edition ;) [15:56:49.0000] <gsnedders> gavinc: Can you aggregate everything I've said and pass it along? [15:56:59.0000] <gavinc> faster, I can edit it ;) [15:58:58.0000] <gsnedders> /me moves to #swig [16:00:16.0000] <espadrine> /join #swig 2012-05-17 [17:03:11.0000] <Bevan> picture++ [17:03:15.0000] <Bevan> img set-- [17:41:53.0000] <zewt> Bevan: not helpful [17:58:39.0000] <Bevan> zewt. Acknowledged. [17:58:42.0000] <Bevan> :) [20:45:59.0000] <inclement_bore> ’twas two nights after srcset, and all through the chat [20:46:09.0000] <inclement_bore> not a creature was stirring, not even a... sig cat [20:46:20.0000] <inclement_bore> the use cases were hung on anything near [20:46:34.0000] <inclement_bore> in hopes that efficient scalable image delivery over the web ... [20:46:41.0000] <inclement_bore> ... would soon be here [20:46:48.0000] <inclement_bore> developers were nestled, all snug in their beds [20:46:55.0000] <inclement_bore> visions of media queries still danced in their heads [20:47:01.0000] <inclement_bore> implementers too, not impressed [20:47:06.0000] <inclement_bore> had just settled down for a long spring rest ... [21:43:34.0000] <paul_irish> gy [21:44:21.0000] <paul_irish> (whoops. IRC doesnt start gyazo. ignore.) [22:50:54.0000] <heycam> what's the reason appendix sections in CSS specs aren't numbered (like A.1, A.2, etc.)? [22:51:15.0000] <heycam> is is that the numbering scheme is ugly? [23:53:16.0000] <hsivonen> heycam: at least some CSS appendix naming forms a joke, but maybe that was only for CSS 2 and 2.1 [23:53:56.0000] <heycam> hsivonen, ah yes, for the appendix letters [23:56:01.0000] <heycam> hsivonen, though notice how if you jump to the bottom of http://www.w3.org/TR/CSS21/cover.html that appendix sections get numbers while those in http://dev.w3.org/csswg/css3-text/#contents don't [23:58:26.0000] <hsivonen> maybe the letter thing is for CSS 2/2.1 only [23:58:49.0000] <hsivonen> Hixie: I've successfully browsed the Web with 1rem == 18px for years [00:36:16.0000] <webben> What is the current pipeline for turning http://svn.whatwg.org/webapps/source into http://wiki.whatwg.org/wiki/FAQ#What_are_the_various_versions_of_the_spec.3F ? [00:44:56.0000] <webben> /me spots http://dev.w3.org/cvsweb/html5/spec/Makefile [00:55:23.0000] <jgraham> webben: Brown paper and glue. And magic. [00:56:31.0000] <webben> jgraham: So I can see. [00:56:35.0000] <jgraham> Hixie has a highly-custom pipeline that is mostly calling various services that others are providing e.g. anolis, the (HTML-specific) spec spliter [00:56:56.0000] <webben> it looks like the w3c bits have makefiles [00:57:10.0000] <webben> is there a makefile for the whatwg "complete" version somewhere? [00:57:20.0000] <jgraham> No [00:57:27.0000] <webben> k [00:57:46.0000] <jgraham> Most of Hixie's pipeline is closed source because it is very environment specific [00:58:43.0000] <jgraham> But the complete version of the spec might be relatively easy to make since it is the most, well, complete [00:59:39.0000] <jgraham> http://pimpmyspec.net/ + the right options might get you a lot of the way there [01:00:05.0000] <jgraham> Of course, the output is conveniently provided, so I wonder what the motivation is ;) [01:36:54.0000] <smaug____> /me doesn't always understand what people mean with 'polyfill' [01:37:49.0000] <smaug____> ahaa [01:37:52.0000] <smaug____> , silly word [02:17:08.0000] <smaug____> We should decide at some point what to do with FileSystem API [02:23:48.0000] <smaug____> hmm, webapps is the right place for that [02:39:54.0000] <AryehGregor> Is there a command-line HTML5 validator that's easy to install and use? [02:44:58.0000] <MikeSmith> AryehGregor: I put https://bitbucket.org/sideshowbarker/vnu together experimentally [02:45:24.0000] <MikeSmith> can download from https://bitbucket.org/sideshowbarker/vnu/get/2759dcb15031.zip [02:45:29.0000] <MikeSmith> or just clone the reop [02:45:33.0000] <MikeSmith> *repo [02:45:49.0000] <MikeSmith> see the "How to use the pre-built jars" part at https://bitbucket.org/sideshowbarker/vnu [02:46:13.0000] <MikeSmith> you can feed it a list of filenames [04:18:49.0000] <Ms2ger> <TabAtkins_> othermaciej: Heh, I can't believe we didn't actually define when MQs are true. ^_^ [04:18:58.0000] <Ms2ger> Really, you couldn't believe that? [05:29:10.0000] <annevk> in an async algorithm, are all observable changes queued tasks? [05:29:24.0000] <annevk> e.g. manipulation of the fullscreen elemetn stack [05:34:03.0000] <annevk> /me has the feeling nobody really cares about task queues [06:39:31.0000] <zewt> well, task queues are really just partial order guarantees, right? [06:39:51.0000] <zewt> (so if an implementation orders *all* queued events, they mostly don't have to care about them) [06:58:02.0000] <EyePulp> 'morning. [to whom it applies] [09:18:42.0000] <dglazkov> good morning, Whatwg! [09:18:57.0000] <dglazkov> I think I need to start a Good Morning CG. [09:19:21.0000] <Ms2ger> /me strangles dglazkov [09:19:26.0000] <Ms2ger> Good morning to you too [09:19:32.0000] <dglazkov> :) [09:31:36.0000] <annevk> hsivonen: curious (re twitter), what should the EU have used instead? [09:32:37.0000] <companyhen> Morning all. :D [09:37:01.0000] <ChrisWilson> /me mumbles something incoherent and goes and gets another cup of coffee [09:48:35.0000] <Philip`> dglazkov: You ought to start a wiki page listing the use cases for having a good morning, before anyone will take you seriously [09:54:12.0000] <EyePulp> heh - I just got that. [09:54:49.0000] <jreading> hee [10:05:28.0000] <dglazkov> Philip`: let's not forget the support forum, er.. mailing list. [10:07:49.0000] <hsivonen> annevk: E for example [10:08:13.0000] <hsivonen> annevk: you know, like FF and DM were plain ASCII before euro [10:12:10.0000] <annevk> heh [10:12:28.0000] <annevk> apparently our former currency didn't have a code point http://en.wikipedia.org/wiki/Florin_sign [10:14:05.0000] <Philip`> On (at least) UK keyboard layouts, the € symbol is stuffed pretty inelegantly into a spare corner of the 4/$ key, which doesn't seem like great design [10:14:52.0000] <Philip`> (The ¬`| key is the only other one that typically follows that pattern) [10:15:24.0000] <annevk> hsivonen: the French had U+20A3 ₣ but also used F/FF indeed [10:18:14.0000] <annevk> some more browsing reveals that http://en.wikipedia.org/wiki/Azerbaijani_manat does not have a Unicode symbol while they do have a distinct sign [10:18:26.0000] <annevk> weird [10:19:24.0000] <Philip`> http://www.tcmb.gov.tr/yeni/iletisimgm/TurkishLira.php - "The upward pointing parallel lines symbolize the "steadily rising value" of the Turkish lira and the Turkish economy." - does that mean they'll redesign the currency symbol if their economy collapses? [10:20:22.0000] <annevk> heh, they got this one because it devalued so much [10:21:35.0000] <hsivonen> before euro, Finland use an ASCII-only currency abbreviation "mk" and it worked great [10:22:23.0000] <gsnedders> /me finds £ annoying with just ASCII — going for GBP as a suffix just seems weird [10:22:37.0000] <Philip`> gsnedders: Use "lb" instead [10:23:32.0000] <annevk> hsivonen: I typically use the three-letter code because that works great in Google :) [10:24:01.0000] <annevk> 1000 pen in eur please [10:25:45.0000] <hsivonen> in other news, it's terribly inconvenient that Ubuntu 12.04 ships Wine64 and the wine binary that ships with CrossOver (which looks like 32-bit Windows even on 64-bit host) is incompatible with winetricks [10:26:13.0000] <hsivonen> of course, CrossOver is supposed to make winetricks unnecessary, but CrossOver is broken [11:43:35.0000] <tantek> /me scrolls up and reads about currency symbols. [11:44:11.0000] <tantek> There's definitely quite a variety of ways that folks publish currency amounts on the web. [11:44:33.0000] <tantek> The three-letter code as annevk suggests is probably the easiest to both publish and search for. [11:44:43.0000] <tantek> But not what every locality is used to seeing - so... [11:46:49.0000] <tantek> anwyay - if anyone is interested in either reading more on currency publishing practices on the web or has further real world examples to add, please take a look at: http://microformats.org/wiki/currency-examples [11:48:53.0000] <tantek> and currency related formats respectively: http://microformats.org/wiki/currency-formats [11:58:10.0000] <gavinc> RDF HTML datatype in terms of DOM4 and HTML5 http://dvcs.w3.org/hg/rdf/raw-file/default/rdf-concepts/index.html#section-html [12:45:37.0000] <annevk> gavin: why do you use <code> for what are <var>s? [12:45:42.0000] <annevk> euh gavinc [12:45:46.0000] <annevk> sorry gavin [12:47:40.0000] <annevk> gavinc: you don't need to invoke normalize() [12:48:08.0000] <annevk> gavinc: and if you don't pass a context element you'll get <html> and <body> etc. implied and you might get quirks mode parsing even... [12:59:52.0000] <gavinc> annevk: isn't normalize needed before using isEqualNode? [13:01:21.0000] <gavinc> annevk: What should the context node be in that case? [13:19:37.0000] <annevk> gavinc: normalizing already happens during parsing afaict [13:19:48.0000] <annevk> gavinc: dunno, kind of depends on what you want I guess [13:20:49.0000] <gavinc> annevk: A floating document fragment [13:21:11.0000] <gavinc> Something that came from, and be put back in a complete HTML document [13:22:21.0000] <annevk> hmm see discussion on public-webapps [13:22:28.0000] <annevk> we don't have a generic thing for that which works [13:22:38.0000] <gavinc> :( [13:23:13.0000] <gavinc> which discussion? [13:24:10.0000] <gavinc> all of it is going to take a good deal of reading... somewhere to start? [13:27:42.0000] <annevk> i guess you could go with what you have now or wait until contextless parsing is defined 2012-05-18 [22:37:26.0000] <AryehGregor> This looks bad: https://tbpl.mozilla.org/php/getParsedLog.php?id=11839094&tree=Try [22:37:26.0000] <AryehGregor> firefox-bin(345,0x106e46000) malloc: *** error for object 0x17ffee000: pointer being freed was not allocated [22:38:18.0000] <padenot> AryehGregor: sure about the channel ? [22:38:25.0000] <AryehGregor> Drat, sorry. [22:38:27.0000] <AryehGregor> Thanks. [22:38:40.0000] <AryehGregor> That was meant for #developers. [22:58:19.0000] <hsivonen> why has http://www.ie6countdown.com/ stuck to February data? Did some countries regress since then and MS is too embarrassed to update the numbers? [22:58:34.0000] <hsivonen> or did the person in charge of the site leave MS or something? [00:15:42.0000] <annevk> hober: cpearce is not subscribed to public-webapps [00:16:04.0000] <annevk> hober: discussing fullscreen there seems bad at least until everyone agreed to subscribe to another mailing list. [00:55:00.0000] <Jaycob> hey, I'm developing an app that uses application storage. how can I bypass that while developing, because it doesn't update anything unless the manifest file gets updated [01:08:08.0000] <moo-_-> Jaycob: I would handle this kind of things by serving development HTML from different URL and disable manifest="" on the server side [01:08:34.0000] <Jaycob> ok! thought about that too. thanks! [01:09:57.0000] <Jaycob> btw chrome gives me an error when it doesn't find the manifest file when offline (naturally), and I'm wondering wether this is supposed to give an error or am I doing something wrong? [01:26:50.0000] <othermaciej> does anyone know if the CSS WG has a bug/issue tracker other than the mailing list? [01:27:16.0000] <othermaciej> I'm wondering if there is some other way to report the issue that Media Queries doesn't define when media queries are true or false in most cases [01:27:53.0000] <othermaciej> it is a PR but it would be nice to at least ensure this is fixed for the next version [01:28:39.0000] <Ms2ger> Yes, but you're not allowed to use them [01:29:09.0000] <Ms2ger> (Depending on the spec, bugzilla, tracker, a text file somewhere, or the editors' heads) [01:29:25.0000] <MikeSmith> they have a bugzilla components [01:29:32.0000] <MikeSmith> 30+ of them [01:29:37.0000] <MikeSmith> I know because I created them [01:29:38.0000] <Ms2ger> But glazou will shout at you if you file anything there [01:29:48.0000] <MikeSmith> heh [01:29:49.0000] <MikeSmith> um [01:29:54.0000] <Ms2ger> They're only for editors to track the www-style threads [01:30:03.0000] <MikeSmith> /me struggles for something constructive to say [01:30:04.0000] <othermaciej> I see [01:30:15.0000] <MikeSmith> that's fucktarded [01:30:20.0000] <othermaciej> so as a non-member of the CSS WG, there is no way to report a comment and be sure it won't get lost? [01:30:21.0000] <Ms2ger> No comment [01:30:40.0000] <MikeSmith> myself, I would just use the bugzilla anyway [01:30:51.0000] <MikeSmith> dude is going to find something to shout about regardless [01:31:27.0000] <othermaciej> I am not sure I care about this issue enough to suffer the wrath of glazou [01:31:31.0000] <jgraham> Maybe tatoo it onto TabAtkins? [01:31:35.0000] <MikeSmith> heh [01:32:12.0000] <othermaciej> but I will be amused if the spec goes to REC while still not actually defining the MQ processing model at all (which seems very likely to happen) [01:32:20.0000] <othermaciej> and more so if the same is true for the next version [01:32:24.0000] <MikeSmith> wouldn't be the first time [01:32:45.0000] <Ms2ger> Or the last [01:33:07.0000] <MikeSmith> othermaciej: speaking of trying to find something constructive to say, I have the same feeling about the chairs decision on meta generator [01:33:30.0000] <Ms2ger> s/meta generator/*/ [01:33:32.0000] <othermaciej> MikeSmith: do you think we should have agreed to reopen it based on the info provided? [01:33:34.0000] <MikeSmith> for now I'll just say I feel zero motivation to actually ever again spend any time contributing to any more change proposals [01:33:40.0000] <MikeSmith> othermaciej: yeah, I do [01:33:41.0000] <othermaciej> MikeSmith: or made a different decision in the first place? [01:33:46.0000] <MikeSmith> that too [01:34:29.0000] <othermaciej> the thing we pretty clearly asked for, both in the original decision, and in response to the first reopen attempt, was specifics and evidence [01:34:46.0000] <othermaciej> I don't feel like it is that hard to find a few scraps of specific data [01:35:05.0000] <MikeSmith> I just don't want to do the monkey dance, man [01:35:23.0000] <othermaciej> well, the bar for reopening an issue has to be higher than the bar for opening it in the first place [01:35:27.0000] <jgraham> /me wodners if you are really supposed to get wyciwyg URLs in firefox history UI when you document.open/document.write/document.close [01:35:29.0000] <othermaciej> or nothing will even be settled [01:35:29.0000] <MikeSmith> I think the information in that CP makes it quite clear what's wrong with the exception [01:35:33.0000] <jgraham> *wonders [01:35:56.0000] <Ms2ger> jgraham, in general, I wouldn't think so [01:35:59.0000] <Ms2ger> hsivonen, ^ [01:36:17.0000] <Ms2ger> othermaciej, are you suggesting anything will ever be settled now? [01:36:27.0000] <othermaciej> I will add that I myself am not even sure the exception is a good idea, but the HTML WG process is optimized for procedural fairness, not for producing the outcome that is most technically optimal in the opinion of the chairs [01:36:35.0000] <jgraham> (iirc, per spec, doing document.write on an existing document shouldn't create a new history position at all)] [01:36:46.0000] <jgraham> (need to write some tests for that one day) [01:37:18.0000] <MikeSmith> othermaciej: it's certainly become optimized that way in practice [01:37:22.0000] <othermaciej> I would much prefer to see a way to declare alt as intentionally unavailable due to wysiwygosity that is scoped instead of global to the document, but no one proposed such a thing [01:38:19.0000] <othermaciej> really on all these reopen requests, whichever way the chairs decide it will make a significant group of people mad [01:38:29.0000] <othermaciej> people who lost one time will be super annoyed to do extra work and still lose [01:38:43.0000] <othermaciej> people who won the last time will be super annoyed to have that jerked out from under them [01:38:46.0000] <MikeSmith> actually I meant it seems in practice to have become optimized for procedural stuff rather for evaluation of what's technically optimal. Period. [01:39:18.0000] <MikeSmith> othermaciej: there is very likely now going to be a formal objection on that issue [01:39:27.0000] <webben> othermaciej: There was never a CP proposing scoped noalt? [01:39:39.0000] <MikeSmith> so I guess it's eventually going to end up going to the Director [01:39:39.0000] <othermaciej> webben: I don't believe so [01:39:56.0000] <webben> would it help if someone produced one now? [01:40:08.0000] <othermaciej> that's one problem with the CP process, it tends to polarize people to the extremes [01:40:22.0000] <webben> It certainly seems like it would have more consensus than page-wide exemptions [01:40:41.0000] <othermaciej> webben: if it included new info that was sufficient to reopen the issue (e.g. argue that it meets all the use cases of generator but with less harm, with at least some specific piece of data to point to), then it would probably make a difference [01:41:57.0000] <MikeSmith> I don't know anybody who has ever actually expressed strong support for the meta generator exception [01:42:08.0000] <othermaciej> MikeSmith: as with past FOs, I think I can live with that given the record (initial survey input, decision, request for particular kinds of additional info to reopen, refusal to provide such) [01:42:30.0000] <MikeSmith> it's not refusal to provide such [01:42:43.0000] <othermaciej> director may overrule, but that is his prerrogative [01:42:48.0000] <MikeSmith> it's disagreement about the need to provide it [01:43:08.0000] <MikeSmith> I regret at this point ever actually implementing the meta generator exception in the validator code [01:43:53.0000] <MikeSmith> and I'm not seriously considering to change the way it's implemented so that it's not on by default, and users have to opt in to it [01:44:12.0000] <MikeSmith> because it's really a very bad idea for validation [01:44:16.0000] <MikeSmith> a misfeature [01:44:21.0000] <Ms2ger> s/not/now/? [01:44:29.0000] <MikeSmith> yeah, now [01:44:29.0000] <othermaciej> trying to find the original ISSUE-31 questionnaire [01:44:43.0000] <othermaciej> do you know if there is a way to look at all questionnaires for a given WG? [01:45:58.0000] <MikeSmith> yeah, lemme find it [01:46:20.0000] <othermaciej> there were three of them, but I can only find the two that are not relevant to generator [01:47:25.0000] <othermaciej> aha, http://www.w3.org/2002/09/wbs/40318/issue-31-80-validation-objection-poll/results [01:47:49.0000] <MikeSmith> othermaciej: https://www.w3.org/2002/09/wbs/40318/closed [01:48:13.0000] <othermaciej> Hixie and hsivonen argued in favor of generator in the survey [01:48:45.0000] <othermaciej> Leif and Laura against [01:48:59.0000] <othermaciej> this covers both survey comments and CPs [01:49:01.0000] <MikeSmith> yeah, I would like to hear more from Henri about it [01:49:20.0000] <othermaciej> to me the use case seems valid but I wish someone had proposed a cleaner mechanism to address it [01:49:43.0000] <othermaciej> global per-doc switches do not strike me as elegant or wise [01:49:49.0000] <MikeSmith> yup [01:50:29.0000] <othermaciej> but the only options that were placed on the table were "don't meet this use case" or "use this somewhat squirrely solution for the use case" [01:51:31.0000] <othermaciej> I think it likely that a well-written proposal for a scoped mechanism to meet the use case (such as noalt) would have beaten both those options [02:01:12.0000] <hsivonen> jgraham: yes, you are really supposed to get wyciwyg URLs in Firefox. [02:01:50.0000] <hsivonen> jgraham: I don't like it, because WebKit get away with non-support for history navigation to document.open()ed and .written content [02:02:12.0000] <jgraham> hsivonen: Doesn't the spec follow webkit? [02:02:35.0000] <jgraham> (on this point :) [02:04:42.0000] <jgraham> Also, I think Opera wants to follow WebKit on this point because in general getting an extra history position for document.write seems like it will create badness from the user point of view [02:05:06.0000] <jgraham> On the basis that the typical use will be something like document.write into an about:blank iframe [02:15:47.0000] <jgraham> Oh, I have now read the spec more closely and I am wrong about the document.open thing [02:16:00.0000] <jgraham> But I still think it makes sense to folloow webkit [02:43:15.0000] <jgraham> hsivonen: I filed https://www.w3.org/Bugs/Public/show_bug.cgi?id=17093 [03:00:44.0000] <hsivonen> jgraham: my vague recollection is that the spec is mostly WebKit-ish here [03:00:55.0000] <hsivonen> except of course the bug you just filed, apparently [03:18:05.0000] <Stevef> othermaciej: how about allowing conformance checkers to provide a switch to suppress no alt for name=generator pages? that would provide the functionality for those who require it without it being silent [03:19:18.0000] <othermaciej> Stevef: interesting idea, I guess it would be sort of like the now-gone "private communications" exception [03:20:14.0000] <Stevef> othermaciej: i discussed this with mike smith with his conformance checker implementor hat on and he said it sounded OK [03:21:52.0000] <othermaciej> I have no problem with it, but if anyone wanted to push it into the HTML WG's HTML5 spec at this point, it would require a Change Proposal with some reasonable level of new information indicating why it's better [03:22:03.0000] <Stevef> othermaciej: making it an opt in provides the point of letting people know that using it only suppresses errors does not make the page conforming [03:22:57.0000] <Stevef> othermaciej: right I am unsure about my desire to write another change proposal, just trying to avoid FO if possible [03:23:32.0000] <Ms2ger> "Are there other cases in HTML where an attribute value contains more than one URI?" [03:23:42.0000] <Ms2ger> head profile=""! [03:23:54.0000] <othermaciej> good of you to do so; unfortunately I have to go to bed so I can't provide more ideas [03:25:05.0000] <Stevef> othermaciej:np [03:32:29.0000] <hsivonen> oh. Stevef left [03:33:44.0000] <hsivonen> Stevef: In case you continue to read the logs: The whole point of the generator thing is to make it so that *by default* validators are silent about missing alt in generated HTML so that generator developers don't feel a need to add bogus alt to silence validators. [03:34:33.0000] <hsivonen> Stevef: So if there was a switch, it would have to default to name=generator content being silent about missing alt and flipping the switch would turn on missing alt reporting. [03:40:51.0000] <hsivonen> Stevef: see http://krijnhoetmer.nl/irc-logs/whatwg/20120518#l-384 [03:40:55.0000] <Stevef> hsivonen: I don't get the logic that it must be silent by default [03:41:16.0000] <hsivonen> Stevef: not getting the logic is the crux of the disagreement, I think [03:41:43.0000] <Stevef> hsivonen:isn't an informed choice better? [03:42:06.0000] <hsivonen> Stevef: the whole generator thing is based on the observation that people who write HTML generators want to make the generated output get a clean report from a validator [03:42:29.0000] <hsivonen> Stevef: so if the validator complains *by default*, they'll change the generator's output to make it not complain [03:42:45.0000] <hsivonen> Stevef: the easiest adjustment being outputting bogus alt [03:43:08.0000] <hsivonen> (if you take a way the option of emitting <meta name=generator> and having that silence validators *by default*) [03:43:48.0000] <hsivonen> Stevef: emitting <meta name=generator> is an informed choice on the part of the person writing the HTML generator [03:44:23.0000] <hsivonen> Stevef: flipping a flag in the validator is a choice by the person invoking the validator [03:44:33.0000] <Stevef> hsivonen: the use of genertaor does not bear that out [03:44:51.0000] <hsivonen> Stevef: not legacy use, sure [03:46:05.0000] <hsivonen> Stevef: anyway, the point is making it so that an HTML generator writer has no incentive to emit bogus alt in order to make it seem that the generator can generate "valid HTML" [03:46:26.0000] <Stevef> many CMS's use name=generator by default any downstream user has to a know about its presence and its effect [03:47:09.0000] <Stevef> hsivonen: i understand the arguement never seen any data to back it up [03:48:14.0000] <hsivonen> Stevef: I've not aware of quantitative data, but I've seen enough anectodal evidence to believe it. Of course, I haven't recorded all the anecdotes when I've seen them. [03:48:42.0000] <hsivonen> It's generally hard to substantiate anecdotal experience of having seen stuff after the fact. [03:49:56.0000] <hsivonen> It would probably be possible to come up with quantitative data by analysing various HTML4 generators and seeing if they generate bogus alts [03:50:23.0000] <hsivonen> (and make the assumption that the obvious explanation that everyone who has ever written a generator knows to be true explains the bogus alts) [03:51:16.0000] <Stevef> hsivonen: i looked at a range of wysywig tools some do some don't [03:51:48.0000] <hsivonen> Stevef: is the "some don't" bucket larger than {Dreamweaver}? [03:52:09.0000] <hsivonen> Dreamweaver has had enough a11y people yelling at them [03:53:33.0000] <hsivonen> also, the "some don't" bucket doesn't disprove the existence of the "some do" bucket [03:53:46.0000] <Stevef> i was stalking about editors that add name=generator by default dreamweaver doesn't appear to do that for me [03:53:48.0000] <hsivonen> the spec tries to empty the "some do" bucket [03:54:14.0000] <hsivonen> Stevef: Dreamweaver is a pro tool. It generally doesn't add stuff you don't ask for. [03:58:59.0000] <Stevef> hsivonen: I guess we have to agree to disagree I made my points here: http://www.w3.org/html/wg/wiki/ChangeProposal/meta_name%3Dgenerator_does_not_make_missing_alt_conforming [03:59:45.0000] <Stevef> hsivonen: its not the end of the world... [04:00:39.0000] <Stevef> hsivonen:as i said the maciej earlier i was looking for a way out of the impasse that leads to FO [04:01:36.0000] <Stevef> hsivonen: but its out of my hands now as it has been taken up formally by judy and the force... [04:01:47.0000] <hsivonen> Stevef: FWIW, I think "WYSIWYG" is immaterial. E.g. mass upload of photos to service that displays them in HTML wrappers faces the issue of not having alt text available without being WYSIWYG [04:02:05.0000] <hsivonen> Stevef: what's "the force"? [04:02:17.0000] <Stevef> a11y taskforce [04:02:21.0000] <hsivonen> ah [04:02:46.0000] <Stevef> i was being mischevious [04:04:10.0000] <Stevef> hsivonen: for the mass upload use case i would suggest using figure/figcaption as it provides way to caption an image with an unambigous semantic that says this a caption, even if that caption is a filename [04:05:57.0000] <hsivonen> It's unclear to me, why it is useful to have an unambiguous semantic for caption if the camption is nonsensical [04:06:07.0000] <hsivonen> s/camption/caption/ [04:08:33.0000] <Stevef> hsivonen: it depends on if you think the filename is nonsensical. note in the vast majority of cases photo uploads are presented inside a link, so AT heaurstucs kick in and the filename is or worse is announced, difference with using figcaption is that at least the user can know its not a text alternative [04:10:05.0000] <hsivonen> Stevef: I mean that a caption like IMG_1234.jpg is only good for getting an identifier for the image. And if you want to grab an identifier, you might as well invoke the context menu and copy the URL of the image. [04:11:20.0000] <Ms2ger> jgraham, did someone file a bug to support the StorageEvent constructor in Opera? [04:11:31.0000] <jgraham> Umm, maybe? [04:11:53.0000] <Ms2ger> Would you mind checking? :) [04:12:07.0000] <Ms2ger> I'd check myself, but, well, you knpw [04:12:08.0000] <hsivonen> Do the broadcasters expect general-purpose browsers to get support for MPEG2-TS? [04:12:43.0000] <hsivonen> or do they intend to use it in off-the-shelf commodity set-top boxes for their walled gardens? [04:13:15.0000] <jgraham> Ms2ger: Turns out that you did [04:13:26.0000] <hsivonen> I hope they realize that MPEG2 visual licensing is even worse than H.264 and is likely to get even more opposition from vendors of general-purpose browsers [04:13:40.0000] <Ms2ger> jgraham, go me :) [04:13:43.0000] <Ms2ger> Did anybody notice? [04:14:42.0000] <hsivonen> or is this about H.264 in .ts because .mp4 doesn't stream? [04:14:57.0000] <hsivonen> (.ogg and .webm FTW for streaming) [04:15:52.0000] <jgraham> Ms2ger: Yeah, but it isn't exactly a showstopper [04:15:55.0000] <jgraham> :) [04:16:32.0000] <jgraham> (if there is some reason that the priority needs to be enexpectedly high, please let me know) [04:17:29.0000] <Ms2ger> jgraham, not really, unless you care about recs [04:17:50.0000] <jgraham> Well I think you know that *I* don't [04:18:52.0000] <Ms2ger> I am not surprised :) [04:23:47.0000] <hsivonen> if I do perl -pi -e 's/foo/bar/', what should I use to use an Unicode apostrophe as foo and an ASCII apostrophe as bar? [04:23:56.0000] <hsivonen> i.e. what escapes? [04:24:03.0000] <hsivonen> when the files are in UTF-8 [04:26:07.0000] <webben> Stevef: Wait … how does <figcaption>IMG_1234.jpg</figcaption> indicate "IMG_1234.jpg" is _not_ a text alternative in a way that <img src=IMG_1234.jpg> does not, given that <figcaption> implies a label (accName) and @src can only be repaired into a label (accName)? [04:29:06.0000] <Philip`> hsivonen: \x{1234} will give you U+1234, and it sounds like doing "... -e 'use open (:std :utf8); s/foo/bar/'" will make it treat STDIN/STDOUT as UTF-8 [04:29:37.0000] <Stevef> webben: in firefox for example figcaption is mapped to IA2 caption role, also figcaption does not map to img accname itt maps to figure (role=group) accname (again in firefox as its the only browser that has implemented figure/figcaption semantics) [04:30:27.0000] <Philip`> /me hadn't seen the 'open' pragma before, and usually does "binmode STDIN, ':utf8'" etc instead [04:32:09.0000] <Ms2ger> /me suggests python [04:32:33.0000] <Philip`> hsivonen: ...or "perl -CSD -pi -e 's/foo/bar/'" [04:33:04.0000] <Philip`> (TIMTOWTDI) [04:33:38.0000] <jgraham> Does perl mainly accept famous mathematical constats as command line arguments? [04:36:38.0000] <Philip`> -pi isn't actually an argument, it's two single-character arguments that can be conveniently smushed together, except it can't be smushed with the -e since the -i consumes the remainder of its token as the extension to use when saving a backup of the files it's modifying, so you can't write "-pie" [04:36:45.0000] <Philip`> if I remember correctly [04:37:47.0000] <Philip`> (A lot like sed, because that's one of the many things Perl copied from) [04:48:56.0000] <hsivonen> Philip`: hooray. Thank you. [05:01:37.0000] <jgraham> hsivonen: Any idea if it is intentional that gecko allows you to edit a script's code from beforescriptexecute? [05:02:36.0000] <hsivonen> jgraham: probably not intentional per se, but probably there's also no intent to prevent it [05:03:20.0000] <jgraham> OK, seems like the spec tries to prevent it and I'm wondering if I should file a bug [05:06:33.0000] <jgraham> I can't really think of a use case for changing the spec other than it seems like it could be simpler to implement and browser-js style site patching, which isn't relevant to the web use case [05:08:31.0000] <jgraham> Hixie: Is there a reason the spec is the way it is? [05:08:41.0000] <hsivonen> jgraham: what's the use case for the spec trying to prevent it? [05:11:15.0000] <jgraham> hsivonen: Dunno, that's why I'm asking Hixie [05:22:50.0000] <hsivonen> I wonder if the creator of this image meant the connotations of the meme: http://a11ymemes.tumblr.com/post/23033757039/a-crying-woman-with-her-hand-over-her-face-says [05:23:30.0000] <hsivonen> It's interesting how a11y memes often don't use the idiomatic meme names [05:23:46.0000] <Ms2ger`> Is that First World Problems? [05:23:52.0000] <hsivonen> e.g. http://a11ymemes.tumblr.com/post/23037268835/philosophical-dinosour-says-if-html5-has doesn't say philociraptor [05:23:55.0000] <hsivonen> Ms2ger`: yes [05:24:18.0000] <hsivonen> and http://a11ymemes.tumblr.com/post/23104907769/action-man-so-i-built-this-bad-ass-website-and says "Action man" instead of "The Rock driving" [05:24:54.0000] <hsivonen> also Scumbag Steve is often called "lazy douchebag" on a11ymemes [05:25:07.0000] <hsivonen> and there's one without a proper text alternative: http://a11ymemes.tumblr.com/post/23091091673/bitch-pleeease [05:26:43.0000] <jgraham> The nice thing about the text alternative on the a11y memes is that it makes them easier to follow for people not so well versed in the reddit culture (i.e. me) [05:27:22.0000] <jgraham> Just as well they're not hidden away in some hard-to-access invisible attribute, really [05:28:57.0000] <jgraham> (although it doesn't help much if the text alternative is missing the same context that I am missing) [05:29:11.0000] <Philip`> /me thought part of the appeal of memes was their obscurity to outsiders [05:29:26.0000] <Philip`> so highly visible text alternatives defeat the point [05:29:31.0000] <jgraham> Appeal to whom? [05:29:50.0000] <Philip`> To the people who use them and want to feel clever and superior [05:30:14.0000] <Stevef> hsivoenen: It's interesting how a11y memes often don't use the idiomatic meme names - they are used when known, like i didn't know that philosoiraptor was called that [05:30:33.0000] <jgraham> Kind of a weird way to feel clever and superior. "I put some text on a picture of a cat" [05:30:44.0000] <Philip`> When memes appear on daytime TV programmes, they're no longer cool [05:31:02.0000] <hsivonen> Philip`: I thought they were about humor and not about superior feelings [05:31:18.0000] <Stevef> hsivonen: feel free to comment on memes you think do not have correct text alternatives [05:31:23.0000] <jgraham> It seems like the more-traditional method of quoting the entirity of uylesses from heart would be a better way to make yourself feel smug [05:31:49.0000] <jgraham> *ulysses [05:32:01.0000] <jgraham> Dammit, I am so inferior I can't even spell the title [05:33:25.0000] <hsivonen> Stevef: I'm not saying they aren't correct. I just think it's interesting how they tend to be more descriptive than idiomatic. [05:34:54.0000] <Stevef> hsivonen:note many of the memes are created by someone for whom english is a second language [05:35:04.0000] <webben> Stevef: I see, yes. [05:36:04.0000] <Stevef> webben: we have a way to go before whats implemented actually works as it should [05:39:12.0000] <Philip`> Maybe I'm confusing different types of memes - there are some that are popular because they're inherently entertaining, and others that are popular in niche communities despite having no inherent value since the context within that community makes them entertaining, and so people will find them entertaining and propagate them in order to be part of that community [05:40:03.0000] <webben> Stevef: @src without alt is sufficient to determine there is no text alternative. It seems like moving the filename to the <figcaption> is just shifting the inaccessibility around. [05:40:28.0000] <webben> A badly labelled group is not much better than a badly labelled image. [05:40:57.0000] <webben> especially when UAs may use the group label to caption the image. [05:42:03.0000] <webben> Stevef: I suppose figcaption allows image hosts to strip some of the gubbins out of the @src, but I'm not sure they can do that much more intelligently than client software. [05:42:04.0000] <Philip`> (e.g. that Rick Astley video doesn't seem particularly entertaining by itself - its only value is to show you're part of a community that understands the wider context of tricking people into watching it) [05:42:52.0000] <Philip`> (and it loses that value once pretty much everyone in the world understands it) [05:43:41.0000] <webben> Stevef: cf. http://www.nvda-project.org/ticket/51 and http://www.nvda-project.org/ticket/1989 [05:43:48.0000] <jgraham> I think I would like to spread the meme that the co-opting of the word "meme" to mean "internet in-joke" is uncool [05:43:52.0000] <Philip`> (whereas photos of cats are always enjoyable) [05:44:28.0000] <Stevef> webben: true except that one at least rpovides and indication that it is something other than a text alternative and provides a method to identify an image, where as now if images don't have an alt they generally are ignored (unless inside a link) the figure/figcaption provides a method to say hey there is an image here, it has a caption of some sort but no text alternative [05:44:48.0000] <hsivonen> btw, will 3rd-party AT be locked out of Metro? [05:44:58.0000] <webben> Stevef: But it's not a meaningful caption : it's basically the same repair text. [05:45:25.0000] <webben> Stevef: @src is not a text alternative therefore that _is_ an indication that @src is not a text alternative. [05:46:06.0000] <webben> Stevef: How client software surfaces the fact that it is repairing the absence of a text alternative is a different matter. [05:51:43.0000] <webben> Stevef: I agree there's a problem with distinguishing significant and insignificant images without text alternatives, although I also think one could apply more intelligent heuristics than tend to be applied. [05:52:08.0000] <webben> e.g. images below a certain rendered size are not significant (ad/tracking pixels) [05:55:04.0000] <Stevef> webben: looking at flickr they provide a caption for all images http://www.flickr.com/photos/ which is usually better than the src value and in other cases there is no disernible source http://www.flickr.com/photos/pitschspics/7221150834/ so the figure/caption technique would be useful [05:58:28.0000] <Stevef> webben or in these cases: http://www.flickr.com/photos/pitschspics/with/7221143506/ the caption while not that useful is a whole lot more useful than the src e.g. 7221058684_abb86bb0d8_m.jpg also note the alt and the caption are the same [06:00:22.0000] <webben> Stevef: I agree that where the original filename is changed to some gibberish in a CDN URL as there, preserving the original filename in some sort of association with the <img> is better than nothing. [06:02:05.0000] <Stevef> webben: what I am trying to say is in practice when looking at how uploaded images on photo istes are captioned, using the figure/figcaption makes sense (to me) [06:09:17.0000] <Stevef> webben: plus i like the figure/figcaption becuase it provides users with an easier way to provide some meaningful text thats programmatically associated with an image, it may not be the best text alternative, but its something and it can be conveyed that is NOT an alt text whereas if you use alt or title the AT does not know the difference (unless it goes in a qureries the DOM directly) as... [06:09:18.0000] <Stevef> ...they both populate the accname. [06:18:22.0000] <annevk> didn't know AT treats alt and title the same [06:18:26.0000] <annevk> that's pretty bad :( [06:21:27.0000] <Stevef> annevk: its not AT as such its how the accessible name calculation works in all browsers (that support it) [06:22:11.0000] <Stevef> annevk:mac works a little differently for images [06:23:49.0000] <Stevef> annevk: essentially if there is no other source for an accessble name then browsers use the title, if alt is present as well as an a title alt is used for the name and title is used for the description [06:24:49.0000] <webben> hsivonen: I think no. http://www.nvda-project.org/ticket/1801 [06:35:48.0000] <annevk> Stevef: mkay [06:42:29.0000] <bjankord> ... [06:53:09.0000] <annevk> hober: first rule about #secret-treehouse is that we don't talk about, definitely not on twitter :p [06:58:41.0000] <jgraham> I thought the first rule of #secret-treehouse was "No smoking"? [06:59:18.0000] <annevk> hahaha [07:12:49.0000] <scott_gonzalez> Is there anything in DOM Events that says whether mouseenter should or should not be triggered as soon as the page loads? [07:14:15.0000] <scott_gonzalez> If you don't move your mouse at all, then load a page, should you get an event for the element that loads into the position that your mouse is in? [07:14:27.0000] <scott_gonzalez> Firefox seems to be the only browser that does this. [07:14:40.0000] <scott_gonzalez> It happens if content is shown after page load as well. [07:17:09.0000] <annevk> mouse events actually be defined? hah [07:17:26.0000] <annevk> "you must be new here" (but I know you're not quite...) [07:17:34.0000] <annevk> being* [07:19:24.0000] <scott_gonzalez> :-/ [07:22:04.0000] <annevk> it's not even tied to things like hit testing [07:22:09.0000] <annevk> wait what am I saying [07:22:12.0000] <annevk> hit testing is not even defined [07:26:03.0000] <smaug____> foolip: the bugs you've been filing are hilarious. [07:26:20.0000] <smaug____> /me can't file audio api bugs, because tracker is locked to wg members only [07:26:20.0000] <foolip> smaug____, glad to be of service! [07:26:35.0000] <annevk> pointers? [07:26:36.0000] <foolip> smaug____, I can file stuff on your behalf if you want [07:26:49.0000] <foolip> but send me an email, right now we're busy filing bugs :) [07:27:01.0000] <annevk> you should file a tracker issue on using W3C Bugzilla [07:27:10.0000] <smaug____> annevk: http://www.w3.org/2011/audio/track/issues/raised [07:27:28.0000] <foolip> annevk, we did suggest using bugzilla, but the chair though issues were fine [07:27:40.0000] <foolip> and we just want to get the issues out into the open to begin with [07:29:07.0000] <annevk> that's a lot of issues [07:29:42.0000] <annevk> Bugzilla is nice because it works across specs [07:33:42.0000] <smaug____> annevk: the number of issues tells a lot about the spec. I feel it is hard to even start reviewing it properly, since nothing is really defined [07:33:56.0000] <smaug____> it is more like a simple API description [07:35:11.0000] <smaug____> after reading that spec, all the other specs feel a lot better though :) Perhaps I'll complain less about them in the future. [07:36:49.0000] <jgraham> So what are the rules about whoch Element-targeted events get corresponding onfoo attributes? [07:37:30.0000] <smaug____> no rules [07:37:52.0000] <jgraham> It's just according to Hixie's whim? [07:38:02.0000] <smaug____> or someone else [07:38:32.0000] <jgraham> Why are there not rules? [07:38:38.0000] <jgraham> I need structure dammit [07:38:57.0000] <smaug____> see the topic [07:39:08.0000] <jgraham> We should have a formal process with proposals and votes [07:39:16.0000] <jgraham> Then this sort of thing will never happen [07:39:22.0000] <annevk> smaug____: nah please keep complaining [07:39:37.0000] <annevk> smaug____: better to know [07:40:02.0000] <annevk> jgraham: I'm sure you can get the W3C to set up a workshop around this topic [07:40:33.0000] <annevk> jgraham: bring in some academics, some people writing software, couple of users, some developers [07:40:51.0000] <jgraham> I could have a community group [07:41:07.0000] <jgraham> The event handler content attribute user's group [07:41:09.0000] <annevk> jgraham: they'd prolly recommend that now yeah :) [07:41:24.0000] <annevk> jgraham: best to just announce it on twitter [07:41:25.0000] <scott_gonzalez> jgraham: It would be nice if all events has corresponding onfoo attributes. [07:41:42.0000] <jgraham> It would be like a rehabilitation centre [07:41:45.0000] <annevk> scott_gonzalez: except for the events we want to drop [07:42:15.0000] <annevk> the other problem is that events are case-sensitive [07:42:18.0000] <annevk> event handlers are not [07:42:31.0000] <scott_gonzalez> annevk: If there are events that don't have corresponding attributes and the events are being dropped, then I think it's fine if they're not added [07:42:38.0000] <annevk> I think all events with casing have no event handler attribute [07:42:53.0000] <annevk> most of those are also candidate for being dropped, though not all [07:43:04.0000] <scott_gonzalez> They should just be lowercased. [07:43:13.0000] <scott_gonzalez> We shouldn't have events with same name but differnet casing. [07:43:19.0000] <scott_gonzalez> We actually do this in jQuery UI. [07:43:24.0000] <smaug____> DOMContentLoaded is perhaps the only one which shouldn't be dropped [07:43:30.0000] <annevk> smaug____: yeah [07:43:36.0000] <annevk> smaug____: and you can't really change casing I think [07:43:38.0000] <scott_gonzalez> We provide callbacks with mixed casing, but trigger associated events lowercased. [07:43:42.0000] <annevk> smaug____: because people check event.type [07:43:46.0000] <smaug____> yeah [07:44:09.0000] <annevk> but you could have ondomcontentloaded I guess [07:44:11.0000] <annevk> ugly as hell [07:44:29.0000] <smaug____> I wouldn't add a special case [07:44:34.0000] <smaug____> we could just add a new event [07:44:40.0000] <smaug____> contentloaded [07:44:52.0000] <smaug____> it would fire right after DOMContentLoaded [07:45:11.0000] <smaug____> then deprecate DOMContentLoaded [07:45:19.0000] <smaug____> whatever deprecation means [07:45:22.0000] <annevk> oh god [07:46:17.0000] <smaug____> though, this wouldn't still let all the events to have onfoo handler [07:46:31.0000] <smaug____> scripts and addons can always dispatch any kinds of events [07:47:04.0000] <annevk> yeah, we should have the ele.on() concept instead [07:47:16.0000] <smaug____> there are still problems with that approach [07:47:39.0000] <annevk> yeah, for one the details are not yet defined :) [07:48:02.0000] <scott_gonzalez> The attributes are useful for testing if the UA supports the event, I don't see how .on() is related. [07:48:57.0000] <annevk> scott_gonzalez: oh solely from that perspective [07:49:16.0000] <annevk> an event being supported is such a weird concept [07:49:26.0000] <bjankord> Does anyone know if the w and h values in srcset are similar to min-width or max-width media queries? [07:49:46.0000] <annevk> bjankord: max-width I think [07:49:54.0000] <bjankord> hmm [07:50:15.0000] <bjankord> I think there is a good case where users would want it to be min-width [07:50:29.0000] <bjankord> Thinking of progressive enhancement/mobile first [07:50:53.0000] <bjankord> Should authors be able to 100minw or 100maxw in srcset? [07:52:16.0000] <annevk> bjankord: there's some mailing list discussion around that [07:52:45.0000] <scott_gonzalez> Heh, "some" [07:53:58.0000] <bjankord> annevk: Has there been any agreement on this topic on the mailing list? [07:54:09.0000] <bjankord> annevk: any link to view the list? [07:54:20.0000] <scott_gonzalez> http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2012-May/thread.html [07:54:41.0000] <scott_gonzalez> There are about a dozen different threads. [07:55:00.0000] <scott_gonzalez> Search that page for <picture>, srcset, media queries. [07:55:33.0000] <bjankord> scott_gonzalez: haha yeah I assumed there would be a lot to dig through, thanks for pointing me in the right direction [07:56:26.0000] <scott_gonzalez> Most of the debate right now seems to be around whether there are other useful media queries for images. [07:56:46.0000] <scott_gonzalez> And how to handle choosing an image based on bandwidth. [07:57:01.0000] <annevk> http://lists.w3.org/Archives/Public/public-whatwg-archive/ might be nicer to browse through [07:57:09.0000] <bjankord> bandwidth is such a changing variable [07:57:20.0000] <bjankord> what might be considered fast for you might be slow for me [07:57:46.0000] <bjankord> I would say the best would be to allow the user to choose their images, regular vs hd [07:57:56.0000] <bjankord> just like the user can choose there default font size [07:58:23.0000] <bjankord> there would be a browser setting the user would click to choose which type of images to display [07:58:42.0000] <annevk> users don't want to do that [07:59:01.0000] <annevk> as much as possible it should just work [07:59:10.0000] <jgraham> annevk: They also don't want the site to suddenly look like crap due to bad heuristics [07:59:27.0000] <jgraham> The whole bandwidth thing is just a bad idea [07:59:40.0000] <bjankord> jgraham: agreed [07:59:49.0000] <annevk> jgraham: I was not suggesting it wasn't... [07:59:56.0000] <jgraham> annevk: Sure [08:00:10.0000] <jgraham> I'm probably preaching to the choir [08:00:54.0000] <jgraham> Hmm, so is it defined anywhere how scripting in SVG works? [08:01:29.0000] <annevk> there is http://dvcs.w3.org/hg/dap/raw-file/tip/network-api/index.html [08:02:01.0000] <annevk> but I'm not sure why that is there... [08:02:18.0000] <annevk> oh, some B2G thing [08:03:22.0000] <annevk> and a whole new object for it? [08:03:23.0000] <annevk> oh well [08:03:53.0000] <jgraham> Yeah, this is the problem with fragmentation. Everyone wants their own little bit of namespace to play in [08:04:22.0000] <jgraham> That API looks horrible [08:08:57.0000] <smaug____> blame volkmar [08:10:41.0000] <annevk> so sad that volkmar could not finish HTML forms instead [08:13:03.0000] <Ms2ger`> I hear there's an intern who's going to work on forms over the summer [08:13:11.0000] <gsnedders> interns++ [08:13:16.0000] <Ms2ger`> But apparently only for mobile :/ [08:13:35.0000] <Ms2ger`> gsnedders, looking for an internship? :) [08:13:47.0000] <gsnedders> Ms2ger`: Nah. [08:14:47.0000] <annevk> Ms2ger`: there have been unsuccessful WF2 internships in the past if I remember correctly [08:15:01.0000] <Ms2ger`> Wouldn't surprise me [08:16:06.0000] <smaug____> forms are hard [08:16:16.0000] <smaug____> certain types at least [08:16:24.0000] <Ms2ger`> Opera did them first, so what's the point? :) [08:17:23.0000] <smaug____> :p [08:46:20.0000] <dglazkov> good morning, Whatwg! [08:51:45.0000] <jreading> hey all, question… [08:53:05.0000] <annevk> Ms2ger`: we even had type=uri before I got it changed all over to type=url [08:53:15.0000] <jreading> our designers don't want to go near html5 input validation, because they are adverse to bubble tooltips. I know you can style them, but is there a roadmap for that, in terms or extending the api to allow custom DOM insertion? [08:53:47.0000] <annevk> you can use setCustomValidity() [08:54:42.0000] <annevk> jreading: might be to start a thread on that though or a wiki page on Forms listing the shortcomings of what we have today [08:55:17.0000] <annevk> jreading: though for the 8 or so years we've been doing this nobody has wanted to take a hard look at styling forms :( [08:55:34.0000] <annevk> it's a rather hard problem [08:56:09.0000] <Stevef> a shortcoming of the validation message bubbles as implemented is that if you have more than one invalid fiield only the first bubble shows [08:57:07.0000] <jreading> annevk: i'm not so concerned with the native styling, but just being able to change the placement and affect behavior [08:58:33.0000] <annevk> Stevef: sounds like a UI bug, yeah [08:58:56.0000] <annevk> jreading: that does sound like changing the native styling... [08:59:24.0000] <Stevef> annevk: from what i can remember when testing its same behaviour across browsers [08:59:26.0000] <jreading> i suppose so, but I'm not asking for anyone to do that legwork, just need hooks for all that stuff [09:00:19.0000] <annevk> Stevef: just like type=file evolves over time so can this; UI is not standardized [09:00:44.0000] <annevk> jreading: I know, defining the hooks is the problem I'm referring to [09:00:58.0000] <jreading> right, ok [09:01:10.0000] <Stevef> yannevk: understand but it limits its practical use currently [09:02:20.0000] <jreading> just like one can do <form novalidate> it would great to do <form message="inline | bubble | group"> [09:02:22.0000] <jreading> or something [09:02:42.0000] <annevk> Stevef: I guess that would be good feedback for that wiki page I was suggesting [09:02:53.0000] <TabAtkins> Ms2ger`: Uh, no. Bugzilla is perfectly fine for random people to file bugs in on CSS drafts. We just don't want *conversation* happening on them - if you file a non-trivial bug, you should also start a thread. [09:02:57.0000] <Stevef> annevk:ok [09:03:27.0000] <annevk> Stevef: although not every browser has every aspect yet, we should probably evaluate again what we have so far and what works and what doesn't [09:05:17.0000] <annevk> jreading: it would be better to have use cases or statements of what does not work and why then some syntax that can mean pretty much anything ;) [09:05:25.0000] <annevk> I'll create a wiki page [09:10:03.0000] <annevk> Stevef: jreading: http://wiki.whatwg.org/wiki/Forms [09:11:09.0000] <Stevef> annevk:cool will provide some detail later [09:13:12.0000] <jreading> annevk: excellent [09:15:33.0000] <jreading> if you can get me an account, I will diligently and conservatively add to this [09:17:45.0000] <annevk> jreading: email? [09:18:24.0000] <jreading> annevk: sent to hixe, that ok? [09:18:43.0000] <annevk> jreading: then you'll have to wait until he gets online [09:18:58.0000] <annevk> if you pm me with it I can do it now [09:19:06.0000] <annevk> (well if you're quick, I have to go) [11:17:47.0000] <Ms2ger`> Huh [11:18:59.0000] <Ms2ger`> 2 and a half hours after being sent, http://lists.w3.org/Archives/Public/www-html/2012May/0001.html got two replies with the same suggestion, within a minute of each other [11:40:34.0000] <eberon_> /join #reddit-dev [11:40:38.0000] <eberon_> bleh [11:49:40.0000] <jreading> http://wiki.whatwg.org/wiki/Forms [11:50:21.0000] <annevk> jreading: cool [11:50:31.0000] <annevk> jreading: fyi, use cases and requirements are separate things ;) [11:51:17.0000] <annevk> jreading: a use case is a scenario of something you want to accomplish; one or more requirements typically follow from that [11:51:47.0000] <jreading> edited [11:53:40.0000] <annevk> anyway, that's pretty awesome [12:01:52.0000] <annevk> Hixie: you around/ [12:02:17.0000] <annevk> Hixie: it would help if you fixed the errors around minimum/maximum in the spec [12:02:59.0000] <annevk> Hixie: srcset spec that is [12:05:40.0000] <tantek> /me reads the wiki page [12:05:53.0000] <tantek> jreading, this is good work http://wiki.whatwg.org/wiki/Forms [12:06:57.0000] <tantek> there's still a lot we have to do to make the "webforms2" features work well in practice - very happy to see someone gathering specific real world examples where the current functionality in HTML is insufficient to achieve the desired effect. [12:07:23.0000] <tantek> I have a few thoughts on more UI elements as well for growing from forms to full on apps. But forms is a good place to focus first. [12:08:51.0000] <annevk> forms are hard :( [12:09:09.0000] <annevk> we should really have sortable tables by now in markup imo [12:09:23.0000] <annevk> but we don't :) [12:10:43.0000] <anatolbroder> I don’t understand your real life examples from http://wiki.whatwg.org/wiki/Hgroup_element#The_Telegraph_.28UK.29 The Telegraph doesn’t use the hgroup element. So why did you place it in the article? [12:11:17.0000] <annevk> "This page documents use cases and existing usage patterns related to marking up subtitles and taglines within headers." [12:12:30.0000] <anatolbroder> annevk: thanks. [12:15:50.0000] <anatolbroder> What is *your* proposed solution for a classic 2-4 sentences long byline between the <h1> element and the text body? A <h2 class="byline"> or a <p class="byline">? [12:18:04.0000] <TabAtkins> the latter [12:18:38.0000] <anatolbroder> TabAtkins: thanks [12:18:44.0000] <jgraham> One problem with forms is that peole aren't really interested in implementing forms features [12:20:22.0000] <annevk> I think if forms/editing/UI events were way better defined it would be less problematic [12:20:39.0000] <annevk> because then at least it would make hacking easier [12:20:49.0000] <annevk> now you're just asking for regressions, in a way [12:21:19.0000] <annevk> "forms are hard, lets go shopping" [12:21:47.0000] <jgraham> Yeah maybe [12:21:53.0000] <jgraham> But also UI is hard [12:22:42.0000] <jgraham> Would be kind of nice for Gecko to imeplement more of the forms already in HTML :| [12:23:10.0000] <annevk> no but you see, Gecko's going to have the best implementation possible, so they're simply not doing it... [12:25:27.0000] <jgraham> Well the main problem seems to be that they don't think it's needed for B2G [12:26:09.0000] <jgraham> Dunno how they plan to get form controls on there. Maybe the who UI will be built with jQuery [12:26:33.0000] <Ms2ger`> jgraham, oh, no, we've got someone who will implement them just for mobile [12:28:14.0000] <annevk> jreading: heh just reading your past tweets [12:28:20.0000] <annevk> jreading: no goat farm? [12:30:18.0000] <jreading> not yet [12:30:58.0000] <annevk> put me down for some cheese if you do [12:31:06.0000] <jreading> my facebook shares didn't hit $1000 yet [12:35:50.0000] <jgraham> Oooh cheese [12:36:09.0000] <jgraham> I would prefer to keep pigs I think [12:36:30.0000] <jgraham> Although goat meat is also delicious [12:36:51.0000] <jgraham> I recommend the Jamacian recipe "Curried Goat" [12:36:56.0000] <jreading> how could you eat this? [12:36:57.0000] <jreading> http://goteaminternet.com/img/docs/57570.jpg [12:37:45.0000] <jgraham> Well first you would let it grow up a bit [12:37:51.0000] <jgraham> Then you would kill it [12:38:04.0000] <jgraham> Then you would butcher it [12:38:06.0000] <gavinc> Facebook, or the goat? [12:38:09.0000] <jgraham> Then you would cook it [12:38:16.0000] <jgraham> Then you would eat it [12:38:29.0000] <jgraham> Nah, facebook will eat itself [12:39:12.0000] <gavinc> Ugh, my new millionare friends are going to be insufferable for weeks [12:39:52.0000] <othermaciej> are you friends with facebookers? [12:40:14.0000] <jgraham> That's so far beyond a first world problem, I don't know how to classify it [12:40:42.0000] <gavinc> yeah, a couple [12:46:17.0000] <Ms2ger`> I'm glad they went public... Now all the Mozillians who sold their souls to facebook can come back [12:47:51.0000] <annevk> Ms2ger`: probably not if they want to cash in on options [12:55:01.0000] <gavinc> annevk: indeed, golden handcuffs better than the normal kind but still kinda suck [12:56:36.0000] <gavinc> gavin: palo alto world problem? [12:56:43.0000] <gavinc> jgraham: palo alto world problem? [13:03:23.0000] <gavin> schrep hired me at Mozilla [13:03:31.0000] <gavin> I should get back in touch now that he's a bazillionaire [13:13:54.0000] <annevk> pablof: I want to retweet, but lack of proper punctuation puts me off [13:14:15.0000] <zewt> twitter is designed to destroy grammar [13:14:30.0000] <zewt> it's pretty absurd that they havn't raised it to something sane, ~500 characters or so [13:14:46.0000] <pablof> annevk: yeah, i know, i should set a timer before pushing the tweet button [13:15:18.0000] <zewt> twitter is pretty much completely different in different languages, depending on the character density, heh [13:15:47.0000] <zewt> way less of a restriction in cjk than western languages [13:16:12.0000] <pablof> well, yeah, but that's like saying that shorthand is very efficient [13:16:22.0000] <pablof> it is, but man, you have to memorize too many symbols! [13:16:32.0000] <pablof> s/too/so/ [13:17:04.0000] <zewt> basically twitter's limit is just too low for decent communication in english, you always end up having to remove content [13:23:00.0000] <TabAtkins> zewt: Maybe learn to communicate better? Otherwise, use tweeplus.com, which is the only longtweet service where your tweets will actually survive the service dying. [13:24:06.0000] <zewt> uh, having more to say than can be expressed in 2.5 bytes of english is not a lack of communication skill [13:24:11.0000] <zewt> it's just a failure of a medium [13:25:38.0000] <TabAtkins> 2.5 bytes of english? Where'd you get that number? [13:27:34.0000] <zewt> that sure is what twitter feels like [13:27:45.0000] <zewt> oh you used an adverb, that's too much for us [13:27:59.0000] <TabAtkins> I very rarely have trouble communicating in Twitter. [13:28:39.0000] <zewt> i have to spend more time trying to find ways to say what I want to say within the absurd length restrictions than saying it in the first place [13:30:56.0000] <TabAtkins> Shrug. Sounds like you're just bad at Twitter. It's a restricted medium on purpose. [13:32:25.0000] <zewt> there is no valid reason for a length limitation that allows English speakers to say 1/2 or 1/3 as much as what a Japanese user can say [13:34:14.0000] <bjankord> I've been reading through the mailing list now for a few hours and havn't seen anything about using minw or maxw in srcset [13:34:39.0000] <bjankord> Has there been any talk about allowing authors the ability to choose if the w value in src set means min-width or max-width [13:38:53.0000] <bjankord> ... [13:39:09.0000] <TabAtkins> bjankord: You haven't seen the talk about changing the syntax to "max-width:NNNpx" instead of "NNNw"? [13:39:28.0000] <bjankord> Not yet, I've been reading through the mailing list [13:39:55.0000] <bjankord> TabAtkins: can you link me there [13:40:37.0000] <TabAtkins> Yeah, gimma a sec. [13:41:06.0000] <bjankord> TabAtkins: Thanks! [13:41:22.0000] <TabAtkins> For example, check uot the "Problems with width/height descriptors in scrset" thread. [13:41:38.0000] <TabAtkins> Which is specifically about that issue. [13:55:38.0000] <annevk> guess I should keep track of when Gecko forgets to raise spec issues to entertain @bz_moz [13:56:51.0000] <annevk> oh but I guess since everyone is a Mozillian me raising the issue still counts [13:57:44.0000] <TabAtkins> Heh. [13:58:27.0000] <Ms2ger`> annevk, shouldn't be too hard... Just look at canvas bugs :) [14:02:36.0000] <bjankord> TabAtkins: So after reading through the mailing list, it's my understanding the srcset spec will likely be updated to use min-width: and max-width: ? [14:06:39.0000] <jgraham> annevk: I am quite in favour of a "name and shame" policy when people implement features but don't make any sharable tests for them [14:07:02.0000] <jgraham> And yes, Opera could do this better [14:07:05.0000] <charlvn> lol @ http://i.imgur.com/d2e5k.png [14:07:14.0000] <jgraham> (but we are actively tring to) [14:08:46.0000] <annevk> /TR/ash and 5hit... [14:09:18.0000] <Ms2ger`> jgraham, you mean, everyone :) [14:10:07.0000] <annevk> runeh would suggest a Tumblr [14:10:15.0000] <TabAtkins> bjankord: It's not decided. We've provided feedback, and we can argue it, but it ultimately falls on Hixie. [14:10:44.0000] <Ms2ger`> s/Hixie/browser vendors/, as he would say himself [14:11:04.0000] <TabAtkins> Of course. [14:11:14.0000] <Philip`> s/browser vendors/users who choose which browser to use/, surely [14:11:21.0000] <jgraham> Ms2ger`: Yeah, lots of shame to go around at the moment [14:12:39.0000] <Ms2ger`> So, I've got some tests I want to submit [14:13:06.0000] <Ms2ger`> Do we have a w3c-test.org subdomain? [14:13:37.0000] <jgraham> Yeah [14:13:45.0000] <jgraham> I don't remember what works though [14:13:58.0000] <Ms2ger`> Different ports? [14:14:14.0000] <Ms2ger`> http/https should work... [14:14:16.0000] <jgraham> I think so. If not we obviously need them :) [14:14:39.0000] <jgraham> I'm pretty sure we set up a few subdomains and a few ports [14:14:48.0000] <annevk> Ms2ger`: http://lists.w3.org/Archives/Public/public-html-testsuite/2011Feb/0022.html [14:15:15.0000] <jgraham> ah, annevk is more full of useful information [14:15:20.0000] <Ms2ger`> Great [14:16:23.0000] <annevk> Ms2ger`: more complete: http://www.w3.org/wiki/Testing/Requirements#The_Web_test_server_must_be_available_through_different_domain_names [14:16:32.0000] <annevk> including different ports [14:16:44.0000] <Ms2ger`> Nice [14:16:47.0000] <annevk> and a fancy looking subdomain http://天気の良い日.w3c-test.org/ [14:17:01.0000] <Ms2ger`> I need to set those up for Mozilla too [14:17:05.0000] <annevk> also one that looks French http://élève.w3c-test.org/ [14:21:23.0000] <Ms2ger`> /me files that [15:05:07.0000] <eddiemonge> why microdata vs microformats? [15:05:26.0000] <tantek> eddiemonge - do you have a specific problem you're trying to solve? [15:05:53.0000] <tantek> that helps with figuring out what solution would work best for you [15:08:36.0000] <tantek> eddiemonge - if you have specific questions about microformats, feel free to join and ask in the #microformats channel as well. [15:08:43.0000] <tantek> much more focused community there on that sort of thing [15:09:16.0000] <tantek> and for discussing vocabularies/schema more broadly, there's the #schema channel - I hang out there as well - as do others who work on vocabularies and things like microformats/microdata/RDFa [15:21:55.0000] <eddiemonge> thanks for the suggestions [16:09:59.0000] <TabAtkins> Hm. I know how to implement compositing, and I know how to implement blending. I'm not sure how to implement them together. :/ [16:10:18.0000] <hober> very carefully :) [16:10:38.0000] <TabAtkins> I'm trying to put together a demo in canvas, though! [16:15:16.0000] <gavinc> eddiemonge: also #swig more RDFa and semweb, less microdata [16:18:26.0000] <TabAtkins> Okay, I think I've deciphered the Compositing spec. It doesn't help that they literally use three different variables for the same thing in nearly-adjacent formulas. [16:27:56.0000] <heycam> jgraham, re scripting in SVG there's no wording hooking into anything HTML defines yet, but I hope to add some at some point soon 2012-05-19 [22:09:09.0000] <heycam> when CSS folks come up with properties that take lengths and percentages, what guidelines do you use for determine whether to have the computed value be the equivalent absolute value, or whether to instead leave the computed value as specified? [22:10:17.0000] <tantek> heycam, I think you meant to ask that in irc://irc.w3.org:6665/css ;) [22:10:32.0000] <heycam> tantek, good idea :) [22:10:50.0000] <tantek> :) [03:37:23.0000] <jgraham> heycam|away: So I should assume that SVG scripts are supposed to fire all the same events as HTML scripts, for example? [03:38:12.0000] <Ms2ger> Or that <script for event> has the same result? [03:38:17.0000] <Ms2ger> (Hint: it doesn't) [03:46:42.0000] <jgraham> Well I hadn't got as far as thinking about that particular unpleasantness yet [04:00:05.0000] <Ms2ger> hober, http://i.qkme.me/3pcrk0.jpg [10:50:27.0000] <annevk> <picture> was first mentioned on the WHATWG list by hsivonen [10:50:32.0000] <annevk> #randomfacts [10:50:36.0000] <annevk> http://lists.w3.org/Archives/Public/public-whatwg-archive/2012Jan/0093.html [10:51:09.0000] <annevk> (found by a search for "<picture>", similar elements have almost certainly been proposed in the past) [10:53:03.0000] <miketaylr> annevk: beat you, http://lists.w3.org/Archives/Public/wai-xtech/2007Jun/0080.html [10:53:29.0000] <miketaylr> oh not whatwg... :( [10:53:34.0000] <miketaylr> /me quits [10:54:56.0000] <annevk> miketaylr: heh nice [11:16:37.0000] <WeirdAl> annevk, Ms2ger: question on DOM4 spec about attribute list... is there a defined ordering to how these attributes are stored? [11:16:56.0000] <Ms2ger> Yes [11:17:08.0000] <WeirdAl> Where can I find that? [11:17:27.0000] <Ms2ger> Hixie refuses to spec what it is for parser-inserted elements, though [11:17:38.0000] <WeirdAl> heh [11:19:07.0000] <WeirdAl> well, I'm thinking of the case: var x = document.documentElement; x.setAttributeNS("http://www.w3.org/xlink", "xlink:href", "#foo"); x.setAttribute("lang", "en"); x.attributes[0] // what should that return? [11:19:51.0000] <annevk> Ms2ger: I talked to Hixie about that, what he needs is setting a bunch of attributes at once [11:20:26.0000] <annevk> WeirdAl: xlink:href I think [11:20:40.0000] <Ms2ger> Yes, first-set [11:21:07.0000] <WeirdAl> again, where's that specified? :) [11:21:10.0000] <annevk> WeirdAl: setAttriubte just appends to the list [11:21:25.0000] <annevk> WeirdAl: and append means to the end of the list [11:22:19.0000] <annevk> Ms2ger: in particular for when you have an element that triggers some kind of algorithm on an attribute being set or on being inserted [11:22:20.0000] <WeirdAl> /me found it [11:22:45.0000] <annevk> Ms2ger: all the attributes associated with it need to be available right away and therefore setAttribute cannot be used [11:23:23.0000] <Ms2ger> That's his claim, yes [11:23:51.0000] <Ms2ger> Gecko manages to implement that with sequential setAttributes, though [11:24:05.0000] <annevk> interesting [11:24:21.0000] <annevk> in what order does it call setAttribute? [11:25:35.0000] <WeirdAl> hmm, this is going to make things interesting for my own DOM design. [11:25:47.0000] <Ms2ger> Still reverse source order, IIRC [11:26:12.0000] <Ms2ger> And the element is told when the parser is done adding attributes [11:28:56.0000] <annevk> Ms2ger: aaah [11:30:51.0000] <WeirdAl> /me frowns: DOM4 is not a nice spec to implement [11:31:15.0000] <Ms2ger> The web is not a nice thing to spec [11:31:46.0000] <WeirdAl> true [11:32:24.0000] <WeirdAl> and besides, I'm only going for limited DOM4 support in my implementation, so maybe this is one of those things I just don't need to be that precise on right away [11:34:12.0000] <annevk> WeirdAl: is it just the somewhat weird requirements or does the spec read weird? [11:34:38.0000] <WeirdAl> neither, really [11:35:38.0000] <annevk> man, Breaking Bad does not come back until July [11:35:51.0000] <WeirdAl> I'm just starting an implementation, and if I go full-bore compliance, I'll be spending far too long on it [11:36:07.0000] <annevk> what language? [11:36:13.0000] <WeirdAl> right now it's "get something resembling the DOM working" [11:36:17.0000] <WeirdAl> in pure JavaScript [11:36:27.0000] <WeirdAl> and yes, I have my reasons [11:36:30.0000] <annevk> you know there is one already? [11:37:01.0000] <WeirdAl> if you mean envjs or dom.js, I'm aware of those projects, and I don't think either one will meet my needs [11:37:12.0000] <annevk> k [11:37:17.0000] <WeirdAl> (which really sucks, but oh well) [11:38:27.0000] <annevk> if you find something in the spec that can be done better be sure to let us know [11:39:09.0000] <WeirdAl> imho, the best thing you, Aryeh and Ms2ger can do is start porting the W3C DOM tests for DOM4 [11:39:23.0000] <WeirdAl> checking the tests against what the spec says [11:40:12.0000] <Ms2ger> I've done the first three or so, a while back :) [11:40:23.0000] <WeirdAl> /me blinks [11:42:33.0000] <WeirdAl> are they hosted somewhere? [11:43:13.0000] <Ms2ger> w3c-test.org/webapps/DOMCore/tests/submissions/Ms2ger [11:43:20.0000] <Ms2ger> If I didn't typo that [11:44:28.0000] <WeirdAl> I'll have to bookmark that :) [11:47:11.0000] <Ms2ger> You may be able to reuse dom.js's testing glue [11:47:48.0000] <WeirdAl> I'll need to look [11:49:32.0000] <WeirdAl> /me is just trying to figure out a good way to implement attributes in his funny shadow content/undo history model [11:52:23.0000] <WeirdAl> I think I've got it though [12:44:35.0000] <Ms2ger> gsnedders, hey, any chance for a patch within a year of your last comment in http://code.google.com/p/html5lib/issues/detail?id=157 ? :) [12:45:25.0000] <gsnedders> Ms2ger: Bah! [12:45:45.0000] <gsnedders> /me wonders what happened to the code he wrote for that… [12:45:55.0000] <gsnedders> /me wonders also why he's suddenly feeling sick [12:48:56.0000] <Ms2ger> Heh, sorry [12:49:14.0000] <gsnedders> (Literally around the second you asked.) [13:03:54.0000] <gsnedders> Ms2ger: Amusingly, I was just thinking about html5lib on PyPy. Really want to get the large-scale changes in first that sort out implicit unicode conversion [16:44:34.0000] <rniwa> Hixie, abarth: are you guys there? [16:44:56.0000] <rniwa> it appears that there is a slight bug in the spec with respect to parser-inserted flag 2012-05-20 [20:36:50.0000] <heycam> jgraham, having SVG scripts firing all the same events as HTML, and having the same attributes -- that's what I'd like to aim for, with SVG2 [20:37:03.0000] <heycam> jgraham, what browsers do right now I have no idea :) [01:11:58.0000] <Ms2ger> "They're not intended to contradict" [01:12:04.0000] <Ms2ger> I'm glad to hear that [03:12:39.0000] <Von_Davidicus> Evening. [03:21:54.0000] <Von_Davidicus> XSLT can get information from multiple XML files, right? [03:33:12.0000] <[tm]> Von_Davidicus: yeah [03:41:57.0000] <Von_Davidicus> Sweet! It works! *Dancedancedance* [03:42:32.0000] <Von_Davidicus> My XSLT file is becoming more and more of a conceptual disaster by the minute! Oh, if only I had a moustache to twirl! [03:42:56.0000] <gsnedders> XSLT seems to often end up like that. [03:43:34.0000] <Von_Davidicus> /me cackles evilly. [03:44:12.0000] <Von_Davidicus> Actually, what I want is a camera--that I might photograph my teacher's face when I show him how I can make the simplest things complicated. [03:45:21.0000] <gsnedders> I'm sure you can find some over-complicated camera. [03:45:30.0000] <jgraham> odinho: Didn't we release our XHR testsuite already? [03:45:48.0000] <Von_Davidicus> Well... no, what I meant is the website I'm making for an assignment. [03:45:59.0000] <gsnedders> jgraham: The XHR/CORS one? [03:46:06.0000] <jgraham> Pretty sure annevk got it released during some spec transition [03:46:09.0000] <jgraham> gsnedders: Yes [03:46:20.0000] <gsnedders> /me has no knowledge of that [03:47:08.0000] <jgraham> http://w3c-test.org/webapps/XMLHttpRequest/tests/submissions/Opera/ [03:48:05.0000] <jgraham> Oh it looks like there are more tests that aren't released yet [03:48:08.0000] <jgraham> That sucks [03:48:17.0000] <gsnedders> Is there anyone it doesn't suck to be? [03:48:27.0000] <Von_Davidicus> I'm in a rather silly situation: I must take a basic HTML course, even though I have done website coding professionally (college rules). My most recent assigment was to create a series of webpages on whatever topic--I chose Rune Words from Diablo II: Lord of Destruction. And, well... I am making things complicated. [03:48:43.0000] <jgraham> gsnedders: I'm not sure I understand the question [03:50:06.0000] <gsnedders> jgraham: http://open.spotify.com/track/1Ns3qh9eOmLZZ8mNbfSPT9 [03:50:54.0000] <Von_Davidicus> Why settle for HTML, when you can use XML + XSLT + WTH? [03:51:44.0000] <gsnedders> Von_Davidicus: Because you'll get marked down for needless complexity the teacher doesn't understand? [03:53:53.0000] <Von_Davidicus> Even if I am, I think I'll have made my point--that I've done this sort of thing before. [04:01:40.0000] <gsnedders> FAILED (errors=2710, failures=1345) [04:01:48.0000] <Von_Davidicus> ??? [04:01:50.0000] <gsnedders> Progress on html5lib on Python 3! [04:05:20.0000] <Von_Davidicus> Ah./ [04:23:50.0000] <gsnedders> UnicodeEncodeError: 'utf-8' codec can't encode character '\ud800' in position 0: surrogates not allowed [04:23:53.0000] <gsnedders> Yay! [04:31:08.0000] <jgraham> nice [04:35:25.0000] <gsnedders> Disallowing invalid Unicode strings does seem like a decent change. [05:06:11.0000] <gsnedders> jgraham: Amazingly, simpletree is the most annoying thing to get working with Py3 [05:08:40.0000] <gsnedders> (the real fun is can I get 2to3 and then 3to2 to produce something working, which should then have all strings as Unicode!) [05:58:05.0000] <Von_Davidicus> Almost done updating my XSLT file. Just one last trick I have to pull off before I can claim it's complete. :) [05:58:32.0000] <Von_Davidicus> Actually two more tricks, but the second one is getting a bit over-the-top. [06:01:04.0000] <Von_Davidicus> Question: Would you agree with the statement "XSLT is easier to understand when you think of it as a programming language, rather than a markup language"? [06:02:24.0000] <AryehGregor> I would also agree with the statement "XSLT is easier to understand when you think of it as a programming language, rather than a disposable paper cup". [06:02:43.0000] <AryehGregor> Generally it tends to be easier to understand things when one thinks about them as they are. [06:04:18.0000] <Von_Davidicus> Well... one guy I know decided that a short essay I wrote about XSLT was full of... erm, leavings... and said that he quit reading after I said that XSLT was a programming language, not really a markup language. [06:06:59.0000] <Von_Davidicus> My response was that it had "if" statements, it had "loops", it had input (an XML file), it had output (an XML DOM), and it was a pain in the butt to figure out, so it felt like a programming language to me. [06:11:00.0000] <AryehGregor> It's definitely a programming language. [06:12:08.0000] <AryehGregor> What's it marking up? [06:17:20.0000] <Von_Davidicus> Well, I'm not sure. I know it's an XML-based language, and the last two letters stand for "Markup Language" [06:17:58.0000] <Philip`> It's much more useful to look at what something is, than at what its name says it is [06:18:10.0000] <Von_Davidicus> Point taken. [06:20:30.0000] <jgraham> Well yeahFor sanity I find it is best not to look at XSLT at all [06:20:42.0000] <jgraham> s/Well yeah// [06:21:13.0000] <Von_Davidicus> How about Schema? do you ever look at XML Schema? :) [06:21:44.0000] <Philip`> Of course not [06:21:49.0000] <Ms2ger> What he said [06:21:54.0000] <jgraham> XML Schema are very important [06:22:11.0000] <Ms2ger> Actually, Philip` might look at XML Schema more often than at his canvas tests [06:22:16.0000] <jgraham> They are the perfect example of why, just because something is in a W3C spec, it doesn't mean it's a good idea [06:22:53.0000] <Von_Davidicus> ... but... but... *Has used schemas several times.* [06:23:30.0000] <Ms2ger> Don't worry [06:23:35.0000] <Ms2ger> You're already lost [06:23:48.0000] <Von_Davidicus> ??? [06:23:56.0000] <Philip`> /me sometimes gets the feeling that people write XML Schema schemas out of a sense of obligation ("we're designing an XML-based language and XML-based languages are meant to have Schemas, right?") rather than because they have any use for it [06:24:39.0000] <Von_Davidicus> Personally, I've found Doctypes more useful. I can specify Entities in them. [06:24:44.0000] <jgraham> Happily EmotionML provides those people with a convenient way to express the frustration this presumably causes [06:24:51.0000] <Ms2ger> The way that people write dtds because they want to use <blink> to make a point and they're employed by the W3C? [06:27:03.0000] <jgraham> Oh actually it turns out that EmotionML doesn't allow yu to express any emotions at all [06:27:17.0000] <Philip`> /me wrote a (internal) DTD because he wanted to use <canvas> in an XHTML 1.0 document without the validator noticing [06:27:20.0000] <jgraham> Because it needs an add on emotion vocabulary [06:27:31.0000] <Philip`> /me subsequently learned the error of his ways [06:27:37.0000] <jgraham> And the WG couldn't agree on a set of emotions to include by default [06:28:06.0000] <jgraham> So instead they published a Note suggesting possible emotions that you might want to express [06:28:29.0000] <jgraham> (using URIs to provide emotion namespacing of course) [06:28:49.0000] <Von_Davidicus> So what do you meant, I'm already lost? [06:32:11.0000] <Philip`> Von_Davidicus: You have strayed from the path of righteousness (i.e. of general scepticism and cynicism about XML and all related technologies) and there is no hope of recovery [06:34:06.0000] <Von_Davidicus> Yes. Yes, there is. And I have followed that path as well. On that path, there is a signpost, and it reads "<!DOCTYPE html PUBLIC "-//IETF//DTD HTML 2.0//EN">" [06:34:10.0000] <Von_Davidicus> :D [06:34:12.0000] <Philip`> (I'm not sure that's true, though - some people have got better) [06:36:12.0000] <Philip`> The IETF is stuck in a layby just off the path to righteousness, so you don't want to go that way either :-) [06:37:01.0000] <Von_Davidicus> Yes, I've done pages in HTML 2.0--as a lark. [06:37:57.0000] <Von_Davidicus> To be honest, the *only* times I've done a webpage that *required* XML technologies was when I did a webpage that was XHTML + SVG (and maybe + MathML). The rest, well--HTML does fine. [06:42:06.0000] <webben> jgraham: EmotionalML is one of the funnier projects to come out of web standards works. :) [06:43:20.0000] <Von_Davidicus> So why are you all so cynical about XML? [06:44:32.0000] <webben> Von_Davidicus: http://wiki.whatwg.org/wiki/Namespace_confusion would be the start of an answer. [06:45:15.0000] <webben> Von_Davidicus: See also http://wiki.whatwg.org/wiki/Why_no_namespaces [06:45:46.0000] <webben> I think someone produced a similar document for why draconian error parsing is not such a great idea [06:45:52.0000] <AryehGregor> Von_Davidicus, you know HTML5 supports inline SVG/MathML, right? [06:45:56.0000] <AryehGregor> So that doesn't require XML. [06:45:59.0000] <AryehGregor> (anymore) [06:46:27.0000] <Von_Davidicus> I do know that. [06:46:57.0000] <webben> Von_Davidicus: http://annevankesteren.nl/2005/11/draconian ... there have been other examples [06:48:32.0000] <webben> Von_Davidicus: There's little more persuasive than XML experts' content becoming inaccessible thanks to serving XML not text/html. [06:49:00.0000] <webben> Of course, browsers increasingly give up and apply a HTML parser to broken XML, but then you might as well author HTML in the first place… [06:49:11.0000] <Von_Davidicus> What I hope to pull off sometime is XHTML + SVG + SMIL + MusicML + MathML (if I need the last)--but that project can wait for now. [06:49:33.0000] <Von_Davidicus> annevk, why does your website say "SVG Sucks"? [06:49:40.0000] <webben> Von_Davidicus: Trying to produce content that can be accessed in zero browsers? ;) [06:51:37.0000] <webben> Von_Davidicus: It's a joke. (The image is SVG: http://annevankesteren.nl/img/daddy ) [06:53:26.0000] <Von_Davidicus> Ah, okay. [06:54:18.0000] <Von_Davidicus> Webben: Oh, I can do that: Just build a webpage whose content *cannot* be stolen, and I'm good. No, I've wanted to see if I could create a game using nothing but markup languages. [06:54:30.0000] <Von_Davidicus> And CSS and JavaScript. [08:35:54.0000] <Tyson_> Any account creators here? [08:36:00.0000] <Tyson_> I need an account on the wiki [08:36:29.0000] <Tyson_> Please ? [08:47:52.0000] <Ms2ger> annevk, maybe? [10:10:37.0000] <jgraham> I'm not sure who the wiki admins are. Perhaps annevk, Hixie, Lachy, AryehGregor? [10:10:58.0000] <jgraham> gsnedders: So, any progress? [10:13:05.0000] <gsnedders> jgraham: 14 errors, 25 failures (8, 13 in Py2), excluding simpletree. [10:13:25.0000] <gsnedders> Video games happened, and not watching Magnolia. [10:13:35.0000] <jgraham> Also, mercurial is making me a bit sad. I obviously use git too much [10:14:04.0000] <jgraham> It seems like not watching magnolia should have given you more time :) [10:15:07.0000] <gsnedders> *now [10:15:26.0000] <gsnedders> (Also, oh Tom Cruise's role in it…) [10:16:56.0000] <jgraham> Is that who you want to be when you grow up? [10:17:36.0000] <gsnedders> No. Hell no. Just, "really, just really?". [10:24:20.0000] <jgraham> To be fair "really, just really?" is what I think of Tom Cruise in real life too [11:11:42.0000] <ShaneHudson> Has there been any discussion of a html element for modal boxes? Currently it has to be done through javascript (or css I suppose) in quite a non semantic way [11:14:52.0000] <webben> ShaneHudson: http://dev.w3.org/html5/spec/commands.html#the-dialog-element and http://www.w3.org/WAI/PF/aria-practices/#modal_dialog [12:09:55.0000] <Ms2ger> jgraham, would you have time to land https://gist.github.com/2636212 at some point? [12:26:34.0000] <gsnedders> jgraham: We have some utterly weird bugs under Python 2. [12:27:27.0000] <Ms2ger> s/under Python 2//, surely [12:28:11.0000] <gsnedders> They're not weird under Python 3, they're obvious hard errors. [12:33:18.0000] <gsnedders> If you'd have told me Simpletree would be the hardest part of html5lib to get working under Python 3, I'd laugh. [13:08:34.0000] <annevk> I created an account for Tyson [13:17:31.0000] <jgraham> gsnedders: Such as? [13:17:46.0000] <jgraham> Also, what's the problem with simpletree, specifically? [13:19:17.0000] <annevk> http://images.4chan.org/v/src/1337537248434.png :) [14:05:06.0000] <gsnedders> jgraham: I eventually realized it was down to __unicode__ [14:05:20.0000] <zewt> heh [14:05:41.0000] <zewt> python 3 is one of the biggest botched transitions ever [14:06:16.0000] <zewt> it's like they were thinking: "everyone else needs a sane transition plan, but we're special and different and we don't need that nonsense" and years later everyone is still on 2.x as a result [14:06:51.0000] <zewt> perhaps only notable because python gets most stuff right, making that one really huge screwup stand out more [14:07:36.0000] <gsnedders> They had a transition plan, which has mostly worked without too much issue. It's just Python 2.x is "good enough" so people don't bother with it so much. [14:07:44.0000] <gsnedders> 2to3 does genuinely work for most stuff. [14:08:16.0000] <zewt> it's pretty much an insane transition plan, heh [14:08:45.0000] <zewt> until every library is working out of the box with 3, i won't touch it (and since everyone else thinks the same thing, people don't rush to support 3.x) [14:10:13.0000] <gsnedders> And because nobody uses Python 3, libraries won't bother because Python 3 doesn't provide much more. [14:11:29.0000] <zewt> if they were patient and released intermediary versions which always allowed code to be compatible with n-1 and n+1 versions, it would have gone much more smoothly [14:12:12.0000] <gsnedders> Eh, not sure they could have. [14:12:21.0000] <gsnedders> The Unicode change was always going to have to be a hard break. [14:12:28.0000] <gsnedders> And 2.6/2.7 include pretty much everything else. [14:13:39.0000] <zewt> not sure about that; might have taken some imagination [14:13:51.0000] <gsnedders> (Of course, post 3.0 releases have contained stuff not in 2.7) [14:15:08.0000] <zewt> well i should say rather than guaranteeing n-1 and n+1 compatibility, just allowing it--each new version might break code in the old one, but always allowing the code to be updated to work in both, so you never have code that requires the bleeding edge version [14:15:55.0000] <gsnedders> Eh, well, that's doable even with 2.x/3.x, just hard. [14:16:05.0000] <gsnedders> Because of Unicode being the default string type. [14:16:11.0000] <zewt> impractically hard and not designed to be done [14:17:49.0000] <gsnedders> It is somewhat designed to be done. [14:18:03.0000] <gsnedders> Support for u"" in Py3 is exactly for that reason. [14:18:09.0000] <gsnedders> Likewise b"" in Py2. [14:21:45.0000] <zewt> for example, a smoother transition would have been to make unicode the default without removing the str type; then leave it that way for a while before the str/unicode/bytes switch [14:22:28.0000] <zewt> that would have reduced the amount of incompatible changes done at once, so would probably have had less resistance [14:23:36.0000] <gsnedders> The *painful* part of the change would be dropping the implicit unicode/str conversion. [14:24:45.0000] <zewt> that never worked worth a damn anyway, heh [14:25:51.0000] <Philip`> It seems like the fundamental problem is that the 2->3 changes introduce more pain than benefit, and if you try to spread the pain over multiple 2.x versions so that people don't notice it, you're not actually making the situation any better - you're just tricking more people into accepting the pain [14:25:56.0000] <zewt> (since it always wants to default to ascii, for some unfathomable reason) [14:26:22.0000] <zewt> Philip`: when you take pain and spread it across a longer time period, there really is less pain :) [14:26:58.0000] <zewt> Philip`: but more importantly is that in practice, you can't have code that works in 2 and 3 (without 2to3, which is evil); a more gradual transition makes that more doable [14:28:00.0000] <zewt> also the fact that as a library developer, I don't want to spend time testing and supporting both 2 and 3, because there's no sign that 2 will go away any time soon; supporting 3.0 and 3.1 is less painful, since it's much more likely that the old version will fade away reasonably soon [14:28:18.0000] <zewt> it feels less like a permanent fragmentation [14:28:48.0000] <Philip`> Perl seems to have had a better approach - tell everyone that Perl 6 is the future and can be massively incompatible, so the language designers experiment with all sorts of crazy ideas in it, but never get around to implementing it properly, so no real user of the language has any difficulty in immediately seeing they should stick with the old stable Perl 5 [14:29:01.0000] <zewt> (theoretical 3.0 and 3.1, that is, with 3.1 introducing some incompatible changes but it being entirely reasonable to update my code to work in both) [14:29:12.0000] <Philip`> and then figure out which of the crazy ideas worked and gradually introduce them in the 5.x series [14:29:16.0000] <zewt> heh [14:29:25.0000] <zewt> i never use perl for anything except shell one-liners anymore [14:29:35.0000] <Philip`> until eventually Perl 5 ends up in a nicer state and Perl 6 can remain dead [14:29:40.0000] <zewt> because almost without exception, any time i write anything in perl I end up regretting it a few months later and rewriting it in python [14:29:41.0000] <Philip`> and everyone is happy [14:31:19.0000] <gsnedders> /me still likes Perl for stuff that is purely string manipulation [14:31:46.0000] <Philip`> /me finds that a surprising amount of stuff is string manipulation [14:36:14.0000] <jgraham> zewt: I am told that 3to2 is a better plan than 2to3 [14:36:30.0000] <jgraham> /me doesn't like perl, particularly for string manipulation [14:36:48.0000] <jgraham> If I wanted to use regexps for basic operations, I would kill myself [14:37:55.0000] <zewt> i'm told that sticking with 2 is a better plan than jumping hoops for a lazy transition plan :) [14:38:18.0000] <zewt> regexes are generally great for writing something quickly that you'll never be able to understand again, ever [14:38:26.0000] <jgraham> Also, 3.3 seems like it will be the release that finally has critical mass [14:38:43.0000] <gsnedders> FWIW, my plan for html5lib is to get everything running on Py3, and then use 3to2 in future for Py2. [14:39:14.0000] <gsnedders> Because then we actually have Unicode strings everywhere we should (unlike currently), and we get hard errors when we screw the Unicode/str difference up. [14:42:12.0000] <jgraham> http://www.perlmonks.org/?node_id=610450 <- this is pretty much what's wrong with Perl [14:42:58.0000] <jgraham> Well I say that [14:43:30.0000] <jgraham> That is pretty much one of a large number of things that is wrong with perl [14:43:44.0000] <Philip`> /me doesn't see what's wrong with that [14:44:02.0000] <gsnedders> Nor I. [14:44:42.0000] <Philip`> /me can write (and read) a regexp like that in far less time than it takes to look up whether the standard library function is called trim() or strip() and whether it trims all whitespace or just space characters etc [14:46:07.0000] <zewt> you have to squint much harder at a regex to make sure of exactly what it's doing; it takes a lot more mental bandwidth to interpret [14:46:33.0000] <Philip`> (Oh, looks like it's actually called foo.strip()) [14:47:11.0000] <Philip`> ((in Python, I mean)) [14:48:03.0000] <jgraham> The fact that about 50% of all sites on the internet are dedicated to the question of how to trim whitespace in perl suggests that it isn't a very satisfactory solution [14:49:23.0000] <jgraham> (OK not 50%, but apparently it was asked so often it made the FAQ) [14:53:01.0000] <zewt> incidentally, it's not like regexes are any harder to use in python for lack of a special-case native syntax [14:53:56.0000] <Philip`> jgraham: That's good since it means new Perl programmers start by asking a nice simple question, which can be easily answered by the FAQ or by other Perl programmers, and the power and flexibility and applicability of regexps is thereby demonstrated to the new programmer, giving them the tools they can use for all subsequent questions they might otherwise ask [14:54:06.0000] <zewt> new programmers should never ever touch regexes [14:54:21.0000] <zewt> not until they have enough experience with saner approaches to know when they're appropriate [14:54:22.0000] <Philip`> whereas in Python the answer is "use str.strip()" which is a special case that is useless for most other manipulations [14:54:32.0000] <zewt> what's it useless for? [14:55:09.0000] <zewt> it's a very common operation, and implemented sanely: strip, lstrip, rstrip with an optional parameter to say exactly what to strip [14:56:12.0000] <Philip`> It's useless for e.g. removing multi-character prefixes/suffixes [14:56:29.0000] <Philip`> which is only a minor variant of the problem [14:57:03.0000] <zewt> a rare one [14:57:11.0000] <zewt> rare problems don't belong in the stdlib :) [14:58:08.0000] <zewt> and you're not actually arguing for perl, since python has regexes too, heh (and makes them clean to use) [14:58:15.0000] <Philip`> Rare problems are (in aggregate) pretty common, so it's good to have a common way of solving them :-) [14:59:25.0000] <zewt> (nothing is quite as nasty as regexes in C, with double-escaping) [15:01:01.0000] <tantek> zewt - my favorite quote about regexes is the one from jwz [15:01:38.0000] <jgraham> Alternatively newcomers will try to do something simple, notice that it requires them to enter something that is barely distinguishable from line noise, and delete the perl binary from their harddrive [15:02:02.0000] <jgraham> Thereby causing a number of critical system scripts to fail [15:02:11.0000] <jgraham> and bricking their machine [15:02:37.0000] <Philip`> When you have to write "baz = re.sub(r'foo', r'bar', baz, flags=re.IGNORECASE)" vs "$baz =~ s/foo/bar/i", the weight of the syntax seems like a not insignificant drawback [15:02:53.0000] <jgraham> (I assume they are not unfortunate enough to be using windows. If they are, the added misfortune of encountering OPerl might require spending time in A&E) [15:03:47.0000] <zewt> i find i need to use regexes so rarely that having a native syntax just feels like core bloat [15:04:14.0000] <zewt> really the only benefit is that (in principle, at least) you can precompile the regex, which python can't do [15:04:35.0000] <Philip`> /me uses them rarely in Python because the syntax is painful and he can never remember it [15:04:49.0000] <zewt> eg. if you have a regex in an inner loop, and you want to avoid compiling the regex every time, you have to do it yourself [15:05:19.0000] <zewt> (it probably caches anyway; not sure, since it rarely matters to me) [15:06:15.0000] <Philip`> (Perl precompiles them in practice, not just in theory) [15:06:54.0000] <Philip`> (so you never have to worry about it) [15:07:22.0000] <Philip`> (unless you happen to use a variable inside the pattern, in which case you can just add the /o flag to say you're not going to change the variable so it can be cached) [15:07:34.0000] <zewt> you mean it caches; perl doesn't even support precompiling [15:08:10.0000] <zewt> (maybe it does now; I havn't touched it in quite a while :) [15:09:31.0000] <zewt> i find reading any $basic-esque language painful now, too [15:09:48.0000] <zewt> perl and php make me feel like someone's spitting punctuation all over the code [15:10:18.0000] <tantek> zewt - how would you compare coffeescript? [15:11:02.0000] <zewt> dunno; don't know what it is [15:15:05.0000] <Philip`> zewt: I'm fairly sure it does precompile them - try e.g. perl -MO=Concise -e's/foo/bar/' vs perl -MO=Concise -e'$x="foo"; s/$x/bar/' [15:15:34.0000] <zewt> .pyc = precompiling [15:15:38.0000] <Philip`> where the latter's syntax tree includes regcomp() stuff to handle the regexp at runtime, while the former doesn't since presumably it's doing it all beforehand [15:16:04.0000] <zewt> (minor, yes :) [15:16:07.0000] <Philip`> .pyc = cached bytecode [15:16:08.0000] <tantek> coffeescript is: http://coffeescript.org/ [15:16:10.0000] <Philip`> Caching != precompiling [15:16:14.0000] <zewt> pyc = compiled bytecode [15:17:02.0000] <zewt> tantek: not sure what the goal is, but first impression is unintuitive, overly novel, out of order expressions [15:17:31.0000] <tantek> goal is code that is more easily and quickly readable/writable than JS [15:17:34.0000] <tantek> but just as fast [15:17:40.0000] <tantek> better maintenance [15:17:45.0000] <tantek> less puncuation [15:17:52.0000] <zewt> call me a conservative when it comes to language design, but a language needs to have *serious* advantages to justify deviating from the style that pretty much all major languages today share [15:17:53.0000] <tantek> *punctuation even [15:17:57.0000] <Philip`> Perl and Python both compile into bytecode in RAM; the only difference is that Python caches it on disk afterwards [15:18:18.0000] <tantek> zewt - developer productivity is perhaps the biggest advantage possible [15:18:34.0000] <zewt> tantek: and needlessly deviating in language style reduces that sharply [15:18:38.0000] <tantek> and that's what coffeescript is apparently targetting. I haven't used it - but I've seen others be very productive with it [15:19:05.0000] <tantek> zewt - apparently the language style deviation is of minimal impact - devs pick up coffeescript very quickly [15:19:42.0000] <tantek> but in general I personally agree with the principle you mention of not unnecessarily deviating in language [15:19:51.0000] <tantek> avoiding NIH and all that [15:19:52.0000] <zewt> the style of this is a pretty massive turnoff to me [15:19:55.0000] <tantek> maximizing re-use [15:20:03.0000] <zewt> at least on first impression [15:20:07.0000] <tantek> ok - fair enough - that's a reasonable opinion [15:20:12.0000] <zewt> eg. out-of-order expressions; i really hate "y if x" [15:20:34.0000] <zewt> (Python makes that mistake with "y if x else z", which I strongly dislike) [15:20:47.0000] <zewt> C gets it right with x? y:z [15:22:30.0000] <zewt> (i'd call that one of the uglier bits of syntax in python; fortunately it's uncommon and isolated enough to not hurt often) [15:22:45.0000] <tantek> zewt - the ternary operator is quite portable: http://tantek.com/2012/090/t1/javascript-php-c-ruby-cassisproject-cassisjs [15:23:11.0000] <zewt> yep, lots of languages copy c's syntax (and I don't go far back enough to know if C copied it from something else) [15:23:43.0000] <zewt> which is another reason I don't like Python's, but the bigger one is it's just in a bizarre order--it's not just flipped (like y if x), it's twisted inside out [15:25:58.0000] <tantek> the non-standard order confuses me too - but I wonder if that's just because we're used to a standard order for such operations in programming languages [15:26:18.0000] <zewt> that's a pretty good reason to be confused by something :) [15:26:34.0000] <zewt> i can read it, i just have to untwist it mentally [15:29:23.0000] <tantek> right, but I wonder if that's just a short-term discomfort [15:29:36.0000] <tantek> like would we become proficient with a just a week or two of use [15:30:09.0000] <zewt> i'm "proficient" with it (python is my primary language these days); i just dislike it [15:30:20.0000] <tantek> oh I thought you meant coffeescript ok [15:30:49.0000] <tantek> I decided to go the other route and just use a more common subset across languages [15:30:52.0000] <tantek> hence CASSIS [15:34:31.0000] <zewt> tantek: i think the fundamental error in that language is the idea that typing is the bottleneck in programming [15:34:46.0000] <zewt> which at least for me couldn't be more wrong [15:35:06.0000] <tantek> it's *a* bottleneck for sure [15:35:25.0000] <zewt> not for me [15:35:27.0000] <tantek> I've been much more productive in weakly typed langs vs strongly typed langs [15:35:31.0000] <tantek> ymmv etc. [15:35:55.0000] <zewt> that's not typing speed, heh [15:35:56.0000] <zewt> afk [15:46:35.0000] <wycats> anyone around who works on Selectors API 2? 2012-05-21 [23:46:42.0000] <Von_Davidicus> I got an interesting question today, and I thought I'd ask everyone here, since I', having difficulties answering it: "What does XML + XSLT do that XHTML cannot"? [23:48:41.0000] <othermaciej> if the context is displaying something in a web browser, then the correct answer is "nothing" [23:51:22.0000] <Ms2ger> Turn you insane [23:54:26.0000] <Ms2ger> wycats, anything in particular you wanted to know? [23:58:50.0000] <Von_Davidicus> What other contexts has XSLT been used in, then? [00:00:29.0000] <Ms2ger> Dunno, the Fukushima reactor? [00:06:20.0000] <karlcow> Von_Davidicus: it can be used sometimes for displaying an XML document in another form in the browser: RDF file as XHTML or HTML, XML feed, etc. but this is not used a lot. [00:07:19.0000] <karlcow> The draconian mode of XML has a nice side effect *for me*, which is to check without too much effort if the content is malformed. [00:20:19.0000] <Von_Davidicus> I like that part of it, myself. [00:22:26.0000] <hsivonen> Von_Davidicus: if the context of the question is displaying stuff in the browser, then, as othermaciej said, there's nothing XML+XSLT can display in a browser that XHTML arriving over HTTP cannot. [00:22:58.0000] <hsivonen> Von_Davidicus: so what XML+XSLT allows you to do is to move CPU burn from your server to the user's device [00:23:02.0000] <webben> You can likely get a validator plugin for your browser to tell you whether content is malformed … [00:23:15.0000] <hsivonen> Von_Davidicus: so you get to burn the battery on the user's device [00:24:20.0000] <webben> Although nowadays people are more likely to push JSON to an HTML frontend page that uses JS to convert the JSON to more HTML, rather than XSLT. [00:25:32.0000] <hsivonen> webben: yeah. I'm sure phone users who are low on battery appreciate designs that move processing to the edge of the network [00:26:25.0000] <Von_Davidicus> I see. Hm. [00:26:59.0000] <Von_Davidicus> How good are phones at displaying graphics on webpages? [00:27:00.0000] <othermaciej> there's plenty of ways to burn battery [00:27:06.0000] <othermaciej> complex CSS can be compute-intensive too [00:27:13.0000] <othermaciej> and sending more data means having the radio on longer [00:27:34.0000] <othermaciej> but generally, client-side XSLT tends to be a particularly inefficient way of doing things [00:27:45.0000] <tantek> certainly seems to be more CPU intensive [00:28:24.0000] <zcorpan> and delays rendering, doesn't it? [00:28:35.0000] <tantek> incremental rendering? who needs that? ;) [00:28:36.0000] <zcorpan> normal XHTML can be rendered incrementally [00:28:43.0000] <tantek> (zing) [00:29:28.0000] <tantek> well here's a question about what can x do in a browser that y cannot [00:29:52.0000] <tantek> what can XHTML do in a browser that HTML as well-formed XML cannot? [00:30:15.0000] <tantek> (per http://tantek.com/2010/302/b1/xhtml-dead-long-live-xml-valid-html5 ) [00:30:37.0000] <othermaciej> it can give you a yellow screen of death [00:31:03.0000] <tantek> oh, like http://www.flickr.com/photos/tantek/5126607972/ ? [00:34:53.0000] <hsivonen> It's nice how "yellow screen" has been abstracted so that a pink box counts as yellow screen [00:35:24.0000] <hsivonen> /me always has to look up the order of the arguments of strcpy :-( [00:37:08.0000] <jgraham> It allows you to be the subject of browser developer wrath [00:37:22.0000] <jgraham> Although I guess HTML has a number of ways to do that [00:42:27.0000] <Von_Davidicus> /me wishes someone, somewhere, would say XSLT is actually *useful* [00:43:17.0000] <webben> Von_Davidicus: Why? [00:43:20.0000] <Ms2ger> jgraham, document.all("foo") [00:43:25.0000] <jgraham> Von_Davidicus: davidc isn't in the channel at the moment, so he is likely to have missed his cue [00:43:37.0000] <Von_Davidicus> Because I spent so much blasted time experimenting with it! [00:43:49.0000] <webben> Von_Davidicus: Did you find it useful? [00:44:00.0000] <hsivonen> Von_Davidicus: if you know XSLT well and need to convert between two XML vocabularies that are not exactly isomorphic but not completely different, either, XSLT can be very useful [00:44:12.0000] <Von_Davidicus> As a client-side processing language for XML files, yes. [00:44:17.0000] <webben> I guess DocBook is an example. [00:44:35.0000] <webben> Von_Davidicus: Well, then you can say it was useful :) [00:44:39.0000] <karlcow> Von_Davidicus: it is useful, but not necessary in the context of live transformation in the browser [00:46:19.0000] <hsivonen> Opera 12 release process reminds me of Firefox 4. I wonder if Opera will move to rapid release after 12... [00:50:33.0000] <hsivonen> Whoa. StatCounter says Silk 1.0 exceeds Opera 11.6 usage in the U.S.! [00:50:43.0000] <karlcow> hsivonen: Opera just synchronized again with the year! [00:50:52.0000] <karlcow> the next one will be 13 in 2013 [00:50:53.0000] <hsivonen> /me didn't expect to see Silk on the charts [00:50:55.0000] <karlcow> :p [00:57:14.0000] <hsivonen> From the outside at least, Opera 12 looks like Firefox 4 in the sense that it has gotten stuck into over-long beta. Presumably there's something holding back release but it means that all the good stuff that's done isn't reaching users. [00:57:26.0000] <hsivonen> the sort of situation rapid release is meant to fix [01:32:02.0000] <Von_Davidicus> /me dances. [01:32:26.0000] <odinho> o/-< o\-< [01:33:16.0000] <Von_Davidicus> My assignment is complete. [01:34:57.0000] <Von_Davidicus> I think I mentioned earlier that I was doing an assignment for an HTML class to build a series of webpages--and instead of doing them in HTML, I've done them in XML + XSLT. [01:41:41.0000] <Von_Davidicus> Anyways, the assignment is done--hopefully I get my point across. And for your enjoyment, the mess starts here: http://www.mrinitialman.com/Coding/D2Runes/Words/ancients_pledge.xml and http://www.mrinitialman.com/Coding/D2Runes/Runes/runeEl.xml [01:48:19.0000] <zcorpan> Von_Davidicus: http://annevankesteren.nl/2004/01/xml-versus-xhtml [01:50:50.0000] <charlvn> yeah i don't know how this is in any way new, we worked past this like 8 years ago [01:52:08.0000] <charlvn> afaik the conclusion was that server-side xslt is cool but client-side is a fail, unless anything changed since then? [01:52:39.0000] <Von_Davidicus> This isn't really meant to be new or edgy--but my HTML teacher has thus far refused to believe that I've done this before (I *have* to take an introductory course in HTML at college, dems da rules), so hopefully something overcomplicated like this will get him to reconsider. [01:54:01.0000] <Von_Davidicus> I -have- done a webpage in HTML5; a professional one: http://sheepmaster.zzl.org/ Some of the differences between HTML 4.01 and HTML 5 forced me to quite being so lazy with my JavaScript. :D [01:54:20.0000] <charlvn> to be honest i don't entirely "get" how this is relevant to the whatwg in 2012 but i guess it's cool as an academic exercise [01:54:41.0000] <jgraham> Von_Davidicus: It might make them believe that you need to be retaught the basics [01:55:33.0000] <charlvn> yeah exactly [02:00:34.0000] <Von_Davidicus> I thought it was fun. [02:05:13.0000] <MikeSmith> so apparently because of a problem with the way we have the W3C bugzilla instance set up, when you have several bugzilla tabs open, it's possible that a comment you submit for one bug in on tab ends up wrongly getting added to a bug open in another tab [02:05:33.0000] <othermaciej> that sounds extremely bogus [02:06:19.0000] <annevk> how is that possible? [02:06:30.0000] <annevk> named windows? [02:07:10.0000] <odinho> Ahhh. Just like some bank websites. I love that. [02:07:41.0000] <odinho> They keep state using cookies with an ID I think, and keep state server side, -- so when you're suddenly in another place it messes up things pretty badly. [02:07:52.0000] <Ms2ger> Sounds like a browser bug to my neighborhood bugzilla developer [02:08:24.0000] <odinho> (I was not speaking of bugzilla btw, -- but evil ways of doing web services, which has frustrated me to no end before :P) [02:08:57.0000] <Ms2ger> "when you submit the comment, the bug id is fetched from the form, not from some other location" [02:09:40.0000] <MikeSmith> othermaciej, annevk - report came from a user who had observed the problem [02:11:43.0000] <MikeSmith> Ms2ger: ok [02:11:49.0000] <MikeSmith> I will let the systems team know [02:12:08.0000] <MikeSmith> I think they had been thinking it had something to do with their mirroring setup [02:27:21.0000] <MikeSmith> foolip: systems team is doing a DB migration right now [02:27:33.0000] <foolip> MikeSmith, thanks, that explain it [02:27:37.0000] <foolip> s [02:27:53.0000] <MikeSmith> it'll probably be a couple more hours at least before they're done [02:28:02.0000] <MikeSmith> I'll try to remember to ping you once they are [02:28:07.0000] <foolip> MikeSmith, thanks! [02:31:14.0000] <Philip`> Von_Davidicus: From your page: xsi:schemaLocation="http://www.mrinitialman.com/Coding/D2Runes/RuneWords/ ../DataFiles/Runewords.xsd" [02:31:28.0000] <Philip`> Von_Davidicus: Surely that's broken (and invalid?) because of the space? [02:32:54.0000] <Von_Davidicus> Oh, thanks for the heads-up. I forgot to edit my files after I changed some directory names. Beg pardon. [02:34:06.0000] <Philip`> Seems a bit odd having all this unwieldy machinery for validation and then letting through errors like that :-p [02:39:36.0000] <Von_Davidicus> Which validator did you use, that I may recheck my coding? [02:40:08.0000] <Philip`> I used my eyes [02:44:35.0000] <Von_Davidicus> Hmmm.... *Tries to remember how to set up the schema location.* [02:45:43.0000] <jgraham> /me wonders which schema format Philip`'s eyes use [02:47:51.0000] <charlvn> eye-based validation... i wonder if there is a spec for that :) [02:50:23.0000] <Von_Davidicus> Nope, that space is supposed to be there. But for some reason, it's not validating. I'll have to figure out why. [02:50:51.0000] <webben> the space is supposed to be there? [02:50:58.0000] <webben> seems very odd [02:56:20.0000] <MikeSmith> Von_Davidicus: maybe you need to make it tab instead [02:56:24.0000] <Von_Davidicus> Yeah, it separates the namespace from the filename of the schema. [02:56:29.0000] <MikeSmith> or maybe it's a non-breaking space [03:00:45.0000] <charlvn> acording to my understanding it's any whitespace character http://www.w3.org/TR/xmlschema-0/#schemaLocation [03:01:52.0000] <zcorpan> each(im.getAttribute('srcset').split(','), function (def) { [03:01:54.0000] <zcorpan> https://github.com/davidmarkclements/Respondu/blob/master/R.js [03:03:24.0000] <Philip`> Von_Davidicus: Hmm, I guess I'm confused - isn't http://www.w3.org/TR/xmlschema-1/#xsi.schemaLocation defining that schemaLocation is a single URL (anyURI)? [03:04:32.0000] <Von_Davidicus> Not in any of the examples I've seen. [03:06:04.0000] <webben> Von_Davidicus: Oh i c. [03:07:31.0000] <hsivonen> so many responsive image emails [03:07:48.0000] <hsivonen> every time I feel I have something to say, I feel I have to catch up with the threads first [03:07:55.0000] <hsivonen> and then there's more to catch up with [03:08:19.0000] <Von_Davidicus> /me has even checked to make sure the schema URIs are correct [03:10:27.0000] <charlvn> Philip`: it's not entirely clear to me either but from my understanding it could be a list http://www.w3.org/TR/xmlschema-1/#variety [03:11:48.0000] <Von_Davidicus> Well, my files follow every example I've seen, so I've no idea what's going on. :( [03:12:31.0000] <charlvn> the example on http://www.w3.org/TR/xmlschema-0/#schemaLocation also shows this [03:14:13.0000] <Von_Davidicus> Oh, I was clicking a box I shouldn't have: "Check As Complete Schema"--the validator was treating my XML file as a schema, which it isn't. *Puts on dunce cap* [03:14:24.0000] <Von_Davidicus> With that box unchecked, things check out fine. :) [03:15:56.0000] <Von_Davidicus> And another validator doesn't do custom schemas. So, all is well. [03:20:08.0000] <Von_Davidicus> I have a question about the HTML5 DOM--something I found a pain in HTML 4.01 [03:20:37.0000] <Von_Davidicus> In HTML5, if you leave out the <tbody> tags, does that element still exist in the DOM? [03:20:55.0000] <webben> yes. [03:21:07.0000] <Von_Davidicus> What about XHTML5? [03:21:30.0000] <webben> the XML serialization has no implied tags [03:21:41.0000] <Von_Davidicus> Okay. [03:22:05.0000] <Von_Davidicus> Boy, were tables a trick to script for. It took me ages to figure out what was happening. >< [03:24:48.0000] <Von_Davidicus> Ladies and Gents, I am headed out. :) [04:03:03.0000] <odinho> hsivonen: Yes. I participated heavily in the beginning, but felt that it didn't really help, because the ones replying didn't really read and understand it anyway... [05:20:39.0000] <david_carlisle> jgraham: Just for you: XSLT is actually *useful* [05:21:51.0000] <Ms2ger> /me sniggers [05:25:25.0000] <jgraham> /me just came acrss an old copy of "Dr Dobbs Journal" from like 2000 or so containing a Microsoft advert featuring someone from Ask Jeeves explaining how XML in visual studio was helping them scale to meet business needs [05:25:56.0000] <jgraham> It was hillarious [05:26:10.0000] <jgraham> No justification of why XML solved any problems [05:26:28.0000] <jgraham> Just the magic letters X M and L were assumed to be a selling point [05:27:47.0000] <david_carlisle> jgraham: well that's always the way. json or html(5) or whatever or sticking "java" at the front of "javascript", people like to have buzzwords [05:30:47.0000] <jgraham> Indeed. I look forward to the future when people laugh at things that talk about "mobile-first responsive HTML5" or whatever today's buzzword bingo entries are [05:31:13.0000] <gsnedders> Web 2.0-enabled HTML5 mobile interface featuring WebGL! [05:33:57.0000] <Philip`> We should fight against the buzzwordification of pragmatic technologies by giving them names that cannot possibly be used with a straight face [05:34:36.0000] <Philip`> Rename HTML5 to something like "SynergML" and then nobody can use it unironically [05:34:39.0000] <jgraham> Like AJAX? [05:34:46.0000] <Philip`> "ECMAScript" was a good one [05:35:00.0000] <gsnedders> W3CML [05:35:09.0000] <gsnedders> WHATML [05:35:33.0000] <jgraham> Really should have gone with the task force in that case [05:39:20.0000] <odinho> I said "responsive" "mobile-first" and a few others at a party, annevk and richt laughed at it IIRC. So given the right people we're already there! [05:40:35.0000] <jgraham> Wait, what? [05:40:39.0000] <jgraham> I don't understand [05:40:43.0000] <jgraham> "party"? [05:43:10.0000] <odinho> jgraham: -> 14:32 < jgraham> Indeed. I look forward to the future when people laugh at things that talk about "mobile-first responsive HTML5" or whatever today's buzzword bingo entries are [05:43:25.0000] <odinho> jgraham: Or is it the party that's hard to understand? :P It's was Lachy's birthday party. [05:47:37.0000] <Ms2ger> odinho, shh, jgraham wasn't invited [05:48:54.0000] <odinho> oh f... [05:49:16.0000] <gsnedders> I guess that's what he gets for not having FB. [05:49:42.0000] <odinho> I didn't either, but I was in the same office as Lachy, so it was OK. I guess I'll be missing out now though... [06:51:41.0000] <zcorpan> // Time in seconds within which the media is seekable. [06:51:41.0000] <zcorpan> var seekableEnd = audio.seekable.end(); [06:51:49.0000] <zcorpan> http://html5doctor.com/html5-audio-the-state-of-play/ [06:52:17.0000] <zcorpan> maybe we should give up and make the argument optional [07:07:15.0000] <zewt> seems wrong that a 5. isn't a "valid floating point number" [07:37:53.0000] <foolip> MikeSmith, still migrating? [07:38:05.0000] <MikeSmith> foolip: ah! [07:38:11.0000] <MikeSmith> no, should be done now [07:38:19.0000] <foolip> ok, I'll try filing that issue again [07:38:29.0000] <MikeSmith> yup [07:39:07.0000] <foolip> MikeSmith, it works now! [07:39:12.0000] <MikeSmith> super [07:43:55.0000] <david_carlisle> Philip`: (schemaLocation is a space separated list of URI (that's one of it's problems) the other one (in general) is that every other URI points at an XSD schema, and in this forum the remaining problem is that the remaining half of the URI are namespaces. Apart from those three minor issues. It's brilliant. [07:44:40.0000] <david_carlisle> its [07:57:16.0000] <Philip`> david_carlisle: Am I missing some part of the spec where schemaLocation is defined as a space separated list, instead of as a single URI, other than examples and notes? [07:57:56.0000] <Philip`> (Not that it actually matters; I'm just curious about to what extent I'm misunderstanding stuff) [08:01:08.0000] <Philip`> Oh, I think see something I didn't see [08:03:12.0000] <Philip`> namely that http://www.w3.org/TR/xmlschema-1/#xsi.schemaLocation says "{variety} list" [08:04:33.0000] <Philip`> so it's less confusing than I thought [08:04:38.0000] <Philip`> so never mind [09:00:23.0000] <gsnedders> I wonder what a reimpl of the jQuery API targetting just nightly builds of browsers would look like. [09:00:35.0000] <gsnedders> (wrt the shrinking libraries thread) [09:22:34.0000] <dglazkov> good morning, WhatWg! [09:27:13.0000] <Ms2ger> Good, dglazkov [09:46:30.0000] <MikeSmith> david_carlisle: you around? [09:46:50.0000] <MikeSmith> I can push the mathml3 validator updates any time [09:47:05.0000] <MikeSmith> but I would like to have help with testing it after I do [09:47:20.0000] <MikeSmith> to make sure I get everything pushed [09:47:55.0000] <MikeSmith> the instance I pointed you to before was code straight from my workspace [09:48:25.0000] <MikeSmith> and experience tells me that sometimes I don't manage to get everything pushed that's needed [09:50:36.0000] <david_carlisle> MikeSmith: yes [09:51:00.0000] <MikeSmith> OK [09:51:06.0000] <MikeSmith> I will give it try now [09:51:28.0000] <david_carlisle> ooh OK I'll stay here for a bit then:-) [09:51:36.0000] <MikeSmith> I don't have the most efficient system for deploying [09:51:54.0000] <MikeSmith> there are 5 validator hosts and I need to ssh into each an manually do updates [10:09:03.0000] <Ms2ger> MikeSmith, dvcs.w3.org appears down [10:09:37.0000] <MikeSmith> Ms2ger: thanks, checking ow [10:09:41.0000] <MikeSmith> *now [10:09:55.0000] <Ms2ger> Ta [10:10:56.0000] <MikeSmith> "Restarting web server: apache2 ... waiting ............." [10:11:09.0000] <Ms2ger> Reported back [10:11:12.0000] <MikeSmith> workign now [10:11:14.0000] <MikeSmith> yeah [10:11:35.0000] <MikeSmith> if it becomes not available again please ping me [10:11:44.0000] <Ms2ger> Will do, thanks [10:23:12.0000] <Ms2ger> Philip`, failed https://www.w3.org/Bugs/Public/show_bug.cgi?id=17141 for you [10:47:51.0000] <jgraham> Doesn't allowing min-* and max-* in srcset significantly complicate the processing [10:47:54.0000] <jgraham> ? [10:48:16.0000] <jgraham> For example at the moment it isn't possible to have gaps [10:49:10.0000] <jgraham> I guess you could use the rule that the src image is used if there's no other candidate [10:49:33.0000] <TabAtkins> jgraham: Yeah, it does. I suspect it's worthwhile, though. [10:51:47.0000] <jgraham> I'm not sure it's not worthwhile, it just takes a conceptually simple model and makes it less simple :) [10:52:05.0000] <jgraham> s/sure/saying/ [11:19:58.0000] <david_carlisle> im back:-) [11:31:26.0000] <TabAtkins> david_carlisle: Yo, I had a question for you a few weeks back. [11:32:25.0000] <TabAtkins> Does MathML have a good definition of "bounding box" of its internal elements? Content: CSS Images 4 has an element() function that lets you use an element as an image. All it needs to be able to do so is the bounding box of the element, so it knows what size the image is. [11:32:48.0000] <TabAtkins> For HTML and SVG this is pretty simple, but I'm not familiar enough with MathML's rendering model to know if it has such a concept. [11:40:06.0000] <david_carlisle> TabAtkins: Probably good enough see perhaps http://www.w3.org/Math/draft-spec/mathml.html#chapter3_presm.mpadded the wording in MathML is a bit wooly as mathml doesn't enforce a css style box model, so eg mathematica or tex or whatever can render it, but you don't care about that in a mathml in html context [11:42:54.0000] <TabAtkins> david_carlisle: Okay, so you do refer to "the bounding box of its content", but that term isnt' linked. [11:44:11.0000] <david_carlisle> TabAtkins: "linking" bit modern for us old timers you know. You should be grateful it's not fixed width plain text ascii:-) [11:45:02.0000] <TabAtkins> Tell that to TimBL twenty years ago. ^_^ [11:46:09.0000] <david_carlisle> TabAtkins: Probably the nearest is the pictures in the next section http://www.w3.org/Math/draft-spec/image/mpadded-resize.png but basically mathml assumes every construct _has_ a bounding box but doesn't always specify what that is as there is quite a bit of leeway in how constructs are laid out. But for element() taht should be OK shouldn't it? [11:47:23.0000] <TabAtkins> Not really. :/ It's not really acceptable if browsers define bounding boxes differently for the same markup. [11:49:21.0000] <david_carlisle> TabAtkins: maybe I misunderstood, but don't you just need to access the bounding box of the markup as rendered in the system applying the function. The bounding boxes will differ on different systems anyway for font reasons if nothing else [11:51:25.0000] <david_carlisle> even something as simple (switching to tex notation) as P_1 will have different bounding boxes on different systems even using the same fonts as different systems will have different ideas about whether to kern the 1 under the P [11:52:04.0000] <TabAtkins> It's fine for them to end up rendering differently, as long as there's a strict definition of what should actually be rendered. [11:52:38.0000] <david_carlisle> TabAtkins: I don't understand. [11:53:58.0000] <TabAtkins> So, like, for SVG the bounding box is "the smallest axis-aligned rectangle that includes the strokes of all the content within the element". [11:54:27.0000] <TabAtkins> For HTML it's the element's border box, modified by border-iamge. [11:59:44.0000] <david_carlisle> http://www.w3.org/Math/draft-spec/mathml.html#appendixd_dt-blackbox and http://www.w3.org/Math/draft-spec/mathml.html#appendixd_dt-blackbox and #appendixd_dt-boundingbox plus the previously mentioned mpadding description are what you have. That seems as definite as the svg definition you gave. If css does need something more proscriptive here we'd certainly be open tp a clarifying note on... [11:59:46.0000] <david_carlisle> ...mathml/css interaction but hopefully there is enough in the spec [12:02:08.0000] <TabAtkins> The "black box" might work. I'm not sure if it's equivalent to SVG's definition, as the definition of how far a stroke extends is rigorous in SVG (it's part of the geometry used for, e.g. pointer events). [12:02:15.0000] <david_carlisle> Although I guess you're not going to like the paragraph immediately before section 3.2 [12:02:22.0000] <TabAtkins> Is the black box what is used for matching :hover, for example? [12:04:24.0000] <david_carlisle> TabAtkins: The only thing MathML says about CSS (so :hover) is that if you have a system supporting mathml and css they may interact. http://www.w3.org/Math/draft-spec/mathml.html#chapter6_world-int-style [12:04:36.0000] <TabAtkins> Yay. [12:04:37.0000] <TabAtkins> ^_^ [12:08:59.0000] <jgraham> TabAtkins: It's not that surprising that the interaction of CSS and MathML is poorly defined [12:09:13.0000] <TabAtkins> jgraham: Oh, definitely. It's just problematic for me. [12:09:18.0000] <jgraham> I mean, the interaction of CSS with itself isn't always that well defined ;) [12:09:32.0000] <jgraham> TabAtkins: The solution is probably to fix MathML :) [12:10:20.0000] <TabAtkins> Yes, here the "right" solution would be for MathML to explicitly say what the bounding box of its elements are, so that other CSS things like :hover and element() could just hook into that. [12:10:20.0000] <jgraham> (I mean it would be surprising if it was the exception to the rule that no spec is good enough) [12:13:04.0000] <david_carlisle> TabAtkins: I'm not sure that it's possible to say what the bounding box is without first fixing a particular layout model such as css box model. Clearly when MathML is implemented natively in a browser it is rendered using a (possibly extended) version of the css layout rules so you can ask the question but what would an "exact" definition of teh bounding box of P_1 be in a purely mathml context? [12:14:18.0000] <TabAtkins> david_carlisle: I disagree. For example, SVG doesn't use the CSS layout model, but it has a precise definition of the bounding box of any element. [12:14:52.0000] <TabAtkins> You can state the definition in terms of whatever layout model you use, as long it's an answer that will be implementable compatibly across browsers. [12:15:10.0000] <TabAtkins> So that, for example, you dont' end up with one browser doing a tight box around glyphs, but another including the spacing on either side. [12:15:11.0000] <othermaciej> TabAtkins: is there anything I should do to ensure that my comment about Media Queries mostly not defining anything gets addressed by the WG? [12:15:28.0000] <othermaciej> TabAtkins: I heard that there is a bug tracker but the CSS WG Chair will get mad at people who use it [12:15:43.0000] <TabAtkins> othermaciej: >_< >_< >_< I yelled at Ms2ger about saying that. [12:15:56.0000] <othermaciej> TabAtkins: is it actually ok to use it? [12:16:11.0000] <TabAtkins> File bugs freely. We don't want *conversation* on the bug tracker, is all - those should happen on the list. [12:16:34.0000] <othermaciej> since no one responded to my email there seems to be little risk of a conversation [12:16:38.0000] <david_carlisle> TabAtkins: yes but it defines a precise layout as that's it's purpose. <msub> in MathM< means a subscript and if you define teh bounding box of P_1 in terms of teh extent of black pixels then TeX is doomed as it doesn't know which pixels are black. If you define it in terms of font metrics it's possible but may be not what you want either. [12:17:35.0000] <TabAtkins> othermaciej: Yeah, it's fine. If discussion is necessary there's the email thread already started, or someone can start a new one. But go ahead and file a bug. [12:18:05.0000] <TabAtkins> david_carlisle: Out of curiosity, why doesn't TeX know that? [12:18:21.0000] <jgraham> david_carlisle: It seems clear that, at the very least, MathML-on-the-web should have a precisely defined layout model [12:18:45.0000] <jgraham> I doubt anyone cares very much about precise layout compatibility with non-web systems [12:18:51.0000] <TabAtkins> But yes, what jgraham says. The limitations of things that aren't browsers shouldn't hold back defining useful things for browsers. [12:18:54.0000] <david_carlisle> TabAtkins: TeX knows nothing about glyph shapes, each character is just 4 integers extracted from the font metrics (height depth width and italic correction) [12:20:02.0000] <david_carlisle> jgraham: as I said above if we restrict to mathml-in-acss-based-rendering model there is definitely more that could be said, and if it neess saying we could drum up a WG not eto say it [12:20:31.0000] <jgraham> Please don't make a Note :) [12:20:41.0000] <TabAtkins> david_carlisle: The rendering engine for TeX certainly knows about black pixels, though. But it's irrelevant anyway, as TeX has no :hover or element() or similar. [12:20:55.0000] <jgraham> The most pointless of all publication types [12:20:57.0000] <TabAtkins> david_carlisle: Yes, please don't make a Note. It needs to be normative, and ideally contained within the main spec. [12:21:42.0000] <david_carlisle> TabAtkins: there isn't _a_ rendering engine for TeX. And even if yiou restrict to a particular one the point is that you can't access it from within TeX. [12:22:02.0000] <TabAtkins> david_carlisle: Exactly, so it doesn't matter for TeX what the definition is. ^_^ [12:22:23.0000] <TabAtkins> (I wasn't implying a single rendering engine for TeX. It's of course rendered by many engines, just as web content is.) [12:23:43.0000] <david_carlisle> TabAtkins: well I think formally we are not chartered to do another rec-track thingy at present, but that's just bureaucracy. If the words get written it could presumably be made a normative document by some combination of math/css/htmlWG [12:24:38.0000] <TabAtkins> You need a charter for errata? [12:24:52.0000] <TabAtkins> (This qualifies as MathML3 errata, I would think.) [12:25:08.0000] <david_carlisle> TabAtkins: No but errata need an error and this is feature creep:-) [12:25:59.0000] <TabAtkins> I disagree. The draft currently defines that MathML can interact with CSS, but it does so incompletely. [12:26:12.0000] <TabAtkins> If you can't even reasonably respond to :hover, that's an error. ^_^ [12:27:59.0000] <david_carlisle> TabAtkins: If a document specifies something and does it wrong, that's an error. If a document explictly says that in order to support multiple rendering systems some aspects of rendering are left unspecified, then it is not an _error_ that they are unspecified. Nor is it surprising that users of one particular system decide to specify it for their case. [12:37:23.0000] <jgraham> It doesn't really matter what the Process is as long as the right text ends up in a document that implementors can use [12:37:37.0000] <david_carlisle> jgraham: yep [12:37:42.0000] <jgraham> Prefereably in a Rec. or something with similar properties [12:40:03.0000] <david_carlisle> TabAtkins: Out of interest what's the bounding box of <span>hello</span> ? that is can you point me to an exact wording somewhere in html spec (or any spec) that I could see what you want me to say for the bounding box of <mtext>hello</mtext> (which should be the same) [12:44:54.0000] <tantek> david_carlisle - that would be inline layout as defined in CSS2.1 [12:45:03.0000] <tantek> (re: the span) [12:45:24.0000] <tantek> and yes, one possible answer is for MathML to simply phrase everything in terms of defined CSS layout terms [12:49:25.0000] <david_carlisle> tantek: I'm looking at that(and Ive read it before:-) but where for example do I find if the width of the box is based on rendered pixels or on the font metrics so including the left sidebearing of h and right of o for example? (and similarly for height) as that seemed to be level of detail TabAtkins was asking for MathML [12:50:28.0000] <tantek> david_carlisle I believe that is defined where stated (CSS 2.1 inline layout) and used by various things such as painting the background, drawing borders etc. [12:50:46.0000] <tantek> (when such spans have backgrounds and borders) [12:50:46.0000] <tantek> f [12:50:47.0000] <tantek> r [12:50:48.0000] <tantek> om [12:51:02.0000] <tantek> from what I recall, that was quite a bit of work in the last few years of CSS 2.1 [12:51:16.0000] <tantek> figuring out inline layout details to get interop background/border rendering [12:53:29.0000] <david_carlisle> "inline layout" doesn't occur in the toc http://www.w3.org/TR/CSS21/cover.html#minitoc I could read the whole spec, or do you mean somewhere else? (clear;y there is a lot of detail about inline and layout but I don't see that but I'm not sure if I'm looking at the right spec:-) [12:54:59.0000] <david_carlisle> Ah found it.. CSS 2.1 says: The 'height' property does not apply. The height of the content area should be based on the font, but this specification does not specify how. [12:55:04.0000] <david_carlisle> hmmmmmm [12:56:20.0000] <david_carlisle> and goes on to say If more than one font is used (this could happen when glyphs are found in different fonts), the height of the content area is not defined by this specification. [12:57:33.0000] <david_carlisle> got to go, back later [13:57:01.0000] <dglazkov> MikeSmith: it's dead again. [14:02:38.0000] <jgraham> {insert Monty Python reference here} [14:03:00.0000] <Ms2ger> I'm not dead yet [14:04:17.0000] <dglazkov> http://downforme.org/is-dvcs.w3.org-down-today-for-everyone [14:04:38.0000] <jgraham> Ms2ger: YOu are an AI not a parrot [14:04:51.0000] <jgraham> Although it seems you are an AI that thinks it is a parrot [14:05:07.0000] <Ms2ger> Excuse me. [14:05:13.0000] <Ms2ger> I would like a fish license. [15:09:01.0000] <splend> Hi! Where can I download the specification for offline reading? [15:16:06.0000] <edwardbc> splend: http://developers.whatwg.org/offline.html [15:18:50.0000] <splend> edwardbc: Oh, I'm sorry I wasn't clear. I want to download the current HTML specification as file(s), so I can read it without a connection to Internet. [15:20:17.0000] <edwardbc> oh I see, can't help you with that (maybe print that to PDF?), but the whatwg site actually uses the offline manifest, so after the first load you should be able to open that same page without internet access [15:20:36.0000] <gsnedders> Or just save the file using your browser of choice. [15:22:24.0000] <edwardbc> indeed, heh [15:26:50.0000] <splend> gsnedders: Thanks for the suggestion, but I think the whole specification as a single HTML file is too much form my low-spec computer to handle easily. [15:27:50.0000] <splend> edwardbc: Printing to a PDF seems like a good suggestion – I think it will do fine. Thank you! [15:30:00.0000] <edwardbc> no problem, you can build a local copy as well: https://github.com/benschwarz/developers.whatwg.org [15:30:40.0000] <edwardbc> (until now I get that you meant the whole spec) [15:32:27.0000] <splend> Either is fine, whole or Web Dev Edition. [15:32:51.0000] <splend> I'll see if I can manage to built it on my own. Thank you! [15:32:58.0000] <edwardbc> :) [16:15:42.0000] <MikeSmith> dglazkov: should be working now [16:16:01.0000] <dglazkov> MikeSmith: \o/ [16:16:06.0000] <MikeSmith> I think the cause may be that the swap partition on that machine needs to be bigger [16:20:36.0000] <benschwarz> splend left, but if people want to try build the developer spec I can help them. [16:20:45.0000] <benschwarz> also, it uses application cache, so they can already read it offline [16:33:12.0000] <ojan> annevk, tantek: i'm about to file a bunch of bugs on chromium's fullscreen implementation and i want to double check that i understand the current state of the spec correctly [16:33:40.0000] <tantek> ok [16:33:41.0000] <ojan> annevk, tantek: for now, we've punted on the nested requestFullScreen situation by saying the requestFullScreen throws if you already have an element fullscreened [16:33:47.0000] <ojan> annevk, tantek: is that correct? [16:34:00.0000] <tantek> no, you should be able to do nested fullscreens [16:34:13.0000] <tantek> but I suppose an implementation could treat that as incremental implementation [16:34:24.0000] <tantek> with Mozilla we're implementing nested fullscreens [16:34:49.0000] <ojan> tantek: i guess i don't understand these lines from the spec then... [16:34:52.0000] <ojan> tantek: The context object's node document fullscreen element stack is not empty and its top element is not an ancestor of the context object. [16:34:52.0000] <ojan> A descendant browsing context's document has a non-empty fullscreen element stack. [16:35:12.0000] <ojan> tantek: from the 1st step of requestFullscreen [16:35:16.0000] <tantek> ojan, annevk will be faster at answering specific spec wording questions [16:35:18.0000] <ojan> tantek: saying to error [16:35:30.0000] <tantek> and I think it's sleeptime in his TZ right now [16:36:04.0000] <tantek> let me check - http://wiki.whatwg.org/wiki/Irc-people [16:36:11.0000] <tantek> hmm - no TZs for annevk [16:36:16.0000] <tantek> ojan - add yourself: http://wiki.whatwg.org/wiki/Irc-people [16:36:17.0000] <ojan> tantek: k. no worries. just curious, is mozilla implementing a stack of fullscreens? e.g. if you hit escape it removes you form the top-level fullscreened element only? [16:36:35.0000] <tantek> no escape pops all the fullscreens off [16:36:46.0000] <tantek> the only situation when it doesn't is when there are intermingled dialogs [16:36:50.0000] <tantek> which we don't implement yet [16:36:59.0000] <tantek> but the model is supposed to be: [16:37:00.0000] <tantek> e [16:37:01.0000] <tantek> s [16:37:02.0000] <tantek> cap [16:37:03.0000] <tantek> e [16:37:16.0000] <tantek> escape pops all the fullscreens off the stack until the next dialog element or there are no more fullscreens [16:37:47.0000] <ojan> tantek: hmm...apparenltly i don't have access to the whatwg wiki and need an administrator to create me an account? [16:38:23.0000] <ojan> tantek: interesting...so is there a way to just pop off the top element? [16:38:33.0000] <tantek> ojan - from API yes [16:38:48.0000] <ojan> tantek: ok...you'd have to provide UI for it in your page [16:38:50.0000] <tantek> however if the user hits the escape key, they're all supposed to pop off until the next dialog element [16:38:54.0000] <tantek> right [16:38:59.0000] <tantek> that's what sites do today [16:39:04.0000] <tantek> provide a fullscreen button [16:39:05.0000] <tantek> an [16:39:05.0000] <tantek> d [16:39:11.0000] <tantek> and an exit fullscreen button [16:39:20.0000] <tantek> (sorry for the extra lines, not sure what Colloquy is doing on my side) [16:39:30.0000] <tantek> for each level [16:39:42.0000] <ojan> tantek: that's true...but escape, in practice would only escape the top of the stack...although the reason is mostly accidental [16:40:05.0000] <ojan> tantek: the case i'm thinking of is having a fullscreened slideshow with a youtube video in it that you then fullscreen [16:40:32.0000] <ojan> tantek: i suppose if you did the slide show and youtube in flash, hitting escape would exit both [16:40:49.0000] <tantek> right [16:40:49.0000] <ojan> it's surprising how complicated fullscreen behavior is. [16:41:05.0000] <tantek> ojan - many new things on the web are apparently complicated [16:41:10.0000] <tantek> until you're used to their intricacies [16:41:11.0000] <tantek> a [16:41:12.0000] <tantek> n [16:41:16.0000] <tantek> and then you take them for granted [16:41:25.0000] <ojan> i suppose that's true of most computer science endeavors...once you really dig into it, all the problems are hard [16:41:46.0000] <ojan> or at least...harder than they look [16:41:57.0000] <ojan> tantek: anyways...thanks... that helps inform my bug-filing [16:42:28.0000] <tantek> ojan - no problem [16:42:41.0000] <tantek> the escape / fullscreen / dialog interaction is non-trivial [16:43:04.0000] <tantek> and it took a whole lunch in-person between myself, hober, Hixie, Tab to work it out. with printouts of use-cases and everything [16:43:28.0000] <tantek> I've also walked folks through the scenarios here on the Mozilla side in our security reviews. [16:50:01.0000] <tantek> ojan - for the wiki you need to ask Hixie for an account [16:54:53.0000] <TabAtkins> Or get annevk to do it for you, I think. [16:55:13.0000] <TabAtkins> tantek: Any idea why sometimes you end up sending multiple single-letter lines in irc? [16:55:31.0000] <TabAtkins> Oh, I see that you notice that already. [16:55:32.0000] <TabAtkins> Never mind. [16:57:21.0000] <tantek> sorry about that TabAtkins [16:57:26.0000] <tantek> I'm trying to track down the problem [16:58:04.0000] <TabAtkins> No problem. As long as you're aware of it I'm okay. 2012-05-22 [17:07:13.0000] <MikeSmith> dglazkov: I set a cron job to check the dvcs.w3.org web server and restart it if it's down [17:08:09.0000] <MikeSmith> and which also sends a naggy message on IRC to the systems team and me each time it restarts it [17:11:06.0000] <ojan> cheesze0 [17:11:14.0000] <ojan> lol [17:11:22.0000] <ojan> now everyone knows my sc2 password [17:11:51.0000] <TabAtkins> brb, hacking ojan's sc2 [17:11:54.0000] <ojan> i should write a chrome extension that never lets me type my password [17:11:59.0000] <TabAtkins> gonna ruin his ladder ranking [17:12:15.0000] <ojan> sigh...i suppose i have to change it now [17:12:17.0000] <TabAtkins> (not sure if we're in a ladder freeze right now or not) [17:21:25.0000] <dbaron> I'm puzzled by the multiple one-letter-on-a-line patterns from tantek given how he was talking about how great his keyboard is. [17:24:59.0000] <tantek> sometimes there's even two letters [17:26:00.0000] <TabAtkins> There's 3 letters in the "escape" one up above. [17:26:10.0000] <TabAtkins> Crazy! [17:26:28.0000] <tantek> :) [17:30:15.0000] <jamesr_> ojan_away, if sc2 allows it, prefix your password with "/!" and your IRC client won't let it go out to channels [17:39:42.0000] <Wilto> Hey WHATWG, I’m here to apologize. [17:40:54.0000] <Wilto> My frustration with processes and such turned into outright hostility at points, and for that I am sorry. [17:41:02.0000] <TabAtkins> Hey Wilto! [17:41:08.0000] <Wilto> Yo. [17:41:14.0000] <TabAtkins> Happy we're all back to useful work. ^_^ [17:42:14.0000] <Wilto> This intrinsic width thing—the more I think about it, the more I think it may have been staring us in the face the whole time. [17:43:07.0000] <Wilto> If an author wants a fixed-width image at 1.5x resolution, wouldn’t height and width attributes solve that neatly? [17:43:32.0000] <TabAtkins> I think the intrinsic width part isn't very important. If you want a fast load you have to give @width and @height anyway, or else you'll produce extra layouts as images gets downloaded. [17:43:50.0000] <Wilto> Meanwhile, authors implementing flexible images with `max-width: 100%` could remove them; no harm, no foul. [17:44:01.0000] <TabAtkins> The fact that the "Nx" modifier changes the intrinsic is thus a good detail, but not an important one. [17:44:30.0000] <TabAtkins> The important bit of the "Nx" component is allowing UAs to decide which one to download, is all. [17:45:13.0000] <Wilto> This isn’t a defense, but wouldn’t `picture` with a height and width attribute make the pixel density MQ viable again? [17:45:40.0000] <TabAtkins> Nope, it doesn't change the reasoning in my blog post. [17:46:24.0000] <Wilto> Bandwidth? [17:47:31.0000] <TabAtkins> That's the whole reasoning for the "Nx" component, yes. [17:48:25.0000] <Wilto> So the idea is that bandwidth detection on the UA side would override the resolution set by the author, yeah? [17:48:37.0000] <Wilto> I’m not baiting you or anything; I’m actually just getting caught up after a few days. [17:48:42.0000] <zewt> jesus christ https://www.khronos.org/webgl/public-mailing-list/archives/1205/msg00199.html [17:48:54.0000] <zewt> The sooner we can achieve media type registration, the sooner interoperable Web-harmonious WebGLSL web services and serialization formats can be deployed and independent developers can build on the common type name. [17:49:11.0000] <Wilto> zewt: ABOUT TIME no I have no idea what any of that means [17:49:15.0000] <TabAtkins> Not quite. The idea is that, once you trim down the images that are invalid according to the width/height components, the browser is then free to use whatever metrics they want to choose between resolutions. [17:49:47.0000] <TabAtkins> So that they can, for example, forgo the high-res image on a 1x screen, or if on a low-bandiwdth connection. [17:51:22.0000] <Wilto> TabAtkins: Couldn’t we just apply that same line of reasoning to media queries? If on low bandwidth, ignore the `source` that matches an explicitly-defined resolution higher than X? [17:51:59.0000] <TabAtkins> Wilto: Not without changing the way that one MQ works, incompatibly with the way that MQs work in general. [17:52:01.0000] <Wilto> I feel like we’re circumventing the author’s intent in either case—which is perfectly logical. [17:53:16.0000] <TabAtkins> Right, it's not a problem of intent. We know what we want to do. It's a problem of purity. That's normally a very low concern, but when you have an equally-powerful solution that doesn't require confusing up the issue, it's better. [17:53:19.0000] <Wilto> Are we operating under the assumption that the UA will always make these decisions, or that there’ll be a “low bandwidth” setting or somesuch that kicks this logic into play? [17:53:47.0000] <TabAtkins> For example, if we use the "device-pixel-ratio" MQ for this use-case, and we define that it can be ignored sometimes, does that apply to general CSS as well? [17:54:34.0000] <Wilto> I feel like we have some wiggle room where there’s not much precedent for this in markup, and something like that would play _very_ well with `video` sources. [17:54:53.0000] <Wilto> I mean, they’re much more bandwidth-intensive, and a solution that covers both as outlined above would be pretty amazing. [17:55:17.0000] <TabAtkins> For video sources we probably want a somewhat more custom solution too, since you want to choose *not* on the attributes of the device, but on the attributes of the stream. [17:55:18.0000] <Wilto> You could make a case that CSS is another beast altogether, I suppose. “This is how things work with `media` attributes.” [17:55:57.0000] <TabAtkins> Yeah, that's confusing. ^_^ Best to avoid it if we can. [17:56:18.0000] <Wilto> I don’t know. Otherwise, we have “this is how things work with images, this with video, and this with CSS.” [17:56:33.0000] <TabAtkins> Really, the issue is just that you're not *actually* caring about the device's capabilities (or rather, you care about a much wider range of things than just the device's pixel ratio, many of which can't be exposed via MQ). You care about some qualities of the image. [17:56:36.0000] <Wilto> Again: not going on the offensive, here. Just throwing that out there. [17:56:45.0000] <TabAtkins> Wilto: Don't worry, I'm not interpreting you as hostile. [17:57:17.0000] <zewt> in my experience, trying to forcibly overgeneralize solutions to things that aren't really the same problem is one of the most common reasons for things being overcomplex and awkward [17:57:37.0000] <zewt> maybe images and video can be done the same way; but be careful not to assume it [17:58:34.0000] <Wilto> No, but that `media` attribute pattern already exists for this purpose. It’s far from ideal, but maybe “responsive images” is an opportunity to improve that to everyone’s benefit. [17:58:48.0000] <TabAtkins> Yeah. Reuse when possible, but don't be afraid to do custom solution when called for. If you have to invent new processing behavior in your new context, it's not really "reusing", it's just "borrowing syntax". [17:59:30.0000] <TabAtkins> Wilto: The @media attribute in video's <source> doesn't seem to actually be very useful, and its removal is being discussed. [17:59:52.0000] <TabAtkins> What you *actually* want to do with video sources is negotiate based on bitrate/size/etc. [18:00:17.0000] <TabAtkins> I don't think there's a strong argument to vary the video based on window width/height, like there is for images. [18:01:52.0000] <Wilto> Well, you could make a case that it isn’t used much where we can’t use it, but I see your point. [18:02:29.0000] <TabAtkins> I'll just say that I find it *very* unlikely that sites would generally record multiple videos for different screen sizes. ^_^ [18:02:32.0000] <Wilto> But having worked on video in a flexible layout, I wish the hell we could. [18:03:02.0000] <zewt> TabAtkins: having separate eg. 720p and 1080p videos is pretty reasonable [18:03:28.0000] <zewt> it'd see much less use than images, though, since it's a lot more expensive [18:03:29.0000] <Wilto> Yeah. We had a handful of videos in play on the Boston Globe site. [18:03:33.0000] <TabAtkins> zewt: Those aren't different sizes, though, right? Different res streams are useful to present, but don't seem to be tied to screen size strongly. [18:04:40.0000] <zewt> sure it is: serve a 720p video to people with a 720p display, 1080p to 1080p [18:05:06.0000] <TabAtkins> zewt: But again, a 720p screen isn't strongly tied to screen size, is it? [18:05:47.0000] <zewt> not the use case I'm talking about [18:06:08.0000] <zewt> 720p for a 720p display, regardless of the screen size [18:06:36.0000] <TabAtkins> Yes yes, that's fine. That's not waht I was talking about. Since you were responding to me talking about multiple screen sizes, I assumed you were talking about the same thing. [18:06:58.0000] <zewt> i'm talking about fullscreen video, where the screen size doesn't matter [18:07:16.0000] <TabAtkins> Okay, sure. I agree. ^_^ [18:07:25.0000] <zewt> (but there are a lot of other parameters to selecting a video, which may change while the video is playing, which is complex) [18:07:52.0000] <TabAtkins> No one's disagreeing that serving multiple res streams is useful. YouTube offers a strong existence proof that it's a good use-case. [18:08:13.0000] <zewt> eg. changing bitrate if the user's stream or CPU can't keep up with the one that was selected--that's pretty severely different from anything static images have to deal with [18:09:22.0000] <TabAtkins> It's kinda similar to a browser deciding "okay, the radio's still awake but nothing's used it for a second, so I can start downloading the hi-res versions of these images now". [18:09:52.0000] <TabAtkins> At least, from the authoring standpoint it's similar, and in broad strokes it's similar from the UA perspective too. [19:02:12.0000] <heycam> MikeSmith, did the certificate on dvcs.w3.org just change? only just got a warning about it now, although I see that it was issued a couple of weeks ago [19:02:54.0000] <MikeSmith> heycam: not as far as I know [19:03:02.0000] <heycam> hmm [19:03:14.0000] <MikeSmith> I seem to recall getting warnings about that for a while now, depending on the UA [19:04:16.0000] <MikeSmith> hmm and now it just seems to be hanging [19:04:44.0000] <heycam> MikeSmith, this is what I got just before: http://pastebin.mozilla.org/1647715 [19:05:12.0000] <MikeSmith> I see [19:05:16.0000] <MikeSmith> also see "remote: abort: HTTP Error 500: Cannot allocate memory" [19:05:24.0000] <heycam> yeah, that doesn't look good [19:06:22.0000] <MikeSmith> Swap: 524280k total, 437012k used, 87268k free, 16068k cached [19:59:22.0000] <niftylettuce> is there a way to specify a webkit location in the /temporary folder for FileEntry API for an <input type="file" /> somehow? [19:59:53.0000] <niftylettuce> e.g. <input type="file" temporaryfile="true" src="http://somesite.com/temporary/somefile.png" /> [20:44:59.0000] <niftylettuce> rather, is there a way to append File Objects to forms? [20:45:01.0000] <niftylettuce> without XHR [22:09:51.0000] <MikeSmith> heycam: I still can't reproduce that cert error [22:10:07.0000] <heycam> MikeSmith, ok -- it seems to have gone away for me now too [22:10:29.0000] <heycam> MikeSmith, I wonder if the HTTP response was truncated due to the memory problem and that caused the cert error [22:10:36.0000] <MikeSmith> ah yeah [22:10:39.0000] <MikeSmith> perhaps [22:10:50.0000] <Von_Davidicus> Heyo, all. [22:11:09.0000] <MikeSmith> heycam: I will try to get some more swap space on that machine. I think that's what's causing the problems [22:11:18.0000] <heycam> MikeSmith, ok cool, thanks for looking into it [22:18:27.0000] <Von_Davidicus> Question about HTML5: Will there be a mechanism to let elements be selected in JavaScript via -class-? [22:19:55.0000] <heycam> can you use IRIs in CSS url() notation? [22:20:49.0000] <zewt> "IRI"? is that just to add to the URL/URI fun [22:20:53.0000] <zewt> heh [22:20:59.0000] <heycam> yeah haven't you heard of it? :) [22:21:04.0000] <zewt> no :0 [22:21:05.0000] <zewt> also :) [22:21:08.0000] <heycam> http://annevankesteren.nl/2005/02/iri [22:22:04.0000] <zewt> The thing is, the 1.0.1 release of Firefox is scheduled for this or maybe next week. [22:22:07.0000] <zewt> sounds contemporary :) [22:22:14.0000] <heycam> :p [22:22:27.0000] <Von_Davidicus> *Checks* Apparently URIs are limited to ASCII characters, while IRIs are intended to work with anything. [22:22:29.0000] <zewt> i don't even bother trying to distinguish URL/URI; nobody cares and I just say URL [22:31:53.0000] <zcorpan> heycam: css3-syntax seems to tokenize it at least [22:33:11.0000] <zcorpan> http://software.hixie.ch/utilities/js/live-dom-viewer/saved/1536 and browsers accept it [22:33:36.0000] <heycam> zcorpan, oh as part of the tokeniser. didn't realise CSS had a html-spec-like tokeniser actually! [22:33:47.0000] <zcorpan> it didn't until recently [22:36:29.0000] <zcorpan> /me wonders what happens with lone surrogates in url() [22:36:54.0000] <heycam> I think http://dev.w3.org/csswg/css3-values/#urls doesn't exactly say to interpret the URI token as a URI rather than anything else [22:38:18.0000] <Von_Davidicus> Hello, hasather. [22:41:47.0000] <zcorpan> http://software.hixie.ch/utilities/js/live-dom-viewer/saved/1537 - why is the ") }" missing in the DOM view in chrome? [22:46:25.0000] <zcorpan> Hixie: live dom viewer throws in encodeURIComponent when using unpaired surrogates [22:46:44.0000] <zcorpan> i'm surprised all browsers agree to throw for that case [22:46:54.0000] <zcorpan> (well haven't tested ie, but anyway) [22:49:34.0000] <zcorpan> seems chrome U+FFFDs a lone surrogate in url() [22:52:41.0000] <Von_Davidicus> ... which character is that? [22:52:44.0000] <zewt> fffd seems much more sane than a really obscure exception [22:56:26.0000] <zcorpan> zewt: the exception was unrelated :-) [22:56:39.0000] <zcorpan> Von_Davidicus: REPLACEMENT CHARACTER [22:56:51.0000] <zewt> unicode is A VERY LOUD STANDARD [22:57:23.0000] <zcorpan> yes, obviously characters are so important they have to be all-uppercase [22:57:34.0000] <zewt> can we have a movement or riot or something to eliminate screaming caps MUST from all standards [22:57:48.0000] <zewt> yes, it's a requirement, we get it, STOP YELLING AT ME [22:57:55.0000] <heycam> small caps works [22:58:05.0000] <zewt> it's ugly and jarring and unnecessary [22:58:26.0000] <zewt> html does just fine without it :) [22:58:30.0000] <zcorpan> i prefer lowercase, though i get much more worked up about specs that just suck [22:58:45.0000] <zewt> well, it'd take more than a riot to make all specs not suck [22:58:53.0000] <zewt> gotta have reasonable goals [22:59:01.0000] <Von_Davidicus> Sorry, small caps do not work for me. I prefer hats that fit properly. [22:59:26.0000] <zcorpan> that's it! give hats to all musts! [22:59:37.0000] <zewt> ____ [22:59:38.0000] <zewt> must [22:59:42.0000] <zcorpan> what's the code point for a hat? [23:00:04.0000] <heycam> I was gonna go with mûst [23:03:09.0000] <zcorpan> U+2229 [23:03:43.0000] <zcorpan> oh, here we go! U+1F3A9 [23:04:22.0000] <zcorpan> also U+1F452 [23:05:47.0000] <zcorpan> one hat for male requirements and one for female requirements [23:06:32.0000] <zcorpan> maybe have one gender for impl requirements and the other for author requirements? [23:11:22.0000] <Von_Davidicus> Sadly, I can't see what those characters are. Alas. [23:13:37.0000] <Von_Davidicus> Well, tomorrow we'll see if my teacher consigns me to the Funny Farm. Or if he merely blows his stack. [23:44:41.0000] <zcorpan> ok here's the "enter" function to be used in the tree walker at the bottom of http://www.whatwg.org/specs/web-apps/current-work/multipage/sections.html to give requirements hats: function(node) { [23:44:41.0000] <zcorpan> if (node instanceof Text) { [23:44:41.0000] <zcorpan> node.data = node.data.replace(/\b(must\s+not|must|may|should\s+not|should)\b/g, function(match) { return match + '\uD83C\uDFA9'; }); [23:44:41.0000] <zcorpan> } [23:44:41.0000] <zcorpan> } [23:45:13.0000] <heycam> zcorpan, you are either up too early or too late [23:45:55.0000] <zcorpan> i had set my clock at 06:00 [23:54:57.0000] <jgraham> That isn't even a real time [23:55:14.0000] <zcorpan> of course it is [23:56:04.0000] <zcorpan> i can't decide if i should send feedback to www-style about url() not caring about non-ascii or if i should wait for the URL spec [23:58:28.0000] <matjas> lol @ http://twitter.com/panic/status/204826332322992128 [23:59:02.0000] <heycam> zcorpan, I already sent a mail [23:59:33.0000] <zcorpan> ah [00:00:15.0000] <annevk> matjas: :/ [00:00:28.0000] <zcorpan> ok, that's it. duck duck go just isn't good enough. it doesn't find me what i'm looking for. back to google. :-( [00:00:31.0000] <Von_Davidicus> matjas? [00:01:15.0000] <matjas> Von_Davidicus: “Yeah, a good suggestion to remove that tumor. Make sure to report that after we stitch.” [00:02:17.0000] <annevk> heycam: yes you can [00:02:30.0000] <annevk> heycam: in addition their URL character encoding is always utf-8, contrary to HTML [00:02:46.0000] <annevk> heycam: not that URL character encoding is a particularly well documented concept... [00:02:53.0000] <heycam> annevk, ok cool, but I couldn't find where it says how to interpret the characteres within the url( ) [00:03:16.0000] <annevk> heycam: probably not well defined; but they're parsed just like in HTML [00:03:28.0000] <annevk> heycam: unless parsing of url() fails of course [00:03:57.0000] <annevk> that is, spaces and such work fine [00:04:03.0000] <heycam> annevk, interpret as in "this sequence of characters in the URI token is a (new-style) URL, or a URI, or a URL" [00:04:14.0000] <heycam> I guess CSS will refer to the new URL spec at some point [00:04:14.0000] <annevk> but you might need to escape the space to make CSS parsing not trip; forgot about that [00:04:25.0000] <annevk> yeah [00:04:43.0000] <annevk> CSS and XHR are the same [00:04:59.0000] <heycam> in that case, $svgspec =~ s/funciri/url/g :D [00:05:04.0000] <annevk> HTML also sets the URL encoding to something weird [00:05:06.0000] <zcorpan> or use url("....") [00:05:13.0000] <Von_Davidicus> What does XHR stand for? [00:05:20.0000] <annevk> heycam: yeah [00:05:41.0000] <annevk> heycam: SVG should do the same as HTML I think, but probably minus the crazy encoding stuff, although given a mixed document... [00:05:48.0000] <annevk> Von_Davidicus: XMLHttpRequest [00:06:29.0000] <heycam> annevk, I haven't looked into what crazy encoding stuff there is in HTML (for interpreting href attributes differently from the documenting encoding or something I guess?) [00:06:49.0000] <zcorpan> annevk: i guess teh URL spec will deal with unpaired surrogates, yeah? [00:06:56.0000] <Von_Davidicus> Okay. Hmm--curiosity question: what, aside from AJAX, uses XMLHttpRequest? [00:07:04.0000] <annevk> heycam: for interpreting them using the document encoding actually [00:07:08.0000] <heycam> oh [00:07:10.0000] <annevk> heycam: cause URLs normally use utf-8 [00:07:13.0000] <heycam> ahh [00:07:15.0000] <heycam> I see [00:07:26.0000] <heycam> so interpreting %nn things differently [00:07:38.0000] <zcorpan> no [00:07:48.0000] <zcorpan> interpreting raw characters differently [00:07:53.0000] <heycam> oh [00:08:13.0000] <heycam> but that's sensible isn't it? interpreting the raw characters according to the document encoding? :) [00:08:24.0000] <annevk> heycam: è becomes two %hh normally, but one %hh for e.g. windows-1252 [00:08:25.0000] <zcorpan> see /topic [00:08:47.0000] <heycam> sensible = illogical for the web platform :) [00:08:53.0000] <annevk> heycam: and only when part of the query string of course [00:08:59.0000] <annevk> heycam: in the path it's always two %hh [00:09:14.0000] <annevk> heycam: see http://annevankesteren.nl/presentations/1F4A9.html [00:09:21.0000] <zcorpan> heycam: it's not really sensible if the url goes to a different server, for instance [00:09:28.0000] <heycam> ok. well I'm going to forget about those details for the moment and hopefully I'll just add some references to toher specs and things will just work! [00:09:39.0000] <annevk> "just work" mwaha [00:10:29.0000] <annevk> zcorpan: haven't played with isolated surrogates; I was kind of hoping someone else would be doing URL stuff, but I guess arv just wanted to update a few small things [00:11:34.0000] <zcorpan> i guess we should just use the webidl code-units-to-utf-8 rules when the input is code units [00:12:16.0000] <zcorpan> (although we can't use that for the query if url encoding isn't utf-8, of course) [00:12:27.0000] <heycam> I don't think I added those rules, just code units to characters [00:12:46.0000] <zcorpan> ah. still [00:13:41.0000] <annevk> we could use those for the query too I think [00:14:12.0000] <zcorpan> yeah [01:04:30.0000] <jgraham> Turns out that having three copies of the complete spec open makes restarting your browser slow [01:45:26.0000] <Stevef> proposal to add <transcript> element to HTML5 http://www.w3.org/html/wg/wiki/ISSUE-194/TranscriptElement [03:27:24.0000] <matjas> annevk: *now* i understand what you liked about my twitter bio [03:28:07.0000] <annevk> took you long enough :p [03:28:26.0000] <matjas> none of you guys told me about that presentation :'( [03:28:53.0000] <annevk> matijs recorded it for you I thought [03:29:08.0000] <annevk> but he prolly never got around to do the work of syncing the various media together [03:29:31.0000] <annevk> it sounded kind of ambitious when he pitched it in the train [03:39:54.0000] <matjas> annevk: huh, i thought that was the big5 presentation [03:40:24.0000] <odinho> annevk: My setup is better then, because it needs no postproduction. Live-switching, cutting and streaming everything while the presentation is going :-) [03:42:12.0000] <annevk> matjas: never presented on big5 in detail [03:42:24.0000] <annevk> matjas: think that would've gone too far :) [03:42:43.0000] <matjas> well, fingers crossed for #fronteers12 [03:42:47.0000] <annevk> Ms2ger`: yo yo [03:42:55.0000] <annevk> Ms2ger`: yo yo yo [03:43:02.0000] <Ms2ger`> Hi [03:43:07.0000] <annevk> Ms2ger`: yay it worked [03:43:18.0000] <annevk> Ms2ger`: so Anolis "NOTE" picks the wrong style sheet [03:43:25.0000] <Ms2ger`> I blame gsnedders [03:43:27.0000] <annevk> Ms2ger`: not sure if we should bother fixing it [03:43:44.0000] <annevk> it uses W3C Note while everything else says W3C Working Group Note [03:43:52.0000] <annevk> (the latter I want) [03:44:08.0000] <annevk> not really fond of this whole Note thing though [03:44:22.0000] <annevk> but I'm happy doing Art another favor [03:44:59.0000] <Ms2ger`> What says "W3C Note"? [03:45:16.0000] <Ms2ger`> Oh, huh [03:45:34.0000] <annevk> there's W3C-Note and W3C-WG-Note [03:45:42.0000] <annevk> prolly also W3C-IG-Note and various others [03:45:54.0000] <annevk> it's kind of silly Notes are not just a global thing [03:46:00.0000] <annevk> because they all mean nothing [03:46:07.0000] <annevk> but I guess W3C likes it complicated [03:46:12.0000] <annevk> keeps people busy [03:46:52.0000] <zcorpan> W3C-CG-Note? [03:47:47.0000] <Ms2ger`> /me doesn't understand this code [03:55:22.0000] <matijsb> annevk: still not got round to it :| [03:57:57.0000] <gsnedders> Ms2ger`: For once blaming me seems to be legit. [03:57:59.0000] <gsnedders> :) [05:18:13.0000] <annevk> Velmont: yo yo [05:18:30.0000] <annevk> Velmont: do you remember what came out of the webapps discussion? [05:18:45.0000] <annevk> Velmont: regarding CORS [05:18:53.0000] <annevk> I guess I have some time to work on that now... [05:21:07.0000] <annevk> /me finds http://lists.w3.org/Archives/Public/www-archive/2012May/att-0022/minutes-2012-05-02.html#item03 [05:23:50.0000] <annevk> okay so that was a comment from the minute taker [05:24:06.0000] <annevk> but instead of working on the technical feedback, he worked on the minutes :p [05:26:27.0000] <annevk> is there an easy way to convert a bunch of HTML to wiki? [05:27:16.0000] <Ms2ger`> Something like http://openfacts2.berlios.de/html2wiki/index.php ? [05:56:45.0000] <jgraham> Hmm, var t0=Date.now(); document.write("<link rel='stylesheet' src='slowloading.css'><script>alert(Date.now() - t0)<\/script>"); [05:57:05.0000] <jgraham> Should that document.written script block on the stylesheet loading? [05:58:16.0000] <jgraham> (afaict browsers don't block, but I can't quite work out why the spec wouldn't block there) [06:02:13.0000] <smaug____> hsivonen: ^ [06:33:07.0000] <webben> hsivonen: What's the holdup on getting latest nu.validator.htmlparser into maven? is that done by a third party? [06:52:06.0000] <gsnedders> /me gets rid of 10/13 failures in html5lib on Py2 [06:52:24.0000] <gsnedders> (Mainly because I can) [07:01:45.0000] <gsnedders> jgraham: Do you have any idea what the expected behaviour of FormFiller is? [07:03:59.0000] <jgraham> gsnedders: No [07:04:10.0000] <gsnedders> jgraham: Greast. [07:04:11.0000] <gsnedders> *Great [07:04:45.0000] <gsnedders> /me is tempted to drop it and see who complains [07:04:57.0000] <gsnedders> html5lib contains too much shit. [07:06:16.0000] <jgraham> gsnedders: I am totally happy for you to do that [07:10:59.0000] <gsnedders> It's breaking in some fairly subtle way under Py3 [07:12:24.0000] <gsnedders> It impls something that has been removed from the spec [07:20:58.0000] <jgraham> Kill it. It's in vc if people want to revive it [07:22:16.0000] <annevk> Ms2ger`: looks interesting [07:23:51.0000] <gsnedders> jgraham: Have killed. Also fixed all but three of the current failures under Py2. [07:24:07.0000] <gsnedders> jgraham: (7 errors, 4 failures under Py3) [07:29:14.0000] <jgraham> Presumably you don't have lxml installed? [07:30:53.0000] <jgraham> I think we should kill the beautifulsoup support also, since BS3 used html5lib as the backend (I think) [07:30:58.0000] <gsnedders> No, I don't. [07:31:00.0000] <jgraham> Dunno what BS4 does [07:31:33.0000] <jgraham> Oh, BS4 can use html5lib [07:32:47.0000] <jgraham> (I assume it doesn't depend on our BS support) [07:35:19.0000] <gsnedders> Hmm, somewhat surprising that Python doesn't fast-case encoding a Unicode string to its underlying repr [07:37:55.0000] <gsnedders> The encoding is special-cased, but not decoding. [07:39:40.0000] <gsnedders> Also: html5lib/constants.py takes a while to run 2to3 on. [08:14:09.0000] <annevk> Ms2ger`: that converter fails horribly :( [08:14:25.0000] <Ms2ger`> Google had more :) [08:15:59.0000] <annevk> found a better one indeed [08:16:27.0000] <jgraham> http://johnmacfarlane.net/pandoc/ ? [08:17:55.0000] <annevk> http://labs.seapine.com/htmltowiki.cgi [08:27:15.0000] <annevk> http://www.w3.org/wiki/CORS [08:32:02.0000] <mhausenblas> in case you haven't seen it yet annevk … http://apiblog.youtube.com/2012/05/unlocking-javascripts-potential-with.html [08:32:46.0000] <annevk> mhausenblas: hadn't seen that yet, sweet [08:33:22.0000] <mhausenblas> yup ;) [08:36:10.0000] <annevk> ojan_away: hey [08:36:15.0000] <annevk> ojan_away: yt? [08:36:39.0000] <annevk> ojan_away: if you guys are not planning on implementing stacked fullscreen I think it would be good to mention that somewhere [08:36:58.0000] <annevk> ojan_away: hober told me Apple was not to keen on that either so maybe it should be removed [08:37:32.0000] <annevk> ojan_away: having said that, you still need the stack on the "top layer" of course for <dialog> [08:40:28.0000] <jgraham> /me wonders which part of _away was unclear ;) [08:41:58.0000] <annevk> jgraham: the part that said "has joined" [08:42:57.0000] <annevk> I wonder how often Julian is going to repeat his point... [08:48:41.0000] <annevk> how is <script crossorigin> tracked? [08:48:53.0000] <annevk> is there a bug on HTML or an unclosed email discussion? [08:53:01.0000] <zewt> clearly what we need is: more threads [08:53:10.0000] <zewt> that would solve everything! [08:53:32.0000] <zewt> mailing list NIH [08:54:00.0000] <dglazkov> good morning, Whatwg! [08:59:16.0000] <Ms2ger`> Good, dgathright [08:59:20.0000] <Ms2ger`> And dglazkov [09:00:47.0000] <dglazkov> :) [09:14:16.0000] <Workshiva> I guess there's nothing even close to a spec for XMLSerializer? [09:16:11.0000] <Ms2ger`> Sure is [09:16:22.0000] <Ms2ger`> Workshiva, http://html5.org/specs/dom-parsing.html [09:16:37.0000] <Ms2ger`> Not terribly close, but it's something [09:17:36.0000] <Workshiva> Looks buggy too [09:18:30.0000] <Ms2ger`> Sure [09:18:55.0000] <Workshiva> Do you want my bug repor there? :) [09:19:50.0000] <Ms2ger`> If zcorpan hasn't filed it yet :) [09:20:40.0000] <Workshiva> There's no escaping of text nodes in the non-cdata case [09:21:47.0000] <Workshiva> Opera 11.5 serializes <> as &lt;> and I was looking for something to confirm that's not right [09:22:04.0000] <Workshiva> (Chrome gives &lt;&gt; as expected) [09:22:32.0000] <niftylettuce> does anyone know of a possible way to have a setter in FileList API? [09:22:38.0000] <Ms2ger`> &lt;> sounds good [09:22:54.0000] <Ms2ger`> In HTML, at least... [09:23:01.0000] <Ms2ger`> Is XML sillier? [09:23:12.0000] <Workshiva> I'm sure it works, but it's inconsistent and silly [09:24:50.0000] <espadrine> Does anyone have an idea for a better scrollIntoViewIfNeeded design to implement in CSSOM? [09:34:02.0000] <gsnedders> jgraham: I'm struggling to see how to handle Unicode files under Python 3 while preserving lone surrogates as we do under Py2. [09:36:11.0000] <gsnedders> jgraham: Most Python codecs disallow lone surrogates, so encode/decode doesn't work (the fact that the UTF-8 one allows them is different to UTF-16 under Py2, for example). [09:39:43.0000] <gsnedders> Sorry, Unicode strings. [09:40:59.0000] <gsnedders> Unicode files are #221 [09:42:10.0000] <gsnedders> s/221/202/ [09:46:55.0000] <zewt> gsnedders: more or less they're logically equivalent to invalid UTF-8 sequences, an encoding error that you typically can't round-trip through any other encoding [09:47:25.0000] <zewt> technically you *can* (unlike most invalid sequences) with other unicode encodings, but whether that's a good idea is another question [09:47:29.0000] <gsnedders> zewt: Yes, I know. [09:47:39.0000] <zewt> i know you know :) [09:47:49.0000] <gsnedders> But at the moment html5lib expects them to round-trip. [09:49:34.0000] <gsnedders> /me thinks detectEncoding is broken under Python 3 and should be asserting, but isn't in any tests [09:49:39.0000] <gsnedders> /me wonders if we have tests for that [09:52:00.0000] <gsnedders> zewt: The HTML spec defines behaviour for all Unicode characters, not just valid ones. [09:52:10.0000] <jwalden> "the whatwg site actually uses the offline manifest, so after the first load you should be able to open that same page without internet access" USEFUL THING I DID NOT KNOW \o/ [09:52:20.0000] <jwalden> /me highlights those lines on http://krijnhoetmer.nl/irc-logs/whatwg/20120522 [09:53:00.0000] <gsnedders> /me thinks someone is a bit too happy [09:54:02.0000] <jwalden> :-P [09:54:15.0000] <jwalden> plz to not be disturbing my happiness [09:54:51.0000] <gsnedders> /me mumbles something about the [[Prototype]] setter [09:54:58.0000] <gsnedders> /me hides [09:55:07.0000] <jwalden> YOU'D BETTER HIDE [09:55:15.0000] <jwalden> :-D [09:56:06.0000] <Ms2ger`> jwalden, Java [09:56:09.0000] <Ms2ger`> /me hides too [09:56:21.0000] <gsnedders> /me needs to sit down and work out what to do about the whole accessor v. weird magic argument [09:56:40.0000] <gsnedders> (for __proto__) [10:00:25.0000] <gsnedders> (I'd rather have it as an accessor even if we posion the property descriptor) [10:01:13.0000] <zewt> gsnedders: where does it require being able to round-trip stray surrogates? in the parser, at least, they should parse straight to FFFD [10:03:04.0000] <gsnedders> zewt: What makes them do that? [10:03:21.0000] <zewt> browsers don't agree on it, but that's Firefox's behavior, and it's the only thing that makes sense [10:04:05.0000] <Ms2ger`> Thanks ;) [10:04:18.0000] <gsnedders> zewt: That's not what the spec says [10:04:39.0000] <zewt> then, at least on first impression, the spec is wrong [10:04:49.0000] <zewt> (last I recall, it *is* what the encoding spec says, FWIW) [10:05:09.0000] <gsnedders> Well, it does for the UTF-8 case. [10:05:14.0000] <gsnedders> For every other case it is unclear. [10:05:23.0000] <zewt> that's what the encoding spec is trying to fix, i believe :) [10:07:04.0000] <zewt> also FWIW, firefox and opera parse it to fffd, ie and webkit just spit out the surrogate [10:07:36.0000] <zewt> (ie8; would have to fire up a VM for ie9) [10:08:26.0000] <gsnedders> \xD8\x00 is the UTF-16BE U+D800, right? [10:08:34.0000] <gsnedders> /me checks he isn't being stupid with endianness [10:09:16.0000] <zewt> yeah [10:10:17.0000] <zewt> http://zewt.org/~glenn/surrogate%20test.html.utf16 [10:10:56.0000] <zewt> editing utf-16 files with a broken surrogate in vim is annoying, because vim goes "oh, I guess this isn't utf-16" and falls back on latin1 [10:11:05.0000] <zewt> and i can never remember how to force reloading in vim with a specific encoding [10:14:22.0000] <gsnedders> Well, it isn't UTF-16. [10:14:36.0000] <zewt> sure it is; it's just UTF-16 with an error [10:14:54.0000] <gsnedders> Well, sure, but it's equally UTF-8 with an error. :) [10:14:55.0000] <zewt> vim just guesses wrong, since it's a heuristic [10:15:22.0000] <zewt> the fact that almost every other byte is \x00 points more strongly to utf-16 with an error :) [10:16:43.0000] <grigs> so… um… about responsive images… [10:16:45.0000] <grigs> /me ducks [10:17:11.0000] <Ms2ger`> /me re-aims and shoots [10:17:15.0000] <grigs> lol [10:17:41.0000] <Ms2ger`> Go join a CG ;) [10:17:59.0000] <grigs> LOL. very nice. [10:18:19.0000] <grigs> Ok, seriously, I have doubts about our ability to reconcile lookahead pre-parser behavior and responsive images as authors want to use them. [10:19:10.0000] <grigs> I wrote a lot about it last night: http://blog.cloudfour.com/the-real-conflict-behind-picture-and-srcset/ [10:19:28.0000] <zewt> can you write less about it for those of us with day jobs :) [10:19:31.0000] <grigs> :-) [10:20:22.0000] <grigs> The gist is the pre-parser wants to know what image exists before layout is calculated, but responsive images don’t want to have that defined until layout is calculated. [10:20:24.0000] <jgraham> I only read the conclusion [10:20:37.0000] <jgraham> But I assumed that the issue was what you just said [10:20:53.0000] <jgraham> This is why we only use the viewport size [10:21:14.0000] <grigs> Yep, but using the viewport size causes a bunch of problems. [10:21:25.0000] <jgraham> And assume the layout is such that a specific viewport width/height would correspond to a specific container width/height [10:21:41.0000] <jgraham> Well it *solves* a bunch of problems [10:21:53.0000] <zewt> ("as authors want to use them"--some authors, not me) [10:21:56.0000] <gsnedders> https://www.w3.org/Bugs/Public/show_bug.cgi?id=17151 [10:22:02.0000] <jgraham> e.g. makes the technique possible to implement in a non-sucky way [10:22:50.0000] <jgraham> If authors really can't cope with the viewport proxy and are willing to kill pageload perf to get something different it will be obvious when they implement something different in javascript [10:23:07.0000] <jgraham> In the meantime assuming the proxy is good enough is the only viable solution [10:23:40.0000] <gsnedders> jgraham: '\ud800a' should produce a parse error according to what? [10:23:42.0000] <gsnedders> hsivonen: ^^ [10:25:26.0000] <grigs> jgraham: i like the phrase 'viewport proxy'. that's a helpful way to describe it. [10:27:20.0000] <grigs> jgraham: the question you pose about coping with the viewport proxy is a good one. as it stands right now, i don't know whether I would choose to implement srcset or not. [10:27:52.0000] <grigs> actually, i know i would choose to implement pixel density srcset, but not sure about the viewport proxy. [10:29:14.0000] <jgraham> grigs: Does it seem worse than the alternative (no feature)? [10:29:56.0000] <grigs> jgraham: i'm honestly torn about that. [10:30:42.0000] <zewt> gsnedders: http://dvcs.w3.org/hg/encoding/raw-file/tip/Overview.html#legacy-utf-16-encodings [10:30:46.0000] <grigs> it's something i've been noodling on a lot in the last week. [10:31:34.0000] <jgraham> I don't see how mot having the feature would be better, really [10:31:40.0000] <jgraham> *not [10:31:48.0000] <grigs> arguments against the feature: [10:32:15.0000] <zewt> that was a short list [10:32:21.0000] <grigs> * It encodes breakpoints into the HTML making markup harder to maintain. [10:32:55.0000] <grigs> * Because it is a proxy, it can be difficult to match the breakpoints defined in CSS. [10:32:57.0000] <zewt> all it does is encode image dimensions into the HTML (and resources), which is nothing new [10:33:45.0000] <grigs> zewt: No, it encodes image dimensions and design decisions (viewport breakpoints) into the markup which is new. [10:34:08.0000] <jgraham> (I really wish "breakpoints" hadn't been co-opted to mean something different from existing use in the context of debugging) [10:34:18.0000] <zewt> image dimensions in markup is not new. <img width=16 height=16> [10:34:45.0000] <grigs> zewt: the dimensions in srcset are viewport dimensions, not image dimensions. [10:35:01.0000] <zewt> the results of the calculation are image dimensions. [10:35:03.0000] <jgraham> grigs: I'm not sure it is a design decision. It is selecting different *content* [10:35:23.0000] <jgraham> So in that sense it should be in the content layer (i.e. HTML) [10:35:25.0000] <zewt> only the *input* is viewport dims [10:35:30.0000] <grigs> jgraham: as an author, how do you pick when to switch from one resolution image to another? [10:36:06.0000] <zewt> jgraham: well, that's true for pixel density, less so if you're selecting images of different sizes [10:36:44.0000] <jgraham> zewt: Images of different sizes are more different than images of different pixel density [10:36:48.0000] <zewt> actually, one thing that's really strange: there's no way in the markup to say what the pixel dimensions of each candidate are (equivalent to @width/@height for @src), which seems sort of bad [10:37:19.0000] <zewt> jgraham: that's what i said [10:37:22.0000] <gsnedders> jgraham: We need more tests :( [10:37:42.0000] <zewt> eg. if you're showing a 1000x100 image for viewports up to 1024x768 and a 2000x100 image for viewports higher than that [10:38:12.0000] <jgraham> grigs: I feel like there is some non-obvious answer to that I am supposed to give (i.e. not "pick some pixel widths to use as my breakpoints). Maybe that isn't really what people do :) [10:38:40.0000] <jgraham> gsnedders: That doesn't seem surprising [10:38:45.0000] <grigs> put differently, how do you decide at which viewport dimensions you should change from a.jpg to b.jpg? there is nothing about the image itself that would tell you that it makes sense to switch at 500px vs. 700px vs. X. [10:40:03.0000] <grigs> So what is most likely to happen is that authors will match the srcset dimensions to their media query “breakpoints” which is why I say design decisions will be in srcset. [10:40:09.0000] <jgraham> grigs: Well clearly in the normal case you would make a design decision that your layout should switch at some width. But the srcset feature would be useful even without CSS [10:40:23.0000] <gsnedders> jgraham: Do we have *any* tests for encoding detection? [10:40:34.0000] <jgraham> gsnedders: Ask sam :p [10:40:37.0000] <jgraham> Or maybe mark [10:40:51.0000] <jgraham> (glwt btw :p) [10:40:52.0000] <grigs> jgraham: agreed that it is useful w/o css. [10:41:43.0000] <jgraham> grigs: If yur point is that you have to sync the markup and the style, I agree, but that's already true to some extent [10:41:46.0000] <gsnedders> jgraham: Mark is especially hard :P [10:41:59.0000] <jgraham> gsnedders: My point exactly :) [10:43:24.0000] <grigs> /me continues arguments against the feature [10:43:52.0000] <grigs> * No support for ems which is where authors appear to be headed with media queries [10:44:55.0000] <grigs> * Seems possible/likely a better long term solution would be a different image format so we're picking a short term solution and it will become canon. [10:45:12.0000] <jgraham> grigs: You can argue it's evil but practically converting pixels->ems as a constant is good for the vast majority of users [10:45:41.0000] <jgraham> (for the rest using full page zoom might be better than increased default font size) [10:46:00.0000] <grigs> jgraham: i doubt i would argue anything in web tech is evil. ;-) [10:46:19.0000] <jgraham> I don't see how a different image format can possibly work [10:47:11.0000] <jgraham> I mean I see how it could solve the pixel density case. But I don't see how it could solve the "different assets at different sizes, only downloading those that are actually used" case [10:48:34.0000] <grigs> jgraham: i think the different assets at different sizes, what I've called art direction, may be complicating things too much. [10:51:06.0000] <jgraham> It seems hard to just ignore one of the major use cases that people have put forward [10:51:21.0000] <grigs> jgraham: I know. I was one of the people pushing it. :-) [10:52:37.0000] <grigs> i'm arguing a 180 from where I was a week and a half ago. but i'm not wedded to an outcome. just trying to make sure we get a good solution. [10:54:53.0000] <annevk> grigs: would the breakpoints not be set at viewport widths of common devices? [10:55:26.0000] <grigs> annevk: no, web designers are moving away from that quickly. [10:55:55.0000] <grigs> annevk: for one, there are no common widths http://stephanierieger.com/the-trouble-with-android/ [10:57:03.0000] <grigs> annevk: second, it makes more sense to pick breakpoints based on where the content or the design starts looking bad. basically resize the browser until the page looks like crap, write down the resolution, create a breakpoint to fix it, then resize again. [11:06:10.0000] <annevk> grigs: yeah mkay, so your proposal is to not have this at all? [11:06:43.0000] <grigs> annevk: i'm not proposing anything yet. i'm testing ideas. [11:07:17.0000] <tantek> a reasonable approach [11:10:55.0000] <annevk> I guess one could argue the alternative representations are presentational [11:10:59.0000] <gsnedders> Ran 25261 tests in 42.323s: FAILED (failures=3) [11:11:10.0000] <annevk> in which case you'd do e.g. img { content:url(smaller) } [11:11:11.0000] <gsnedders> Ran 25261 tests in 40.888s: FAILED (failures=3) [11:11:12.0000] <Ms2ger`> You FAIL [11:11:16.0000] <gsnedders> I'm calling that success! [11:11:36.0000] <annevk> thereby keeping the media queries and such in CSS too [11:11:57.0000] <MattWilcox> Just popping in - I'm with Grigs on this [11:12:08.0000] <MattWilcox> I can't see how responsive images can ever work alongside the pre-parser [11:12:16.0000] <annevk> though 'content' doesn't support the 2x stuff [11:12:47.0000] <MattWilcox> The two ideas are mutually exclusive. [11:13:04.0000] <grigs> I would state it a little differently than MattWilcox did. I hope there is a way they can work alongside each other, but am beginning to fear they can't. [11:13:31.0000] <MattWilcox> I'd love them to work alongside each other - i just don't think it's possible - I've been thinking about this for a few weeks and they just don't gel [11:13:56.0000] <MattWilcox> RespImgs would have been a simple hack to fix with nothing but JS if it wasn't for the pre-parser [11:14:25.0000] <MattWilcox> Admittedly, still a hack, but it'd have worked. Indeed *did* work for the Boston Globe - right up until pre-parsing got turned on [11:14:59.0000] <MattWilcox> Which is why I've almost come full circle and think it's back to sending headers along with requests so the server can do the adaption. [11:15:21.0000] <MattWilcox> Adaptive Images *just works* but it has to hack by using cookies instead of headers [11:15:32.0000] <MattWilcox> Anyway, I have to go - hopefully speak to you all soon. [11:15:43.0000] <grigs> lol [11:15:49.0000] <annevk> server doesn't work for CDN [11:16:03.0000] <MattWilcox> PS I mean http://adaptive-images.com specifically not the general concept [11:16:13.0000] <MattWilcox> Server can work for a CDN when it's headers [11:16:23.0000] <MattWilcox> CDNs are smart these days, not just dumb [11:16:27.0000] <MattWilcox> Anyway, have to go! [11:16:29.0000] <MattWilcox> :) [11:16:48.0000] <annevk> "smart" [11:16:58.0000] <annevk> CORS was already hard [11:17:04.0000] <annevk> asking them to do content negotiation... [11:17:15.0000] <annevk> which as a concept failed pretty badly on the web thus far [11:17:18.0000] <annevk> ... [11:17:37.0000] <grigs> annevk: to your question about alternative representations being presentational, i think the art direction use case is closest to presentational. sometimes the image changes orientation and text should reflow as well at a breakpoint [11:17:58.0000] <hober> annevk: yeah. Someone has a problem and says "I'll use conneg." Now they have N problems. [11:18:44.0000] <annevk> grigs: yeah, you could have your canonical content in markup and the adaptations in CSS [11:18:45.0000] <grigs> annevk: the nice thing about hober's original proposal was it only dealt with pixel density and thus avoided this problem. smart guy. ;-) [11:19:07.0000] <annevk> grigs: also keeps the media queries nicely isolated and for a single purpose [11:20:05.0000] <gsnedders> jgraham: So yeah, we have no tests for the meta pre-parser [11:21:11.0000] <annevk> the problem is of course that if you do that you download too much resources [11:21:13.0000] <grigs> one idea i had based on the current srcset syntax would be to try to teach people not to match the dimensions in srcset with their media queries. Instead to teach people to use some sort of “common” tiers of images that have nothing to do with design decisions. [11:21:17.0000] <annevk> so you'd prolly not use <img> at all [11:21:30.0000] <annevk> and just play with <div> and 'background'/'content' [11:21:34.0000] <gsnedders> jgraham: Fairly certain it's broken under Py3. [11:21:50.0000] <jgraham> gsnedders: That sounds like my fault :\ [11:22:09.0000] <annevk> I think people do that already as animating <img> is more expensive than <div> + background... [11:24:21.0000] <grigs> i'm not sure what the “common” ones would be or how to convince people to implement srcset that way, but it would mean that the different versions of the image were only there to give the browser options, not because of design decisions. [11:26:09.0000] <annevk> grigs: I don't really follow that idea [11:27:14.0000] <grigs> annevk: let me try again. right now when I see the dimensions listed in the srcset syntax, i automatically translate that into media query syntax and try to figure out how to match what is defined in srcset with what is defined in css. [11:27:54.0000] <grigs> annevk: i suspect that many designers have done the same (see Jeremy Keith's points on the list) and will do so in the future. [11:28:18.0000] <annevk> gotta eat brb [11:29:35.0000] <grigs> annevk: matching srcset dimensions to media query breakpoints is only necessary in the art direction use case. if we're simply dealing with the same image available at different sizes, then it would behoove the document author to avoid tying the different sizes of the image (the content) to the presentation (media query breakpoints). [11:31:26.0000] <gsnedders> /me snarls at people manually editing html5lib py3 [11:31:45.0000] <Ms2ger`> /me snarls back [11:32:16.0000] <gsnedders> jgraham: We seem to have both fixed the nose noise. [11:32:20.0000] <gsnedders> bah! [11:32:23.0000] <grigs> annevk: so instead, if authors simply picked three sizes of the image and defined three random viewport dimensions in which they should be used per the srcset syntax, then when a redesign comes along, the srcset code likely won't need to change. the srcset code wouldn't be tied to the design (as expressed in media query breakpoints) [11:35:37.0000] <WeirdAl> dglazkov: ping [11:35:57.0000] <dglazkov> WeirdAl: pong [11:36:32.0000] <WeirdAl> I sat with rniwa this weekend, and he suggested I talk with you about your HTML template spec... it's very interesting and parallels one of my own ideas [11:37:00.0000] <rniwa> dglazkov: yeah sorry i was going to tell you today in person (i'm in mtv today) [11:37:09.0000] <dglazkov> I didn't know rniwa knew Weird Al Yankovic! [11:37:12.0000] <rniwa> dglazkov: basically WeirdAl is creating an XML editor [11:37:14.0000] <WeirdAl> 8-) [11:37:22.0000] <dglazkov> can you write a song about me? [11:37:24.0000] <rniwa> dglazkov: that lets you edit templates [11:37:28.0000] <WeirdAl> <== Alex Vincent [11:37:54.0000] <rniwa> dglazkov: and use that as sort of building builds to build XML documents [11:37:59.0000] <dglazkov> From Child's Play?! That's even better! [11:38:05.0000] <WeirdAl> arrrgh [11:38:12.0000] <dglazkov> :) [11:38:41.0000] <rniwa> dglazkov: so and he uses things like <repeat> from XBL [11:38:44.0000] <rniwa> dglazkov: to accomplish this [11:39:10.0000] <WeirdAl> uh, I don't think XBL has a repeat element, unless I'm out of date (which wouldn't surprise me) [11:39:20.0000] <grigs> Wait… so WeirdAl is Alex Vincent the child actor? [11:39:26.0000] <WeirdAl> no, I am not [11:39:37.0000] <rniwa> dglazkov: it appeared that a lot of things he invented for his xml editor were very closedy related to the component model & templating [11:39:44.0000] <dglazkov> rniwa: neat! [11:40:21.0000] <WeirdAl> dglazkov: except that I have only gotten bits and pieces to work over the years, and I'm writing a fresh DOM impl in JS to nail that once and for all :p [11:40:21.0000] <grigs> WeirdAl: sorry. couldn't resist. :-) [11:40:28.0000] <WeirdAl> /me is a bit nuts [11:40:45.0000] <dglazkov> WeirdAl: you're finally home then. [11:40:53.0000] <dglazkov> welcome to #whatwg [11:41:21.0000] <WeirdAl> hehe, I've been hanging out here for a while... my repeat element's inspired by an early HTML5 draft that had repetition blocks [11:41:37.0000] <dglazkov> WeirdAl: any code I can look at? [11:41:52.0000] <WeirdAl> /me thinks [11:42:13.0000] <WeirdAl> I've got code on my sourceforge repo, but it's extremely old [11:42:23.0000] <WeirdAl> I'm not even sure if I have a working demo anymore [11:42:43.0000] <dglazkov> I am sorry, I can't fund your venture without a working demo [11:42:49.0000] <dglazkov> /me pretends to be rich [11:42:52.0000] <WeirdAl> /me chuckles [11:43:35.0000] <WeirdAl> basically, my repeat element does start out a lot like your template element, but with the added benefit of a for loop [11:44:07.0000] <dglazkov> WeirdAl, this sounds like MDV. Are you familiar with that? [11:44:19.0000] <dglazkov> http://code.google.com/p/mdv/ [11:44:27.0000] <WeirdAl> I took a glance at it on Saturday afternoon, but it was a lot to absorb at once [11:44:30.0000] <dglazkov> rafaelw_ is the spiritual leader there. [11:44:51.0000] <gsnedders> jgraham: Py3 stuff pushed [11:45:13.0000] <dglazkov> WeirdAl: where are you based? We can chat over lunch if you'd like [11:45:21.0000] <WeirdAl> I'm in the SF Bay Area [11:45:50.0000] <WeirdAl> and as I told rniwa, I'd like that very much [11:46:11.0000] <dglazkov> what a coincidence! are you in any particular part of the bay area? [11:46:25.0000] <dglazkov> /me is in Mountain View [11:46:35.0000] <WeirdAl> I live in San Leandro and work in Oakland [11:46:43.0000] <WeirdAl> but no car - public transit's my friend [11:47:50.0000] <dglazkov> whoa. Carless people. What will they think of next? [11:48:14.0000] <Ms2ger`> AI's wishing you a good morning? [11:48:14.0000] <WeirdAl> carless engineers? :) [11:48:25.0000] <tgecho> peopleless cars [11:48:59.0000] <dglazkov> I'd feel kind of bad asking you to find your way down to Mountain View, but the drive to Oakland seems also a bit daunting [11:49:01.0000] <dglazkov> ooh [11:49:06.0000] <WeirdAl> dglazkov: I have some extremely poorly documented code at http://sourceforge.net/p/verbosio/code/ci/c0049422968454a8e6869b4cb24317eccd756473/tree/experimental/floor13/chrome/res/modules/MarkupRepeatBindings.jsm?force=True [11:49:08.0000] <dglazkov> G+ Hangouts an option? :) [11:49:28.0000] <WeirdAl> /me is a heathen, he doesn't use Google+ [11:49:52.0000] <dglazkov> I thought only heathens were on G+ [11:49:58.0000] <WeirdAl> but SF is easy for both of us to get to [11:50:04.0000] <dglazkov> I can never remember [11:50:06.0000] <Ms2ger`> I would, but the Empire doesn't let me :( [11:50:38.0000] <dglazkov> sure. I should be in SF at least one day next week. [11:50:49.0000] <WeirdAl> well, Monday is Memorial Day [11:51:11.0000] <dglazkov> so, not Monday :) [11:51:33.0000] <WeirdAl> :p the only other options really are weekends [11:51:42.0000] <WeirdAl> since this project is *not* part of my day job [11:52:29.0000] <dglazkov> ah. my weekends and holidays have been donated to The Family. [11:52:43.0000] <WeirdAl> at least you have one. [11:52:44.0000] <dglazkov> that didn't come out right [11:52:53.0000] <dglazkov> :) [11:53:03.0000] <WeirdAl> no, I think that came out exactly right. Family comes first. [11:53:42.0000] <dglazkov> WeirdAl: one thing to consider: rafaelw_ is in SF. Maybe you should catch him over lunch? [11:54:06.0000] <dglazkov> WeirdAl: he is just like me, except less annoying. [11:54:54.0000] <WeirdAl> hmm [11:56:13.0000] <WeirdAl> ... I guess part of the reason I hesitate is that frankly, my project is not in a very good state right now [11:57:11.0000] <WeirdAl> that's what happens when it's a spare-time project [11:57:32.0000] <WeirdAl> with MDV, I have to admit I don't grok it [11:58:22.0000] <WeirdAl> I really think a lunch won't be enough time :) [12:00:17.0000] <tantek> grigs et al re: responsive/adaptive images - with so much conversational traffic (email, IRC), it's a bit shocking that more use-cases haven't been fleshed out - questions answered etc.: http://www.w3.org/wiki/Images [12:02:27.0000] <Wilto> tantek: Is the W3C wiki or the WHATWG wiki the most appropriate place for those? [12:02:49.0000] <grigs> tantek: the wiki didn't seem like the right place for my thoughts about the pre-parser and once that idea got into my head, well, it wouldn't go away until i wrote about it. [12:02:58.0000] <zewt> step 1 is mailing list; nobody will ever see stuff on the wiki [12:03:08.0000] <tantek> Wilto - at the time Shane Hudson volunteered to write them up, I said up to him, either is fine, and he chose the URL [12:03:26.0000] <tantek> zewt - no, step one is capture content on the wiki. step 2 is email a URL to said content. [12:03:28.0000] <dglazkov> WeirdAl: then your best bet is probably ramping up on Web Components and MDV and then condensing questions into emails [12:03:46.0000] <tantek> no one sees content on the email list after a week - it's effectively dead and unfindable [12:04:10.0000] <grigs> /me thinks we had this conversation about wikis and emails the last week. [12:04:25.0000] <tantek> grigs - pre-parser sounds like a solution approach rather than talking about use-cases to solve. [12:05:02.0000] <tantek> grigs, folks that primarily use email tend to forget the past and make the same errant statements repeatedly (see: every email list). [12:05:28.0000] <tantek> Though I'll add zewt's misconception to the wiki so it can be referenced by URL. [12:05:43.0000] <grigs> lol [12:08:25.0000] <grigs> tantek: i would describe talking about the pre-parser as understanding the technical problem, not talking about a solution approach. the technical problem is why the solutions seem disconnected from the use cases to me. understanding that is helpful. [12:08:27.0000] <zewt> no, send it to the list and stop wasting time with a wiki [12:08:45.0000] <zewt> replace the wiki with tantek's misconception that duplicating stuff on the wiki is useful [12:09:12.0000] <tantek> zewt - if your point is that duplication is bad, I'll accept that [12:09:21.0000] <grigs> /me sighs [12:09:35.0000] <Wilto> Ditto. [12:09:37.0000] <tantek> however note that using only/primarily email = duplication over and over [12:09:55.0000] <tantek> whereas if you put it on the wiki, you maybe duplicate once, and then subsequently send URLs instead. [12:10:03.0000] <tantek> = less duplication [12:10:08.0000] <zewt> it's a useless way to discuss use cases. [12:10:53.0000] <zewt> tons of use case discussion happens on the list and it works fine. every time someone goes "i'll start a wiki page!" everything gets sidetracked and people waste time. [12:13:56.0000] <Wilto> It doesn’t really seem sustainable as more and more people join in these talks, but I guess this is a little out of my wheelhouse. [12:14:17.0000] <tantek> irc and email are good for asking the question of were/are there any existing discussions of a use-case / topic / feature (since humans can still often find things better conceptually than search engines) [12:14:27.0000] <tantek> so in that regard, email/irc is a good step 1 for asking questions. [12:17:26.0000] <WeirdAl> dglazkov: wrt MDV's bindings, have you guys looked at Harmony proxies for that? [12:17:45.0000] <grigs> Oh hey everyone, in case you’re wondering, I’m apparently at odds with you: http://www.webmonkey.com/2012/05/browsers-at-odds-with-web-developers-over-adaptive-images/ [12:17:58.0000] <grigs> /me shakes his head. [12:18:51.0000] <zewt> snr in this topic is miserable, heh [12:18:57.0000] <tantek> Wilto - you're exactly right, email doesn't scale. [12:19:15.0000] <dglazkov> WeirdAl: yep. You really need to talk to rafaelw_ -- he wrote it [12:19:25.0000] <WeirdAl> /me sighs [12:19:53.0000] <tantek> fewer and fewer people have the time to keep up with so many conceptual diffs - so the conversation typically devolves into taking sides / opinions, rather than reasoned discussion based on established assumptions (use-cases). [12:19:53.0000] <Wilto> BREAKING NEWS: grigs wants to physically fight everyone in #whatwg; news at 11:00. [12:20:15.0000] <Wilto> (I call first.) [12:20:34.0000] <miketaylr> tabloids gonna tabloid [12:20:45.0000] <grigs> Wilto: i can take you little man. it's the rest of the room i worry about. ;-) [12:21:10.0000] <Wilto> grigs: You at me, at the flagpole, after school. [12:21:22.0000] <WeirdAl> /me looks for his .50 cal [12:21:47.0000] <grigs> ok, WeirdAl both just lived up to his namesake and won this fight. [12:22:18.0000] <WeirdAl> you haven't been paintballed until you've been fifty-cal paintballed. [12:22:27.0000] <Wilto> I _am_ deathly allergic to bullets. [12:22:38.0000] <tantek> zewt - evidence that disproves your assertion that "every time someone goes "i'll start a wiki page!" everything gets sidetracked and people waste time." - the Time element enhancements were primarily developed on the wiki, grown quite effectively over time, and used as the source for decision making. http://wiki.whatwg.org/wiki/Time_element [12:22:41.0000] <tantek> Email on the issues were minimal. There was some IRC, and some in-person discussion at TPAC. [12:23:56.0000] <zewt> you're making up a problem that doesn't exist and telling whatwg to change how it works to solve it. sorry; no thanks [12:25:40.0000] <zewt> (and since there isn't actually a problem to solve--email works fine--I don't feel like dedicating much more time debating this at the moment) [12:26:27.0000] <Wilto> Wow. [12:27:37.0000] <tantek> zewt - your behavior seems to contradict your statements. you say "don't feel like dedicating much more time debating this", yet you're the one that brought up the debate with your statement "step 1 is mailing list; nobody will ever see stuff on the wiki". [12:27:43.0000] <grigs> Found it. The url to the last time this channel discussed the merits of email and wikis. http://krijnhoetmer.nl/irc-logs/whatwg/20120515#l-2223 [12:28:32.0000] <zewt> i'm telling people to keep doing things the way they're normally done on whatwg. that's not debating anything; it's just saying how things are normally done. [12:28:56.0000] <grigs> /me thinks it must be lunch time somewhere. [12:29:04.0000] <tantek> zewt - the example of the Time element enhancement experience proved that it doesn't matter if a group works primarily by email, it's *still* more efficient to document research on a wiki and then use email only for brief descriptions with links to the wiki documentation. [12:29:27.0000] <zewt> it proves nothing of the kind, but I'm not wasting more time on this today. [12:29:57.0000] <tantek> zewt, perhaps when you have time for more than simple contradiction. [12:30:08.0000] <jgraham> Oh look, in all this metacrap, I almost missed the interesting stuff [12:30:17.0000] <zewt> now you're just being childish (even more reason for me to do something else) [12:30:44.0000] <zewt> jgraham: should we start bolding the interesting stuff? :P [12:30:48.0000] <tantek> zewt, your statement "proves nothing of the kind" is simple contradiction, not an actual argument. [12:31:24.0000] <zewt> it's a statement of fact, which is so straightforward it needs no further explanation. [12:31:43.0000] <jgraham> gsnedders: So, what state is the stuff you pushed in? [12:31:49.0000] <tantek> your statement about "not wasting more time on this" implied you didn't have time for anything more than your previous statement which was merely a simple contradiction. my statement was merely a summary. there was nothing in my statement that indicated any particular age-bias (e.g. "childish"). [12:32:17.0000] <zewt> seems to me that you're just setting out to waste my time, so I'm not replying any further [12:32:44.0000] <tantek> zewt - by raising the debate, you alone are responsible for wasting your time. [12:36:59.0000] <Ms2ger`> /me sighs [12:37:37.0000] <zewt> /me ROTT god mode yawns [12:40:51.0000] <gsnedders> jgraham: Passes equal number of tests on both Py2 and Py3, excluding a few tests that I think are wrong. [12:41:43.0000] <jgraham> gsnedders: Do you think it's more broken on py3? [12:41:59.0000] <gsnedders> jgraham: I think encoding detection is. [12:44:35.0000] <jgraham> OK. You planning to fix that? [12:45:05.0000] <jgraham> (when you say encoding detection, do you mean meta prescan? Or the other stuff?) [12:45:24.0000] <gsnedders> (The larger thing, inc. meta prescan) [12:45:29.0000] <gsnedders> And BOM detection, etc. [12:46:28.0000] <jgraham> So... is all the python 3 code autogenerated from python 2? [12:46:34.0000] <gsnedders> Yes. [12:47:03.0000] <jgraham> Can we make sure that is documented? :) [12:47:24.0000] <gsnedders> If you're using the royal we ;) [12:47:37.0000] <gsnedders> I'll fix encoding detection when someone gives me tests :) [12:47:40.0000] <jgraham> Not quite :p [12:47:50.0000] <jgraham> Also, can we fix the prescan stuff and still work entirely using 2to3? [12:48:04.0000] <gsnedders> Yes. [12:48:16.0000] <jgraham> What semantics are you assuming for the encoding detection? [12:48:32.0000] <gsnedders> It'll be even more code that'll force us to be Py2.6 only, but it should work fine. [12:48:33.0000] <jgraham> It only gets used if you pass a bytes object? [12:48:50.0000] <gsnedders> I don't know. Whatever we do now. [12:49:17.0000] <jgraham> What we do now probably doesn't make sense [12:49:23.0000] <gsnedders> If we're given a Unicode object, we have encoding ("utf-8", "certain"), having re-encoded it. [12:49:26.0000] <gsnedders> Oh, it certainly doesn't. [12:49:47.0000] <gsnedders> We should probably split stuff out so we have a separate class for the Unicode case. [12:49:57.0000] <gsnedders> And work out how to handle bytes/unicode files. [12:51:40.0000] <jgraham> Yes [12:51:55.0000] <jgraham> I thought I had looked at this when I did the original hand port [12:52:00.0000] <jgraham> But maybe npt [12:52:07.0000] <jgraham> I thought about it at least [12:52:11.0000] <gsnedders> Really I was just doing whatever got us working on Py3 quickest. [12:52:53.0000] <gsnedders> Once I think everything works on Py3, should then move to trying 3to2. [12:54:21.0000] <jgraham> Well we might as well do it more or less right [12:55:07.0000] <jgraham> Anyway, this looks pretty nice [12:55:09.0000] <jgraham> Thanks! [12:55:45.0000] <gsnedders> I mean, I agree we *should* do stuff right. [12:55:55.0000] <gsnedders> But that's more work, and work for another time. [12:56:16.0000] <gsnedders> If I was going to have to rewrite large amounts of stuff, then sure, it would've made sense to do it right, but I didn't, [13:16:37.0000] <tantek> Wilto, it would be good to work on aggregating the use cases on one wiki page - do you have a preference? [13:26:40.0000] <Wilto> tantek: I mean, I’m flexible: whichever is going to be the best received. I’ve got a WHATWG wiki account already, for what that’s worth. [13:27:44.0000] <tantek> Wilto, cool. I'm flexible too - have developed pages in both places. Shane Hudson expressed a preference for the w3.org wiki so I've been most recently contributing there for the responsive image use-cases. [13:28:37.0000] <tantek> So I guess step one - could you take a look at http://www.w3.org/wiki/Images#Responsive_Images_Use_Cases and see if all the use cases you have are covered? [13:28:41.0000] <tantek> (and if not, add more?) [13:29:12.0000] <Wilto> Sure thing. [13:33:34.0000] <Wilto> It doesn’t sound like it’s going to be… especially well-received, is all. [13:34:53.0000] <Wilto> But, “be the change you want to see,” and all. [13:36:37.0000] <tantek> Wilto - nevermind those discouraging adding to wikis. regardless of opinions, wiki pages provide URLs you, I, and anyone else can use to provide citations. [13:36:48.0000] <tantek> (citations in communications, rather than duplicating text) [13:38:27.0000] <Wilto> Sounds good to me. [15:37:21.0000] <gsnedders> Ah, benchmarks. Aren't you wonderful? "ms (slower is better)" [16:28:44.0000] <tantek> and in this week's episode of needlessly reinvented standards, we bring you rel="logo" http://relogo.org/ - subhead: what was wrong with rel="icon" ? [16:29:36.0000] <tantek> or is this perhaps the result of not understanding that you can have multiple rel="icon" links with different mime/content types including SVG? 2012-05-23 [17:25:16.0000] <tantek> yikes - rel-logo was apparently added to the rel-registry 2011-11-24. snuck-in on last year's US Thanksgiving holiday. [17:36:57.0000] <tantek> I've written up the reasons why rel=logo is unnecessary and moved it to the "rejected" table in existing-rel-values: http://microformats.org/wiki/existing-rel-values#rejected [17:37:17.0000] <tantek> interesting that the notion of linking to a logo has been *twice* reinvented (pavatar, logo) [18:38:34.0000] <Druide_> lol [18:38:34.0000] <Druide_> [02:37] <josiah> What is the best way to generate a secret for a SF2 application? [18:38:34.0000] <Druide_> [02:39] <wiistriker> drop face on keyboard [20:31:09.0000] <MikeSmith> roc: you around? I wanted to ask you about something related to the HTML WG DRM video stuff [20:31:18.0000] <roc> yes [20:34:51.0000] <MikeSmith> roc: the chairs are planning to have a weekly telcon to discuss the DRM proposal and the other for enabling adaptive-streaming use cases [20:34:52.0000] <MikeSmith> what I wanted to know is, would you be interested in calling in [20:35:37.0000] <roc> my immediate reaction is "urk" [20:35:38.0000] <MikeSmith> if so, there's a survey about the call time which I hope can respond to [20:36:29.0000] <roc> I don't feel like I have anything to contribute beyond the policy issues we've already discussed on the list [20:36:31.0000] <MikeSmith> because otherwise, the call time is likely going to be pretty bad for NZ and Australia too [20:36:31.0000] <MikeSmith> heh [20:36:31.0000] <MikeSmith> yeah [20:36:31.0000] <MikeSmith> well, the survey is here: https://www.w3.org/2002/09/wbs/1/html-media/ [20:36:32.0000] <MikeSmith> in the current results, Tuesday, 1500 UTC is the time that's leading [20:36:34.0000] <MikeSmith> OK [20:37:36.0000] <roc> it's basically impossible to schedule a telecon that works for both Europe and me [20:38:21.0000] <roc> so if I don't absolutely need to be on it, better that I'm not [20:38:37.0000] <MikeSmith> I see [20:40:01.0000] <MikeSmith> 2100 UTC would work if Europe participants were willing to call in at 11pm [20:40:04.0000] <roc> (it is possible, but it's never fun) [20:40:05.0000] <MikeSmith> the thing is, what always ends up happening if we do have a telcon is that the New Zealand and Australia and Asia people always get the short end of the stick as far as call times [20:40:11.0000] <roc> yep [20:40:12.0000] <MikeSmith> yeah, I hear you [20:40:38.0000] <roc> small price to pay for the privilege of living here, I say :-) [20:40:49.0000] <MikeSmith> indeed [20:41:02.0000] <MikeSmith> I'd love to visit there some time [20:41:33.0000] <roc> if people wonder whether some specific proposal would meet with my approval, of course I can answer that via email or even dial in to a particular meeting if necessary [20:42:44.0000] <MikeSmith> OK [20:44:39.0000] <MikeSmith> I'll make an effort personally to give you a heads-up if there's anything on the calls that seems worth alerting you about [20:45:24.0000] <MikeSmith> also, I guess the call time is likely to end up amenable to hsivonen being able to call in -- though I'm not sure he'll want to [20:46:42.0000] <MikeSmith> I know I don't want to.. [20:48:56.0000] <roc> heh [23:20:31.0000] <zcorpan> https://www.w3.org/Bugs/Public/show_bug.cgi?id=17151 isn't this something for the encoding standard? [23:29:54.0000] <othermaciej> zcorpan: maybe - which spec defines other types of UTF-16 encoding errors, such as unpaired surrogates, nulls, etc? [23:30:12.0000] <othermaciej> oh, this is the case of unpaired surrogate [23:36:38.0000] <zcorpan> seems the encoding standard already does "emit a decoder error" for lone surrogates, which for HTML means U+FFFD [23:36:52.0000] <zcorpan> (and for XML means fatal error) [23:53:06.0000] <zcorpan> TabAtkins: should css3-syntax U+FFFD surrogates in escapes? [00:24:23.0000] <annevk> https://twitter.com/awbjs/status/204962858113564672 didn't [tm] do that [00:24:26.0000] <annevk> ? [00:24:48.0000] <annevk> also, cool how twitter now lets you copy permanent URLs and they don't include the hash bang nonsense [00:26:44.0000] <zcorpan> nice [00:28:14.0000] <annevk> so the differences is that it's a normative copy http://ecma-international.org/ecma-262/5.1/ [00:28:35.0000] <smaug____> rafaelw_: since I don't know how to update chromium on linux, can't test this right now, but does webkit handle mutationobserver correctly when using documentfragments? [00:28:38.0000] <annevk> but apparently has less information on bug reports and such when compared to http://es5.github.com/ which seems logical [00:29:14.0000] <annevk> oh "The PDF version is the definitive specification." [00:29:16.0000] <annevk> meh [00:30:28.0000] <zcorpan> The D in PDF stands for Definitive [00:32:50.0000] <smaug____> interesting, I got chromium into a state where it doesn't load anything [00:34:01.0000] <annevk> http://w3cmemes.tumblr.com/post/23559522267/http-w3cmemes-tumblr-com how is this foul bachelor frog? [00:34:25.0000] <othermaciej> doesn't make sense to me either [00:34:52.0000] <othermaciej> maybe Lazy College Senior but even that does quite fit [00:35:06.0000] <othermaciej> maybe even a Good Guy Greg [00:35:22.0000] <annevk> yeah, GGG would be it I think [00:35:32.0000] <annevk> who doesn't want a more readable spec [00:46:36.0000] <othermaciej> I guess there is implied laziness to not trying harder to understand it [00:46:51.0000] <othermaciej> depends on if the rewrite is easier to understand for everyone, or only the author [00:48:29.0000] <annevk> true [01:00:31.0000] <annevk> via rniwa http://discuss.joelonsoftware.com/default.asp?joel.3.219431.12 [01:00:38.0000] <annevk> pretty excellent [01:00:51.0000] <rniwa> annevk: yeah. it's via othermaciej [01:01:08.0000] <annevk> :) [01:02:22.0000] <othermaciej> that pieces is pretty well known, I can't take much credit for knowing about it [01:02:24.0000] <othermaciej> but it is funny [01:02:32.0000] <othermaciej> and explains why DOM2/3 sucks as bad as it does [01:02:41.0000] <rniwa> annevk: and then othermaciej posted http://www.facebook.com/l.php?u=http%3A%2F%2Fcode.google.com%2Fp%2Faxis-ssl%2Fsource%2Fbrowse%2Ftrunk%2Fsrc%2Forg%2Fapache%2Faxis%2Fcomponents%2Fnet%2FSocketFactoryFactory.java&h=VAQGMbw_NAQEG5nvKnT2Shq9fjmmeHTHVR7k9nQHoV4gjRA&enc=AZP-XNPofm9WLTtopsES2fGSbWkXS_7jVMv6HKWGt2J3fiOJmWxlBjPCaK_-IvXSrcOVvK1DdqgYPA3peVIltXdEVoL2ATD4W8hLzQfhe1E5og [01:02:52.0000] <rniwa> annevk: ugh... i mean http://code.google.com/p/axis-ssl/source/browse/trunk/src/org/apache/axis/components/net/SocketFactoryFactory.java [01:03:07.0000] <annevk> ah thanks [01:03:17.0000] <othermaciej> if you do a Google search for Java FactoryFactory you will find lots of real cases of factory factory classes [01:03:22.0000] <annevk> I'm afraid of following facebook.com URLs in case they never heard about GET not having side effects [01:04:22.0000] <annevk> I'm glad we don't have factories [01:04:24.0000] <heycam> the "though you may also have to deploy a few configuration files to make it all work" is the funniest part [01:05:09.0000] <annevk> the follow up is great too [01:05:13.0000] <annevk> "Luckily, 70% of the workers in the Tool-Oriented Metafactory Union are certified against this version of the spec." [01:06:10.0000] <heycam> that is why I write all my web apps in bash [01:43:59.0000] <annevk> kind of sad even new drafts such as http://people.mozilla.org/~bbirtles/web-animations/web-animations.html format everything per the old DOM standards... [01:44:31.0000] <annevk> and thereby define things in a crappy way, e.g. http://people.mozilla.org/~bbirtles/web-animations/web-animations.html#widl-TimedTemplate-animate-TimedItem-Element-target-float-startTime [01:45:28.0000] <annevk> oh, it doesn't even use RFC 2119 [01:45:34.0000] <annevk> *sigh* [01:48:25.0000] <othermaciej> hmmm [01:48:31.0000] <othermaciej> it has "must" in notes [01:48:42.0000] <othermaciej> "must" in at least one conformance requirement [01:48:55.0000] <othermaciej> and then a colored small-caps MUST in another place [01:48:56.0000] <jgraham> Tell this to the people writing the spec, perhaps :) [01:49:08.0000] <othermaciej> so it's actually much worse than not using RFC2119 [01:50:03.0000] <othermaciej> so far, the proposed API seems kinda bad too, not just the spec [01:50:13.0000] <othermaciej> so not yet worth investing energy in [01:50:14.0000] <zcorpan> do we have a wiki with guidelines about how to write conformance requirements correctly and how to format specs in a good way? [01:50:24.0000] <zcorpan> s//page/ [01:50:54.0000] <zcorpan> it seems this comes up time and time again with "new" editors [01:51:14.0000] <zcorpan> having new editors is awesome; having specs that suck is not :-) [01:51:35.0000] <othermaciej> this spec has at least a couple of non-new editors [01:51:51.0000] <othermaciej> http://wiki.whatwg.org/wiki/How_to_write_a_spec [01:52:23.0000] <jgraham> We have http://ln.hixie.ch/?start=1140242962&count=1 ofc [01:53:42.0000] <othermaciej> oh, that one is also good, too bad it's not in that whatwg wiki page [01:55:25.0000] <annevk> I always wonder why people haven't read existing specs... [01:55:55.0000] <annevk> e.g. if you have read HTML, XMLHttpRequest, DOM, ... you should know the problems [01:56:22.0000] <annevk> but I guess my expectations don't meet reality, so we should indeed have some kind of wiki or document explaining that [01:56:45.0000] <annevk> and tell everyone "you're awesome, but take a look at <URL>" [01:57:27.0000] <zcorpan> /me expanded a bit on rfc2119 usage [01:58:29.0000] <smaug____> annevk: "haven't read existing specs"? well, depends on which specs [01:58:34.0000] <smaug____> there are tons of different specs [01:58:48.0000] <smaug____> HTML, XHR, DOM use similar style [01:59:02.0000] <smaug____> and have their problems [01:59:14.0000] <annevk> smaug____: everyone writing APIs should be at least familiar with DOM/HTML/IDL imo [01:59:25.0000] <smaug____> DOM/IDL yes, [01:59:27.0000] <smaug____> and some HTML [01:59:43.0000] <smaug____> reading and understanding full HTML... [01:59:50.0000] <smaug____> no one does that :) [01:59:56.0000] <othermaciej> annevk: are you expecting people to just be generally familiar with those technologies, or to understand in depth how they are specced, why they are specced that way, what problems you could run into building on their specs, etc? [02:01:00.0000] <annevk> othermaciej: I expect them to see that the way they define methods is radically different (and has holes); I expect them to look for patterns in existing methods with respect to exception handling, event dispatching, etc. [02:01:18.0000] <annevk> othermaciej: my expectations however seem unrealistic [02:01:52.0000] <annevk> e.g. quite some specs that clearly require use of the event loop don't use it [02:02:08.0000] <annevk> (some started using it after it being pointed out) [02:02:26.0000] <zcorpan> http://wiki.whatwg.org/wiki/How_to_write_a_spec doesn't mention the event loop [02:03:03.0000] <othermaciej> many people writing web platform specs are not expects on other web platform specs or necessarily even the other relevant technologies at all [02:10:10.0000] <roc> or in some cases, not even familiar with what a spec is [02:11:08.0000] <zcorpan> maybe the respec tool should have a "read this before use" sticker [02:12:25.0000] <roc> annevk: email your gripes to Brian, I'm sure he'll take them on board [02:13:14.0000] <annevk> zcorpan: I guess we should expand that wiki page with "Patterns" of some kind; examples of how to do various things [02:13:35.0000] <annevk> zcorpan: including dispatching events asynchronously [02:14:11.0000] <zcorpan> annevk: sounds like a good idea [02:14:30.0000] <zcorpan> annevk: also explicitly say that the old DOM specs is a bad pattern and why it is [02:15:54.0000] <zcorpan> annevk: it could point to real-world examples in actual specs and discuss what it does [02:42:39.0000] <roc> hmm, why is the From-Origin spec stalled? [02:43:20.0000] <annevk> roc: I updated it yesterday to include a reason [02:44:17.0000] <annevk> roc: it doesn't mean much though, if people want to implement it, I'll keep updating it and we'll go back to REC-track [02:44:36.0000] <roc> is the reason here? http://www.w3.org/TR/from-origin/ I can't see it [02:44:46.0000] <annevk> roc: http://dvcs.w3.org/hg/from-origin/raw-file/tip/Overview.html [02:45:08.0000] <annevk> roc: TR/ is never up-to-date [02:45:41.0000] <roc> sorry, I couldn't find the real version :-) [02:45:58.0000] <roc> I wonder if I can teach google to simply not show me TRs [02:46:14.0000] <annevk> TR/ will be updated next week with http://dvcs.w3.org/hg/from-origin/raw-file/tip/TR.html fwiw [02:46:26.0000] <annevk> I created that yesterday too per request from Art [02:49:17.0000] <othermaciej> are canvas gradients new? [02:49:48.0000] <othermaciej> I have a hard time remembering which are the new bits [02:50:02.0000] <annevk> they're not new [02:50:05.0000] <annevk> see http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2012-March/035239.html for what's new [02:51:30.0000] <othermaciej> thanks! [02:52:24.0000] <jgraham> So, when's Hixie back? (I assume he is away) [02:55:44.0000] <othermaciej> he's on vacation, I forgot when back [02:57:10.0000] <jgraham> It is curiosuly unsettling to be trapping ina buggy part of the spec without the Dear Leader around for guidance [02:57:32.0000] <zcorpan> jgraham: about <script src>, does it await a stable state before fetching src (when setting src), in case the script also sets crossorigin? (like <img>) [02:58:30.0000] <zcorpan> the spec doesn't have crossorigin on script yet, it seems [02:58:45.0000] <annevk> jgraham: would be more fun if Hixie took a three-month break :) [02:59:43.0000] <zcorpan> OMG WE NEED A NEW EDITOR THAT DOESN'T DO VACATIONS [02:59:58.0000] <annevk> jgraham: I guess we should count our blessings, it used to be so much worse: http://ln.hixie.ch/?start=1172653243&count=1 [03:00:43.0000] <jgraham> zcorpan: We should jsut be glad he relocated to the US where no one does vacations (almost :p) [03:01:48.0000] <annevk> so my non-n network seems to perform better outside [03:02:00.0000] <jgraham> annevk: I guess with a three month break you would quickly adjust your expectations [03:02:11.0000] <jgraham> s/you/one/ ^c. [03:15:47.0000] <annevk> in wiki markup, what's the best way to markup some kind of example? [03:19:27.0000] <gsnedders> annevk: Woah, that was five years ago. [03:19:48.0000] <othermaciej> is Selectors API level 2 still being worked on? [03:19:54.0000] <othermaciej> last draft looks old [03:19:57.0000] <zcorpan> annevk: <blockquote>? :-) [03:21:49.0000] <zcorpan> or <div class="..."> maybe; are there any classes with default styling? [03:22:49.0000] <annevk> how does http://wiki.whatwg.org/wiki/How_to_write_a_spec#Patterns look? [03:23:03.0000] <annevk> gsnedders: you're getting old man [03:23:32.0000] <annevk> the Patterns I wrote down now is a start btw, I plan on adding the exception stuff and such [03:24:04.0000] <zcorpan> <</var> [03:24:39.0000] <annevk> thanks, though not quite the feedback I was looking for [03:24:53.0000] <zcorpan> heh :-) looks good [03:25:30.0000] <zcorpan> maybe bake in id="" pattern recommendation [03:28:43.0000] <smaug____> does anyone know where to download chromium for linux [03:28:51.0000] <smaug____> the latest snapshot [03:28:57.0000] <smaug____> (whatever it is called) [03:29:27.0000] <annevk> smaug____: http://www.chromium.org/getting-involved/dev-channel#TOC-Linux [03:30:42.0000] <smaug____> /me was hoping .tar.* packages [03:31:09.0000] <smaug____> or .zip [03:35:16.0000] <MikeSmith> david_carlisle: I just pushed the rest of the MathML3 updates to http://validator.w3.org/nu/ [03:35:52.0000] <david_carlisle> MikeSmith: ohh thanks will try [03:36:24.0000] <MikeSmith> including that openmath fix [03:36:41.0000] <MikeSmith> but really I would rather we just rip out all of the openmath stuff altogether [03:36:51.0000] <david_carlisle> yo! one valid mathml spec:-) (could you chnage the identification string on the schema Using the schema for HTML5 + SVG 1.1 + MathML 2.0 + to say 3.0 [03:36:59.0000] <MikeSmith> ah yeah [03:37:04.0000] <MikeSmith> will change that now [03:37:35.0000] <annevk> is the validator not supporting post-HTML5 things? [03:37:37.0000] <MikeSmith> hmm, I wonder of hsivonen wants it specified or not [03:37:56.0000] <annevk> it seems better to have "valid HTML/SVG/MathML" or just "this content is valid" [03:38:02.0000] <MikeSmith> yeah [03:38:09.0000] <MikeSmith> that's what I meant [03:38:21.0000] <MikeSmith> I think hsivonen might now prefer that we don't specify the versions at all [03:39:09.0000] <david_carlisle> MikeSmith: Really I would rather there were not any openmath special cases in the schema. I'd prefer that the attribute value be open as it says in the spec, but if you/heri think that's too risky and just want to allow mathml/svg/html in teh annotations I'll just grumble and not inline the openmath bits in that version of the spec [03:39:17.0000] <MikeSmith> but in the case of SVG, we have it that way to make sure the people realize we're not supporting SVG 1.2 [03:39:31.0000] <smaug____> hmm, am I confident enough that MutationObserver is a good API and that the implementation is good enough for un-prefixing... [03:39:41.0000] <david_carlisle> apply usual typo correction, especially to Henri's name [03:40:16.0000] <MikeSmith> david_carlisle: yeah, I think it's better not to make it open and I think Henri still thinks so as well. We don't leave anything open anywhere else like that. [03:40:32.0000] <annevk> smaug____: no, but if you don't start unprefixing we'll never get rid of the much worse thing [03:41:19.0000] <smaug____> true [03:41:52.0000] <david_carlisle> MikeSmith: I think it;'s a mistake but I'm not going to argue just grumble:-) but as I said either way it shouldn't special case openmath [03:42:41.0000] <MikeSmith> OK [03:43:12.0000] <david_carlisle> because people who want to use ChemML (say) should be able to look what I did for OpenMath and I don't want them to find that what you have to do is mail you and Henri and ask to be whitelisted [03:46:37.0000] <MikeSmith> david_carlisle: ah yeah [03:46:43.0000] <MikeSmith> so I will yank the openmath stuff [03:48:23.0000] <david_carlisle> MikeSmith: Thanks for all the work on this, appreciated. [03:48:32.0000] <MikeSmith> no problem [03:48:37.0000] <MikeSmith> sorry for taking so long [04:41:37.0000] <annevk> rename proposal "Howto_spec" [04:42:01.0000] <annevk> reasoning: easier to remember and search for [04:50:11.0000] <zcorpan> what does a Test Facilitator do? [04:50:50.0000] <annevk> I think the idea is that similar to an editor, it's the person responsible for making sure the test suite is adequate [04:51:03.0000] <annevk> it's kind of lame name though [04:51:13.0000] <annevk> but then so is "editor" [04:51:58.0000] <zcorpan> Test Suite Adequate Making Surer [04:52:11.0000] <zcorpan> +Is [04:52:51.0000] <gsnedders> And dealing with test case approval etc. [05:01:17.0000] <annevk> http://wiki.whatwg.org/wiki/Howto_spec#Patterns added attributes [05:09:08.0000] <jgraham> The way that gecko implements alert is funny [05:09:35.0000] <jgraham> I particularly like the way you can make the background get darker and darker as you layer alerts one on top of the other [05:09:45.0000] <annevk> is "Legacy DOM" a good name for the wiki page describing the problems with respec's default output? [05:10:17.0000] <annevk> or should it be Legacy DOM-style which is kind of annoying to remember and has a dash [05:13:22.0000] <annevk> mah I'll just make it a separate heading at the end of "howto spec" [05:21:39.0000] <annevk> http://wiki.whatwg.org/wiki/Howto_spec#Legacy_DOM-style [05:44:17.0000] <matjas> “ES5 claims the global scope "this" is the same as the global object, which is not always true in HTML5.” — http://wiki.whatwg.org/wiki/Web_ECMAScript#Global_scope [05:44:22.0000] <matjas> can haz more info on this? ^ [05:45:00.0000] <annevk> see Window versus WindowProxy [05:49:34.0000] <zcorpan> annevk: awesome [05:51:56.0000] <matjas> /me tries to understand [05:52:08.0000] <matjas> hmm, so when is global this != window? [05:52:46.0000] <annevk> matjas: prolly when navigating [05:53:32.0000] <matjas> heh [05:53:48.0000] <annevk> ah yeah, the spec says so to [05:53:51.0000] <annevk> o [05:54:25.0000] <annevk> yeah so each Document has a Window [05:54:50.0000] <annevk> but each browsing context (e.g. <iframe>) has a WindowProxy [05:56:17.0000] <jgraham> zcorpan: So, do you want to be Web Workers test facilitator? [05:56:58.0000] <zcorpan> jgraham: i guess i could [05:57:37.0000] <jgraham> Excellent [05:57:49.0000] <jgraham> I think that makes more sense than me doing it :) [05:58:19.0000] <zcorpan> i just need to find that email again so i can reply to it [05:59:01.0000] <annevk> if you do work you can break the rules [05:59:29.0000] <annevk> I mean, create a new thread :) [06:00:10.0000] <zcorpan> ""ext James Graham" <jgraham⊙oc>"? [06:00:28.0000] <jgraham> ?? [06:00:57.0000] <annevk> some email software inserts "ext" [06:01:03.0000] <annevk> prolly an enterprise feature [06:01:08.0000] <annevk> sorry, "feature" [06:01:11.0000] <zcorpan> i'm as confused as you are. the email i'm replying to has "James Graham" in the To field, but when i was going to reply to it, it became that [06:01:39.0000] <jgraham> Maybe M2 became an enterprise mail client :) [06:01:52.0000] <annevk> or your stored contacts have been garbled by someone else [06:02:35.0000] <zcorpan> oh wait, the email i'm replying to actually had To: "ext James Graham" <jgraham⊙oc> [06:02:53.0000] <annevk> from Art? [06:02:59.0000] <zcorpan> it's just opera's UI uses the contact name rather than what the email header said [06:03:03.0000] <zcorpan> yeah [06:09:34.0000] <Ms2ger> Yeah, Art's got lots of "ext"s [06:12:02.0000] <zcorpan> should <label><a href=x> be allowed? it seems to be allowed in the spec currently [06:52:19.0000] <matjas> annevk: ok, so it’s still safe to rely on global `this === window`, as it will always be the case; be it through a Window or a WindowProxy [06:52:25.0000] <matjas> phew [06:53:40.0000] <zcorpan> this === windows is always true, but window and this don't always refer to the ES global object (instead they both refer to the WindowProxy) [06:55:24.0000] <gsnedders> zcorpan: When do they refer to different things? [06:56:38.0000] <annevk> gsnedders: navigation [06:57:23.0000] <gsnedders> But when do you have a global existing after a navigation? [06:58:08.0000] <annevk> not an ES global, but you can still have a reference to a WindowProxy object [06:58:20.0000] <annevk> it just proxies to a different ES global [06:58:51.0000] <zcorpan> say you have an iframe, the parent saves a reference to its window, then it navigates [06:59:19.0000] <smaug____> hmm, is this===window when doing document.open() [06:59:21.0000] <smaug____> /me tests [07:00:15.0000] <zcorpan> smaug____: document.open() should reuse the Window per spec, IIRC [07:00:29.0000] <Ms2ger> Parsing scientific notation in CSS is still undefined? [07:00:42.0000] <smaug____> zcorpan: really [07:00:43.0000] <smaug____> hmm [07:00:49.0000] <zcorpan> Ms2ger: see css3-syntax [07:01:00.0000] <zcorpan> /me has gotta go [07:01:15.0000] <Ms2ger> I don't want to [07:02:33.0000] <smaug____> window===this is true even in the case I was thinking [07:02:40.0000] <smaug____> ofc [07:02:47.0000] <smaug____> because of outer window [07:02:49.0000] <smaug____> silly me [07:26:14.0000] <hsivonen> webben: You can follow along the Maven stuff in http://bugzilla.validator.nu/show_bug.cgi?id=921 . I landed the OSGi stuff from http://bugzilla.validator.nu/show_bug.cgi?id=919 . [08:27:20.0000] <webben> hsivonen: ta [08:39:04.0000] <jgraham> Oh look! A new mailing list! [08:41:07.0000] <Ms2ger> Needs ski instructor meme [08:42:06.0000] <dglazkov> good morning, Whatwg! [08:42:25.0000] <Ms2ger> Good, you [08:47:29.0000] <jgraham> Ms2ger: +God + again? [08:47:47.0000] <Ms2ger> No comment [09:09:03.0000] <Ms2ger> dglazkov, sure, as long as we decide in favour of Anolis :) [09:22:57.0000] <jgraham> I propose that we settle it by a fight to the death. rberjon vs gsnedders. [09:23:14.0000] <jgraham> Although rberjon would win [09:23:21.0000] <jgraham> So Ms2ger would be unhappy [09:23:39.0000] <Ms2ger> OTOH, no more gsnedders... [09:24:08.0000] <Ms2ger> That would be doubleplusungood [09:28:21.0000] <gsnedders> /me wonders what he's done around #whatwg for, well, years [09:28:31.0000] <Ms2ger> py3! [09:28:31.0000] <Workshiva> Talk [09:28:43.0000] <gsnedders> Ms2ger: Oh yeah. [09:28:53.0000] <Ms2ger> And accept one of fantasai's patches after like three years :) [09:29:48.0000] <jgraham> dglazkov: Anyway afaict the tradeoff between respec and anolis is that people prefer writing specs using respec, presumably due to the lack of a compile step, but it encourages a worse style of spec writing and has a worse user experience (due to scripts running when the document loads) [09:30:50.0000] <gsnedders> jgraham: You could always set up respec to build for each commit [09:30:59.0000] <gsnedders> Instead of only doing it for TR [09:31:32.0000] <jgraham> gsnedders: Then it would have a compile step [09:31:53.0000] <gsnedders> jgraham: But eliminating that for actual work before you commit is probably the big gain anyway. [09:32:07.0000] <gsnedders> You can automate it through commit hooks to get all the gain for no effort. [09:33:44.0000] <jgraham> gsnedders: For what? [10:18:22.0000] <TabAtkins> Man, zcorpan always disappears just as I come in for the day. It's hard to remember to continue a discussion in the evening with him. :/ [10:19:03.0000] <Ms2ger> Get up earlier :) [10:19:17.0000] <TabAtkins> Getting to work at 8am should be early enough for anybody. [10:19:29.0000] <Ms2ger> Relocate [10:19:39.0000] <TabAtkins> Already did. It made the problem worse. [10:19:51.0000] <Ms2ger> Relocate in the right direction [10:20:00.0000] <TabAtkins> Man, nothing makes you happy. [10:20:11.0000] <Ms2ger> No [10:23:51.0000] <jwalden> TabAtkins: you must be new here :-) [10:28:53.0000] <nights> TabAtkins: lots of your demos doesnt work. [10:29:01.0000] <nights> and if they do work they only work in Chrome... [10:29:31.0000] <nights> the live image processing is cool though [10:35:14.0000] <TabAtkins> nights: It's very possible that they might require Chrome, because I was lazy and authored them solely for my own use in a presentation, where I know I'll be using Chrome. [10:41:26.0000] <nights> I see [10:41:29.0000] <nights> :) [10:41:45.0000] <TabAtkins> All of the demos on xanthir.com/demos work, though - I jsut tried thema ll. [10:48:08.0000] <nights> ok cool! :) [10:48:31.0000] <nights> its nasty when things doesnt work in a presentation [10:51:05.0000] <Ms2ger> I find it fascinating that other glenn claims to have a hybrid spec tool for cssom, yet hasn't actually ever edited cssom [10:56:01.0000] <zcorpan> TabAtkins: you summoned? [11:01:42.0000] <tantek> Ms2ger, perhaps you can ask him in irc://irc.w3.org:6665/css - looks like he's even there now. [11:04:04.0000] <TabAtkins> zcorpan: Oh, I was just going to say that your suggestion sounds like a good idea. [11:04:33.0000] <TabAtkins> Ms2ger: Yes he has. What makes you think otherwise? [11:05:15.0000] <Ms2ger> Apart from putting his name on it [11:05:46.0000] <nights> TabAtkins: you work at google? [11:05:50.0000] <TabAtkins> I dont' know whether he's commited the relevant changes yet, but I know for a fact that he and Shane are working on it. [11:05:53.0000] <TabAtkins> nights: Yeah. [11:06:06.0000] <nights> TabAtkins: Cool...i actually rejected an offer from them once :) [11:06:14.0000] <TabAtkins> Bad move. ^_^ [11:06:20.0000] <Ms2ger> Good move [11:06:24.0000] <nights> now im cofounder of a company making USD -10000 every month [11:06:29.0000] <Ms2ger> Come join us at The Rebellion [11:06:31.0000] <nights> hehe [11:06:31.0000] <Ms2ger> Er, Mozilla [11:07:12.0000] <TabAtkins> I can't make much revenue directly from doing what I love, which is working on standards. [11:07:38.0000] <Ms2ger> TabAtkins, http://dev.w3.org/csswg/cssom/ <-- this is cssom. He's never actually edited it [11:08:03.0000] <nights> TabAtkins: i wrote minus 10000 usd [11:08:08.0000] <nights> so im sure you make more than me ;) [11:08:11.0000] <TabAtkins> Oh, hahahaha [11:08:17.0000] <TabAtkins> I read that as a ~ [11:08:22.0000] <nights> :) [11:08:27.0000] <nights> that time [11:08:37.0000] <nights> i started working for an another american company [11:08:43.0000] <nights> i regret that move very much :) [11:11:10.0000] <nights> google was a few floors down [11:11:23.0000] <nights> their office here in tokyo is very cool [11:12:06.0000] <zcorpan> TabAtkins: what suggestion? [11:12:43.0000] <TabAtkins> zcorpan: To make the parser replace escapes of surrogates with U+fffd [11:12:59.0000] <TabAtkins> nights: Agreed! I've been to the tokyo office before. [11:13:20.0000] <zcorpan> ah [11:14:03.0000] <nights> TabAtkins: the entrance looks like a bath house :) [11:14:19.0000] <nights> a few friends of mine work there at the android department [11:14:59.0000] <nights> once my company bleed up all our investment i might give it another shot. hope they have no hard feelings :) [11:15:23.0000] <TabAtkins> We're generally blind to previous attempts, and encourage people to try again. [11:15:57.0000] <nights> actually i didnt respond [11:16:02.0000] <nights> so maybe the offer is still hot [11:16:03.0000] <nights> :p [11:16:06.0000] <nights> been two years [11:16:09.0000] <TabAtkins> Then there's definitely no hard feelings. Go for it. ^_^ [11:16:57.0000] <nights> what do you guys think about still running a business when applying? [11:17:12.0000] <nights> maybe that wouldnt look good? [11:17:12.0000] <TabAtkins> I have no idea what our policies might be about that. [11:17:43.0000] <TabAtkins> I suspect it's problematic to be *working* at another company at the same time. But just working/running one while applying is probably fine? [11:18:09.0000] <nights> its essentially bootstrapped [11:18:59.0000] <smaug____> nights: there are other (less evil ;) ) companies than the big G. [11:19:05.0000] <nights> hehe [11:19:06.0000] <TabAtkins> LIES [11:19:13.0000] <nights> i worked for the most evil company in the world [11:19:20.0000] <nights> so my perception is quite damaged [11:19:24.0000] <Ms2ger> Facebook? [11:19:25.0000] <TabAtkins> Sony? [11:19:31.0000] <nights> actually ive worked for sony [11:19:32.0000] <smaug____> /me has also background in an evil empire [11:19:37.0000] <nights> but that wasnt the company i had in mind [11:19:41.0000] <nights> an american bank [11:19:43.0000] <Ms2ger> Mordor? [11:20:13.0000] <smaug____> Nah, never visited Mordor, though Morgoth was a nice boss [11:20:32.0000] <nights> mordor [11:20:40.0000] <nights> that must be much nicer than where i went [11:20:49.0000] <smaug____> :) [11:20:52.0000] <nights> :) [11:21:57.0000] <nights> the dark lord at that company basically had a will to dominate all life [11:22:07.0000] <nights> so there was similarities [11:22:39.0000] <nights> over me he poured his cruelty, his malice and a shitty bonus [11:35:46.0000] <annevk> am I missing something or is http://daringfireball.net/2012/05/bigger_display_iphone_thing_wwdc not discussing landscape mode weird? [11:45:24.0000] <annevk> about that email problem I had earlier [11:45:34.0000] <annevk> just learned of http://sparrowmailapp.com/mac.php via http://www.metafilter.com/116164/Comin-like-a-ghost-town#4358181 [11:45:41.0000] <annevk> apparently it has support for the Gmail label concept [11:48:08.0000] <smaug____> rafaelw_: hey, does the webkit implementation of MutationObserver support DocumentFragment? [11:48:53.0000] <Ms2ger> A test! [11:49:19.0000] <smaug____> I tested, but only using C19 [11:49:43.0000] <smaug____> haven't figured out where to download newer build for Linux (.zip/tar.* package, not rpm) [11:49:59.0000] <smaug____> C19 didn't pass the test [11:57:08.0000] <smaug____> TabAtkins: do you happen to know if there are .zip builds of chromium available ? [11:57:18.0000] <smaug____> I do have a .zip for C19 [11:57:25.0000] <smaug____> but don't remember where I downloaded it [11:57:44.0000] <smaug____> linux builds [12:00:58.0000] <Wilto> Not gonna lie, guys: I’m pretty pumped about Florian’s compromise proposal. [12:01:17.0000] <Wilto> Also: “Florian’s Compromise” would be a great name for either a bar or a scientific theory. [12:01:19.0000] <Wilto> http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2012-May/036162.html [12:01:46.0000] <Wilto> Seems like it covers everyone’s bases. [12:09:59.0000] <sicking> annevk: why aren't we adding the "chunked-text" and "chunked-arraybuffer" responsetypes to XHR? [12:10:50.0000] <annevk> pretty sure I emailed about that [12:12:13.0000] <annevk> so Sparrow does not do bottom posting [12:12:28.0000] <annevk> other than that it's pretty neat... :/ [12:26:15.0000] <TabAtkins> smaug____: No clue. [12:26:46.0000] <TabAtkins> annevk: "Does not", or "does not automatically"? [12:27:51.0000] <annevk> does not as far as I can tell [12:28:08.0000] <annevk> I could not find an option that inserts the signature at the bottom by default and leaves no space at the top [12:28:44.0000] <TabAtkins> Oh, okay. That counts as "not automatically" as far as I'm concerned. You can always put your signature down there yourself. [12:28:49.0000] <annevk> initially I was afraid plain text was also doomed, but that appears to be a simple setting (which is not accessible via the preferences panel, but whatever) [12:28:50.0000] <TabAtkins> Benefit of having a trivial signature. [12:31:22.0000] <annevk> yeah sure, I could type the whole response myself including all the quotes, too [12:31:47.0000] <TabAtkins> That's a silly extrapolation. [12:40:51.0000] <dglazkov> jgraham, Ms2ger: I actually prefer writing specs in just HTML. [12:41:16.0000] <Ms2ger> So do I [12:44:07.0000] <TabAtkins> I write my specs in a combination of HTML and Spanish. [12:46:04.0000] <annevk> dglazkov: you don't use any scripts to generate table of contents and such? [12:46:18.0000] <annevk> dglazkov: or cross-specification cross-references? [12:46:36.0000] <annevk> dglazkov: because I prefer HTML too and I use Anolis for some of the boring stuff [12:47:15.0000] <dglazkov> annevk: I wrote some simple scripts for the most annoying stuff. [12:47:51.0000] <annevk> dglazkov: what language did you use? [12:48:09.0000] <dglazkov> annevk: Jabba Script [12:48:18.0000] <annevk> dglazkov: I might be interested in that, maybe we don't need all of Anolis [12:48:23.0000] <annevk> dglazkov: heh [12:49:18.0000] <dglazkov> :) [12:53:26.0000] <dglazkov> annevk: I need to look at Anolis too. the Pythoney smell took me back at first. [12:54:08.0000] <dglazkov> another thing I've been experimenting with is using c9.io for spec-editing. [12:54:22.0000] <dglazkov> but it's still pretty wonky [12:55:03.0000] <annevk> so currently Shadow DOM is still using the legacy DOM-style [12:55:11.0000] <dglazkov> who is bjoern hoehrmann? [12:55:20.0000] <annevk> I take it that's because of Jabba Script? [12:55:35.0000] <annevk> dglazkov: http://thebjoernhoehrmannproject.org/ [12:55:39.0000] <dglazkov> no, that's because I stole the style from respec [12:56:19.0000] <annevk> dglazkov: student/ex-student? from Germany; not much else is known :) [12:56:38.0000] <annevk> a Ms2ger with an actual name and slightly different location within Europe [12:57:01.0000] <Ms2ger> Also, I don't care for Frisian [12:57:21.0000] <annevk> and when you troll it's never subtle :p [12:57:39.0000] <Ms2ger> Oh, no [12:57:44.0000] <TabAtkins> dglazkov: I don't know, but on CSS issues he's a troll. [12:58:30.0000] <dglazkov> GMail needs a troll hat feature. [12:59:10.0000] <Ms2ger> He's the only person who ever got a W3C spec rescinded, afaik [12:59:21.0000] <TabAtkins> Wow, which one? [12:59:27.0000] <TabAtkins> Also: how? [12:59:47.0000] <Ms2ger> "Spiderman and the XHTML Kindergarten" / http://lists.w3.org/Archives/Public/www-archive/2009May/0029.html [13:00:11.0000] <Ms2ger> They republished anyway soon after, but still [13:01:09.0000] <annevk> he also once submitted 300-500 emails about issues in SVG [13:02:10.0000] <Ms2ger> I also rather enjoyed http://lists.w3.org/Archives/Public/www-archive/2012Apr/0072.html [13:02:39.0000] <annevk> oh sweet [13:02:43.0000] <annevk> smaug____ went ahead [13:02:46.0000] <annevk> http://hg.mozilla.org/mozilla-central/rev/e54a85233701 [13:02:48.0000] <annevk> yay smaug____ [13:03:04.0000] <TabAtkins> smaug____++ [13:03:12.0000] <dglazkov> smaug____++ indeed [13:03:31.0000] <smaug____> waiting for webkit to unprefix too :) [13:05:08.0000] <annevk> Ms2ger: agreed, updated http://thebjoernhoehrmannproject.org/ [13:05:27.0000] <Ms2ger> \o/ [13:05:37.0000] <Ms2ger> Now I really accomplished something :) [13:08:53.0000] <annevk> hmm [13:09:05.0000] <annevk> maybe I should use some prefixes in that XML document [13:09:12.0000] <annevk> Google seems to badly cheat [13:09:20.0000] <gsnedders> Ideally in such a way as to expose browser bugs. [13:09:36.0000] <annevk> gsnedders: are there still browser bugs with prefixes? [13:09:49.0000] <annevk> gsnedders: other than Opera applying html:class or some such? [13:12:11.0000] <gsnedders> annevk: I dunno. [13:12:35.0000] <gsnedders> annevk: That Opera bug is the only one I know of off hand, but I know our bugs better than anyone else's. [13:12:36.0000] <annevk> gsnedders: don't you learn anything at uni? :p [13:12:47.0000] <gsnedders> annevk: Pretty much, no. [13:14:01.0000] <annevk> btw [13:14:05.0000] <annevk> the stuff about layout [13:14:14.0000] <annevk> that does not actually apply to media queries afaik [13:14:30.0000] <annevk> the whole point about media queries is that they don't depend on layout [13:14:58.0000] <annevk> the font-size preference from the user (defaulting to 16px) is known way before layout [13:15:13.0000] <annevk> and that's the only font-size 'em' in media queries depends upon [13:15:25.0000] <annevk> same for 'rem' [13:15:34.0000] <annevk> and 'ex', etc. [13:35:40.0000] <jgraham> dglazkov: I write specs in blood myself [13:35:53.0000] <jgraham> Tony Wilson style [13:36:10.0000] <gsnedders> Next you'll be documenting how to open the chamber of secrets… [13:36:26.0000] <dglazkov> jgraham: that's why they are always called First Public Working Draw [13:41:54.0000] <Ms2ger> /me hisses "Open" [13:43:02.0000] <jgraham> Ms2ger is a parselmouth? [13:43:52.0000] <Ms2ger> There have been precedents of non-parselmouths opening the chamber of secrets [13:44:49.0000] <dglazkov> hey, Harry Potter is a fellow standards wonk! He has specs! [13:45:16.0000] <TabAtkins> Boooooo [13:45:35.0000] <dglazkov> :P [13:46:04.0000] <jgraham> Ms2ger: Indeed. Notably Ron Weasley. [13:46:25.0000] <Ms2ger> dglazkov, wow [13:46:29.0000] <Ms2ger> That was pretty bad [13:46:51.0000] <dglazkov> so bad it's good? [13:46:59.0000] <jgraham> The other kind [13:46:59.0000] <Ms2ger> No [13:47:01.0000] <Ms2ger> Afraid not [13:47:04.0000] <dglazkov> :D [13:51:59.0000] <WeirdAl> oh, Harry Potter. :| [13:52:08.0000] <dglazkov> speaking of Ron Weasley, how do we define HTML interfaces from JavaScript perspective? For example, what is the distinction between HTMLElement and any object that implements this interface? [13:52:23.0000] <dglazkov> magic hosted objects aside [13:52:53.0000] <WeirdAl> testability? [13:53:07.0000] <dglazkov> both document.createElement('div') and window.HTMLDivElement are just objects that have HTMLElement in their prototype chain [13:53:36.0000] <TabAtkins> document.createElement('div') creates an element with HTMLDivElement as its prototype. [13:53:38.0000] <dglazkov> window.HTMLDivElement.prototype that is [13:53:55.0000] <TabAtkins> HTMLDivElement has HTMLElement in its prototype. [13:55:46.0000] <dglazkov> right. But from JS perspective, they are really not special in any way. There's no such thing as an "interface" in JS. So how do we get around this in DOM specs? [13:56:02.0000] <Ms2ger> What do you mean, get around this in DOM specs? [13:56:07.0000] <Ms2ger> What are you trying to do? [13:57:45.0000] <dglazkov> darn it, gotta run [13:57:55.0000] <dglazkov> bbiab [14:04:49.0000] <TabAtkins> dglazkov: Something is an "interface" if it doesn't have a constructor. 2012-05-24 [17:29:24.0000] <dglazkov> TabAtkins: are there things like that in JS? [17:29:59.0000] <TabAtkins> Not in real JS, no. [17:30:07.0000] <TabAtkins> It's something magical that only pops out of IDL. [17:30:36.0000] <TabAtkins> (And which imo should be avoided as much as possible in favor of constructible things.) [17:40:32.0000] <Wilto> oh what the crap is this [17:40:34.0000] <Wilto> http://axis.yahoo.com/ [17:40:43.0000] <Wilto> —Hah! Wrong window. Hi. [17:40:52.0000] <Wilto> Question stands. [21:30:56.0000] <dglazkov> TabAtkins: IDL sucks then. It must be terminated. [22:52:10.0000] <annevk> abarth: the allowseamless stuff should probably include some advice about how to prevent being embedded by the wrong third party [23:51:42.0000] <[tm]> crockford left Yahoo? [23:53:17.0000] <[tm]> ah, paypal [23:53:53.0000] <tantek> yeah - http://techcrunch.com/2012/05/13/paypal-gets-its-own-share-of-the-yahoo-diaspora-hires-java-icon-douglas-crockford/ [23:54:15.0000] <tantek> ironic since the recently deposed Yahoo CEO came from paypal [23:54:24.0000] <tantek> the one that lied on his resume [23:54:29.0000] <tantek> about having a CS degree [23:56:33.0000] <tantek> http://www.sfgate.com/cgi-bin/article.cgi?f=/c/a/2012/05/13/CMKC1OHF1B.DTL - scroll down to "Resume irregularity" [23:56:51.0000] <tantek> love this part: "Thompson sent employees a memo apologizing for the impact that the disclosure of his true credentials had on Yahoo, without apologizing for the error itself." [23:57:11.0000] <tantek> apologizing for how it looked, rather than apologizing for lying. [00:06:36.0000] <annevk> o_O [01:25:09.0000] <annevk> ideas for http://wiki.whatwg.org/wiki/Howto_spec [01:25:18.0000] <annevk> merge Content and Patterns [01:26:43.0000] <annevk> add more patterns: callbacks, class=domintro, cancelable events [01:27:24.0000] <jgraham> It is very about:blank. You are likely to be eaten by a load. [01:29:32.0000] <annevk> jgraham: I don't get it :( [01:30:17.0000] <jgraham> annevk: Oh don't mind me, I'm just experiencing the joy of about:blank today [01:30:31.0000] <jgraham> My comment was apropos nothing else in particular [01:32:52.0000] <jgraham> (I think I was thinking of "blank" as an alternative to "dark") [01:44:18.0000] <annevk> man [01:44:29.0000] <annevk> about:blank sure causes a lot of pain [02:01:08.0000] <jgraham> with that in mind [02:01:12.0000] <jgraham> hsivonen: yt? [02:08:15.0000] <annevk> so WebKit's location.origin is still not defined... [02:08:19.0000] <annevk> meh [02:09:55.0000] <othermaciej> was it ever officially proposed? [02:10:12.0000] <othermaciej> if not, I can complain to the person who added it to the code [02:11:16.0000] <annevk> pretty sure abarth proposed it [02:12:07.0000] <annevk> the way to clean this up is probably by defining the URL specification properly and then have Location and HTMLAnchorElement use it [02:12:34.0000] <annevk> I think [tm] already did some work in that direction [02:12:43.0000] <othermaciej> is there anything about it currently that would prevent it from being used as a supplemental interface for Location and HTMLAnchorElement? [02:13:27.0000] <hsivonen> jgraham: ? [02:13:46.0000] <annevk> othermaciej: yeah, the attributes are defined directly on the URL interface [02:14:11.0000] <annevk> othermaciej: we should probably have a [NoInterfaceObject] interface URLAttributes thingie [02:15:02.0000] <othermaciej> annevk: would that make a difference to the ability of Location to reuse it? (I might not sufficiently understand the Web IDL issues) [02:16:12.0000] <jgraham> hsivonen: So, I am looking at about:blank. Please feel sorry for me :) Also gecko seems to have some odd behaviour if you create an iframe, insert it, and then set the src [02:16:22.0000] <jgraham> Let me put a TC somewhere [02:16:22.0000] <othermaciej> in fact I am still super confused about the difference between sequence<Foo> and Foo[] and why both exist [02:16:45.0000] <annevk> othermaciej: interface URL has a bunch of other stuff (such as a constructor) that would not make sense elsewhere [02:17:02.0000] <annevk> actually, maybe it's just that constructor... [02:17:25.0000] <annevk> see http://dvcs.w3.org/hg/url/raw-file/tip/Overview.html#url [02:17:26.0000] <hsivonen> jgraham: Please don't try to make Opera behave like current Gecko behaves [02:17:58.0000] <hsivonen> jgraham: unfortunately, I've been blocked from making progress on changing Gecko yet again [02:18:07.0000] <hsivonen> jgraham: but the plan still is to change it [02:18:08.0000] <jgraham> hsivonen: I'm not trying to :) [02:18:24.0000] <jgraham> I'm trying to understand what behaviour is actually web compatible [02:18:40.0000] <hsivonen> webben: The deps of http://bugzilla.validator.nu/show_bug.cgi?id=923 now track Maven stuff [02:19:00.0000] <jgraham> (the spec is not) [02:19:02.0000] <hsivonen> webben: I have almost no clue about Maven and depend on people who want Maven support to contribute pom.xml patches [02:19:13.0000] <jgraham> I am told that it is relatively easy to change Opera to make it Webkit-like [02:19:34.0000] <hsivonen> jgraham: firing onload synchronously? [02:19:53.0000] <hsivonen> events that fire sync with parse are evil [02:20:12.0000] <hsivonen> I still want to make the events for about:blank to fire async [02:20:13.0000] <jgraham> Why? [02:20:45.0000] <hsivonen> jgraham: they can run scripts, so they require putting Gecko in a state where it's safe to run scripts [02:20:58.0000] <hsivonen> which add both code and perf overhead [02:21:14.0000] <hsivonen> also interesting document.write() call opportunities [02:21:47.0000] <jgraham> http://hoppipolla.co.uk/tests/about_blank/002.html [02:22:53.0000] <jgraham> In Opera that currently claims to be 2 async events, but apparently it is sure that the events will run before any other script [02:23:02.0000] <jgraham> In webkit it is one sync/one async [02:23:09.0000] <jgraham> In gecko there is only one event [02:28:13.0000] <othermaciej> annevk: I just read the spec and complained about the parameter stuff [02:28:27.0000] <othermaciej> annevk: anyway - are constructors inherited? I think they are not, so it doesn't matter [02:28:50.0000] <othermaciej> annevk: only issue would be whether prototype-hacking the URL interface object would affect HTMLAnchorElement [02:31:25.0000] <annevk> othermaciej: what you want is implementing them and defining them in a single place, but the effect should be that they appear to be on HTMLAnchorElement, Location, and URL independently I think [02:31:46.0000] <annevk> I don't really care how we go about that, I guess heycam|away can figure that out :) [02:32:24.0000] <jgraham> hsivonen: So, what behaviour are you proposing, and how sure are you that it's web compatible? [02:32:24.0000] <annevk> it might be that "implements" does exactly that, but I don't know what happens to Constructors [02:32:30.0000] <Ms2ger> That's 'implements', AIUI [02:34:05.0000] <othermaciej> annevk: I think it might just work for Location and HTMLAnchorElement to both inherit from URL [02:34:34.0000] <Ms2ger> No [02:34:43.0000] <Ms2ger> There is no multiple inheritance in IDL [02:35:03.0000] <othermaciej> ok, whatever the alternative to multiple inheritance is to add a mixin (is there such a thing?) [02:35:19.0000] <Ms2ger> implements copies "constants, attributes or operations" [02:35:38.0000] <Ms2ger> So, implements should be fine [02:35:41.0000] <othermaciej> Location could just inherit, since it has no base class [02:35:52.0000] <othermaciej> I think HTMLAnchorElement could use "implements" [02:36:14.0000] <Ms2ger> It would be silly to make them different, IMO [02:37:12.0000] <othermaciej> Location could use "implements" too despite not having any other base class [02:37:22.0000] <othermaciej> I do not know if there is any other comparable case in the HTML DOM [02:37:37.0000] <othermaciej> where an element and a non-element share an interface in common [02:39:01.0000] <annevk> HTMLElement and Document for a few members [02:39:07.0000] <annevk> but they're just duplicates [02:39:17.0000] <annevk> duplicated* [02:40:21.0000] <Stevef> othermaciej: hi, what's the status on the <hgroup> issue? [02:40:35.0000] <othermaciej> Stevef: I do not know offhand [02:40:50.0000] <othermaciej> do you remember the issue number? [02:41:48.0000] <othermaciej> there were draft reviews of all the proposals in progress, I do not recall if they have been posted yet [02:41:51.0000] <othermaciej> if not, that is the next step [02:42:05.0000] <Stevef> othermaciej: http://www.w3.org/html/wg/tracker/issues/164 [02:43:07.0000] <Stevef> othermaciej: haven't been posted yet, paulc soad he was going to post them about 2 weeks ago [02:43:41.0000] <othermaciej> ok, I'll check in with him [02:43:52.0000] <Stevef> othermaciej: thanks [02:45:47.0000] <Stevef> can anybody tell me what ted oconnors nick is? [02:45:51.0000] <Ms2ger> hober, [02:46:13.0000] <Stevef> Ms2ger:thanks [02:46:17.0000] <Ms2ger> Np [02:47:17.0000] <Stevef> hober: any response to this http://lists.w3.org/Archives/Public/public-html/2012May/0008.html ? [02:58:37.0000] <webben> hsivonen: I'm looking into the Maven issue btw. How do you build and test htmlparser yourself? I want to verify my pom.xml doesn't break your workflow? [04:06:17.0000] <hsivonen> webben: I use a plain Eclipse project setup with its intregated incremental builder [04:06:50.0000] <hsivonen> webben: pom.xml isn't part of my workflow at all except for making release bundles for the Maven Central Repository [04:07:00.0000] <hsivonen> webben: hence, I don't grok pom.xml [04:09:57.0000] <zcorpan> http://forums.whatwg.org/bb3/viewtopic.php?f=10&t=5000#p7899 anyone have suggestions about what to do about this? [04:12:21.0000] <ashemedai> MathML in Opera Next gets better and better, I have to say [04:15:50.0000] <jgraham> ashemedai: Interesting. Examples of things that improved? [04:18:56.0000] <ashemedai> jgraham: http://www.in-nomine.org/~asmodai/3d.html [04:19:06.0000] <ashemedai> jgraham: scroll to alpha blending [04:19:22.0000] <ashemedai> The entire layout of that part is just tight now, as it should [04:19:29.0000] <ashemedai> of course, the other ones still can use some work :) [04:20:08.0000] <Ms2ger> Heh, that looks bad on my opera-next [04:20:24.0000] <jgraham> Yeah, maybe I am missing some fonts [04:20:34.0000] <jgraham> But I see [r] instead of [r,g,b] [04:20:36.0000] <Ms2ger> ^ [04:20:58.0000] <Ms2ger> Fortunately, Gecko is there when you need a good MathML implementation ;) [04:22:37.0000] <ashemedai> oh wait [04:22:41.0000] <ashemedai> Good point jgraham [04:22:48.0000] <ashemedai> Totally missed that [04:25:28.0000] <ashemedai> Ms2ger: And you are right, just still have this pipedream of seeing Chrome and Opera implement more of MathML [04:26:00.0000] <zcorpan> publish more mathml content so it becomes a sitecompat issue :-) [04:26:20.0000] <Ms2ger> Acid4 ;) [04:26:24.0000] <jgraham> Convince Google to make google.com depend on MathML :p [04:26:52.0000] <annevk> ashemedai: I'm afraid that as long as we use CSS and CSS cannot describe all of MathML displaying Math in Opera will be bad :( [04:27:06.0000] <jgraham> Ms2ger: Well given acid 3 that would probably give you stubs for MathML [04:28:00.0000] <Ms2ger> Acid4: The entire MathML test suite? :) [04:28:14.0000] <ashemedai> mmm, opera next updater gives me a cannot validate file when it tries to grab the update file, at least manual update worked [04:28:28.0000] <zcorpan> each pixel is a test [04:28:37.0000] <ashemedai> annevk: Yea, but I'm happy with the intermediate work already [04:28:45.0000] <Ms2ger> zcorpan, ... and they animate ;) [04:28:56.0000] <ashemedai> annevk: I should just use Mathjax in the meantime [04:29:59.0000] <annevk> ashemedai: nah, please publish MathML [04:30:49.0000] <Ms2ger> Does annevankesteren.nl have any? [04:30:54.0000] <annevk> ashemedai: and write about the problems you encounter [04:31:02.0000] <ashemedai> annevk: I'll be adding more formulae to that page in the weekend. [04:31:04.0000] <annevk> Ms2ger: I suck at math [04:31:06.0000] <ashemedai> annevk: as in blog? [04:31:15.0000] <annevk> ashemedai: for instance [04:31:22.0000] <ashemedai> Can do that :) [04:31:35.0000] <annevk> you can also create a wiki page on the WHATWG wiki [04:31:55.0000] <annevk> we try to track/document shortcomings of all web technology [04:32:11.0000] <annevk> so whenever someone gets a chance to improve it, they know where to start [04:32:17.0000] <Ms2ger> annevk, I'm sorry to hear that :) [04:32:49.0000] <annevk> Ms2ger: the people who wanted me to define matrix-aware layout APIs were too [04:33:12.0000] <Ms2ger> Meh, that's not real math [04:33:35.0000] <ashemedai> annevk: Ah, sure, I can do the MathML page on Whatwg wiki [04:34:00.0000] <ashemedai> So who do I bother for an account? :P [04:34:09.0000] <annevk> ashemedai: you message me an email address [04:34:46.0000] <annevk> ashemedai: oh and give me a desired username [04:37:34.0000] <ashemedai> Thanks! [04:39:09.0000] <ashemedai> annevk: Is there a specific subsection where status/shortcomings are discussed? [04:39:17.0000] <ashemedai> Where I hook a new page under [04:39:54.0000] <annevk> [[Category:Proposals]] maybe [04:39:55.0000] <annevk> hmm [04:40:27.0000] <Ms2ger> [[Category:ToBeAssimilated]] [04:40:30.0000] <annevk> oh or maybe http://wiki.whatwg.org/wiki/Category:Spec_coordination [04:40:34.0000] <ashemedai> Ms2ger: :P [04:40:42.0000] <annevk> Spec coordination is probably good [04:42:30.0000] <ashemedai> annevk: ok, done. Got the page up, linked under Spec coordination [04:43:12.0000] <annevk> "Spec coordination is what keeps #whatwg busy at night." [04:43:15.0000] <annevk> heh [04:43:27.0000] <annevk> oh I wrote that [04:43:54.0000] <ashemedai> lol [04:44:08.0000] <jgraham> Well at least you can laugh at yourself [04:45:11.0000] <hsivonen> ashemedai: convince Wikipedia to go all-in with MathML without MathJax fallbacks [04:45:44.0000] <annevk> yeah Wikipedia really ought to switch [04:46:20.0000] <ashemedai> hsivonen: I can try asking some of my Wikipedia-Unicode contacts about that [04:46:37.0000] <ashemedai> Think that would be incentive for the browser makers to step up the game? [04:47:13.0000] <hsivonen> ashemedai: If Wikipedia went all-in with Presentation MathML, yeah. [04:47:32.0000] <hsivonen> ashemedai: (if they did Content MathML, they'd be too far out there) [04:48:07.0000] <annevk> is it even agreed we should keep Content MathML? [04:48:21.0000] <hsivonen> annevk: no [04:48:56.0000] <hsivonen> annevk: we decided to make Gecko not corrupt Content MathML alternatives in clipboard operations, though [04:49:07.0000] <annevk> mkay [04:49:08.0000] <hsivonen> annevk: so we're trying to be nice about it [04:49:33.0000] <annevk> I guess the MathML wiki page should at some point define the subset we're interested in [04:49:34.0000] <hsivonen> annevk: but no plans to really support it, AFAIK [04:49:43.0000] <annevk> if the Math WG is not going to do that [04:50:03.0000] <hsivonen> annevk: are there parts of Presentation MathML that we're *not* interested in? [04:50:14.0000] <hsivonen> annevk: I thought the subset was "Presentation MathML" [04:50:39.0000] <annevk> what is <annotation-xml> part of? [04:51:01.0000] <zcorpan> lala land? [04:51:06.0000] <gsnedders> We've pretty much said we're only really interested in http://www.w3.org/TR/mathml-for-css/ [04:51:07.0000] <hsivonen> annevk: it's an integration point on the edge of Presentation MathML [04:51:12.0000] <jgraham> hsivonen: So re: about:blank it seems like we will go for something WebKit-like unless we are persuaded otherwise (e.g. by a gecko implementation that is web compatible and doesn't have synchronous events or task queue mangling) [04:51:19.0000] <hsivonen> annevk: you walk off the cliff when you walk into that subtree [04:51:35.0000] <jgraham> Or by finding out that "webkit like" is actually hard to implement in detail [04:51:39.0000] <annevk> hsivonen: right so maybe that should be "disallowed" [04:51:40.0000] <gsnedders> i.e., only really interested in the subset of MathML that can be described by CSS. [04:51:46.0000] <hsivonen> jgraham: :-( [04:51:47.0000] <gsnedders> Of course as CSS grows that may expand. [04:52:30.0000] <hsivonen> annevk: depends on how you view use cases for alternative semantic content for copying and pasting into symbolic algebra software [04:52:34.0000] <jgraham> hsivonen: Yeah, sorry :( But it is hard to argue against something that I'm told is - at a high level - simple in our implementation and has known web compat properties [04:52:48.0000] <hsivonen> jgraham: I understand [04:53:08.0000] <hsivonen> jgraham: I fail for what I want being continuously vaporware [05:24:23.0000] <ashemedai> So feel free to let me know what you think such a page on MathML needs - http://wiki.whatwg.org/wiki/MathML [05:24:39.0000] <ashemedai> I'll add some implementation notes/details as soon as I get home [05:25:54.0000] <annevk> ashemedai: feel free to phrase it more direct [05:26:34.0000] <ashemedai> annevk: Phrase what more direct? [05:26:35.0000] <annevk> ashemedai: like "Content MathML should not be used and is not going to be implemented by browsers." [05:27:10.0000] <annevk> though maybe mention the copy-and-paste exception hsivonen was talking about [05:27:16.0000] <ashemedai> Ah [05:28:22.0000] <annevk> if we communicate our intentions clearly, there's less room for surprise [05:28:40.0000] <ashemedai> Gotcha [05:28:50.0000] <annevk> (and more room for debate, people might disagree with our stance) [06:05:39.0000] <annevk> /me is making modest progress with http://wiki.whatwg.org/wiki/Howto_spec [06:06:58.0000] <annevk> defined a bunch of event patterns [06:07:02.0000] <annevk> and callbacks [06:11:29.0000] <Ms2ger> Anyone have a link for the previous readyState string/short discussion? [06:19:11.0000] <annevk> you're lazy [06:19:15.0000] <annevk> took me a minute: http://lists.w3.org/Archives/Public/public-script-coord/2012JanMar/thread.html#msg166 [06:20:02.0000] <annevk> but it's really my fault for responding here :) [06:32:05.0000] <matjas> http://lachy.id.au/dev/markup/tests/html5/charref/ lists 1449 named character references, while http://whatwg.org/html/named-character-references.html lists 2232 [06:32:19.0000] <matjas> would there be any interest in an updated test case for this? [06:32:44.0000] <matjas> i was thinking of making one that automates the testing through JS [06:33:48.0000] <Philip`> matjas: That test case has 1449ish lines, but lists multiple named character references on a single line [06:33:56.0000] <Philip`> <tr><td>&#x00026;<td>&amp; &amp &AMP;<td>U+00026 AMPERSAND [06:33:58.0000] <Philip`> etc [06:33:59.0000] <matjas> right [06:34:04.0000] <Philip`> which might make up some of the difference [06:34:35.0000] <Philip`> (or all of it) [06:35:26.0000] <matjas> not all of it; there are only 3581 occurences of & in the source [06:35:43.0000] <matjas> 3581 / 2 = 1790,5 < 2232 [06:36:33.0000] <matjas> anyhow, just wanted to know if it hasn’t been done already, or if lachlan had an update planned [06:37:37.0000] <Philip`> Hmm, looks like it excludes at least the multi-codepoint references (&acE; etc) [06:38:37.0000] <matjas> zcorpan: without checking the spec i’d guess A [06:41:40.0000] <zcorpan> matjas: http://i.imgur.com/Emi0l.gif [06:41:54.0000] <Philip`> There's http://code.google.com/p/html5lib/source/browse/testdata/tokenizer/namedEntities.test which would hopefully be up to date, but I don't know whether there's a way to run those tests inside a browser (rather than a standalone parser) [06:42:19.0000] <Ms2ger> C [06:42:32.0000] <matjas> Philip`: okay, I’ll just give it a whirl then — can’t hurt, right? [06:42:32.0000] <zcorpan> Ms2ger: http://i.imgur.com/Emi0l.gif [06:42:49.0000] <Ms2ger> Nice [06:43:05.0000] <Philip`> matjas: More tests never hurt :-) [06:45:40.0000] <jgraham> http://test.w3.org/html/tests/submission/Opera/html5lib/tests/test_tokenizer_namedEntities.html perhaps? [06:46:18.0000] <matjas> jgraham: woah @ source [06:47:00.0000] <jgraham> You can't decode uri components in your head? [06:47:08.0000] <jgraham> :) [06:47:50.0000] <odinho> Ms2ger: Seems test-runner doesn't run in IE10? I just wanted to check something, but seems it never really starts :-) Is it known, or an authoring fault from my side on making the testrunner.htm file? [06:48:24.0000] <jgraham> matjas: It was supposed to be a very general purpose way of converting the html5lib tests that didn't optimise for readability [06:49:10.0000] <odinho> Ms2ger: fsck, it was (my :P) authoring error as I had a hunch it might be. [06:51:50.0000] <Ms2ger> odinho, does it work in IE10, then? I never tried :) [06:52:20.0000] <odinho> Ms2ger: Well, I'll see shortly, I need to stop reading IRC backlog first :P I'm easily distracted it seems. [06:53:25.0000] <odinho> Ms2ger: Aaaand, it works! :-) [06:53:36.0000] <Ms2ger> \o/ [07:12:12.0000] <jgraham> It would be nice if Bryan Sullivan's name wasn't always written in loudcase [07:14:37.0000] <odinho> Hehe. [07:20:38.0000] <zcorpan> nobody guessed the right answer for this quiz (other than by eliminating the wrong answers), i think that's a first [07:38:34.0000] <matjas> Hixie: why does whatwg.org/html/named-character-references.html#entity-DotDot have \u25CC\u20DC in the glyph cell rather than just \u20DC? [07:41:51.0000] <zcorpan> file a bug [07:42:01.0000] <tkadlec> e [07:42:18.0000] <Philip`> matjas: I expect it's a placeholder so that the combining character is sensibly visible [07:42:50.0000] <matjas> zcorpan: it doesn’t seem like a bug/typo, as other cells have it too… i just don’t see why — it doesn’t make it clearer [07:43:13.0000] <matjas> Philip`: i think the extra circle just makes it harder to see the combining char [07:43:40.0000] <matjas> guess it’s just me then [07:43:55.0000] <Philip`> If you didn't have any character in front of it then it'd appear too far to the left, which would also look unclear [07:44:44.0000] <Philip`> http://www.fileformat.info/info/unicode/char/25cc/index.htm has a comment relating to combining character indications [07:45:08.0000] <Philip`> (I presume that comment is originally from the Unicode standard itself) [08:12:41.0000] <matjas> other mismatches between code points + glyphs on that page: https://www.w3.org/Bugs/Public/show_bug.cgi?id=17170#c1 [08:16:02.0000] <matjas> ^ writing those tests pays off already :) [08:47:44.0000] <matjas> test results: WebKit currently fails on 1 named character reference (&AElig doesn’t match Æ), the WebKit in latest stable Safari has 94 errors, Presto and Gecko pass all tests [08:49:05.0000] <jgraham> matjas: Nice, but you should test one of the many parts of the spec that *doesn't* already have tests. Then you should submit the tests to the W3C :) [08:50:39.0000] <Ms2ger> matjas, or review the submitted tests, if you like that :) [08:51:13.0000] <jgraham> s/that/torture/ [08:52:26.0000] <Ms2ger> shh [08:52:39.0000] <jgraham> Uh right [08:52:48.0000] <jgraham> Testcase review is really fun! [08:52:54.0000] <jgraham> You should do it! [08:53:01.0000] <Ms2ger> Hah [08:53:29.0000] <matjas> thanks, i’d rather write my own :') [08:53:39.0000] <jgraham> Problem is, if people start reviewing TCs, they will quit before they get to writing them [08:53:56.0000] <jgraham> And I would rather have people writing TCs [08:54:08.0000] <jgraham> Even though that means more reviewing... [09:00:17.0000] <odinho> Funny, Firefox does 1 test better than IE10 in MS' idb testsuite. But they both fail the same number (the diff is not-run). [09:00:46.0000] <odinho> But both more or less lies at 84% pass. [09:01:09.0000] <odinho> /me wonders when WebKit will start updating their idb implementation... [09:01:25.0000] <annevk> matjas: the &lang; thing is a problem in the publishing pipeline [09:01:40.0000] <annevk> matjas: there's a bug on it somewhere, we haven't quite located the source [09:02:06.0000] <matjas> annevk: hmm, can’t you just use numeric character references instead of the raw characters, or are those unreliable as well? [09:02:43.0000] <annevk> matjas: you mean instead of named entities? [09:02:56.0000] <matjas> no, i mean for the “glyph” cell [09:03:02.0000] <annevk> matjas: I think numeric character references are used, but the publishing pipeline fucks it up [09:03:14.0000] <matjas> annevk: oh, i see [09:04:50.0000] <dglazkov> good morning, WhatwG! [09:05:30.0000] <odinho> good evening dglazkov :-) [09:06:26.0000] <annevk> dglazkov: still trouble spelling names I see :p [09:06:43.0000] <dglazkov> annevk: I am mixing it up. [09:07:59.0000] <tantek> good morning dglazkov [09:08:23.0000] <tantek> /me is debating whether to join the HTMLWG telcon this morning. [09:08:27.0000] <annevk> dglazkov: GoOd [09:08:31.0000] <tantek> perhaps I should take a straw poll [09:08:36.0000] <annevk> tantek: lol [09:08:43.0000] <dglazkov> tantek: yes, but only if you do it ironically [09:08:52.0000] <annevk> tantek: I can tell you how it goes: http://annevankesteren.nl/2010/12/html-wg-teleconference [09:09:10.0000] <annevk> tantek: maybe you can write down your experience as well so we can compare notes and see how things have progressed [09:09:15.0000] <tantek> ok, I leave it to #whatwg - please express +1/-1/0 [09:09:30.0000] <TabAtkins> -1 [09:09:33.0000] <odinho> +1 (plus notes) [09:09:35.0000] <annevk> +1 if you write it down [09:09:40.0000] <annevk> otherwise waste of time [09:21:39.0000] <tantek> annevk - your post is similar to most of my past experiences on the telcon [09:22:22.0000] <tantek> however, when I've had an issue (as in an Issue in tracker) to discuss, and have specifically brought up questions, typically I've had very productive discussions with the chairs. [09:23:15.0000] <tantek> so I think the telcons are mostly what you make of them. if just lurking, it's perhaps not that useful. if you have specific business to raise/discuss with the chairs, it can be useful. and anyone can do that during the "any other business" part of the telcon. [09:26:08.0000] <annevk> not sure why you'd dial in for that, but then I'm not sure why I ever dialed into teleconferences anyway [09:28:10.0000] <tantek> annevk - mostly I've dialed in to help move along the various <time> and <data> element change proposals that I made to officially make the changes happen that I discussed/debated/designed with Hixie (and somewhat in this channel) ages ago - after we got rough consensus at the TPAC HTMLWG f2f last year. [09:29:35.0000] <annevk> the <time> stuff was handled in such a fucked up way [09:29:44.0000] <annevk> W3C HTML was not updated for months [09:33:32.0000] <Ms2ger> Mm, we support rest parameters now [09:34:30.0000] <annevk> Ms2ger: meaning? [09:34:31.0000] <tantek> annevk - W3C HTML5 still lacks <data> :( [09:34:46.0000] <Ms2ger> function foo(...rest) {} [09:36:44.0000] <othermaciej> tantek: it was decided to restore it but it probably won't happen now til there are stable draft editors in place [09:38:40.0000] <annevk> tantek: not the only thing http://www.whatwg.org/specs/web-apps/current-work/multipage/introduction.html#how-do-the-whatwg-and-w3c-specifications-differ? [09:38:58.0000] <annevk> Ms2ger: is that the automatic array extraction? [09:39:18.0000] <annevk> Ms2ger: maybe it's time we update DOMTokenList and friends then [09:39:53.0000] <Ms2ger> Wfm [09:41:22.0000] <annevk> maybe add a note to the bug? [09:41:37.0000] <annevk> I'm gonna play with Howto spec a while longer I think [09:42:09.0000] <annevk> and then maybe review a bunch of API specs and point Howto spec out [09:43:38.0000] <tantek> othermaciej - is there a list somewhere of such pending edits? [09:44:22.0000] <tantek> (where a decision was made to make an edit, but the edit itself has not been made) [09:44:28.0000] <othermaciej> tantek: I think most pending decisions have an open bug with WGDecision keyword, but this case may be an exception since it was a CFC not directly related to a bug [09:46:42.0000] <tantek> /me is having some trouble following HTMLWG process intricacies. [09:54:31.0000] <TabAtkins> I'd be glad to see Content MathML dropped, since it contains the only two tagname conflicts in the language. Presentation MathML was careful (or lucky?) not to collide with SVG or HTML in any way. [09:55:32.0000] <Ms2ger> Careful [09:55:47.0000] <TabAtkins> Good to know. [10:03:17.0000] <tantek> interesting, an actual IETF draft to define rel="profile" - http://tools.ietf.org/html/draft-wilde-profile-link-01 [10:24:08.0000] <dglazkov> in custom elements spec, I am very tempted to be just describe how things happen from JS perspective (i.e. prototype chain, objects in it, etc.). However, WebIDL/DOM4 and most other specs tend to specify things in terms of interfaces and generic inheritance concepts. Should do the former and not worry about consistency with existing specs, or should I do the latter and try to fit the concept of cu [10:24:08.0000] <dglazkov> stomizable DOM interfaces into the WebIDL world? [10:24:48.0000] <Ms2ger> Sounds like you're doing something wrong [10:25:35.0000] <dglazkov> Ms2ger: yes, but what? :) [10:25:55.0000] <Ms2ger> Well, I don't know what you're trying to accomplish [10:27:24.0000] <dglazkov> trying to capture http://dvcs.w3.org/hg/webcomponents/raw-file/tip/explainer/index.html#custom-element-section as a spec. [10:28:20.0000] <dglazkov> specifically, describing the process of building a custom DOM element with a prototype chain, which includes a DOM object and an arbitrary JS object. [10:29:08.0000] <Ms2ger> I don't think I want to lat you do that, dave [10:29:30.0000] <dglazkov> Ms2ger: it's too late for that, R2D2 [10:30:55.0000] <dglazkov> actually, http://www.w3.org/TR/WebIDL/#dfn-interface I think is workable [10:32:28.0000] <dglazkov> it's just such a weird transpiler operation, having to speak in WebIDL when really everyone knows it's JS. [10:32:50.0000] <dglazkov> can we just rewrite all specs to be in JS? :) [10:33:14.0000] <Ms2ger> It isn't [10:33:16.0000] <Ms2ger> It's C++ [10:40:39.0000] <dglazkov> Ms2ger: that's just an implementation detail. C++ tries darn well to pretend it's not there. And Microsoft peeps are leading the charge to make it even less visible. [10:41:19.0000] <dglazkov> in IE10 that is [10:52:30.0000] <annevk> dglazkov: is there agreement that we want to do such a thing? [10:56:38.0000] <annevk> Ms2ger: stumbled upon http://lists.w3.org/Archives/Public/spec-prod/2012AprJun/0075.html [10:57:00.0000] <Ms2ger> Heh [11:00:35.0000] <scott_gonzalez> smaug____ (or any other Mozilla devs): Do you know why scrollable elements are in the tab order in Firefox but no other browser? [11:00:47.0000] <scott_gonzalez> Is it an accessibility concern? [11:00:54.0000] <smaug____> yes [11:01:55.0000] <smaug____> scott_gonzalez: https://bugzilla.mozilla.org/show_bug.cgi?id=755766#c11 [11:02:04.0000] <scott_gonzalez> Target audience of sighted keyboard users? [11:02:05.0000] <smaug____> (that bug is invalid) [11:47:13.0000] <rniwa> AryehGregor: yt? [12:31:32.0000] <gsnedders> "Ecma International now hosts a normative HTML copy of Ecma-262, Edition 5.1 The ECMAScript Language Specification." [12:33:20.0000] <Ms2ger> It's not actually normative [12:36:17.0000] <cbright6062> /me tries to read over the recent posts here but gets a headache and decides to not bother. [12:40:10.0000] <cbright6062> I'd also love to see the day where things didn't seem to get so overcomplicated with everything. Lol. [15:51:43.0000] <alecflett> hey - is anyone around who can clarify something for me in the DOM4 spec? [15:52:00.0000] <alecflett> specifically what looks like a spec bug in the section about DOMException [15:55:01.0000] <gsnedders> Just ask. [15:56:44.0000] <alecflett> well the problem I have is that there's no specifically called-out field for the string-based exception type in DOMException [15:57:05.0000] <alecflett> so if a caller wants to check the exception type using the newer string-based mechanism, there's no way to do it [15:57:29.0000] <alecflett> i.e. rather than if (ex.code == DOMException.IndexSizeError) …., [15:57:40.0000] <alecflett> there's no equivalent for the string "IndexSizeError" [15:57:41.0000] <alecflett> ie. [15:57:58.0000] <alecflett> if (ex.<???> == "IndexSizeError") {…} [15:58:19.0000] <alecflett> ex.name? (consistent with DOMError) or ex.type (alluded to in the spec but not called out in the IDL) [16:02:02.0000] <heycam> alecflett, it's .name [16:02:27.0000] <heycam> alecflett, it's in Web IDL that it's defined that exception objects get a name property [16:03:12.0000] <alecflett> ahh.. .thanks for the clarification [16:03:44.0000] <alecflett> might be nice to have some non-normative reference to that somewhere in the DOM4 spec as I spent a good deal of time looking for it! 2012-05-25 [18:51:31.0000] <ian128K> Just wanted to put in my 2¢ regarding responsive images: I much prefer the proposed <picture> element over adding a "set" attribute to the existing <img> element. [19:14:13.0000] <zewt> (clearly what we need are: more unsubstantiated opinions) [21:32:06.0000] <Hixie> i have a utf-8 validator function (given input bytes, give a thumbs-up or thumbs-down on whether it's valid utf-8) [21:32:12.0000] <Hixie> anyone got any good test data for that? [22:22:55.0000] <asmodai> Hixie: http://www.cl.cam.ac.uk/~mgk25/ucs/examples/UTF-8-test.txt [22:23:19.0000] <asmodai> http://www.cl.cam.ac.uk/~mgk25/ucs/examples/UTF-8-demo.txt [23:26:43.0000] <zcorpan> asmodai: nice tests [23:59:10.0000] <zcorpan> TabAtkins: http://www.zurb.com/playground/responsive-tables [00:17:02.0000] <Hixie> asmodai: yeah, i know about those, but they don't list the utf-8 bytes nor indicate valid vs not-valid so they're hard to use [00:17:03.0000] <asmodai> zcorpan: Yeah, useful at least [00:17:25.0000] <asmodai> Hixie: So what exactly are specifically looking for? [00:17:47.0000] <asmodai> Hixie: Because I am sure some of my acquaintances will have these tests lying around [00:17:56.0000] <Hixie> something like "0xC0 0x01 INVALID" "0x01 0x01 VALID" [00:18:19.0000] <asmodai> ah [00:18:26.0000] <asmodai> ok, let me dig around for you [00:18:59.0000] <Hixie> or "0x01 0x01 => U+00001 U+00001" "0xC0 0x01 => U+0FFFD U+00001" [00:19:15.0000] <Hixie> (since i can derive the former from the latter) [00:19:31.0000] <Hixie> (by looking for FFFD in the output side) [00:31:02.0000] <matjas> http://mathias.html5.org/tests/named-character-references/ Opera and Firefox pass all tests, WebKit nightly fails one, IE10pre fails 3. IE9 fails 1869 of them. [00:36:07.0000] <smaug____> I wonder why "If current node is a pre, textarea, or listing element, and the first child node of the element, if any, is a Text node whose character data has as its first character a U+000A LINE FEED (LF) character, then append a U+000A LINE FEED (LF) character." is needed [00:40:33.0000] <Ms2ger> Because otherwise the one LF is discarded when reparsing [00:41:31.0000] <smaug____> ahaa [00:58:39.0000] <smaug____> hmm, does innerHTML work differently for data documents [00:59:20.0000] <smaug____> seems like [01:00:49.0000] <smaug____> I doubt that is right [01:02:03.0000] <zcorpan> smaug____: data documents? [01:02:17.0000] <smaug____> zcorpan: I mean something like createHTMLDocument() [01:02:41.0000] <zcorpan> how is it different? [01:03:57.0000] <smaug____> "or if the parent of current node is noscript element and scripting is enabled for the node, then append the value of current node's data IDL attribute literally." [01:04:24.0000] <smaug____> I don't see what defines that scripting is enabled for a node in a data document [01:05:43.0000] <smaug____> yes, "(A Document created using an API such as createDocument() has no browsing context." [01:05:51.0000] <smaug____> so, different serialization [01:06:13.0000] <annevk> how can scripting be disabled if you're doing scripting? [01:06:19.0000] <zcorpan> but the spec doesn't say that scripting should be disabled for such documents, does it? [01:06:39.0000] <zcorpan> (scripts don't run in documents without a browsing context, but that's different) [01:06:59.0000] <smaug____> "Scripting is enabled for a node if the Document object of the node (the node itself, if it is itself a Document object) has an associated browsing context, " [01:07:16.0000] <zcorpan> aha [01:08:33.0000] <zcorpan> yeah that seems like a bug then [01:08:58.0000] <smaug____> and I was convinced that spec's html-fragment algorithm might be bug-free... [01:09:14.0000] <smaug____> but good to go through [01:09:34.0000] <zcorpan> who convinced you? there are open bugs on it :-) [01:10:14.0000] <smaug____> well, perhaps no one said bug-free [01:10:21.0000] <smaug____> but not full of bugs [01:11:38.0000] <zcorpan> i guess that's about right [01:14:31.0000] <zcorpan> would be nice if the spec had annotations about open bugs [01:14:48.0000] <zcorpan> Hixie: ^ [01:16:10.0000] <smaug____> that would be awesome [01:19:41.0000] <rniwa> smaug____: hi smaug____! [01:20:01.0000] <rniwa> smaug____: slowing improving webkit's fragment parsing APIs :D https://bugs.webkit.org/show_bug.cgi?id=87454 [01:20:35.0000] <rniwa> smaug____: please file bugs & cc me if you find more bugs w.r.t. createContextualFragment, etc... [01:21:28.0000] <rniwa> oops I misread the long [01:21:33.0000] <rniwa> matjas: ^ what I just said to smaug____ [01:21:45.0000] <smaug____> ok [01:21:51.0000] <smaug____> uh [01:22:04.0000] <smaug____> I'll file some spec bugs [01:22:22.0000] <rniwa> smaug____: yeah, it'll be nice if Ms2ger's spec becomes more mature [01:22:30.0000] <matjas> zcorpan: +∞ [01:22:39.0000] <rniwa> smaug____: there are a lot of vaguely defined stuff in the spec at the moment :\ [01:22:46.0000] <smaug____> rniwa: this is in HTML spec [01:22:53.0000] <smaug____> http://www.whatwg.org/specs/web-apps/current-work/multipage/the-end.html#html-fragment-serialization-algorithm [01:23:03.0000] <rniwa> smaug____: oh, sure browser context, etc... stuff is in, yes [01:23:17.0000] <smaug____> looks like noscript handling is all wrong there [01:23:19.0000] <rniwa> smaug____: but that alone doesn't do much [01:23:24.0000] <rniwa> smaug____: because it doesn't define APIs for them [01:23:29.0000] <rniwa> smaug____: oh yeah? [01:23:44.0000] <smaug____> spec defines something that browsers don't do [01:23:51.0000] <rniwa> :( [01:24:24.0000] <rniwa> scumbag spec monsters. [01:25:26.0000] <rniwa> smaug____: one of these days, i'll go through that part of the spec and give a more comprehensive feedback. [01:28:38.0000] <rniwa> smaug____: btw, https://plus.google.com/u/0/105748986001435560355/posts/63GRsrNgGK3 [01:29:04.0000] <rniwa> smaug____: i made a sample code for obtaining mutation records between events [01:29:41.0000] <rniwa> the specific use case i was thinking of was for execCommands since those will how fire input event. [01:31:59.0000] <smaug____> rniwa: your createMutationObserver doesn't probably work in Nightly [01:32:06.0000] <smaug____> which has MutationObserver [01:32:15.0000] <rniwa> smaug____: yeah, saw just your prefix removal. [01:32:23.0000] <rniwa> smaug____: i'll find sometime to update it this weekend [01:33:21.0000] <smaug____> rniwa: hey, btw, since I haven't managed to figure out how to update Chromium, does the latest build handle documentFragments correctlyl [01:33:27.0000] <smaug____> correctly [01:33:36.0000] <smaug____> I mean DOM changes to documentFragment [01:35:53.0000] <rniwa> smaug____: what do you mean by "correctly"? [01:36:36.0000] <smaug____> rniwa: I mean, I couldn't get any mutation records [01:36:44.0000] <rniwa> smaug____: oh i see. [01:36:45.0000] <smaug____> when observing documentfragment [01:36:50.0000] <rniwa> smaug____: it should be fixed. [01:36:52.0000] <smaug____> and adding new child nodes [01:36:53.0000] <smaug____> ok [01:37:11.0000] <rniwa> smaug____: http://www.chromium.org/getting-involved/download-chromium [01:37:17.0000] <smaug____> /me really needs to figure out how to get up-to-date C [01:37:41.0000] <smaug____> hey, that is what I need :) [01:37:44.0000] <smaug____> rniwa: thanks! [01:37:49.0000] <rniwa> smaug____: this is the latest trunk build [01:37:53.0000] <rniwa> of chromium [01:37:57.0000] <rniwa> smaug____: but which means it won't self-update :( [01:38:07.0000] <smaug____> I don't care about that [01:38:23.0000] <smaug____> but I couldn't find that link to .zip file [01:38:25.0000] <rniwa> smaug____: if you want chromium equivalent of aurora, then go to http://www.chromium.org/getting-involved/dev-channel [01:38:38.0000] <rniwa> smaug____: http://download-chromium.appspot.com/ ? [01:38:43.0000] <smaug____> yes [01:38:47.0000] <rniwa> smaug____: no link? [01:39:05.0000] <smaug____> er, I mean, there was .zip file somewhere [01:39:10.0000] <rniwa> smaug____: oh [01:39:15.0000] <smaug____> but couldn't find the page [01:39:19.0000] <rniwa> :\ [01:39:30.0000] <rniwa> smaug____: which OS do you use? [01:39:35.0000] <smaug____> linux [01:39:37.0000] <smaug____> fedora [01:39:44.0000] <rniwa> smaug____: x64? [01:39:50.0000] <smaug____> yup [01:39:56.0000] <rniwa> smaug____: http://commondatastorage.googleapis.com/chromium-browser-snapshots/index.html?path=Linux_x64/ [01:40:01.0000] <rniwa> smaug____: (warning takes forever to load) [01:40:29.0000] <rniwa> smaug____: http://commondatastorage.googleapis.com/chromium-browser-snapshots/index.html?path=Linux_x64/139006/ is better :D [01:41:56.0000] <smaug____> (all the browsers are getting so fat nowadays) [01:43:04.0000] <rniwa> smaug____: yeah i know :\ [01:43:17.0000] <rniwa> smaug____: we need to get rid of all the features [01:43:45.0000] <zcorpan> first, get rid of this thing "HTML5" [01:43:58.0000] <smaug____> HTML2 was enough [01:44:12.0000] <smaug____> hmm, it had JS [01:44:15.0000] <smaug____> HTML1 [01:44:36.0000] <smaug____> brb [01:44:43.0000] <zcorpan> responsive images? how about no images?? [01:45:03.0000] <annevk> shit would be fast [01:45:07.0000] <rniwa> zcorpan: THAT SOLVES EVRYTHING :D [01:45:27.0000] <rniwa> zcorpan: welcome to my website: https://rniwa.com/ [01:45:55.0000] <zcorpan> <!DOCTYPE html dir="ltr" lang="en-US"> [01:45:59.0000] <rniwa> (except i still use images for math eqn :( ) [01:46:19.0000] <rniwa> zcorpan: huh, that's weird. [01:46:35.0000] <rniwa> zcorpan: it seems like my DOCTYPE and html element got smashed together :\ [01:46:51.0000] <zcorpan> nice :) [01:48:36.0000] <rniwa> /me fixes the stupid bug [01:48:47.0000] <rniwa> fixed. [01:49:12.0000] <rniwa> zcorpan: anyway, no image is the future. [01:49:38.0000] <rniwa> zcorpan: with inline svg and all that jazz, i don't know why we need images. did you also see github's latest ui change to use custom fonts for icons? [01:50:07.0000] <rniwa> zcorpan: https://github.com/blog/1135-the-making-of-octicons ? [02:45:14.0000] <david_carlisle> rniwa: There is an alternative to images for math you know... [02:45:54.0000] <Ms2ger> No way! [02:47:25.0000] <jgraham> Pencil + paper + the postal service? [02:52:52.0000] <zcorpan> use a custom font: one glyph for each formula you want to use [02:54:45.0000] <jgraham> Ah, the github solution [03:01:28.0000] <[tm]> david_carlisle: I removed the openmath stuff from the schema in my workspace but doing that makes your doc invalid [03:10:02.0000] <david_carlisle> That's OK I was waiting to see exactly what went and what you're allowing and then I shall squirrel away the openmath in some plausibly valid way:-) [03:10:39.0000] <david_carlisle> better do that now.... [03:12:52.0000] <david_carlisle> [tm]: have you pushed that out yet? is saying my existing file is valid? [03:13:26.0000] <david_carlisle> [tm]: meant to say http://validator.w3.org/nu says my existing file is valid [03:14:40.0000] <david_carlisle> sigh you did say "in my workspace " which might have been a clue it wasn't the public copy, if I could read:-) [03:24:04.0000] <david_carlisle> [tm]: Can you try now, I just updated mathml.html [03:25:28.0000] <smaug____> no, I don't understand the newline insertion [03:25:30.0000] <smaug____> hsivonen: ping [03:26:03.0000] <jgraham> smaug____: What's the question? [03:26:15.0000] <smaug____> in http://www.whatwg.org/specs/web-apps/current-work/multipage/the-end.html#html-fragment-serialization-algorithm [03:26:34.0000] <smaug____> why we want extra newline [03:26:46.0000] <smaug____> for pre and textare and listing [03:26:52.0000] <smaug____> textarea even [03:26:57.0000] <smaug____> or how should it work [03:28:35.0000] <Ms2ger> So if you parse <pre>\n\n [03:28:51.0000] <Ms2ger> You get a pre element with a text child whose data starts with one \n [03:29:29.0000] <Ms2ger> And in that case, you want to serialize two \ns [03:30:05.0000] <smaug____> if I do document.body.appendChild(document.createElement("textarea")); document.body.firstChild.appendChild(document.createTextNode("\nhello")); document.body.innerHTML [03:30:10.0000] <smaug____> what should be the result? [03:31:09.0000] <Ms2ger> <textarea>\n\nhello</textarea> [03:31:11.0000] <Ms2ger> No? [03:31:29.0000] <smaug____> it should, but that is not what browser seem to do [03:31:54.0000] <smaug____> I mean, per spec it should [03:32:03.0000] <Ms2ger> That doesn't surprise me all that much :) [03:34:47.0000] <jgraham> Anyone got any opinion on whether javascript URIs should be dereferenced in a sync way or async way? [03:35:06.0000] <jgraham> The spec seems to say async like other URIs [03:35:15.0000] <jgraham> and gecko seems to mostly do that [03:35:30.0000] <jgraham> But WebKit seems to make it sync in at least some cases [03:35:57.0000] <jgraham> Sync has the advantage of being less racy [03:36:05.0000] <jgraham> But the disadvantage of being more magic [03:36:52.0000] <smaug____> wasn't javascript: sync in gecko and then changed to async [03:36:57.0000] <smaug____> bz would know that [03:37:02.0000] <jgraham> Maybe? [03:37:13.0000] <smaug____> /me checks the blame [03:37:16.0000] <jgraham> Inconveniently, bz isn't here [03:37:49.0000] <smaug____> it would be quite early for bz [03:38:24.0000] <jgraham> Or quite late :p [03:38:45.0000] <jgraham> In any case he doesn't use #whatwg [03:38:53.0000] <smaug____> jgraham: https://bugzilla.mozilla.org/show_bug.cgi?id=351633 [03:40:22.0000] <jgraham> smaug____: Thanks [03:52:02.0000] <AryehGregor> "The font element must override the color of any text decoration that spans the text of the element to the used value of the element's 'color' property. Note: This applies in all modes." zcorpan, if it applies in all modes, then why is it in a spec entitled "Quirks Mode"? [03:52:48.0000] <Ms2ger> At least Gecko still has it in quirks only [03:54:44.0000] <zcorpan> AryehGregor: it was quirks-only in the spec first but i changed it [03:55:27.0000] <zcorpan> AryehGregor: ideally all requirements in the quirks mode spec are moved to the "main" specs that define the features involved [03:55:41.0000] <jgraham> /me hopes that note isn't normative [03:55:43.0000] <AryehGregor> I'm against any legacy HTML feature not aligning exactly with an equivalent CSS feature if we can possibly avoid it, so I vote that it be quirks-only. Does it really cause compat problems if it's not? [03:55:55.0000] <AryehGregor> jgraham, no -- every other thing in the spec says "In quirks mode, . . ." [03:56:00.0000] <AryehGregor> So the note is redundant. [03:56:03.0000] <jgraham> Good :) [03:56:20.0000] <jgraham> /me would be very surprised if zcorpan had not got that right [03:56:32.0000] <Ms2ger> I object to it being quirks-only :) [03:59:10.0000] <zcorpan> AryehGregor: it applies in all modes in webkit, which suggests we can get away with it not being quirks-only :-) [03:59:45.0000] <AryehGregor> Ah, I see -- having legacy features not match CSS is bad, but having differences between quirks and standards is worse. [03:59:58.0000] <AryehGregor> Okay, I changed my mind, I'm in favor of it being true in all modes. [04:00:27.0000] <zcorpan> it's a trade-off, of course. depends on how ugly the quirk is [04:00:52.0000] <zcorpan> for instance mozilla want window.foo to be limited to quirks mode [04:01:03.0000] <Ms2ger> Pff [04:01:05.0000] <zcorpan> or some mozilla people [04:01:12.0000] <zcorpan> i know y'all aren't one entity [04:01:13.0000] <Ms2ger> We can want all we want, but we gave up [04:01:29.0000] <Ms2ger> Thank you, Google! [04:07:07.0000] <jgraham> So is the whole "Mozilla are individuals" thing supposed to make me imagine that Mozilla are starfleet and everyone else are the Borg? Should I imagine that Mozilla HQ is like the enterprise with Gary Kovacs as Picard and Brendan Eich as Geordi Laforge? When your senior crew, I mean staff, leave for Facebook do you talk about them being "assimilated"? [04:07:45.0000] <Ms2ger> You know what our meeting rooms are called, right? [04:07:51.0000] <jgraham> And, most importantly in this scenario, who is Wesley Crusher? [04:08:04.0000] <jgraham> Because they deserve to be punched [04:08:42.0000] <jgraham> Ms2ger: I assume given context that they must be star trek based, although iirc the smaller ones are internet memes [04:09:05.0000] <Ms2ger> http://25.media.tumblr.com/tumblr_m4gatutgsA1rrf1eeo1_400.jpg [04:11:00.0000] <jgraham> So basically, I *am* supposed to have this picture [04:11:19.0000] <Ms2ger> No comment [04:11:24.0000] <jgraham> /me assumes Ms2ger is a random redshirt [04:15:13.0000] <AryehGregor> Does anyone here have ready access to Word? If so, could you do a quick test for me? [04:15:41.0000] <AryehGregor> 1) Type "foo\nbar\nbaz". 2) Select all and make bold. 3) Delete the three letters "bar", so the middle line is empty. 4) Move the cursor someplace else, then move it back to the empty middle line. 5) Type "quz". 6) Is "quz" bold or not? [04:15:58.0000] <AryehGregor> (also tell me what version) [04:16:10.0000] <AryehGregor> (my wife has Word on her laptop, but I can't figure out how to turn the laptop on . . .) [04:30:32.0000] <jgraham> AryehGregor: Look for a button on the outer casing? [04:30:51.0000] <AryehGregor> jgraham, I see a power button! But pushing it doesn't seem to do anything . . . [04:30:56.0000] <AryehGregor> Thanks for the tip, though! I appreciate it. [04:31:03.0000] <Philip`> Is it plugged in? [04:31:34.0000] <AryehGregor> I mean, my grandfather would reportedly call my father sometimes to get a step-by-step walkthrough on how to turn his computer on. [04:31:42.0000] <AryehGregor> No reason for you to think I'm any different. [04:31:43.0000] <Philip`> Could you ask your wife? [04:31:52.0000] <AryehGregor> I will when she gets back. [04:32:23.0000] <AryehGregor> However, I was hoping that maybe someone here had a copy of Word handy, which might be faster. [04:33:04.0000] <Philip`> /me doesn't :-( [04:34:00.0000] <Philip`> (It's strange being in a community where most people don't have access to probably the second most widely installed piece of software in the world) [04:34:33.0000] <jtcranmer> most people I know don't have it installed [04:34:50.0000] <asmodai> What about Word? [04:34:54.0000] <asmodai> I have 2010 installed [04:35:02.0000] <AryehGregor> [120525 14:16:48] <AryehGregor> Does anyone here have ready access to Word? If so, could you do a quick test for me? [04:35:02.0000] <AryehGregor> [120525 14:17:16] <AryehGregor> 1) Type "foo\nbar\nbaz". 2) Select all and make bold. 3) Delete the three letters "bar", so the middle line is empty. 4) Move the cursor someplace else, then move it back to the empty middle line. 5) Type "quz". 6) Is "quz" bold or not? [04:35:02.0000] <AryehGregor> [120525 14:17:33] <AryehGregor> (also tell me what version) [04:35:06.0000] <AryehGregor> asmodai, thanks. :) [04:35:54.0000] <asmodai> yes, bold [04:37:14.0000] <asmodai> on 2010 [04:37:49.0000] <asmodai> I think the begin and end mark for bold are at f and z, respectively, so deleting the middle text doesn't change that [04:40:33.0000] <AryehGregor> Right. [04:43:13.0000] <AryehGregor> asmodai, https://bugs.webkit.org/show_bug.cgi?id=81656#c10 [04:43:39.0000] <asmodai> Awesome [05:19:07.0000] <annevk> Ms2ger: is there a way for Anolis to recognize [VERSION] and such but not complain about references being in a single section? [05:19:35.0000] <annevk> Ms2ger: also, having a flag for First Public Working Draft would be nice [05:19:47.0000] <annevk> although maybe it's not worth optimizing for [05:27:34.0000] <gsnedders> annevk: What's different for FPWD? [05:30:34.0000] <annevk> gsnedders: SotD contains "First Public Working Draft" [05:30:49.0000] <annevk> (yes it is insane) [05:41:30.0000] <david_carlisle> fpwd don't have to have a "previous version" (obvious but true:-) [05:51:02.0000] <odinho> gsnedders: Oh, that's sooo nice. http://ecma-international.org/ecma-262/5.1/ Didn't see it till now. [05:52:09.0000] <annevk> first have to see what they're releasing ES6 in [05:52:28.0000] <annevk> it seems like this was a transformation done by someone, similarly to http://es5.github.com/ [05:52:34.0000] <annevk> not done by the editors of the spec... [05:53:32.0000] <gsnedders> annevk: That's what was said in the email to es-discuss [05:53:42.0000] <gsnedders> The *big* deal here is that ECMA agreed to publish it. [05:54:02.0000] <odinho> It's on their site. [05:54:19.0000] <gsnedders> Still, in the event of any dispute, the PDF version is normative, which makes sense. [05:54:21.0000] <odinho> one small step ... etc [05:54:59.0000] <gsnedders> (Reviewing a the transformation to HTML would be a lot of work to ensure there are no differences) [05:54:59.0000] <Ms2ger> annevk, doesn't --w3c-compat-substitutions get you that? [05:57:45.0000] <annevk> Ms2ger: ah good to know [06:01:01.0000] <matjas> annevk: it’s https://github.com/jorendorff/es-spec-html and it looks like this will be used for ES6 as well [06:01:28.0000] <matjas> i actually configured the ecma-international.org server the other day — they were sending the SVG figure as text/plain which broke things [06:03:22.0000] <gsnedders> The interesting question is whether anyone can make a stylesheet for the HTML copy that meets the publishing format rules, of cover pages and such like. [06:04:14.0000] <Ms2ger> Well, the more interesting question is whether ECMA will ever make those rules available [06:04:59.0000] <Ms2ger> /me gets his towel out [06:05:14.0000] <gsnedders> But that'd allow non-members to participate! [06:05:37.0000] <gsnedders> And how are we meant to stay financially viable unless members get something for their money!? [06:06:09.0000] <Ms2ger> In the case of ECMA? [06:06:14.0000] <Ms2ger> I suspect "not" [06:06:45.0000] <Philip`> gsnedders: Make everything free and open, but encourage bribery [06:07:06.0000] <gsnedders> Ms2ger: Yeah, in the case of Ecma. [06:07:30.0000] <gsnedders> The only thing you get, in the case of ECMAScript, is to get to go to F2F meetings. That is literally the only gain of being an ECMA member. [06:07:48.0000] <jgraham> And you can submit tests [06:08:30.0000] <gsnedders> Though Mozilla people have in principle agreed to submit third-party tests. [06:08:39.0000] <gsnedders> Provided they meet the licensing requirements. [06:08:45.0000] <gsnedders> So even that gain is questionable. [07:35:18.0000] <odinho> Who did/knows IDLharness? [07:35:58.0000] <jgraham> AryehGregor [07:36:05.0000] <odinho> Trying to make a interface test for IDB, but it's really not all roses. [07:36:28.0000] <odinho> AryehGregor: // XXX: [17:25:40.864] uncaught exception: Window implements IDBEnvironment, but one is undefined or not an interface [07:36:33.0000] <odinho> Window implements IDBEnvironment; [07:37:25.0000] <odinho> It doesn't like that. And I have an [NoInterfaceObject] interface IDBEnvironment { readonly attribute IDBFactory indexedDB; }; [07:37:54.0000] <odinho> Tried removing [NoInterfaceObject] and other stuff, but I don't really know where to start at all. [07:38:19.0000] <AryehGregor> odinho, I'm about to run now -- could you send an e-mail? I should be able to look at it Monday (UTC+0300). At a glance, I'd guess the problem is that "Window" isn't defined -- you have to include the declarations of any interfaces you depend on too. [07:38:30.0000] <AryehGregor> You can include them in a fashion that they don't get tested themselves. [07:39:01.0000] <odinho> AryehGregor: Ah, I can look at that. It's a way I haven't explored, it might fix it, yea :] [07:39:06.0000] <odinho> AryehGregor: Thanks. [07:40:09.0000] <AryehGregor> odinho, look here for inspiration: http://dvcs.w3.org/hg/html/file/0bc4307c200a/tests/submission/AryehGregor/interfaces.html Note use of "add_untested_idls" -- this is to avoid double-testing the DOM IDLs, which the HTML IDLs depend on. [07:45:47.0000] <odinho> AryehGregor: Wohoo! :D [09:30:16.0000] <TabAtkins> zcorpan: Yeah, responsive tables of that variety have been seriously discussed before, but didn't have sufficient momentum to get one of us to work on them. [09:30:57.0000] <TabAtkins> zcorpan: But the idea of pinning rows/columns is very cool and obviously useful. Mobile presentation of tables just drops the threshold for how large of a table it's useful on. [09:36:08.0000] <Hixie> zcorpan: patches welcome. :-) [09:37:06.0000] <tantek> welcome back Hixie [09:37:14.0000] <Hixie> still on vacation til tuesday [09:37:20.0000] <Hixie> but back in the area and on the net [09:37:36.0000] <tantek> good for you for taking vacation [09:37:50.0000] <Ms2ger> And bad for us :) [09:55:34.0000] <Ms2ger> Anyone around here who cares about <script type>? [09:59:02.0000] <TabAtkins> "cares" in what sense? [09:59:59.0000] <Ms2ger> About the supported values [10:00:03.0000] <TabAtkins> Nope. [10:13:07.0000] <Philip`> Is there any good reason why someone would send dozens of packets per second to a server's HTTPS port, for many hours, when that server doesn't have anything responding to that port? [11:38:24.0000] <TabAtkins> Ms2ger: D'oh, I had a good snarky comment for that bug. Damn your quick AI reflexes! [11:38:43.0000] <Ms2ger> :) [11:42:06.0000] <Hixie> which bug? :-) [11:42:15.0000] <TabAtkins> A webapps bug. [11:42:25.0000] <TabAtkins> Entitled: "Looking for the music I downloaded" or similar. [11:42:31.0000] <Hixie> aah [11:42:36.0000] <annevk> Ms2ger: is bz saying Gecko wants to implement the spec or something else? [11:42:58.0000] <Ms2ger> The spec [11:43:13.0000] <annevk> k [11:43:19.0000] <Ms2ger> But since nobody else does, we'd like to hear if anybody objects [12:06:06.0000] <Ms2ger> Hmm, wasn't [ArrayClass] interface NodeList not going to fly? [12:06:44.0000] <annevk> I believe WebKit was trying [12:07:37.0000] <Ms2ger> Also [12:07:38.0000] <Ms2ger> http://www.whatwg.org/specs/web-apps/current-work/multipage/common-dom-interfaces.html#htmlpropertiescollection-0 [12:07:39.0000] <annevk> and I believe they did run into issues, but not so much that people were admitting defeat... [12:08:02.0000] <annevk> Ms2ger: ? [12:08:04.0000] <Ms2ger> Would you consider the "represents" in the first paragraph there as overriding http://dvcs.w3.org/hg/domcore/raw-file/tip/Overview.html#represented-by-the-collection [12:08:07.0000] <Ms2ger> ? [12:09:10.0000] <annevk> Ms2ger: no, but it's confusing [12:09:26.0000] <annevk> Ms2ger: represents in DOM applies to collection; in HTML it applies to interface... [12:09:38.0000] <Ms2ger> Yeah, I thought so [12:09:44.0000] <Ms2ger> It confused bz as well [12:27:23.0000] <Ms2ger> ojan_away, s/you're feeling/your feeling/ :) [12:27:43.0000] <annevk> btw [12:27:45.0000] <annevk> did bz test [12:27:52.0000] <annevk> language=javascript1.0 [12:28:13.0000] <Ms2ger> We're not touching language right now [12:28:30.0000] <Ms2ger> Does anybody forward language to type like the spec claims? [12:28:32.0000] <annevk> the whole point of the list in HTML is that it's the same for type and language [12:28:41.0000] <annevk> I think we might [12:28:49.0000] <annevk> not sure though [12:30:26.0000] <Ms2ger> http://software.hixie.ch/utilities/js/live-dom-viewer/saved/1541 suggests Opera doesn't [14:05:18.0000] <zewt> nothing speaks of competence in web design as much as phone number entry boxes that swallow keystrokes, breaking browser hotkeys [14:08:45.0000] <Hixie> am i missing something here? UTF-8 byte 0xED followed by two continuation bytes 0x80 0x80 is equivalent to U+D000 right? [14:10:26.0000] <gsnedders> That's non-shortest form, no? [14:10:41.0000] <Hixie> no? [14:11:24.0000] <gsnedders> Yeah, that is U+D000, except 0xED 0x80 0x80 is not a valid UTF-8 string. [14:11:36.0000] <Hixie> how can it be anything but shortest form if all the continuation bytes are 0x80 and the first byte has its second nibble set to anything but 0 [14:11:46.0000] <gsnedders> What, that's wrong. [14:11:56.0000] <gsnedders> Ignore me, ignore me! [14:11:59.0000] <Hixie> ok [14:12:00.0000] <Hixie> phew [14:12:01.0000] <gsnedders> I'm not even going to try and answer that. [14:12:02.0000] <TabAtkins> Yes, it's U+d000 [14:12:07.0000] <Hixie> i'm confused enough already as it is :-P [14:12:29.0000] <gsnedders> I seem to just be speaking nonsense so answering a harder question will likely result in more of that. [14:14:00.0000] <TabAtkins> Hixie: Why do you ask? [14:21:38.0000] <jgraham> zewt: Half the fucking web thinks it's OK for "/" to be used for "focus site search box" [14:28:27.0000] <zewt> jgraham: this is worse--probably cancelling onkeydown if it's not 0-9 [14:28:48.0000] <zewt> which means every browser hotkey stops working (why browsers allow that I can't fathom) [14:29:12.0000] <zewt> not to mention pasting, which is ... sort of basic for an "enter bank phone number" field [14:37:49.0000] <jamesr_> man i really wish you could declaratively say which keys you care about when registering a JS key listener [14:38:02.0000] <jamesr_> so the browser could avoid having to bounce through the page and hope that it didn't fuck up the preventDefault() [14:38:07.0000] <TabAtkins> a;lksdjf;al WANT IT SO BAD [14:38:32.0000] <jamesr_> window.addEventListener("keydown", flobberize(), false, ['/']); ? [14:38:44.0000] <jamesr_> s/flobberize()/flobberize/ [14:38:59.0000] <TabAtkins> Hey, you don't know. Maybe flobberize() returns a function. [14:41:26.0000] <jamesr_> i want it for all event types [14:41:33.0000] <jamesr_> like mouse listeners should say that they only care about right click or whatever [14:41:35.0000] <zewt> jamesr_: well, it doesn't really fit the dom event model, but could have element.addKeyListener(func, keySpecifier, capture) that's functionally equivalent to an event listener with a key check (except easier for the browser to optimize) [14:42:07.0000] <jamesr_> yeah, the model is you fire everything and the handler figures out if it cares or not [14:42:17.0000] <jamesr_> which means if we want to optimize it in the browser we have to do it on an event type level [14:42:30.0000] <zewt> in a sense make it so you can attach a sub-event-level filter to each event handler [14:42:33.0000] <jamesr_> i.e. define a new set of events for each interaction [14:42:56.0000] <jamesr_> even applying an after-the-fact filter might work [14:43:08.0000] <zewt> elem.addEventListener("keydown", func, false, {filter: {keyCode: someKindOfFilterLanguage}}) [14:44:11.0000] <zewt> personally i think it's inefficient but mostly only academically, so I'm not too bothered by it [14:45:10.0000] <jamesr_> it's not quite theoretical for things like pagedown [14:45:25.0000] <jamesr_> we would really like to scroll the page when you hit page down without having to wait for the thread running JS/DOM/etc [14:45:34.0000] <jamesr_> but if the page listens to key events, we dunno if it's going to preventDefault() the pagedown [14:46:05.0000] <jamesr_> if the page really just wants to hijack '/' - merits of doing that aside - it'd help us out to know that in the browser [14:48:13.0000] <rafaelw_> Ms2ger: question for you about your Document.parse() WebKit patch comments.... [14:48:51.0000] <jwalden> rafaelw_: gone [14:48:59.0000] <jwalden> (nn, to be precise) [14:48:59.0000] <rafaelw_> ah. thanks. [14:49:01.0000] <jamesr_> it's probably just a pipe dream though :( [14:49:26.0000] <TabAtkins> I have no idea why you think it's a pipe dream. [14:49:38.0000] <TabAtkins> People have watned it for years, just no one's cared enough to make it happen. [14:49:40.0000] <TabAtkins> MAKE IT HAPPEN. [14:49:41.0000] <zewt> jamesr_: if you can think of a static filter language that would work... [14:49:42.0000] <TabAtkins> Easy. [14:49:56.0000] <zewt> i mean, i know of some, like mongodb's queries, but it's not the cleanest [14:50:03.0000] <jamesr_> zewt, that's one tricky part of the problem [14:50:08.0000] <jamesr_> list of keycodes? [14:50:14.0000] <zewt> eg. {$or: [{keyCode: 1}, {keyCode: 2}]} [14:50:19.0000] <zewt> i mean, to be generic for any event [14:50:34.0000] <jamesr_> it gets messy quickly [14:51:15.0000] <jamesr_> and someone always goes "can i just pass a function?" then you have to explain how it picks up the global context and the thread implications, then someone wants a JS isolate, then you're in a world of hurt [14:51:32.0000] <zewt> well, mongo's query language is static [14:51:54.0000] <zewt> the above can be optimized (and even if it's not, it's a lot faster to convert it to a native object and just test them all, since you don't have to fire off JS) [14:52:21.0000] <TabAtkins> Yes, take some object with values. Eagerly parse them to dispose of getters. Done. [14:52:47.0000] <zewt> (idl dictionary) [15:16:03.0000] <Hixie> TabAtkins: my utf-8 validating code isn't vorking [15:16:24.0000] <TabAtkins> Hixie: Interesting. Should be trivial! [15:16:40.0000] <Hixie> specifically, it thinks 0xED 0x9F 0xBF is invalid, because it decodes it to a bogus value [15:17:52.0000] <Hixie> my code for decoding sequences starting with 0xED is equivalent to 0xD000 + ((c1-0x80) << 6) + (c2-0x80) [15:17:56.0000] <Hixie> am i missing something? [15:19:14.0000] <TabAtkins> Assuming that c1 and c2 are guaranteed to be in [00,ff], and you're already making sure that the second bit in each is 0, that should be fine. [15:19:31.0000] <TabAtkins> Better would be to & each with 0x00111111. [15:19:45.0000] <Hixie> each one is known to be in the range 0x80 .. 0xBF [15:20:22.0000] <TabAtkins> Just to be safe, try swapping your "-0x80" with "&63". [15:23:39.0000] <Hixie> that (thankfully) made no difference [15:24:06.0000] <TabAtkins> Okay, good. Just sanity-checking. [15:24:16.0000] <TabAtkins> What's the bogus value it's decoding to? [15:24:56.0000] <Hixie> 0xDB5F [15:25:23.0000] <Hixie> oh i see the problem [15:25:29.0000] <Hixie> off by one error [15:25:35.0000] <Hixie> don't mind me [15:25:56.0000] <TabAtkins> uh, okay. [15:26:24.0000] <Hixie> c2 contained c1 and c1 contained c0 [15:26:30.0000] <Hixie> the leading byte [15:27:15.0000] <Hixie> that resulted in disturbingly few of my tests failing [15:27:25.0000] <Hixie> but that's another story [15:27:27.0000] <Hixie> ok, bbiab [15:29:15.0000] <TabAtkins> Oh, duh, I should have noticed that as soon as I looked at it. [15:33:59.0000] <rniwa> hsivonen: yt? [15:34:30.0000] <rniwa> Hixie: do you know how well tested the part it says to throw SYNAX_ERR in http://html5.org/specs/dom-parsing.html#extensions-to-the-range-interface ? [15:34:40.0000] <rniwa> when the element name is html, body, etc...? [15:53:29.0000] <rafaelw_> Hixie; Where are you WRT Document.parse()? [15:54:27.0000] <rafaelw_> I'd really like to get this in. I'm wondering if you still have strong objections or not. [16:00:13.0000] <espadrine> Firefox cannot redefine Element.prototype.scrollIntoView, while webkit can. Firefox's wrong, right? [16:22:00.0000] <jwalden> espadrine: wrong assuming you consider webidl authoritative now, yes [16:22:14.0000] <jwalden> espadrine: DOM bindings are being rewritten now and should end up fixing that [16:22:41.0000] <espadrine> jwalden: oh ok, thanks! 2012-05-26 [23:52:09.0000] <tantek> who has admin access to the wiki? [23:52:12.0000] <tantek> e.g. to install extensions [23:52:52.0000] <tantek> perhaps annevk? [23:53:09.0000] <tantek> would be great to get the SemanticHTML extension installed [23:53:11.0000] <tantek> https://github.com/microformats/mediawiki-semantic-html [23:53:39.0000] <annevk> I don't think I can ssh into the wiki [23:53:57.0000] <tantek> it allows use of more semantic HTML markup tags on MediaWIki pages, useful for new tags in particular like <time> and <data> [23:53:58.0000] <tantek> ah ok [23:54:14.0000] <tantek> we're running it live on the microformats wiki if anyone wants to experiment [23:54:51.0000] <tantek> annevk, ok, np. I'll wait til Hixie is official back from vacation and ask him (though I'm sure he's going to be overrun by a lot of pent-up requests). [23:56:02.0000] <tantek> *officially [23:56:04.0000] <annevk> Lachy might have access too [23:56:13.0000] <annevk> but waiting for Hixie works [23:56:23.0000] <tantek> not in a rush so that's fine [23:56:33.0000] <tantek> just a nice convenient incremental upgrade [23:56:41.0000] <tantek> plus, hey, more HTML5 :) [23:57:18.0000] <tantek> is there an implementations or tools page somewhere where we keep track of HTML5 implementations / tools? [23:57:33.0000] <annevk> yeah [23:57:43.0000] <tantek> (I figured I would ask before creating one) [23:57:51.0000] <annevk> http://wiki.whatwg.org/wiki/Category:Implementations [23:58:00.0000] <tantek> ok - so any page tagged with that category [23:58:23.0000] <annevk> all major things are listed from the front page [23:58:24.0000] <tantek> or would it be better to make a publishing page [23:58:32.0000] <tantek> for HTML5 publishing tools [23:58:35.0000] <annevk> at some point I turned a few into categories to make it easier to group stuff [23:58:50.0000] <tantek> whatever grouping/organization is your preference [23:59:21.0000] <annevk> I don't think we have quite enough implementation pages to start separating them further [23:59:29.0000] <tantek> ok [00:02:10.0000] <Ms2ger> AryehGregor probably also has access [01:12:34.0000] <Stevef> tantek: a heads up, I mentioned <hgroup> to you at TPAC last year, seems like the HTML WG chairs will soon provide some feedback on Issue 164 http://dev.w3.org/html5/status/issue-status.html#ISSUE-164 if you have any interest in keeping hgroup in HTML5 you may want to keep an eye on this issue. there are 5 proposals 4 of which proposae replacing hgroup. As there has been no real data or... [01:12:36.0000] <Stevef> ...argument from the WHATWG side on keeping hgroup I would be surprised if it survives. [01:25:17.0000] <Stevef> hober: ping re: http://lists.w3.org/Archives/Public/public-html/2012May/0008.html [03:05:32.0000] <Ms2ger> Velmont, http://dvcs.w3.org/hg/webapps/rev/ec3740e46700#l8.1 doesn't look like it should have been there [07:05:04.0000] <Velmont> Ms2ger: wtf! grrrrrr.... i am not happy with hg [07:05:37.0000] <Velmont> sorry about that [07:08:19.0000] <gsnedders> Velmont: Deep breaths. [07:57:21.0000] <Ms2ger> Velmont, (reverted it, btw) [08:21:44.0000] <Velmont> Ms2ger: thanks. yea I am always surprised by hg behaviour. have to be more careful [09:18:11.0000] <GPHemsley> If there is a runaway <a> element that is inside a block element like <p> or <div>, shouldn't the closing of that element close the <a>? [10:19:49.0000] <gsnedders> Hate __proto__. Hate it. [10:25:20.0000] <Ms2ger> Who does e4x again. [10:25:20.0000] <Ms2ger> ? [10:25:50.0000] <gsnedders> SpiderMonkey. [10:26:32.0000] <Ms2ger> How silly of us [10:27:34.0000] <gsnedders> Non-strict only, IIRC [10:27:52.0000] <gsnedders> Ms2ger: You and all your ES extensions! [10:27:55.0000] <Ms2ger> Possible [10:28:15.0000] <Ms2ger> Pff [10:28:25.0000] <Ms2ger> I prefer to think of ES as a JS extension ;) [10:28:47.0000] <gsnedders> ES is a subset of JS :P [10:28:50.0000] <Ms2ger> After all, Ope^W Mozilla did it first [10:29:07.0000] <gsnedders> Operabook^W ? :) [10:37:30.0000] <gsnedders> Why am I writing emails to es-discuss on a Saturday? [10:45:37.0000] <Ms2ger> You've got exams coming up? [10:46:38.0000] <gsnedders> Nah, just going out soon. And it's the weekend, a break is nice at times. [13:12:29.0000] <cbright6062> /me reads to see if he's missed anything interesting. [13:12:30.0000] <cbright6062> hm. 2012-05-27 [01:39:18.0000] <annevk> MikeSmith: https://www.w3.org/Bugs/Public/show_bug.cgi?id=17205 seems like actual spam [01:39:33.0000] <annevk> MikeSmith: different from the contributor⊙wo spam that is [03:06:51.0000] <sorenso> list [03:06:58.0000] <sorenso> Sorry :) [05:34:00.0000] <gsnedders> webben: That's similar to extensions breaking if you disable JS :P [05:48:09.0000] <webben> can't reproduce in opera next [05:48:18.0000] <webben> (the broken inspector) [09:23:08.0000] <matjas> some-element::before { content: '\3A' } → what should getComputedStyle(element, '::before').content say? ':', "':'", or something else? browser behavior varies [09:24:25.0000] <matjas> Gecko _always_ wraps the return value in double quotes [09:24:28.0000] <Ms2ger> Hmm [09:25:08.0000] <matjas> WebKit usually doesn’t, but seems to wrap the return value in _single_ quotes when a simple char escape is used (e.g. `\:`) [09:26:14.0000] <Ms2ger> Gecko is right per CSSOM [09:27:16.0000] <matjas> http://dvcs.w3.org/hg/csswg/raw-file/tip/cssom/Overview.html#dom-window-getcomputedstyle [09:28:34.0000] <matjas> Ms2ger: so, http://dvcs.w3.org/hg/csswg/raw-file/tip/cssom/Overview.html#resolved-value? “The resolved value is the computed value.” [09:28:41.0000] <Ms2ger> http://dev.w3.org/csswg/cssom/#serialize-a-string [09:28:43.0000] <matjas> where does it say the quotes are part of the computed value [09:28:46.0000] <matjas> ah [09:29:21.0000] <matjas> Ms2ger: thanks! [09:29:25.0000] <Ms2ger> Np [09:29:30.0000] <Ms2ger> Now, a test! ;) [09:30:28.0000] <matjas> Ms2ger: yeah, i was working on some :) [09:31:27.0000] <Ms2ger> Going to submit them? :) [09:31:54.0000] <matjas> Ms2ger: sure, if anyone’s interested [09:32:21.0000] <Ms2ger> I am, does that count? :) [09:32:31.0000] <matjas> haha, sure [09:38:08.0000] <matjas> Ms2ger: where does it say that computed values must be serialized when returned through `getComputedValue`? [09:53:44.0000] <matjas> Ms2ger: http://mathias.html5.org/tests/css/escape-sequences/ [09:55:16.0000] <matjas> suggestions for additional test strings welcome [10:53:40.0000] <kennyluck> Given that CSSOM is an unmaintained spec, I wouldn't say it's trustworthy. [11:38:00.0000] <Ms2ger> matjas, it is implied in "The getPropertyValue(property) method must ...." [15:57:39.0000] <smaug____> where is the spec which defines innerHTML [15:58:18.0000] <smaug____> (Ms2ger isn't here) [15:58:45.0000] <smaug____> ah, http://html5.org/specs/dom-parsing.html 2012-05-28 [00:47:24.0000] <smaug____> hsivonen: ping [02:15:06.0000] <jgraham> New theory: document.write is only the second worst part of the platform and window.open is the worst [02:16:38.0000] <Ms2ger> Can I enter HTML serialization as a contestant? [02:17:14.0000] <Ms2ger> If I have a DOM like <div><input><span></span></input></div>, what is div.innerHTML? [02:19:39.0000] <jgraham> If I window.open a window and there is an alert in the opener window, what happens in the new window? What about if the user has explicitly navigated the opened window? [02:35:18.0000] <zcorpan> Ms2ger: <input> [02:36:06.0000] <Ms2ger> Per spec and in Opera, yes [02:36:24.0000] <zcorpan> do all other browsers disagree? [02:36:33.0000] <Ms2ger> Gecko has "<input><span></span>" [02:36:43.0000] <Ms2ger> Chrome has "<input></input>" [02:37:06.0000] <zcorpan> i don't object to changing the spec what gecko has. chrome's serialization seems weird though [02:37:43.0000] <Ms2ger> I would object to changing to what Gecko has :) [02:37:49.0000] <zcorpan> why? [02:38:19.0000] <Ms2ger> Because I think it's rather weird :) [02:39:17.0000] <zcorpan> not weirder than non-text nodes in <style> [02:39:27.0000] <Ms2ger> Hmm, IE10 seems to match Gecko [02:39:41.0000] <Ms2ger> /me looks at what style does [02:40:29.0000] <zcorpan> we can't make innerHTML be "correct" for all cases because the dom can contain lots of things that are not serializable [02:40:35.0000] <Ms2ger> Yeah [02:47:59.0000] <Ms2ger> /me wanders off for a bit, before innerHTML gives him a headache [07:50:32.0000] <annevk> todo: add event handler attributes to http://wiki.whatwg.org/wiki/Howto_spec [07:59:18.0000] <annevk> context: http://lists.w3.org/Archives/Public/public-device-apis/2012May/thread.html#msg283 [14:39:36.0000] <asmodai> I long for a day where user agent strings are meaningful again [14:44:30.0000] <Philip`> The only reason for them to be meaningful is if you're trying to extract meaning from them, in which case you're presumably trying to use that meaning to distinguish browsers and handle them differently in some way, in which case a future browser will benefit from lying and saying something that it doesn't really mean but that produces the desired behaviour [14:44:51.0000] <Philip`> so your longed-for day will not be a stable equilibrium [14:45:21.0000] <Philip`> so there will be no long-term gain and plenty of short-term pain [14:45:32.0000] <Philip`> so we are doomed :-( [14:47:12.0000] <asmodai> heh [14:48:36.0000] <espadrine> wow. That was the most perfect answer possible. [14:50:50.0000] <Philip`> (The only stable equilibrium is when every browser is identical in every way that web pages can distinguish, so the user agent string must convey zero bits of information, e.g. by being the union of every UA string used by every other UA in the history of the web) [14:51:40.0000] <wilhelm> Speaking of useless headers – we should do a purge of both request and response headers. [14:53:02.0000] <asmodai> In my case with meaningful I just meant name + version and no using it for anything special. [14:54:01.0000] <Philip`> What's the point of knowing that information if you're not going to use it? [14:55:17.0000] <asmodai> Philip`: Only for seeing adoptation rates and such. [15:46:00.0000] <gsnedders> wilhelm: annevk was trying 2012-05-29 [00:04:32.0000] <Ms2ger> jgraham, yt? [00:12:27.0000] <matjas> zcorpan: wtf :'( [00:14:13.0000] <matjas> zcorpan: is XML allowed? [00:17:29.0000] <zcorpan> matjas: it's not #xmlpubquiz :-P [00:17:40.0000] <matjas> fffffuuuu [00:17:53.0000] <zcorpan> matjas: but i could go there i suppose... [00:18:10.0000] <matjas> data:text/xml,<x/> [00:18:11.0000] <matjas> document.documentElement.innerHTML = '<head/><head/><body/><body/>'; [00:18:32.0000] <Ms2ger> Hah [00:19:50.0000] <matjas> the Siamese DOM doesn’t show up in Dragonfly, but it does in the Chrome dev tools: http://i.imgur.com/xa9L5.png [00:20:34.0000] <Ms2ger> document.head.outerHTML = "." [00:20:34.0000] <Ms2ger> document.body.outerHTML = " " [00:21:43.0000] <matjas> Ms2ger: what the actual fuck [00:22:36.0000] <Ms2ger> /me bows [00:23:16.0000] <Ms2ger> zcorpan, do you know who I should bug about http://dvcs.w3.org/hg/html/file/0bc4307c200a/tests/submission/Opera/microdata/001.html ? [00:26:07.0000] <zcorpan> Ms2ger: you mean, who wrote the tests? [00:26:29.0000] <Ms2ger> Or who could review changes [00:27:07.0000] <MikeSmith> matjas: merged your pull request to add links to the TC39 HTML-formated spec [00:27:10.0000] <zcorpan> i've reviewed that test before so i guess i can do it again [00:27:19.0000] <MikeSmith> matjas: sorry for taking so long [00:27:24.0000] <matjas> MikeSmith: yay! thanks [00:28:25.0000] <zcorpan> Ms2ger: the last two checkins? [00:28:37.0000] <Ms2ger> zcorpan, no, a patch from David Zbarsky in our bug [00:28:51.0000] <Ms2ger> https://bugzilla.mozilla.org/show_bug.cgi?id=591467 [00:31:14.0000] <zcorpan> there's a David Zbarsky? did dbaron and bz merge? [00:33:59.0000] <padenot> bz's brother [00:34:36.0000] <zcorpan> i see [00:34:39.0000] <jgraham> Although the idea of dbaron and bz merging is quite terrifying [00:35:37.0000] <zcorpan> Ms2ger: is https://bugzilla.mozilla.org/attachment.cgi?id=627852&action=diff what i should be looking at? [00:35:48.0000] <Ms2ger> Yep [00:36:27.0000] <Ms2ger> jgraham, not the least for us, do you know how much work they get done? :) [00:36:33.0000] <zcorpan> line 6 and 7 is not OK [00:36:41.0000] <Ms2ger> Yeah [00:37:25.0000] <jgraham> Ms2ger: Maybe if they merged it would create someone who was 4 times as productive [00:37:35.0000] <jgraham> A super browser-engine god to rule over us all [00:38:28.0000] <Ms2ger> I think trying to rope in bz's two other brothers might be safer :) [00:38:32.0000] <matjas> zcorpan: the old line 6 and 7 weren’t okay as per http://www.w3.org/html/wg/wiki/Testing/Authoring/#Javascript_tests either (no type="text/javascript") [00:39:12.0000] <matjas> (all this is new to me) [00:40:42.0000] <jgraham> type is really not a requirement [00:44:22.0000] <Ms2ger> jgraham, the wiki page can be read as forbidding type [00:44:23.0000] <jgraham> matjas: (but I don't see what on that page you are referring to) [00:44:35.0000] <jgraham> Oh [00:44:43.0000] <matjas> jgraham: “You need the exact following code in your source:” [00:45:05.0000] <matjas> reading some other pages on the wiki, i get the idea they’re pretty strict about the test harness stuff [00:45:07.0000] <jgraham> That is a lie to children [00:45:26.0000] <jgraham> You need something that is semantically equivalent to that exact code [00:45:33.0000] <gsnedders> Well, thankfully we have no children contributing tests. [00:45:56.0000] <jgraham> I have better things to do in life than complain at people because they put useless legacy type attributes on their script elements [00:48:27.0000] <jgraham> gsnedders: The notion that lies to children only apply to children is itself a lie to children [00:49:54.0000] <Ms2ger> gsnedders, you don't write tests? ;) [00:50:56.0000] <jgraham> Turns out that gsnedders is a <del>criminal</del><ins>adult</ins> in the eyes of the law. [00:51:05.0000] <jgraham> Who knew? [00:56:44.0000] <Ms2ger> matjas, something like assert_equals(result, glyph, 'Expected ' + reference + ' to match ' ...) should work [00:57:36.0000] <matjas> Ms2ger: but e.g. in Trident you’d get "Expected ` ` to match ` `", which is not helpful at all as the glyphs look alike [00:58:53.0000] <matjas> Ms2ger: currently i print “Expected &nbsp; to match U+000A0 ( ); got U+00020 ( ) instead.” which is more useful [00:58:59.0000] <Ms2ger> Right [00:59:00.0000] <zcorpan> just do assert_equals(result, glyph, 'ALIENS') [00:59:32.0000] <Ms2ger> You should be able to pass "Expected &nbsp; to match U+000A0 ( ); got U+00020 ( ) instead." to assert_equals [00:59:35.0000] <jgraham> matjas: I am still somewhat confused as to how this is different from the html5lib test, aprt from (probably) being much faster to execute and easier to read [01:00:27.0000] <matjas> jgraham: it seems the html5lib test didn’t detect the &AElig bug in WebKit, nor the 3 Trident bugs (?) [01:01:42.0000] <jgraham> That is quite worrying [01:02:28.0000] <matjas> not sure how that happened. my test is a simple 1:1 copy of this table http://www.whatwg.org/specs/web-apps/current-work/multipage/named-character-references.html [01:02:36.0000] <jgraham> Random feedback: please try to stick to one standardised property to get information; in general tests shouldn't have UA-specific codepaths [01:02:52.0000] <matjas> perhaps html5lib is missing a few? [01:03:00.0000] <jgraham> Even if that means that some browsers fail for reasons unrelated to the original test [01:04:48.0000] <matjas> jgraham: ok, I’ll keep that in mind when submitting. (in this case, it’s not really UA-specific, but rather feature-specific code (no UA sniffing)) [01:05:07.0000] <jgraham> Right, but it's still bad [01:05:19.0000] <jgraham> Ideally each browser should execute exactly the same code [01:05:35.0000] <matjas> /me “ideally” :) [01:05:40.0000] <jgraham> http://code.google.com/p/html5lib/source/browse/testdata/tokenizer/namedEntities.test is the source data for html5lib [01:05:58.0000] <zcorpan> Ms2ger: it's quite a big diff :-/ [01:06:35.0000] <jgraham> matjas: "ideally" as in "I will loudly complain if that isn't the case" :) [01:07:51.0000] <gsnedders> And when jgraham complains, he goes on and on and on. [01:07:59.0000] <gsnedders> You don't want him complaining at you. [01:08:33.0000] <matjas> gsnedders: noted :) [01:09:59.0000] <matjas> jgraham: aren’t the html5lib tests included in WebKit’s layout tests? I wonder how the &AElig bug went unnoticed, not to mention the 93 other named char ref bugs that latest stable Safari has [01:10:36.0000] <jgraham> matjas: I am not sure what the WebKit people do. I thought they used the html5lib data but in a different harness [01:11:29.0000] <annevk> matjas: are those 93 by any chance references for two code points? [01:11:34.0000] <matjas> annevk: yeah [01:11:38.0000] <annevk> matjas: because that's a bug that was fixed fairly recently in WebKit [01:12:13.0000] <matjas> annevk: i know… but i wonder why they didn’t fix it sooner, given that the html5lib tests from 2010 show the errors [01:12:25.0000] <matjas> annevk: same for &AElig which only got fixed the other day [01:12:43.0000] <annevk> matjas: because not all issues are considered of equal importance? [01:12:58.0000] <annevk> matjas: returning two code points prolly required an updated data structure [01:13:00.0000] <matjas> annevk: maybe, but no bug reports either? seems weird [01:13:20.0000] <matjas> (no bug reports for &AELig, that is) [01:13:33.0000] <annevk> matjas: oh, dunno about &AELig [01:13:44.0000] <annevk> abarth might know [01:14:03.0000] <abarth> hi [01:14:06.0000] <gsnedders> https://bugs.webkit.org/show_bug.cgi?id=87465 ? [01:14:16.0000] <gsnedders> Ah, your bug. [01:14:30.0000] <abarth> yeah, i screwed up my python falsy checks [01:14:46.0000] <abarth> people really like Python, but falsy is tricksy [01:14:54.0000] <abarth> anyway, it's fixed now :) [01:15:03.0000] <abarth> if you find these sorts of bugs, please feel free to fix them [01:15:35.0000] <matjas> abarth: we were wondering how this didn’t get detected by the layout tests sooner… aren’t the html5lib tests included in the WebKit layout tests? [01:15:42.0000] <gsnedders> The question was why the existing tests for entities didn't find it [01:16:00.0000] <abarth> i'm sure it was detected, but there are many, many bugs to fix [01:16:04.0000] <abarth> and not enough folks to fix them [01:16:11.0000] <abarth> if you'd like to help out by fixing bugs [01:16:20.0000] <abarth> i'm happy to find some that you might be interested in [01:16:40.0000] <matjas> abarth: i assumed that if a test from 2010 shows the error, _someone_ would have logged it in the bug tracker (even if it wouldn’t get fixed right away, which would be understandable) [01:17:05.0000] <abarth> oh, we have bugs on file from 2005 that haven't been fixed too [01:17:14.0000] <abarth> i tried fixing all the bugs in the order they were reported once [01:17:15.0000] <matjas> abarth: yeah, on file [01:17:20.0000] <abarth> that was fun for a week :) [01:17:26.0000] <abarth> I fixed some really crazy stuff [01:17:29.0000] <matjas> abarth: but this one wasn’t reported before, and i just wonder how that happened [01:17:41.0000] <abarth> but that turns out not to be the most productive order [01:17:55.0000] <abarth> i guess I don't understand the question you're asking [01:18:01.0000] <abarth> have you worked on a large software project before? [01:19:07.0000] <gsnedders> abarth: I think the point is that tests in the html5lib testsuite failed yet there was no bug report. i.e., WebKit failed its own tests without having any bug report about doing so. [01:20:13.0000] <zcorpan> Ms2ger: some tests are dropped. they should instead be changed to match what the spec expects (e.g. an exception for caller) [01:20:52.0000] <abarth> gsnedders: what's the point of filing a bug when there's a test that clearly shows the problem? [01:21:16.0000] <abarth> that's redundant, no? [01:21:50.0000] <jgraham> abarth: (depends on your testing setup and processes, really) [01:22:10.0000] <abarth> the truth is that this bug was vastly less important that many thousands of other bugs [01:22:20.0000] <abarth> so no one cared enough to file a bug about it or fix it until now [01:22:33.0000] <zcorpan> Ms2ger: after those two changes, i approve, but with reservation that we will defer more careful review until we update our impl to match the spec [01:22:38.0000] <matjas> abarth: ok, thanks! [01:23:12.0000] <abarth> if there are other bugs that you'd like to see fixed, please feel encouraged to file bugs about them and/or write patches to fix them [01:25:43.0000] <zcorpan> matjas: you could output your own log that you're using now as well as using the standard testharness output [01:26:16.0000] <matjas> zcorpan: i see, thanks! [01:28:50.0000] <zcorpan> though it would be nice to be able to implement your own custom message in the output table, maybe by using a function as the message argument [01:49:43.0000] <smaug____> huomenta [02:29:41.0000] <AryehGregor> matjas, I've checked in my editing spec tests to Gecko, so they're run as regression tests. If you think that means anyone has sat down and classified every single expected failure, you're crazy. :) [02:29:51.0000] <AryehGregor> Likewise richtext2 has been part of the Gecko test suite for some time. [02:30:05.0000] <AryehGregor> But there are tons of expected failures, no one has looked at all of them or filed bugs on all of them. [02:30:16.0000] <AryehGregor> It's not clear in many cases why they're failing or whether the test suite is even correct. [02:30:26.0000] <matjas> AryehGregor: not necessarily a separate bug for every single failure, but I would expect a “master tracking bug” of some sort, yeah [02:30:40.0000] <matjas> I guess I’m just naive. [02:31:01.0000] <AryehGregor> Or maybe you're volunteering to review all expected failures in Gecko/WebKit's test suites and file bugs on all of them? :) [02:31:24.0000] <AryehGregor> Generally you can't even file a useful bug unless you know the spec and/or code well and spend a bunch of time debugging -- the failure is often nontrivial. [02:33:21.0000] <AryehGregor> There are tons of low-hanging editing test failures that I haven't looked at because I have lots of other stuff to do. [02:33:50.0000] <AryehGregor> And in my case, I'm even the one who's paid by Mozilla both to maintain the spec/tests and to write editing code for them. [02:34:18.0000] <AryehGregor> Although I wrote the tests while I was working for Google -- my CSS Transforms tests, which I wrote while at Mozilla, have bugs filed for all Gecko failures. [02:34:29.0000] <AryehGregor> (a number of which are still open) [02:39:26.0000] <Ms2ger> zcorpan, thanks! [03:03:02.0000] <jgraham> matjas: Yeah, it would be a huge amount of work to make a bug report for every single failing test, particularly from imported test suites which can be of very dubious quality [03:03:49.0000] <zcorpan> Ms2ger: welcome. in return, maybe you could convert the web worker testsuite to use testharness? :-P [03:04:36.0000] <MikeSmith> does anybody know what the term "browse-by Web" is supposed to mean? [03:04:54.0000] <matjas> gsnedders: did Steven Levithan end up filing a bug detailing the non-standard regex additions in Opera? if not, here’s a list: http://kangax.github.com/es5-compat-table/non-standard/ [03:07:06.0000] <annevk> matjas: maybe link that from http://wiki.whatwg.org/wiki/Web_ECMAScript#RegExp ? [03:07:38.0000] <zcorpan> there's already a link [03:08:44.0000] <annevk> oh right [03:08:51.0000] <matjas> ah, at the bottom, damn [03:08:53.0000] <matjas> too late [03:08:53.0000] <Philip`> MikeSmith: Maybe the parts of the web where you only stop at a page for a short time and glance over it and then browse on by, unlike e.g. Gmail or Facebook which are more like major destinations you'll spend hours at [03:09:30.0000] <MikeSmith> Philip`: yeah, that would seem like an intuitive definition, given those words [03:09:54.0000] <gsnedders> matjas: He did. [03:10:09.0000] <annevk> matjas: reverted [03:10:10.0000] <MikeSmith> Philip`: but looking at http://www.w3.org/2012/05/sysapps-wg-charter.html I don't see what connection "browse-by Web" has to the scope of that proposed work [03:10:25.0000] <gsnedders> matjas: And that's nothing compared with what we support [03:10:34.0000] <annevk> matjas: also sorry :) [03:10:53.0000] <jgraham> MikeSmith: Maybe it's the bits of the web you would like to shoot? [03:11:03.0000] <MikeSmith> heh [03:11:05.0000] <matjas> gsnedders: now i’m curious [03:11:19.0000] <MikeSmith> jgraham: that would be a lot of bits [03:11:31.0000] <gsnedders> matjas: Named capturing groups, nested character classes, etc. [03:13:35.0000] <Philip`> MikeSmith: Yeah, in that case it sounds like they're saying "the "browse-by" web" for what everyone calls "the web", but they don't want to admit that their adoption of web-originated technologies for use as the core platform of new devices with very different characteristics is not really the web [03:14:20.0000] <MikeSmith> yeah [03:14:30.0000] <MikeSmith> exactly [03:15:41.0000] <MikeSmith> fwiw, I suggested using "Web-application-unsafe APIs" instead [03:15:46.0000] <MikeSmith> http://lists.w3.org/Archives/Public/public-sysapps/2012May/0028.html [03:15:55.0000] <MikeSmith> or "non-SOP APIs" [03:16:04.0000] <annevk> MikeSmith: did you see my comment earlier about https://www.w3.org/Bugs/Public/show_bug.cgi?id=17205 being actual spam? [03:16:21.0000] <MikeSmith> annevk: yeah [03:16:29.0000] <Philip`> I suppose the problem is that "web" has become a generic brand name for a collection of technologies, rather than being about a distributed network of interlinked resources or whatever the original idea was [03:16:58.0000] <annevk> MikeSmith: though it seems kind of weird spam looking at it closer; URL returns some JSON [03:17:02.0000] <Philip`> and those technologies happen to be just about good enough that people want to reuse them in totally different contexts [03:17:09.0000] <annevk> MikeSmith: fine if I close that bug? [03:17:13.0000] <MikeSmith> annevk: you want me to mark that account as a spammer? Looking at that bug, it's not clear to me that it's intentional malicious spamming [03:17:20.0000] <MikeSmith> annevk: yeah, please close it [03:18:34.0000] <annevk> MikeSmith: can only find that bug for that user [03:18:46.0000] <annevk> but yeah, lets wait a bit [03:20:21.0000] <MikeSmith> OK [03:22:18.0000] <Philip`> MikeSmith: I think "non-SOP APIs" is good, because then the rest of the web can be called "soppy APIs" [03:25:09.0000] <gsnedders> https://developer.mozilla.org/en/JavaScript/Reference/Global_Objects/WeakMap — I wonder how much effect that warning will really have… [03:26:29.0000] <zcorpan> non-SOAP APIs? [03:33:29.0000] <Ms2ger> zcorpan, :) [03:34:30.0000] <Ms2ger> zcorpan, might be easier if you threw jsframework.js my way ;) [03:35:22.0000] <zcorpan> that's probably doable [03:36:05.0000] <Ms2ger> Also [03:36:22.0000] <Ms2ger> odinho, about that assert_throws patch... [03:36:29.0000] <Ms2ger> jgraham, about that timout patch... [03:38:19.0000] <jgraham> Ms2ger: Yeah, I should push that [03:38:30.0000] <jgraham> You had some feedback that I've forgotten [03:39:00.0000] <Ms2ger> Philip`, about that 2d.pattern.repeat.null bug... [03:39:49.0000] <odinho> Ms2ger: wawawa, I thought about that when I came to work today actually :P [03:39:57.0000] <Ms2ger> jgraham, You're inconsistent about if () { vs if ()\n{ [03:40:27.0000] <Ms2ger> Dunno what else I'd complain about [03:40:56.0000] <Ms2ger> odinho, I'm glad to hear that :) [03:42:46.0000] <zcorpan> Ms2ger: http://simon.html5.org/dump/jsframework.js [03:43:53.0000] <Ms2ger> Thanks, I'll have a look at some point [03:44:01.0000] <zcorpan> Ms2ger: awesome, thanks! [03:44:02.0000] <odinho> IDLharness is cool although I don't really trust it all that much yet. It's just so ... automatic and stuff. :P [03:45:33.0000] <zcorpan> Ms2ger: one thing i recall is that the web worker testsuite at places has nested tests, which is not considered a good practice. instead it should create all tests up front (as async tests). that way, the number of tests don't change based on the result of the tests [03:45:51.0000] <Ms2ger> Mm [03:46:03.0000] <Ms2ger> matjas, `\0` (U+0001) < eh? [03:46:46.0000] <matjas> Ms2ger: what? [03:46:52.0000] <AryehGregor> /me wants nsINode::AsContent [03:47:05.0000] <Ms2ger> AryehGregor, wrong window, and sure, file a bug :) [03:47:08.0000] <AryehGregor> Drat. [03:47:17.0000] <AryehGregor> /me usually has #developers here, gets mixed up when it's #whatwg [03:47:31.0000] <Ms2ger> matjas, surely that's U+0000 [03:47:46.0000] <matjas> Ms2ger: yeah; where did i make that mistake? [03:47:57.0000] <Ms2ger> The twattersphere [03:48:44.0000] <matjas> whoops, thanks! [03:50:55.0000] <AryehGregor> odinho, I don't trust it so much either, and I wrote it. :) [03:50:57.0000] <AryehGregor> But it's cool! [03:51:06.0000] <matjas> zcorpan: on http://simon.html5.org/htmlpubquiz, could you please link to the explanation as well? [03:52:03.0000] <zcorpan> matjas: it shows up if you click on correct answer [03:52:28.0000] <matjas> zcorpan: not on a mobile device [03:52:48.0000] <matjas> (do you really want to rely on the current twitter web layout?) [03:52:49.0000] <zcorpan> blame twitter [04:07:08.0000] <annevk> zcorpan: relying on twitter to keep those relations indefinitely seems unwise though [04:07:48.0000] <annevk> zcorpan: anecdotal evidence suggests replies, retweets, etc. get lost over time [04:08:25.0000] <zcorpan> ok. i guess i should just move over all the data [04:11:11.0000] <odinho> Hmz, sicking did send an email not too long ago. But no IRC. [04:11:34.0000] <odinho> So, there's a double ;; in one of the IDL-fields in the idb spec, -- should I really make a bug for that? Feels like a real waste of bytes. [04:13:37.0000] <annevk> bugs are filed for typos all the time [04:13:42.0000] <annevk> the real waste is spam [04:14:08.0000] <Ms2ger> <meta name='revision' content='$Id: Overview.html,v 1.41 2010/03/24 17:58:17 nmehta3 Exp $'/> [04:32:23.0000] <Stevef> Philip: any data on mathml on the web? [04:39:26.0000] <david_carlisle> Stevef: what kind of data are you looking for? (just passing through and the last comment in the channel is on mathml:-) [04:59:14.0000] <Stevef> david_carlisle: looking into providing accessible math, wondering how much mathml is used and if prose form of math is provided alongside for AT users [04:59:51.0000] <annevk> shouldn't AT support math? [05:00:09.0000] <annevk> seems better than require everyone to do make work [05:01:46.0000] <Stevef> annevk: not that simple even if it is supported still a problem for vision impaired to understand complex equations, also browsers need to support math, support is spotty [05:02:09.0000] <jgraham> No one is going to type out a verbal description of all their formulae [05:02:58.0000] <jgraham> Since there is support for TeX->speech it seems like presentation mathml to speech should be just as possible [05:03:26.0000] <jgraham> Anyway, what I actually wanted to say is "does anyone know anything about navigation"? [05:04:15.0000] <Stevef> jgraham: for educators it is an issue, there is extensive research on the topic, simply saying "no one will" is not helpful [05:04:46.0000] <jgraham> It seems like Chrome and Firefox abort navigations triggered from inside the "prompt to unload" algorithm i.e. in the beforeunload event [05:05:05.0000] <Stevef> there is editing software that outputs MathSpeak i.e. equations in prose form [05:05:22.0000] <jgraham> But I can't see any requirement for this in the spec. Anyone know if there's a reason for the difference or if it is just a bug? [05:06:55.0000] <jgraham> Stevef: Using software to automate the translation is equivalent to having it built in to the browser. I think working from the assumption that no one will write out verbal descriptions is an excellent position because it forces you to consider solutions that will work with the vast majority of content rather than just solutions where people put in special additional effort [05:06:59.0000] <david_carlisle> Stevef: There is a lot of work being done at present in DAISY and NIMAS groups on refining the guidelines on how MathML should interact with AT software, If you ask on www-math list rather than here you should be able to hook up with other people working on that [05:07:24.0000] <Stevef> david_carlisle: thanks [05:12:42.0000] <Stevef> jgarham: and thats what I am trying to undesrstand what are the mechanisms for providing accessible math and are prose transaltions provided (from automated or other source) [05:13:14.0000] <Stevef> jgraham: because MathML is not a complete solution [05:16:19.0000] <Stevef> jgraham: saying 'nobody' will is incorrect, saying the vast majority won't is plausible [05:18:34.0000] <david_carlisle> Stevef: Have you tried MathPlayers AT cpabilities? That's probably the most generally available system in that area, the version in the browser has an option to read the expression, but I understand that that is really just a demo on how it could hook into a general AT framework (but it isn't really my area) [05:18:49.0000] <jgraham> I agree that "nobody" isn't strictly accurate. But acting as if it is seems more likely to produce good results than optimising for the difference [05:21:16.0000] <Stevef> david_carlisle: have been playing with MathPlayer, the built in speech capability is good, supports many languages. limitation is that it only runs in IE, JAWS/Window YES/zoomtext reader/magnifier works well with it. [05:22:51.0000] <Stevef> jgraham: i am not advocating that all math must be accompanied by hand written prose form of equation, I am trying to work out how people do or don't do math on the web [05:23:18.0000] <Stevef> jgraham: in particular accessible math [05:25:25.0000] <david_carlisle> Stevef: Yes I know it's windows only, but having _an_ implementation is good to validate the general premise that it is _possible_ to get a good accessible behaviour from the markup. After that it's "just" a matter of applying peer pressure (or citing relevant legislation:-) to encourage others to follow suit. [05:26:51.0000] <jgraham> I don't really know how accessible maths is different from any other kind. I mean I imagine if I was a blind physicist, for example, I would want/need to have access to all the same maths content as my peers [05:27:20.0000] <Stevef> david_carlisle:true, something is better than nothing, just trying to work out the practicalities for a online education company who need to provide accessible math [05:28:30.0000] <Stevef> jgraham:point being? also I am not talking about blind physicist I am talking about providing content to a school kid [05:28:31.0000] <david_carlisle> jgraham: True enough but the mathml spec does try to give general guidance on how to make the markup more accessible including things that have no effect on the visual presentation, things like the invisible unicode characters fro invisible times or function application [05:32:07.0000] <david_carlisle> Stevef: we have members of the math WG who work full time on these aspects, but I'm not one of them so I can only pass on second hand information I've picked up over the years, which is why I suggested www-math might be a better forum if you want to get first-hand reports of what people are doing [05:32:55.0000] <Stevef> david_carlisle: sure, just asked philip as i know he collects data on stuff [05:34:13.0000] <david_carlisle> Stevef: Yes if you want to reach people with real data collected from real web sites this probably is the place:-) [05:37:41.0000] <Stevef> david_carlisle: also seems that what is published is presentational mathML rather than content mathml, so meaning is not fully convyed in the markup exposed to AT/ [05:40:23.0000] <david_carlisle> Stevef: mostly perhaps yes, by coincidence I got pointed to this message this morning which is relevant: the thread starts here, but includes a quote from Nemeth argguing presentation is what you want: http://host.nfbnet.org/pipermail/blindmath_nfbnet.org/2012-May/005435.html [05:42:29.0000] <Stevef> david_carlisle: cheers, the subject of accessible math is a real can of worms, it is difficult to know what advice to give to content producers [05:43:33.0000] <jgraham> annevk, Ms2ger: DOM should define the term "empty text node" [05:44:41.0000] <david_carlisle> Stevef: It's probably not fair to say meaning isn't conveyed to AT as that implies that meaning is being lost. If the same markup is being used for visual rendering and for AT then that isn't the case. It will always be the case that there is more presentation mathml (even if browsers supported content mathml natively) as it is much easier to produce, you can give a hand drawn expression to... [05:44:43.0000] <david_carlisle> ...a suitably trained typist and get presentation mathml, but to produce content mathml you need to know something about what it means., which means production has to be a lot more controlled [05:45:09.0000] <Ms2ger> jgraham, file a bug, please? [05:45:24.0000] <Stevef> david_carlisle:ok thanks for the clarification [05:46:55.0000] <annevk> jgraham: also list a use case in that bug, please [05:47:27.0000] <jgraham> annevk: For the definition? HTML uses it [05:47:47.0000] <Ms2ger> HTML also defines it, fwiw [05:49:05.0000] <jgraham> Where? [05:50:03.0000] <Ms2ger> At the end of 2.1.3 DOM trees [05:50:56.0000] <jgraham> Interesting [05:51:02.0000] <jgraham> That should really be in DOM I think [05:51:12.0000] <jgraham> But more importantly Hixie should link to the definition [05:53:25.0000] <annevk> I'm happy to have the collection of useful terms related to the DOM in DOM [05:53:29.0000] <annevk> that makes a lot of sense [05:53:55.0000] <annevk> I similarly asked dglazkov to file a bug on getting DOM to define subtree [07:40:05.0000] <Stevef> jgraham: Tex typesetter and text translation http://www.math.union.edu/~dpvc/transfer/mathjax/speech-lab.html [07:48:06.0000] <jgraham> Stevef: Seems pretty cool [07:48:19.0000] <david_carlisle> Stevef: given the URI that's presumably mathjax based so would work for MathML too, as MathJax has two input parsers, one for TeX-like syntax and one for MathML. [07:50:27.0000] <Stevef> right, got the link from this guy https://twitter.com/#!/pkrautz [07:53:23.0000] <david_carlisle> Stevef: Davide has a description of it here: http://lists.w3.org/Archives/Public/public-fx/2012AprJun/0126.html [08:06:25.0000] <Stevef> david_carlisle:cheers interesting thread [08:35:01.0000] <annevk> roc: fwiw, http://www.w3.org/TR/from-origin/ is now updated [08:35:52.0000] <annevk> Ms2ger: are you going to review DOM 3 Events? should I? [08:36:00.0000] <Ms2ger> I'm planning to [08:36:10.0000] <Ms2ger> But the more the better :) [08:37:01.0000] <annevk> "HTTP living standard" would sure be welcome [08:37:29.0000] <annevk> seven part spec with god knows how many registries and extensions sucks monkey balls [08:37:46.0000] <Ms2ger> "IETF" [09:16:01.0000] <dglazkov> good morning, Whatwg! [09:16:26.0000] <dglazkov> annevk: please steal the file a bug button script from me for your spec, so that I can file bugs easily :) [09:18:31.0000] <Ms2ger> dglazkov, you mean the "file a bug" link he has at the top of all his specs? :) [09:20:01.0000] <odinho> Ms2ger: Prolly the just point and write and press enter. [09:31:14.0000] <davidb> at [09:31:29.0000] <Ms2ger> dot [09:31:33.0000] <odinho> bash: at: command not found [09:32:12.0000] <Ms2ger> Eh, let's discuss top-posting [09:33:15.0000] <odinho> ... crickets ... [09:33:20.0000] <odinho> and that's that. [09:47:40.0000] <davidb> /me chuckles [10:15:00.0000] <jgraham> irony alert: ifette complained about *other people's* mail formatting [10:15:17.0000] <jgraham> His is generally pretty terrible [10:15:39.0000] <jgraham> Not the worst I have ever seen but often hard/impossible to follow in alpine [10:16:16.0000] <Ms2ger> Also fun: nonstandard quoting characters [10:16:25.0000] <Ms2ger> (Hi, Alex Mogilewski) [10:17:36.0000] <jgraham> The worst offenders are 1) people who use lotus notes (i.e. IBM employees) and 2) people that top post (generally gmail users working at Google) [10:17:41.0000] <TabAtkins> Heh, yeah, I don't know what's that about. [10:17:56.0000] <jgraham> Seems like the bigger your company is the worse your mail habits are likely to be [10:18:23.0000] <zewt> the worst are people who post in light blue in a giant font [10:18:34.0000] <jgraham> (pretty sure it is notes that does the quotes-distinguished-by-colour thing) [10:18:55.0000] <zewt> people posting in large fonts always seem to me like they think their posts are more important than everyone else's [10:19:00.0000] <zewt> heh [10:19:11.0000] <jgraham> Well alpine doesn't do light blue or coloured fonts and I think I turned that off in thunderbird too [10:19:31.0000] <jgraham> So generally when people do that I have to disregard their message due to it being unintelligable [10:19:54.0000] <zewt> the colored-quoting thing is definitely worse, but fortunately fairly rare [10:19:59.0000] <zewt> at least on public lists [10:20:00.0000] <TabAtkins> That seems like a legitimate effect, actually. In a big company, most of your email will be internal. Most people aren't picky about their email, so they'll use whatever is the default and easiest. Thus, in a large company you're exposed much more to "bad" practices caused by monoculture. [10:20:22.0000] <jgraham> Sometimes I have had to unpick threads and only been able to work out who was saying what by looking for the place where the opinion inverted [10:20:43.0000] <zewt> also messages where people inline quote, but leave ten pages of quotes above *and* below their reply [10:20:46.0000] <TabAtkins> Whereas someone in a smaller company has to deal more with external emails in a variety of formats, so they get pickier and politer intheir formatting. [10:20:55.0000] <TabAtkins> zewt: Heh, see: every single internal thread at Google. [10:21:01.0000] <TabAtkins> It makes me die inside. [10:21:04.0000] <zewt> gmail will collapse both, but then replying is a complete mess [10:21:32.0000] <jgraham> Yeah, it seems that gmail encourages terrible habits [10:21:46.0000] <zewt> to be honest, i don't think it's worth trying to get non-technical users to inline quote if they don't do it intuitively, because they tend to make a hash of it [10:22:00.0000] <jgraham> Pretty sad coming from a company so closely associated with the internet [10:22:03.0000] <TabAtkins> zewt: I'd expect everyone on one of our lists to be technical, though. [10:22:10.0000] <zewt> sure [10:23:06.0000] <hober> All of this is made that much worse by having Ahem installed [10:23:20.0000] <TabAtkins> s/worse/better/ [10:23:26.0000] <TabAtkins> Because then you get to do fun ascii art. [10:23:41.0000] <hober> TabAtkins: :) [10:23:54.0000] <zewt> another side benefit of html mail: you can explicitly declare when text is meant to be fixed-width [10:24:08.0000] <zewt> rather than assuming everyone reads mail in fixed width (which most people don't) [10:24:27.0000] <TabAtkins> That, and inline images, are the only two reasons I ever switch to sending HTML email. [10:24:33.0000] <jgraham> hober: Why? Unless you only have ahem installed [10:24:44.0000] <zewt> italics and bold are pretty useful when you're discussing spec text [10:25:00.0000] <hober> jgraham: email from outlook read in mail.app on a machine with ahem installed is, shall we say, square. [10:25:18.0000] <zewt> reading specs without keywords marked is a headache [10:25:23.0000] <TabAtkins> Oh, so now it's not *cool* enough for you, hober? [10:25:25.0000] <TabAtkins> Jeez. [10:25:48.0000] <hober> jgraham: outlook's html email contains an invalid font-family rule which gets matched with ahem in webkit [10:25:57.0000] <jgraham> hober: Wow. Mind you, I like to imagine that Björn Höhrmann only has ahem installed [10:26:06.0000] <jgraham> and uses it for all things [10:26:08.0000] <hober> hahahahahahaa, that would explain a few things [10:26:25.0000] <TabAtkins> Hahahahaha [10:26:48.0000] <TabAtkins> Unrelated: how do I politely suggest to someone that they don't understand any of the details of what they're talking about? [10:27:03.0000] <Ms2ger> You don't, on the interwebs [10:27:06.0000] <TabAtkins> Specifically, John Daggett, in his discussion of Variables, keeps making assertions about CSS grammar that are simply wrong. [10:27:24.0000] <Ms2ger> Well, he thinks the same about your assertions :) [10:27:36.0000] <TabAtkins> I understand that not everyone can read FLEX or whatever easily, but still. [10:28:34.0000] <jgraham> TabAtkins: Make an implementation of the grammer that you can feed his mistakes to and demonstrate that they don't match? [10:28:41.0000] <jgraham> *grammar [10:29:07.0000] <TabAtkins> I was planning on doing that anyway, but I don't particularly feel like doing so *right now*. ^_^ [10:29:44.0000] <TabAtkins> Plus I'd do the implementation according to my Syntax draft, so I've misread the FLEX grammar, I'll still get it wrong. [10:30:49.0000] <jgraham> Yeah, but you could ask him to point out the bug :) [10:31:48.0000] <TabAtkins> Well, I'm pointing out his mistake anyway, so that's equivalent. [10:32:04.0000] <jgraham> It wasn't an entirely serious suggestion. If you can't do things in the normal way (i.e. say "this doesn't match the spec; as we se in section X.Y {explaination of the mismatch}" then I don't know what will work [10:32:39.0000] <TabAtkins> That's working individually, he's just made several incorrect assertions that are plainly wrong if you have a good understanding of the grammar. [10:33:16.0000] <TabAtkins> Like assuming that "counter\(foo\)" parses as a function (it's an ident). [10:33:23.0000] <jgraham> Ah, so the problem is how to say "I think you need to reread the grammar section of the spec" [10:33:29.0000] <TabAtkins> Yes. [10:33:38.0000] <TabAtkins> However, I said almost exactly that in the email I just sent, so whatever. [10:49:48.0000] <zewt> is there any difference visible to scripts between {} and {a: undefined}? [10:50:24.0000] <zewt> guess hasOwnProperty("a") returns true [10:51:38.0000] <jgraham> Yeah [10:52:33.0000] <TabAtkins> Yup, that's it. [10:53:06.0000] <TabAtkins> Pair that with a prototype walk just in case the property is further up the chain, and you can robustly detect it. [10:53:24.0000] <zewt> undefined is ... defined D: [10:53:41.0000] <jgraham> Well "a" in foo is just that [10:53:42.0000] <TabAtkins> However, good code shouldn't treat the two differently. [10:54:07.0000] <TabAtkins> If you really need a "this property exists but doesn't have a value", use null. [10:54:20.0000] <TabAtkins> That's what it's there for - to be the same as undefined, but detectably different. [10:54:29.0000] <zewt> well, yeah [10:54:34.0000] <jgraham> TabAtkins: That's like saying you should never use hasOwnProperty withotu also checking the value isn't undefined [10:54:39.0000] <zewt> (the question comes from something unrelated to web stuff) [10:54:41.0000] <jgraham> Which is crazy [10:55:02.0000] <TabAtkins> jgraham: You just shouldn't use hasOwnProperty. ^_^ for-in is a broken construct that encourages bad habits. [10:55:03.0000] <jgraham> For for...in [10:55:13.0000] <jgraham> s/For/Or/ [10:55:32.0000] <TabAtkins> Which is why we did it all again with for-of, which acts the way you'd *expect*, rather than the crazy way that for-in does. [10:55:50.0000] <jgraham> Or... I forget hwo array iterators work with sparse arrays [10:56:37.0000] <TabAtkins> Depends on the details of the iterator. Which one are you talking about? [10:58:09.0000] <Ms2ger> new Iterator(), duh [10:58:23.0000] <TabAtkins> Ah, thanks, that's useful. ^_^ [12:06:27.0000] <jgraham> Sigh. ifette claims that the fact that his mail appears broken in the archives and in other mail clients means that everyone else should switch clients/archiving software [12:07:35.0000] <Ms2ger> Duh? [12:16:41.0000] <annevk> jgraham: Gmail should really fix its text/plain handling imo [12:17:23.0000] <annevk> jgraham: I noticed the problem too; Opera's pretty awesome at handling quotes in emails (and reformatting them) [12:25:14.0000] <Hixie> i am back [12:25:19.0000] <Hixie> and have about 1000 e-mails to read [12:25:37.0000] <Hixie> anything urgent for me to do first? [12:25:53.0000] <zewt> the bathroom's that way --> [12:33:33.0000] <annevk> Hixie: nothing has collapsed so far :) [12:39:38.0000] <Hixie> cool [12:41:02.0000] <Ms2ger> The HTMLWG has requested that you add a link to a 404 error to the spec, if you were gone already when they did that [12:46:21.0000] <Hixie> come again? [12:47:05.0000] <Hixie> (264 e-mails on responsive images? jesus) [12:49:15.0000] <Wilto> Hixie: Florian Rivoal has a pretty great proposed compromise at the end. Y’know, after the first 263 chapters. [12:58:47.0000] <Hixie> Wilto: i will be reading all the e-mails as always :-) [12:59:34.0000] <Wilto> Hixie: Oh, definitely didn’t mean it as in “skip the rest.” Just throwin’ that out there. [13:03:33.0000] <Hixie> Ms2ger: i don't see any decisions in recent times [13:04:13.0000] <Ms2ger> https://www.w3.org/Bugs/Public/show_bug.cgi?id=11204#c34 [13:08:24.0000] <Hixie> lol [13:09:57.0000] <jgraham> Hixie: Nothing urgent, but the document loading / navigation parts of the spec turn out to have quite a few differences from implementations, and we would appreciate it if you could look at those relatively soon [13:10:16.0000] <Hixie> jgraham: is there mail? [13:10:25.0000] <jgraham> Hixie: Bugs mostly [13:10:32.0000] <Hixie> jgraham: i mean, i know the browsers all differ from each other, that's nothing new... [13:10:35.0000] <jgraham> Although there are some older mail threads [13:10:54.0000] <Hixie> can you send me a mail with links to all the things you want me to process on the topic? [13:11:05.0000] <jgraham> Hixie: There seem to be places where the spec is different from any browser and the browsers are relatively consistent [13:11:14.0000] <Hixie> well that seems bad [13:11:31.0000] <jgraham> Yeah, I will put together a mail tomorrow [13:11:44.0000] <jgraham> There is probably more to discover of course [13:11:45.0000] <Hixie> cool, thanks [13:53:25.0000] <WeirdAl> Hixie - I love the big black box on the XBL2 note :) [13:54:43.0000] <WeirdAl> voice of God, with Broadway lights [13:54:55.0000] <Hixie> url? [13:55:08.0000] <WeirdAl> http://www.w3.org/TR/2012/NOTE-xbl-20120524/ [13:56:04.0000] <Hixie> ah, that was below the fold when i brought it up, heh [13:56:07.0000] <Hixie> so i didn't see it [13:59:28.0000] <WeirdAl> it's a nice touch [14:12:01.0000] <Hixie> if anyone is interested in participating in this year's http://www.jseverywhere.org/ please let me know [14:13:59.0000] <Hixie> in other news, i do not understand how so many people manage to get an account and subscribe to parts of the HTML spec and yet not understand why they get e-mails when i change the spec [15:23:29.0000] <Hixie> abarth: i have some e-mails from you regarding https://bugs.webkit.org/show_bug.cgi?id=44641 -- should i just discard them? as you say in the bug, nobody else really got particularly excited about it [15:28:53.0000] <zewt> fwiw, not really sure how it's better than just <?= htmlspecialchars($untrusted_string) ?>, it's just more opaque and not human readable [16:23:03.0000] <abarth> Hixie: yeah, that's fine [16:24:16.0000] <Hixie> k, thanks 2012-05-30 [23:41:13.0000] <annevk> jgraham: btw, GMail itself cannot handle the incorrectly formatted messages it emits either [23:46:41.0000] <annevk> "The open*Cursor exception should be enforcable using WebIDL by using an enum. The only reason we don't have it that way in the spec is because ReSpec doesn't support that." o_O [23:48:46.0000] <sicking> does any browser support AnonXMLHttpRequest yet? [23:49:34.0000] <sicking> I'm wondering if we should change the constructor syntax to something like |new XMLHttpRequest({ anon: true})| to make it easier to expand for the future [23:50:58.0000] <sicking> annevk: rewriting the spec to not use ReSpec would be a pretty major undertaking. Especially to do without introducing regressions. But hey, if you are offering to put in the time... :-) [00:16:31.0000] <sicking> annevk: https://bugzilla.mozilla.org/show_bug.cgi?id=692677#c39 [00:19:59.0000] <Ms2ger> sicking, file a bug on XHR, please? [00:20:42.0000] <jgraham> annevk: Neither can thunderbird, it seems [00:26:38.0000] <sicking> Ms2ger: done [00:27:01.0000] <Ms2ger> Thanks [00:58:45.0000] <odinho> 09:51 < sicking> does any browser support AnonXMLHttpRequest yet? <<< Opera does, but more importantly, does anyone use it? Not too sure about that :P [01:00:11.0000] <zcorpan> we do? [01:02:01.0000] <MikeSmith> anybody know what "RC buggy" means in Debian discussions? [01:02:49.0000] <MikeSmith> hmm, RC means release critical, I guess [01:03:18.0000] <odinho> zcorpan: Uh, Dragonfly isn't working for me so it seems we support everything actually. [01:03:30.0000] <odinho> Even aoeu [01:04:51.0000] <MikeSmith> anybody have suggestions for alternatives to hypermail? [01:06:26.0000] <jgraham> MikeSmith: Why? [01:06:49.0000] <MikeSmith> because Fette is right about it being dead [01:07:15.0000] <odinho> sicking, zcorpan: Oh, no, we don't :P Actually. I remember writing a test for it, and somehow thought that meant we did more to it :] [01:07:29.0000] <jgraham> Oh well I guess you have to choose one of his four approved mail clients then [01:07:35.0000] <MikeSmith> heh [01:07:37.0000] <jgraham> We could have a shared gmail account [01:07:43.0000] <MikeSmith> heh [01:07:44.0000] <jgraham> No way that could go wrong [01:07:48.0000] <MikeSmith> I use mutt man [01:07:59.0000] <MikeSmith> compile it myself from sources [01:08:07.0000] <MikeSmith> that's the way everybody should do [01:08:39.0000] <sicking> odinho: what about this page: http://fr.opera.com/docs/specs/presto2.10/ [01:09:26.0000] <jgraham> I like to imagine that Philip` reads emails like he reads PDFs: cat and the mbox specification [01:09:57.0000] <MikeSmith> I used to use mh, which isn't far from that [01:10:40.0000] <MikeSmith> I assumed Fette's list of approved mail clients was some kind of Dadaist joke [01:11:57.0000] <odinho> sicking: My tests are green, so yes, we do support it. I don't know why it didn't work in the console. [01:12:03.0000] <reggna> sicking: Yes, it's included in 2.10.232 and should work in Opera Next at least. [01:12:16.0000] <odinho> sicking: I switched to using DragonFly experimental track yesterday, I think I maybe shouldn't do that. [01:12:39.0000] <sicking> ok [01:12:40.0000] <odinho> sicking: IE10 doesn't implement it. Fx does not. [01:12:51.0000] <sicking> yeah, we don't [01:13:22.0000] <odinho> sicking: Neither the Chromium I have does. [01:34:48.0000] <annevk> sicking: everything trumps spec writers man [01:34:54.0000] <annevk> sicking: better fix it up [01:35:43.0000] <annevk> sicking: about AnonXMLHttpRequest -> XMLHttpRequest(object); I like that, but that does make it less detectable for code validating libraries [01:36:01.0000] <annevk> well, it gets more complicated at least [01:36:38.0000] <annevk> but if you could do new XMLHttpRequest({url:..., } etc. that'd be pretty neat [01:39:44.0000] <odinho> What do you think about the sugar cg idea? -- It would probably make an easy version like that, but without the long name. [01:41:33.0000] <annevk> sugar cg? [01:41:58.0000] <annevk> guess I should read the logs [01:43:24.0000] <annevk> not in the logs odinho [01:43:25.0000] <odinho> Hmm. Like django.shortcuts, having a way for 90% to do 90% of what they want using much easier, less flexible shortcuts. -- It's something that a library should do, at least at first, as a proof of niceness (and to keep the bikeshedding down, because it'll basically only be about bikeshedding). [01:44:08.0000] <annevk> the danger with aliasing is that often you also start to want to "fix" things [01:44:10.0000] <odinho> They can even build it into extensions in the different browsers and put it on window, and see how it works :P [01:44:24.0000] <annevk> a problem with aliasing is that you need tests x2 [01:44:42.0000] <odinho> annevk: Yeah, - it was like, it should only do json and process it and only give you the object back. [01:44:55.0000] <annevk> maybe even tests ^2 for combinations [01:52:32.0000] <matjas> re: http://krijnhoetmer.nl/irc-logs/whatwg/20120524#l-737 is the makefile / whatever it is you’re using to publish open source? [02:00:02.0000] <annevk> matjas: not everything is public [02:00:23.0000] <annevk> matjas: the spec goes through http://pimpmyspec.net/ which is open source [02:00:59.0000] <annevk> matjas: http://code.google.com/p/html5/source/browse/trunk/spec-splitter/spec-splitter.py is used to generate the multi-page copy but as you observed it's a problem with the single-page copy as well [02:01:19.0000] <matjas> thanks [02:01:42.0000] <annevk> matjas: I think the problem is with Anolis (which pimpmyspec.net uses), but I'm not a 100% sure and have not looked at it in a while [02:01:54.0000] <annevk> there's an open bug with confused comments about it [02:02:18.0000] <matjas> i suppose the source file that’s fed to pimpmyspec.net to generate the HTML spec isn’t public? [02:03:27.0000] <annevk> there is http://svn.whatwg.org/webapps/ [02:03:50.0000] <annevk> but how source and the .inc files are combined is not public I'm afraid [02:03:54.0000] <annevk> you'll have to ask Hixie [02:04:04.0000] <matjas> aha, https://www.w3.org/Bugs/Public/show_bug.cgi?id=12539 [02:04:28.0000] <annevk> I wish he'd publish more of that code so everyone can have a go, but apparently it's a somewhat hackish script [02:04:54.0000] <annevk> matjas: your Bugzilla fu is better than mine :) [02:05:52.0000] <annevk> matjas: my comments there are bullshit as the problem is in the single-page version too reportedly [02:06:44.0000] <annevk> hmm although currently that does validate... [02:07:11.0000] <matjas> annevk: comment #2 by [tm] says it’s only a problem in the multi-page spec [02:07:50.0000] <matjas> i only noticed in the multi-page version, but then again i hadn’t looked at the single-page version [02:08:38.0000] <annevk> I think david_carlisle pointed that out at one point [02:08:51.0000] <annevk> meh [02:12:30.0000] <matjas> ok, i’ll just wait for you to fix it then :p [02:12:57.0000] <annevk> looks like people have been waiting since January :) [02:13:25.0000] <annevk> I'll appoint Philip` [02:13:29.0000] <annevk> he's always timely [02:14:07.0000] <annevk> time to fly [02:18:05.0000] <matjas> TIL html.org is owned by the W3C now — that’s new, right? [04:29:15.0000] <Ms2ger> The note on http://dev.w3.org/csswg/css3-3d-transforms/ could stand out more... [04:32:05.0000] <jgraham> The "please test" junk on the new spec makes me unhappy [04:33:35.0000] <Ms2ger> I don't get that, use a tested browser ;) [04:39:29.0000] <Philip`> I like how the new spec uses SVG images that have what looks like ugly raster scaling artifacts in Firefox [04:39:35.0000] <Philip`> (Particularly the text is ugly) [04:40:04.0000] <Philip`> /me just be misinterpreting a lack of hinting or something [04:56:28.0000] <annevk> matjas: yeah, was a donation apparently [04:57:04.0000] <annevk> matjas: http://lists.w3.org/Archives/Public/www-archive/2012May/0013.html [04:58:44.0000] <matjas> cool [05:07:23.0000] <matjas> i wonder what they’ll use it for [05:30:42.0000] <odinho> MikeSmith: If you want to research, and don't need actual suggestions, you could look at Postorius, and Hyperkitty. I think Posterius is tied to the Mailman 3.0 REST API, so guess it won't work. I don't know much about hyperkitty, it kinda seems like an earlier name of postorius, so it will probably be just as unusable :P [06:42:29.0000] <zcorpan> AryehGregor: having the stack in the message in testharness.js is quite verbose and most of the trace is about the harness itself, which isn't overly helpful [06:46:27.0000] <zewt> pretty annoying that we still don't have any legitimate stack trace API in JS [06:46:57.0000] <jgraham> zewt: But it might be used for *evil* [06:47:08.0000] <jgraham> And no one is allowed to use javascript for evil [06:47:14.0000] <jgraham> Crockford forbids it [06:47:19.0000] <zewt> heh [06:47:28.0000] <zewt> well, we can already get stack traces (in most browsers), just badly :) [06:48:40.0000] <gsnedders> ES7 may well define it. [06:53:46.0000] <jgraham> Will ES7 come out before or after flying cars go mainstream? [06:56:17.0000] <gsnedders> Well, ES6 will be this year. [07:02:11.0000] <zcorpan> ES3 1999, ES5 2009, ES5.1 2011, ES6 2012, ES7 ??? [07:02:56.0000] <zcorpan> ES1 and ES2 were 1997 and 1998 [07:03:02.0000] <gsnedders> I dunno. Many of the proposals are beginning to stablize. [07:03:24.0000] <zcorpan> so they burst out three editions every decade [07:04:11.0000] <zcorpan> which means ES7 will be some time around 2020 [07:04:42.0000] <Ms2ger> Or as early as 2017 [07:04:52.0000] <Ms2ger> Still 5 years ahead of HTML5 [07:08:37.0000] <jgraham> There is no HTML5, there is only HTML [07:13:01.0000] <volkmar> Ms2ger: we should have HTML 2020, HTML 2021, HTML 2022 :) [07:21:34.0000] <Philip`> I guess HTML 2020 would be the best and then they'd get progressively more shortsighted [07:25:42.0000] <jgraham> (tumbleweed) [07:39:05.0000] <Philip`> (Optometry jokes are hard :-( ) [07:40:05.0000] <annevk> uploading a large file to DreamHost is no fun [07:40:16.0000] <annevk> starts at MiB/s [07:40:21.0000] <annevk> now KiB/s [07:40:27.0000] <annevk> oh now done [07:40:28.0000] <annevk> yay [07:50:43.0000] <odinho> annevk: So why the uploading of big files to DreamHost òÓ [07:52:07.0000] <annevk> I made a one-minute video for someone and I don't think 70MiB works as an email attachment [07:52:15.0000] <annevk> but maybe it does these days [07:52:21.0000] <jgraham> That's not a big file [07:52:33.0000] <annevk> jgraham: yes it is [07:52:37.0000] <odinho> Heh [07:52:45.0000] <Philip`> /me thought everyone used Youtube for videos nowadays [07:52:47.0000] <annevk> most of my files are several KiBs [07:52:59.0000] <odinho> I normally use BitTorrent for it, -- but those files are normally several hundreds of gigabytes. [07:53:01.0000] <annevk> Philip`: doesn't work great for sharing the raw data [07:53:24.0000] <Philip`> Ah [07:53:38.0000] <karlcow> Philip`: vimeo, dailymotion too [07:53:48.0000] <jgraham> youporn [07:54:14.0000] <odinho> jgraham: ... Well, now we know what kind of videos you share. [07:54:29.0000] <karlcow> hmm… if we go into hipsters video hosting sites [07:54:32.0000] <jgraham> I don't think I have ever shared a video of anything [07:54:41.0000] <odinho> s/share/make/ [07:54:52.0000] <zcorpan> "you wanna see a video of my cat? i uploaded it to youporn. here's a link." [07:55:01.0000] <jgraham> I guess I probably can't claim I have never made a video of anything [07:55:05.0000] <odinho> lol, would be cool to try. [07:55:14.0000] <odinho> See how long before it is taken down for not being explicit. [07:55:25.0000] <jgraham> I don't think I have ever made anything that some one surfing youporn would be looking for :) [07:55:28.0000] <annevk> it would be disturbing if it was odinho [07:55:45.0000] <odinho> annevk: If what was? [07:56:01.0000] <jgraham> the porographic video annevk is uploading [07:56:07.0000] <jgraham> That's what's happening, right? [07:56:08.0000] <odinho> :o [07:56:24.0000] <annevk> I was referring to the cat video [07:56:54.0000] <odinho> annevk: So, it'd be disturbing if I was zcorpan's cat. Yes, I can agree with that. Thankfully I'm not :] [07:57:10.0000] <Philip`> Whose cat are you, then? [07:57:32.0000] <odinho> Philip`: I'm my own cat. [07:57:58.0000] <odinho> Just like the tomcat in aristocats. [07:58:14.0000] <annevk> an employed wild cat? [07:58:41.0000] <odinho> /me sees his story starts crumbling [07:58:52.0000] <jgraham> Mmm, crumble [07:59:00.0000] <annevk> apple crumble [07:59:11.0000] <annevk> yumyum [07:59:22.0000] <odinho> I bought yumyum yesterday. [07:59:36.0000] <jgraham> /me has quite some rhubarb at the moment. Makes very nice crumble [07:59:48.0000] <karlcow> https://en.wikipedia.org/wiki/I_Am_a_Cat [08:00:04.0000] <jgraham> (not as much as I would like of course. It is not possivle to have too much rhubarb) [08:00:42.0000] <odinho> jgraham: True dat. [08:03:22.0000] <jgraham> I wonder if I will be brave enough to split the crowns later in the year [08:03:40.0000] <jgraham> Truly the most republican-sounding of all gardening tasks [08:04:38.0000] <Ms2ger> Not in this jubilee year, of course [08:11:58.0000] <odinho> Ms2ger, jgraham: Waddaya think about my last comment? [08:13:15.0000] <Ms2ger> Er, yes, rhubarb makes good crumble for you humans [08:14:01.0000] <jgraham> Indeed. I recommend homemade custard [08:14:17.0000] <jgraham> to go on the side [08:14:24.0000] <jgraham> Or homemade ice cream [08:14:44.0000] <odinho> Ms2ger: On github :P [08:14:51.0000] <jgraham> (I have also made rhubarb crumble icecream) [08:14:52.0000] <Ms2ger> What's github? :) [08:14:52.0000] <odinho> Making a commit now. [08:14:58.0000] <jgraham> (that was nice) [08:15:12.0000] <jgraham> Ms2ger: The clue's in the namew [08:15:17.0000] <jgraham> -w [08:15:21.0000] <odinho> This is no environment to get any work done :P [08:15:56.0000] <Ms2ger> /me did not get email [08:19:23.0000] <odinho> Ms2ger: https://github.com/jgraham/testharness.js/pull/1#r900500 << this is new stuff [08:20:49.0000] <odinho> Ms2ger: Think I'll make it into a patch now, -- and dump it on jgraham as long as it's OK. [08:22:02.0000] <Ms2ger> odinho, what was "I actually think adding || required_props.code == 0..." in reply to? :) [08:22:42.0000] <odinho> This one --> Right now, it's possible to just throw code = 0, and no name and it'll pass for e.g. IndexedDB spec's DataError or TransactionInactiveError. [08:24:22.0000] <Ms2ger> Please update the documentation for assert_throws too? [08:25:16.0000] <odinho> Ms2ger: Smart. Hmm. [08:25:38.0000] <Ms2ger> And explain what we're checking if you pass in an object [08:27:27.0000] <Ms2ger> odinho, in the part where you check required_props.code === undefined, could you use `in` there too? [08:28:12.0000] <odinho> Ms2ger: Sure. [08:28:43.0000] <Ms2ger> And in the if (e.name !== e.name.toUpperCase() || required_props.code === 0), add a comment along the lines of "New style exception: also test the name property." [08:37:10.0000] <matjas> probably waaaay too easy for you guys, but here goes: http://twitter.com/mathias/status/207858124802097154 [08:38:11.0000] <Wilto> matjas: A REGEX-BASED HTML PARSER???? [08:38:20.0000] <Wilto> [cue: Zalgo] [08:38:37.0000] <matjas> Y̸͉͚̘̯̥̠̋̌̋̈́̏̃͒͆ͦ̓ͧ͛ͫ̀͐͗̇ͮ͘e̸̴̛͉̳̼̼̯ͥ̑̏͆ͩ̃̆̅ͬ̌̍͗̀̚͘sͦ̒̔̏̾̆̇ͮ͊ͤͣͮ̂҉̶̹̪̘̘̳̞̞͍̞̝̭̪̻̕ͅ,̴̥͎͇̲̠̻̱͇͓̯̭̦̯̟̣̟̤̅ͩͩ͌ͦ͗ͪ̇̒͋͂ͧͩ̚͜ ͪ̾̿ͩ̓ͫ͐ͮ̓҉̶͔͉̯̫̙̞̟̤̗̹̹̕͠ͅW̶̸̶̝̪̞̩̤̰̉̇͗͛͒͌͛̽ͦ͛͑̕i̔̒͆͐ͪ͗ͩ͐̌̑̓̒̋ͥ̈͑̕҉͟͏̢̱͖͈͉͓͙l͑̒ͧͩ̾̒̆̇ͣͦ͠ [08:38:38.0000] <matjas> ̸̨͇̘̩̦̣̼ṯ͎̤̱̺̤̜̜̯̦͖͇͓̋̉̂́ͮͦ̑̃͂͆͟͡ͅö̷̔͐ͬ̓͏̨̖̦̙͖͇͎̹̠͔͘.̧̥͓̹̜̹͖̩̙̭̯̪̣̺̘̅͗̎ͩ̔̄ͬ͑͟ͅ ̧̢͉̱͈̫͚̬ͮ͐ͮ͝A̡͍͇̥͖̣̜͎͙̦̖̻̼̅̈̆ͭͬͧ̔͐ͩ̏̍̍ͫ̀͘͡ ̨̛̪̲͈͇͖̯̫͔̟͈̒ͧ͋ͬͧ̎ͤ͂͑͐̔̚͠r̪͚͔̝̩̐͌̽͗ͨ̏̇̿ͣ͆̉̈̽͊̈́̅̒ͧ͘͢͜͜͠ͅę̬̘̰̘̤̣̱̫̟͖͐̔̋̃ͤ̾̾̓̔́̎̔͡ [08:38:38.0000] <matjas> ̘g̡͖̹̙̣͉̭̥̤̝̯͈̝̣̳̬͂͊̄̽ͤ͑͢͠͡e͔̙͖͍͎̟̣̬̰̟͓̍͂͌͗̃̄̇̑ͭ̉̉ͦͥ͗ͣ̐̈́͢͠x̬͚͙̣̥͖̲̞̣̠̞̍͒ͫ̒ͮ̐̽̇͆̂͂͊̚͜͜-̛̟͓͙̹̱͚̙̟̣̭̙̤̳̖̤̥̜̐ͮ̉͆̏̽̊̌̓̀̀͢͝b̴̸̸̢͚̼̭̞̥̭̬̣̘̮̮̣͇͈̪̹̜̓̇͗̽á̺̥͎̰̲̺͖͉̬̗͓̪̫͉̘̯ͣ͆̄ͪ̈̓̒͐̄̍ͩͨ́͆́͢ͅͅs̨͓͔̲̜̗̥̦̦͚̖̻̫͍͈̍̈ͩ̏͐ͤ̃e͑ [08:38:38.0000] <matjas> ̷̙̗͇̬̭͕̘̞̯̣͕̟̩̑͂̒͐͋ͨ͊͆͆̇̏ͩ͛͋ͯ̕͟͝͞d̡ͤ̃ͧ̋͒̎̌̌ͫ҉͕̩̰̭̻͎̠̗̙̭̼̪̀ ̝̥̰͎̼̙̮̫̮͚͖͖͎̫̐̎̈́ͤ͒ͪ̏ͦ͗͛ͣ͑͋͠͡ͅH̶̢̛̠̪̼͎͖͇̠̲͕͓̥͇͍̬̟̘̟ͨ̂ͪ͌̋̈͌̋͗̑̆͊͒̔̌̀͝T̛̘̝̬̜̜̝͈̣̗͍̤̝̤̤̣̎͋͛̿͛ͧ͒ͤ̃̄̀̒̚͘͠͠͠M̵͈̼͈̜͔̗̟̯͖͎̜̦̈́̒ͧ͗ͩ́̾ͨ̉ͪͨ͛ͦ͟͝Ļ͐̒̃̃͏̣̘̠̝͖̜ [08:38:38.0000] <matjas> ̤̗̫̥̣͙͇̭̘ ̶̶̵̡͙̰̼͈͔̩̹̻̻̫̪̓͊́̿ͪͮ͆͐ͭ̄̔ͪͨ̾͋́̿̎̍͝p̴͑ͮ̏̓̅̌̉ͭ̀͐̄ͧ̽҉̴̶̭̖̮̦̣͈͚͓̦̠̤͍̫̱̩͎͢ă̷̶̪̮̼̹̬̯̫͖͈̻̦̪͈̭͚̩͇̞̔̃̄ͭ͆͗̄̂͋͌́͜r̢̛̫͓͓͕̫͙͈̟͇͚̮͔̜̰̥̄̋̌̔ͨ͑͐̏̔͗ͨ̔̀͘͠ş̸̸͔̥̩̰̦̹̝̳͖̲̻̣̮̭̲͔͈̅̔̈̾͊̔́ͣ̓͋̿̋̓̚͘e̵ͩ̅͊ͮͣ͗̊̽̏̀͗̒͒̓̊ͩͬ̊ [08:38:38.0000] <matjas> ̴̶͍͍̲̱̟̺̘̥̀͡ͅr̢̟̳͉͈̫͇̒ͮ̒͛ͬͣ͐ͩ̈́ͯ͒ͧ͘͡.̛͋ͨ̓͌͡͏̨͓̱̺̠̪̭͈̩ [08:38:51.0000] <miketaylr> matjas: my answer will fail if you come from an enterprise shop, most likely [08:39:14.0000] <matjas> go for it [08:39:27.0000] <miketaylr> just @'d you [08:39:36.0000] <matjas> ah, no i don’t think that’d work [08:39:40.0000] <matjas> if it does i just got owned [08:39:54.0000] <matjas> hahahaa omg i suck [08:40:01.0000] <matjas> i had something completely different in mind [08:40:01.0000] <miketaylr> :D [08:40:14.0000] <miketaylr> hmm what's the other way [08:40:16.0000] <miketaylr> /me looks [08:40:20.0000] <annevk> matjas: more like #dompubquiz [08:40:49.0000] <matjas> annevk: yeah, i was considering adding multiple hashtags!!1 [08:41:02.0000] <TabAtkins> Hah, miketaylr, that was my first thought too. I just wasn't certain that document had the node methods. [08:41:40.0000] <matjas> sometimes i miss document.innerHTML [08:41:48.0000] <annevk> matjas: better question: why ever access document.doctype? :) [08:42:06.0000] <matjas> fo’ funsies [08:42:22.0000] <annevk> "because you can and because Acid3 made it so" [08:42:24.0000] <odinho> :] Writing annoying analytics/stats package? [08:43:25.0000] <miketaylr> i was using it w/ jsdom to modify some html files, http://miketaylr.com/post/4e85dc30.png [08:43:32.0000] <miketaylr> but then i removed jsdom because i hate it [08:43:43.0000] <Philip`> Surely the best approach is to tell everyone to use <!DOCTYPE html> and then never worry about doctypes ever again [08:43:50.0000] <matjas> odinho: no, just learned about //node() in XPath today, and noticed it even gets the DOCTYPE node [08:44:01.0000] <matjas> (i’m a complete XPath n00b too, fwiw) [08:44:16.0000] <annevk> data:text/xml,<!DOCTYPE html><svg/> [08:44:24.0000] <odinho> matjas: Ah, I was answering annevk what you could do with it. :] [08:44:30.0000] <Ms2ger> BSOD [08:44:38.0000] <Ms2ger> Well, YSOD / RBOD [08:44:46.0000] <annevk> Ms2ger: not in Opera [08:44:55.0000] <Philip`> Solution: data:text/html,<!DOCTYPE html><svg/> [08:45:03.0000] <Ms2ger> Hence, the Red Box Of not-quite-Death [08:49:38.0000] <TabAtkins> annevk: No need to ever use a doctype for SVG. [08:50:16.0000] <matjas> (^ for XML in general) [08:52:11.0000] <miketaylr> matjas: i think i found the other way, document[document.toString().slice(12,15).toLowerCase()+TypeError.toString().slice(9, 13).toLowerCase()] [08:52:36.0000] <TabAtkins> Booooo [08:52:43.0000] <tantek> what are folks' opinions on lowercase doctype vs uppercase DOCTYPE? e.g. <!doctype html> (faster to type, no shiftkey) vs. <!DOCTYPE html> ? [08:52:43.0000] <matjas> miketaylr: that’s cheating! [08:52:57.0000] <TabAtkins> tantek: <!DoCtYpE hTmL> [08:53:10.0000] <matjas> tantek: I like <!DOCTYPE html> as SGML always uppercases DOCTYPE [08:53:13.0000] <miketaylr> matjas: ok, but it was a good way to kill 2 minutes [08:53:14.0000] <tantek> TabAtkins - that's only because you can't use upside down letters [08:53:19.0000] <TabAtkins> tantek: True. [08:53:25.0000] <jgraham> I think my opinion is that time spent debating it would be better spent making crumble [08:53:29.0000] <TabAtkins> But anyway, I'm lazy so now I always just write <!doctype html> [08:53:32.0000] <matjas> miketaylr: brilliant use of TypeError, though [08:53:35.0000] <odinho> /me agrees with jgraham [08:53:46.0000] <odinho> (But I do as TabAtkins) [08:54:11.0000] <tantek> which do you use in practice odinho jgraham? (not a debate, more of a poll) [08:54:36.0000] <TabAtkins> Same reason I usually use single quotes when I have to quote an attribute. [08:54:38.0000] <TabAtkins> No shift. [08:54:39.0000] <Philip`> tantek: How do you write the "<!" without the shift key? [08:54:58.0000] <tantek> good catch Philip :) [08:55:04.0000] <odinho> tantek: Looking over all my tests at w3c-test.org, they're all <!doctype html> if I wrote them myself (not porting other people's :P) [08:55:07.0000] <Ms2ger> Philip`, French keyboard [08:55:15.0000] <miketaylr> i use <!DOCTYPE html> because it looks cooler [08:55:40.0000] <Philip`> tantek: Surely it's better to keep the shift key held down for all of "<!DOCTYPE" because then you've got the entire duration of the " " to release the shift key, whereas with "<!doctype" you've got to release it very quickly else you'll end up with "<!Doctype" [08:56:29.0000] <TabAtkins> Philip`: I'm not sure what kind of keyboard you use, or how you type, but I can't "keep the shift key held down" for all of those letters. [08:56:58.0000] <Philip`> I always do <!DOCTYPE html> largely because KDE's syntax-highlighting text control thing complains about <!doctype ...> in underlined red [08:57:02.0000] <Ms2ger> How many fingers do you use? :) [08:57:21.0000] <Philip`> TabAtkins: I have multiple fingers, so I can use my right little finger for shift while typing with the rest of my hands [08:57:54.0000] <TabAtkins> It feels supremely unnatural to try and hold down a shift on the same side as the letter I'm typing. [08:58:08.0000] <TabAtkins> When I do happen to write it in caps, I just use caps lock. [08:58:13.0000] <Philip`> I only do the O and P with my right hand [08:58:33.0000] <TabAtkins> Wait, the P is done with your pinky though, right? [08:58:48.0000] <jgraham> Presumably not [08:58:50.0000] <Philip`> No, that's done with my middle finger if I'm holding shift [08:59:00.0000] <TabAtkins> Okay, so I'm back to that just being weird. [08:59:11.0000] <Philip`> or with the next outermost finger if I'm not holding shift [08:59:24.0000] <TabAtkins> That's the ring finger. [08:59:44.0000] <TabAtkins> And that's still really weird. It's quite difficult for me to hit P with my ring finger. [08:59:59.0000] <TabAtkins> While it's directly above the ; key, so quite natural to use my pinky. [09:00:06.0000] <matjas> Philip`: better answer: use a custom keyboard layout [09:00:17.0000] <Philip`> /me doesn't have any kind of consciously consistent pattern for what fingers he uses - he just hits the keys with whatever happens to be nearest at that instant [09:00:23.0000] <Ms2ger> Use a hotkey for <!DoCtYpE hTmL>? [09:00:31.0000] <odinho> This is getting unweild. [09:00:48.0000] <matjas> DOCTYPEs are too mainstream. [09:01:13.0000] <matjas> stop using them & you get box-sizing: border-box for free!!11 [09:01:28.0000] <tantek> lol [09:12:38.0000] <tantek> annevk - on the topic of fullscreen - are you ok with the suggested edits from glazou? http://lists.w3.org/Archives/Public/www-style/2012May/1129.html [09:23:10.0000] <tantek> and fantasai's comments are constructive/helpful as well: http://lists.w3.org/Archives/Public/www-style/2012May/1131.html [09:23:10.0000] <miketaylr> matjas: oh thought of another one, document.all.item(0).previousSibling [09:26:09.0000] <Ms2ger> Objection! [09:26:37.0000] <Ms2ger> Also, not necessarily with whitespace [09:29:07.0000] <miketaylr> :'( [09:29:30.0000] <dglazkov> good morning, Whatwg! [09:30:21.0000] <Ms2ger> Reasonable afternoon, dglazkov [09:30:35.0000] <dglazkov> \o/ reasonable is almost good [09:32:01.0000] <tantek> good morning dglazkov [09:51:38.0000] <odinho> https://github.com/Velmont/testharness.js/commit/0cba0a7a006cb9493a9f9d3eea5a64d0e606832a < docs look reasonable, anyone? [09:52:56.0000] <Ms2ger> odinho, it checks .name for the object, no? [09:54:32.0000] <odinho> Ms2ger: Yeah. But I think it should do what I wrote, although I see that is rather stupid to do that. I guess I thought a bit too much about it and wrote it in. [09:55:05.0000] <Ms2ger> I mean, I'd be happy to have the code do what you documented [09:55:25.0000] <Ms2ger> But I want the docs and the code to match [09:55:38.0000] <odinho> I was all like, "it does that? yeah, strange, should maybe fix that. How to do it? Well, could do that. ... <long break>. So, writing this docs stuff. write/write totally wrong stuff, write." [10:43:21.0000] <Ms2ger> odinho, and now the { type: "TypeError" } example :) [10:43:40.0000] <odinho> WÆÆÆ :P [10:43:59.0000] <odinho> So just remove the object example I guess. [10:44:03.0000] <Ms2ger> W���� [10:44:05.0000] <odinho> Because I can't manage to make a sane one. [10:44:22.0000] <Ms2ger> Make it `new TypeError()`, because that's what's actually used :) [10:46:31.0000] <odinho> There, pushed. [10:50:07.0000] <Ms2ger> <Ms2ger> odinho, in the part where you check required_props.code === undefined, could you use `in` there too? [10:50:18.0000] <odinho> Ms2ger: I tried. [10:50:30.0000] <odinho> Ms2ger: Didn't work. :P Because it does set code on required_props [10:50:35.0000] <odinho> Ms2ger: But it sets it to undefined. [10:50:38.0000] <Ms2ger> Yeah [10:52:21.0000] <Ms2ger> I was thinking of var name_code_map = { ... }; if (!(name in name_code_map)) { throw... } required_props.code = name_code_map[code]; [10:57:19.0000] <odinho> Ms2ger: Ah, you want to expand that one. Sure that's needed? It was like that before. ... It might be a bit clearer, yea. [10:57:34.0000] <Ms2ger> I would prefer it :) [10:58:24.0000] <odinho> Ms2ger: Will have a code_name_map and a name_code_map then, where code means different things in the two different ones :P [10:58:53.0000] <odinho> assert_throws taking "code" is a bit strange now already. [10:58:55.0000] <Ms2ger> We already overload "code", no? [10:58:57.0000] <Ms2ger> Right [11:05:29.0000] <odinho> Ms2ger: Refactored. [11:06:59.0000] <Ms2ger> Hmm, you could make it var required_props = { code: name_code_map[name] }; [11:07:03.0000] <Ms2ger> Either way, push it [11:07:09.0000] <Ms2ger> (As one patch) [11:09:08.0000] <odinho> Ms2ger: Yeah, I was mindlessly moving code :P [14:37:26.0000] <TabAtkins> Oof, just spent two hours working through a nerd-snipe. [14:37:45.0000] <TabAtkins> On the good side, Animations now has much cleaner markup. [14:38:47.0000] <tantek> TabAtkins, you know often you have to rework markup to "clean it up" in MediaWiki? Like never. [14:39:06.0000] <TabAtkins> MediaWiki markup is horrible and I hate it. [14:39:11.0000] <tantek> I stand by the assertion that specs as we develop them today have more markup than we need. [14:39:33.0000] <tantek> no argument about MediaWiki markup opinions - I'm neutral on it. [14:39:41.0000] <tantek> how do you feel about MarkDown? [14:39:55.0000] <TabAtkins> Shrug. You're wrong. If your assertion is that you would prefer a stronger preprocessor so you do more compact syntax instead of the occasional tag, okay, but you can't actually reduce the amount of syntax we have significantly. [14:40:04.0000] <TabAtkins> Beyond attaching a Markdown processor to the front of the preprocessor. [14:40:11.0000] <TabAtkins> Markdown is great. [14:40:54.0000] <tantek> Markdown is not bad. Except how it does links seems bass-ackwards. [14:41:08.0000] <tantek> which seem quite critical [14:41:12.0000] <TabAtkins> I think doing links in plain text is a hard problem no matter what you do. [14:41:26.0000] <TabAtkins> But I'm okay with how it's done in Markdown. [14:41:32.0000] <TabAtkins> I'd also be okay with the brackets swapped. [14:42:59.0000] <tantek> what gets me about the markup we use in spec development aren't the occasional things like emphasis, keywords, but rather the common text [14:43:00.0000] <tantek> li [14:43:00.0000] <tantek> ke [14:43:15.0000] <tantek> paragraphs, lists all become a noisier because of the markup [14:43:24.0000] <Hixie> welcome to HTML [14:43:24.0000] <TabAtkins> I can sympathize with that. [14:43:47.0000] <tantek> welcome Hixie [14:44:30.0000] <tantek> TabAtkins - so if wikis accepted Markdown - you'd think they're ok? [14:44:59.0000] <tantek> who "owns" the Markdown spec btw? like if someone wants to suggest syntax additions... [14:45:05.0000] <TabAtkins> Gruber. [14:45:05.0000] <Hixie> /me wishes wikis would just use HTML [14:45:22.0000] <karlcow> Gruber owns markdown [14:45:27.0000] <karlcow> and not in a liberal way [14:45:28.0000] <tantek> /me has documented Hixie's preference. [14:45:28.0000] <TabAtkins> tantek: I think Markdown is a technically superior syntax, so yeah. [14:45:52.0000] <tantek> karlcow - do you have a URL to the live spec? [14:46:12.0000] <karlcow> ah no! my bad [14:46:19.0000] <karlcow> Markdown is free software, available under a BSD-style open source license. See the License page for more information. [14:46:20.0000] <TabAtkins> http://daringfireball.net/projects/markdown/ [14:46:28.0000] <karlcow> http://daringfireball.net/projects/markdown/ [14:46:37.0000] <karlcow> I remember reading a stricter license in the past [14:49:50.0000] <karlcow> and really my bad because the files on Web archives show a BSD license at least since 2004. [14:54:42.0000] <tantek> very nice, so thanks to liberal licensing anyone can fork it and improve it. [14:54:54.0000] <tantek> whether the syntax (spec) or the implementation (code) [14:54:56.0000] <TabAtkins> Yes, and has been done a few times. [14:55:07.0000] <tantek> is there a community evolving it? [14:55:27.0000] <TabAtkins> No clue. But I use an extended Markdown parser written in PHP for my blog. [14:55:39.0000] <TabAtkins> (And have been off-and-on working on my own parser.) [14:55:56.0000] <tantek> is there a spec for the extended Markdown syntax that you use? [14:56:14.0000] <TabAtkins> http://michelf.com/projects/php-markdown/extra/ [14:58:03.0000] <TabAtkins> Oh man, I forgot that it had footnote support. [14:59:57.0000] <tantek> /me reads - hey that's pretty slick. [15:00:39.0000] <TabAtkins> I mainly use it just for the code block guards. Indenting my code blocks like regular Markdown wants is annoying. [15:00:40.0000] <kamals> Any girls here [15:00:48.0000] <TabAtkins> Oh man, so many girls, kamals. [15:00:51.0000] <TabAtkins> You wouldn't believe. [15:00:57.0000] <TabAtkins> Girls coming out my fucking ears. [15:01:09.0000] <TabAtkins> None of them like you, though. [15:01:18.0000] <kamals> I dont know how to fing the girls [15:01:50.0000] <TabAtkins> First suggestion: try the real world. Plenty of girls out there. I hear estimations that upwards of 50% of people are girls. [15:01:59.0000] <Hixie> more than 50% [15:02:06.0000] <TabAtkins> Thus my phrasing, Hixie. [15:02:27.0000] <Hixie> i thought "upwards of x" meant "up to x", my bad [15:02:37.0000] <TabAtkins> ...I don't *think* so, but I"m not sure. [15:02:38.0000] <dglazkov> /me shakes head [15:02:43.0000] <Hixie> no, no, you're right [15:02:58.0000] <Hixie> i looked it up before saying "my bad" :-P [15:03:19.0000] <TabAtkins> Haha [15:08:45.0000] <dglazkov> poor kamals. His quest was unsuccessful. [15:11:09.0000] <TabAtkins> tantek: Oh man, the <abbr> support in Markdown Extra is great too. So much I never actually paid attention to and then forgot. [15:12:13.0000] <tantek> dglazkov, indeed, why aren't more women involved in WHATWG? [15:12:41.0000] <Hixie> or indeed the IT industry in general [15:12:50.0000] <dglazkov> I would say we are definitely failing the upwards to 50% criteria. [15:13:09.0000] <tantek> well there's been plenty of articles written about IT industry in general - with various theories. simple web search will suffice there. [15:13:16.0000] <dglazkov> I blame Hixie. He started WHATWG. Surely it's his fault. [15:13:18.0000] <tantek> but does WHATWG have *any* active women? [15:13:24.0000] <Hixie> /me hopes his effors with FIRST will help in the future, but that's a long-term solution [15:13:34.0000] <Hixie> tantek: dunno. i thought anne was a woman for a while. [15:13:36.0000] <tantek> FIRST(citation?) [15:13:38.0000] <Hixie> turns out not [15:13:46.0000] <Hixie> tantek: http://www.usfirst.org/ [15:13:54.0000] <tantek> Hixie, we could get anne to wear a dress perhaps? [15:15:07.0000] <Hixie> /me actually has no idea what gender people in the whatwg have (other than those he's met), and pays no attention to it [15:15:43.0000] <tantek> Hixie: re: FIRST agreed. Similarly that's why I help co-organize ScienceHackDay (when it's in SF) http://sciencehackday.com [15:15:49.0000] <Hixie> cool [15:16:19.0000] <tantek> and yes, inspiring kids (of all genders/backgrounds) is a good long term approach. [15:16:27.0000] <Hixie> (i expect the reason for the imbalance in web standards is the same as for the IT industry as a whole, i don't think there's anything special about web standards) [15:16:45.0000] <tantek> web standards appear to be worse than the IT industry in general [15:17:03.0000] <tantek> just the small sample during the WG f2f at MS-SVC the other week [15:17:12.0000] <Hixie> do we have data on that, or is that just an impression? [15:17:15.0000] <tantek> there were a few women in the HTML WG. none in the WebApps WG. [15:17:20.0000] <Hixie> cos impressions on this kind of thing can be really misleading [15:17:28.0000] <gavinc> Err, I don't think FIRST has any women/girl specific mission does it? [15:17:31.0000] <Hixie> (in both directions) [15:17:57.0000] <Hixie> gavinc: it's about bringing engineering up in general, for all genders, nothing female-specific as far as i know. [15:17:58.0000] <tantek> Hixie, I have yet to find a WG with a % of women even close to what averages in IT are. [15:18:02.0000] <gavinc> There is Girls Connect, but I thought that was a small part of FIRST [15:18:33.0000] <gsnedders> /me doesn't even know how many girls are in his CS course at uni, as lecture attendence is so low [15:19:15.0000] <tantek> I do think W3C has tried to be specific about outreach, e.g. in attendance at TPACs. [15:19:33.0000] <tantek> And I vaguely remember the CEO quoting improving statistics year over year [15:19:46.0000] <tantek> but that's TPAC attendance, not WG membership/participation [15:20:27.0000] <Hixie> gavinc: i dunno what the %s are, but (a) the proportion of girls to boys is much higher in FIRST events than in the IT industry as a whole, and (b) since the problem with the IT industry seems to be about girls losing interest before picking a career, anything that reaches children at a young age is huge for this kind of thing [15:20:27.0000] <tantek> still, unless someone wants to stand up and provide a counter-example, from all outside (and even semi-inside) perspectives, it does appear WHATWG has no women. [15:20:39.0000] <tantek> Hixie +1 [15:20:40.0000] <Hixie> gavinc: (there are numerous girl-only teams every year, just like there are boy-only and mixed teams) [15:21:03.0000] <Hixie> tantek: what do you mean by "WHATWG"? [15:21:14.0000] <tantek> Hixie, what do you? [15:21:21.0000] <gavinc> Hixie: Yeah, same was true of Odyssey of the Mind [15:21:21.0000] <Hixie> tantek: if you mean the list as a whole, there's over 1000 people subscribed, and i've no idea what the proportions are [15:21:48.0000] <tantek> sure, but can you provide any (even one) counter example? [15:22:02.0000] <tantek> or in irc [15:22:06.0000] <tantek> or wiki contributors [15:22:24.0000] <Hixie> tantek: if you mean spec editors, which is where i think we really need good people, then there's a lack of people of all sexes, and i'm not sure the numbers are big enough that we can tell what the proportions are relative to the industry as a whole [15:22:28.0000] <Hixie> tantek: statistically [15:22:47.0000] <Hixie> tantek: but certainly i agree that there aren't enough active female editors [15:23:02.0000] <gsnedders> tantek: Are there even that many women working for browser vendors? As that'll have some knock-on effect. [15:23:03.0000] <Hixie> tantek: silvia, e.g. [15:23:13.0000] <Hixie> tantek: if you just want an existence proof :-) [15:23:24.0000] <tantek> gsnedders - yes, plenty of women working at Mozilla for example, and on other working groups, e.g. fantasai on CSS. [15:23:31.0000] <Hixie> (nessy on irc) [15:23:48.0000] <TabAtkins> Yeah, within the browsers themselves, the women fraction is probably around IT average. [15:23:56.0000] <gavinc> I admit part of my introduction to standards and the W3C was Jeni Tennison and I assumed there were plenty more like her... and was very surprised. [15:24:18.0000] <tantek> Hixie - good to know there's at least one exception. [15:24:18.0000] <wilhelm> Hixie: Reaching out to kids is a great approach. However, our industry (and its bridgeheads within education) has fundamental cultural problems we need to solve for them to remain interested. [15:24:32.0000] <Hixie> gavinc: there's plenty of other people like jeni in the web standards world, they're just mostly male. :-P [15:24:41.0000] <tantek> wilhelm - the dropoffs occurs at various levels in school before they even get to industry. [15:24:48.0000] <wilhelm> tantek: Yes. [15:24:52.0000] <gavinc> Hixie: Yes. [15:24:53.0000] <gsnedders> tantek: Mainly saying that because — at least within the Core team — at Opera there are very few. Get out into mobile delivery teams and other such things and there are far more. [15:24:55.0000] <tantek> women in STEM drop-off starts in highschool/college. [15:25:05.0000] <Hixie> wilhelm: there may well be cultural issues too [15:25:17.0000] <Hixie> wilhelm: but i don't think those are female/male-specific [15:25:36.0000] <TabAtkins> Well, a lot fo the IT industry (like most industries, unfortunately) is very sexist. [15:25:47.0000] <Hixie> wilhelm: e.g. i find the w3c culture abominable, and i'm a guy :-P [15:25:51.0000] <TabAtkins> You see this pretty easily at conventions. [15:25:54.0000] <tantek> Hixie, cultures that tend to be against women literacy are problematic. [15:27:13.0000] <wilhelm> Hixie: I disagree. I don't have anything else than anecdotal evidence, but I've observed that in mostly male collectives, males may behave like if they were participating in a college fraternity. [15:27:14.0000] <gavinc> Mmm, how many staff contacts at W3C are women? [15:27:49.0000] <Hixie> wilhelm: if you see anything even approaching such a thing in the whatwg list, please let me know so i can stamp it out [15:28:13.0000] <Hixie> wilhelm: (i am culturally insensitive to that kind of thing) [15:28:37.0000] <wilhelm> Hixie: That's the difficult part. In formal discussions, people generally behave well. It's in the informal settings things go bad. [15:29:23.0000] <Hixie> wilhelm: well the only informal setting for the whatwg is here, really, and i'm not sure i'd say that #whatwg's culture was particularly friendly to anyone, male or female, who isn't already in the "in crowd" :-P [15:30:09.0000] <wilhelm> Indeed. I haven't observed anything particularly bad here either. (c: [15:31:01.0000] <TabAtkins> If you see anyone chanting BROS BROS BROS in a bar, go ahead and confront them there too. [15:33:08.0000] <wilhelm> You don't have to venture very far to see the bad stuff, though. I've seen this happen at Opera, at the University of Oslo and a number of other places. Heck, even the “why are there no women in this industry” discussions at Hacker News are a disgrace. [15:33:31.0000] <gsnedders> TabAtkins: Even if I do it sarcastically? [15:33:51.0000] <wilhelm> gsnedders: Actually, yes. [15:34:54.0000] <Hixie> especially so [15:35:02.0000] <tantek> gsnedders, yeah - it still creates a unfriendly environment, even as humor/sarcasm. [15:35:24.0000] <tantek> also, "being ironic" or humor is a frequent defense of that kind of sexist behavior [15:36:34.0000] <wilhelm> gsnedders: Just take a look at your mail archive! Dig up the “rape joke” thread on nonsense⊙oc That was probably intended to be ironic. It's a perfect example of the frat boy culture that plagues our industry. [15:36:48.0000] <Hixie> o_O [15:36:58.0000] <wilhelm> That's what I thought too. [15:37:00.0000] <gsnedders> I guess this is rather how I end up often being percieved as homophobic, which as plenty of people who've met me can say is laughable. [15:37:09.0000] <gsnedders> wilhelm: There's a reason I'm not subscribed to nonsense. :) [15:37:28.0000] <tantek> for more on this, here's a Google News search: https://www.google.com/search?tbm=nws&q=brogrammer [15:37:43.0000] <tantek> (and that's just recent phenomena) [15:39:35.0000] <wilhelm> It's a real pity. There's an equally large subset of the female populatuion that could be active contributors here. But they've been weeded out in cultural funnels long before they could get here. [15:39:36.0000] <gsnedders> I think certainly that plenty of people who make such jokes do so deliberately (not necessarily consciously) seeing the victimized group as "less good" in whatever way. [15:39:38.0000] <TabAtkins> gsnedders: I was your age only a few years ago, so I know how funny being sarcastically offensive is. But it's still usually offensive. :/ [15:40:07.0000] <Hixie> what's with the styles on http://dev.w3.org/csswg/selectors4/ ? they make the summary table really much harder to read on a wide monitor than is necessary [15:40:26.0000] <Hixie> or http://dev.w3.org/csswg/selectors3/ for that matter [15:40:38.0000] <Hixie> styles on http://www.w3.org/TR/css3-selectors/ don't suffer from the same problem [15:40:42.0000] <TabAtkins> Which styles? [15:40:48.0000] <gsnedders> TabAtkins: I think the other issue that a lot of people have with me is I'm often deliberately sarcastic and making jokes about really quite horrific things as a way of coping with them myself. [15:40:57.0000] <Hixie> TabAtkins: seems to be forcing a column width or something [15:41:08.0000] <Hixie> TabAtkins: but the forced width is narrower than the table's intrinsic width [15:41:15.0000] <Hixie> TabAtkins: so it forces lots of wrapping [15:41:18.0000] <wilhelm> gsnedders: I love sarcastic jokes at the expense of victimized groups. I just try to be careful about _where_ I say those jokes out loud. [15:41:56.0000] <TabAtkins> Hixie: Interesting. [15:42:16.0000] <Hixie> (forwarding this complaint from someone who found the TR/ page version easier to read because of this) [15:44:34.0000] <Hixie> in other news, the web interface to mercurial is a disaster [15:44:42.0000] <TabAtkins> Yes. [15:44:55.0000] <TabAtkins> I have no idea how someone actually made that and thought they did a good job. [15:44:57.0000] <Hixie> can anyone work out for me what url i should go to to see the changelog for http://dev.w3.org/csswg/selectors3/ ? [15:45:07.0000] <Hixie> i've been trying for like the last 10 minutes [15:45:23.0000] <Hixie> https://dvcs.w3.org/hg/csswg/log?rev=selectors3 didn't seem to be it [15:45:43.0000] <Hixie> aha! [15:45:44.0000] <Hixie> got it [15:45:44.0000] <Hixie> https://dvcs.w3.org/hg/csswg/log/27544a0821fd/selectors3/Overview.src.html [15:46:35.0000] <TabAtkins> Yeah, that's it. [15:47:02.0000] <tantek> I agree. Mercurial web UI is a step down from the CVS one. [15:47:21.0000] <tantek> I am curious how people make these thing so much worse from one tool to the next. [15:47:35.0000] <gsnedders> Well, CVS didn't provide any UI [15:47:36.0000] <tantek> Are people really not smart enough to just copy first, invent later? [15:47:51.0000] <Hixie> gsnedders: he means cvsweb, i expect [15:47:51.0000] <tantek> gsnedders, the paths were nicer to tip [15:47:53.0000] <TabAtkins> I have no idea how I might interact with Systeam about this kind of thing. Last time we tried to make a UI improvement to something (adding line anchors for the mail archives), they balked for a long time and never actually did it. [15:47:58.0000] <tantek> Hixie, right [15:48:23.0000] <TabAtkins> tantek: The paths problem is the whole reason plinss put together his redirects. [15:48:34.0000] <gsnedders> tantek: My point is in some sense they are moving forward — just not equaling what third-party tools for CVS achieved. [15:48:35.0000] <Hixie> tantek: well the good thing about mercurial is that it's distributed, so you could do it on another server somewhere [15:48:44.0000] <tantek> TabAtkins, I've been hanging out in irc://irc.w3.org:6665/sysreq and dropping increasing difficulty requests to see what happens. [15:48:55.0000] <TabAtkins> Ah, cool. [15:49:09.0000] <tantek> using the friendliest of polite language of course :) [15:50:00.0000] <tantek> TabAtkins - it's fairly low traffic, feel free to auto-join it. [15:50:21.0000] <tantek> do you have a URL to the request for the email line anchors? [15:50:55.0000] <tantek> usually timeless is in there too and can be quite supportive and a good interpreter of how w3c's systems work [15:51:33.0000] <TabAtkins> tantek: I'd have to go look it up in the sysreq email archives. [15:51:53.0000] <tantek> yeah [16:02:48.0000] <Hixie> huh [16:02:53.0000] <Hixie> well here's an interesting question [16:02:59.0000] <Hixie> what task source should we use for the task in http://www.whatwg.org/specs/web-apps/current-work/#scroll-to-fragid ? [16:03:41.0000] <Hixie> dom manipulation, i guess? [16:07:44.0000] <zewt> http://www.furnituredude.com/ i wonder if this guy hired htmldude to make his website [16:09:20.0000] <tantek> zewt - glasses houses? http://zewt.org/ [16:10:44.0000] <tantek> *glass even [16:12:03.0000] <zewt> <- not a store [16:14:41.0000] <tantek> still, you can do better than "Forbidden ¶ You don't have permission to access / on this server." certainly if you're going to pick on others' sites. [16:15:09.0000] <zewt> why? [16:15:38.0000] <Hixie> MikeSmith: cvs is being particularly slow currently, fwiw [16:15:51.0000] <tantek> zewt … unless you like being an armchair critic, all talk, no code. [16:16:23.0000] <TabAtkins> tantek: Are you actually suggesting that zewt needs to provide an example of better visual design for that page before he can criticize it? [16:16:30.0000] <zewt> seriously. heh [16:17:05.0000] <zewt> (so sorry for providing passing amusement to people bored on IRC) [16:19:01.0000] <tantek> TabAtkins - no, just something better than "Forbidden ¶ You don't have permission to access / on this server." [16:19:14.0000] <tantek> and yes, a simple visual design is pretty easy, so yeah, add that too. [16:19:26.0000] <TabAtkins> tantek: His own homepage is irrelevant. (Though horrible.) But that store page is unarguably horrible as well. [16:19:38.0000] <tantek> It's ok, it's useful to distinguish talkers vs. doers. [16:19:47.0000] <Hixie> zewt.org isn't his home page, it says right there that you're not allowed to see it! [16:19:56.0000] <tantek> TabAtkins - someone's own homepage provides a good indication of self-dogfooding. [16:20:05.0000] <TabAtkins> tantek: Still irrelevant. [16:20:11.0000] <Hixie> i don't think making "access denied" messages pretty is a high priority for anyone [16:20:21.0000] <zewt> (a page with nothing on it might indicate to you that it's not a homepage; it's a server used for other purposes that doesn't serve a public webpage) [16:20:53.0000] <TabAtkins> Oh man, I never realized what the old display:compact value was supposed to do. [16:21:07.0000] <tantek> TabAtkins - that depends on whether you value self-dogfooding or not. I think self-dogfooding demonstrates both skill and commitment. [16:21:14.0000] <TabAtkins> I love the crazily underspecified stuff in CSS2. [16:21:21.0000] <tantek> and thus quite relevant. [16:21:46.0000] <tantek> TabAtkins - you wouldn't say that if you'd spent years trying to implement it. [16:21:50.0000] <zewt> ("you don't have a hobby of making pointless webpages for nobody? how dare you laugh at comically bad websites!") [16:22:00.0000] <tantek> it starts out funny and interesting sure, but after a few years, you're done. [16:22:01.0000] <TabAtkins> tantek: Oh, I'm sure. [16:22:36.0000] <tantek> it's ok zewt, Twitter is good for that too. [16:23:40.0000] <tantek> And yeah, I expect anyone who wants to speak seriously about evolving the web (as most people in #whatwg seem to be here to do) to be able to at least make a simple decent home page for themselves. If not, then yeah, it kind of undermines credibility. [16:24:26.0000] <zewt> (also, i think i make enough of a case for my own competence in here to not need to put up a content-free website to prove myself; if you disagree; oh well) [16:24:53.0000] <zewt> (comma one of those semicolons) [16:25:24.0000] <Hixie> i think one's feedback ot the list should speak for itself, you don't have to build credibility first [16:25:49.0000] <Hixie> especially since i completely ignore the source of feedback so the "credibility" part would have no effect... [16:26:26.0000] <Hixie> (case in point, before my vacation i was replying to one e-mail rejecting some guy's proposal as inane and almost sent the e-mail before i noticed i was the one who had sent the e-mail in question) [16:26:31.0000] <zewt> one's feedback to the list is what builds credibility, and that does matter for those of us with different day jobs, who have to pick and choose what we spend time on [16:26:46.0000] <Hixie> that's fair [16:27:18.0000] <Hixie> i agree that if you want to be considered by anyone other than me then it makes sense to worry about your credibility [16:27:19.0000] <TabAtkins> Hixie: Hahahahahaha [16:28:05.0000] <zewt> Hixie: it's a bit weird when I search for something, start reading a mailing list post, then go "???" about halfway through and realize it's something I wrote a decade earlier [16:28:11.0000] <tantek> Hixie - good that that method of inbox processing works for you (per the list). I've found it to be more productive to prioritize feedback from creators over talkers. I might eventually get to the talkers (people that only send messages), but then I also have other things I'd rather do with my time. [16:28:40.0000] <zewt> also, one of those most discouraging things that can happen when researching a problem--and this has happened to me more than once--is to search for it and to have the first hit be me asking about the problem [16:29:02.0000] <Hixie> tantek: to each his own method, indeed [16:29:26.0000] <Hixie> tantek: (i find that if i allow myself to dismiss feedback based on who sent it, i start dismissing feedback just because it's hard to fix, not because it's not valid) [16:29:42.0000] <dcheng> 4 [16:29:46.0000] <dcheng> Oops sorry [16:29:48.0000] <zewt> an awesome number [16:30:54.0000] <tantek> Hixie, I get more hard to fix feedback from creators that tend to be based on real world use-cases (that they ran into while creating). And again helps with prioritization, as unblocking creators helps them find more problems. Whereas who knows if theoretical hard fixes will ever block anyone. [16:31:09.0000] <tantek> Again, just a different approach to productivity and prioritization. [16:31:58.0000] <zewt> (afk) [16:32:16.0000] <Hixie> the kind of hard problem i'm talking about is something like "the entire spec's organisation makes a flawed assumption X which results in the spec being wrong in edge case Y", which i could just ignore, but not while keeping the spec in line with reality [16:32:30.0000] <Hixie> (assuming Y is something browsers can't change, however minor it is) [16:32:49.0000] <tantek> /me will keep an eye out for such problems. [16:33:49.0000] <Hixie> MikeSmith: cvs seems stalled (it hasn't progressed since my last comment on the matter 18 minutes ago) [16:33:58.0000] <Hixie> MikeSmith: i'm going to comment out cvs checkins for now, will try again later [16:48:11.0000] <dbaron> Hixie, if you want a permalink to the changelog, you want https://dvcs.w3.org/hg/csswg/log/tip/selectors3/Overview.src.html [16:48:26.0000] <Hixie> ah, cool [16:48:27.0000] <TabAtkins> dbaron: Is that link discoverable through the interface? [16:48:29.0000] <Hixie> is there a link to that? [16:48:31.0000] <Hixie> what tab said [16:48:41.0000] <TabAtkins> I can never seem to find the permalinks. :/ [16:48:49.0000] <dbaron> TabAtkins, afaict, no [16:48:54.0000] <TabAtkins> Awesome. 2012-05-31 [17:02:24.0000] <heycam> Philip`, that SVG-in-<img> is not scaling nicely when zooming the page is https://bugzilla.mozilla.org/show_bug.cgi?id=600207 [17:33:00.0000] <MikeSmith> Hixie: looking into the cvs problem now [17:59:57.0000] <MikeSmith> Hixie: cvs problem should be fixed now [18:06:29.0000] <Hixie> MikeSmith: k thanks [21:11:11.0000] <Hixie> is there any equivalent of the fillText() method's last argument in CSS? something that says "try really hard to fit this text in this width"? [21:11:27.0000] <Hixie> i have a heading that i'd really like to have fit on one line, even if it has to be condensed to fit, regardless of screen width [21:11:49.0000] <Hixie> (it's only two words and on desktops it just looks fine but on some phones with particularly narrow screens it ends up wrapping to two lines which just looks ugly) [22:44:51.0000] <MikeSmith> looks like CSAIL has a major scheduled network outage on the weekend of June 9 and 10 that I think will mean bugzilla and probably a lot of other stuff will not be available [22:45:00.0000] <MikeSmith> mail too I guess [22:45:30.0000] <Hixie> k [22:57:53.0000] <Hixie> does JS have some mechanism by which an object can respond to method calls for methods i haven't yet defined? [22:58:07.0000] <Hixie> like perl's autoload? [22:58:32.0000] <Hixie> i have an object that represents something on the server but at the time the js object is instantiated i don't yet know what the object's type is cos i'm still waiting on the server [22:58:41.0000] <Hixie> the object's methods are all async [22:58:57.0000] <Hixie> so i want the methods, if called before i have the data, to just queue up the info and wait until we have it and then call back [23:01:17.0000] <MikeSmith> Hixie: google finds http://stackoverflow.com/questions/5422754/javascript-equivalent-of-perls-autoload [23:01:27.0000] <MikeSmith> no idea if that's helpful or not [23:02:04.0000] <Hixie> yeah, __noSuchMethod__ is what i want [23:02:10.0000] <Hixie> wonder if they're adding this to new ES versions [23:02:28.0000] <MikeSmith> I thought that was already in 5.1 [23:02:45.0000] <Hixie> https://developer.mozilla.org/en/JavaScript/Reference/Global_Objects/Object/NoSuchMethod (first hit on google) says "non-standard" [23:03:05.0000] <MikeSmith> oh [23:03:33.0000] <Hixie> hm, there's something called "proxies" in some version of JS [23:04:48.0000] <Hixie> blimey [23:05:01.0000] <Hixie> "simplicity" is the first thing on their requirements list [23:05:06.0000] <Hixie> but it's not clear that they managed it [23:06:19.0000] <MikeSmith> speaking of simplicity the powers that be in the CSS WG seem to be doing a great job making it a really good illustration of Conway's Law [23:06:30.0000] <Hixie> fullscreen? [23:06:37.0000] <MikeSmith> bingo [23:08:01.0000] <Hixie> yeah ok this proxies thing is going so far above my head i can't even feel it [23:08:12.0000] <Hixie> i guess i'll try something dumber [00:22:23.0000] <dbaron> MikeSmith, what has the CSS WG done to fullscreen? [00:25:12.0000] <jgraham> In the interests of what journalists like to call "balance" http://www.zdnet.com/blog/btl/women-in-tech-manuela-hutter-sees-endless-possibilities/68158 [00:28:32.0000] <annevk> http://tumbledry.org/2012/05/30/5_things_about_television o_O [00:30:18.0000] <MikeSmith> dbaron: discussion on the chairs list [01:13:29.0000] <annevk> odinho: haha, that IDL typo bug [01:13:43.0000] <odinho> annevk: heh, yeah, embarrasing [01:13:54.0000] <annevk> notices double ;, does not notice glaring typo [01:14:23.0000] <annevk> that's a meme for the "IDL junky" face [01:18:26.0000] <odinho> annevk :P [01:18:45.0000] <odinho> annevk: I can said that was implied, too easy, anyone could see it :P [01:19:24.0000] <wodemaye__> how do i get at the hostname of an href the DOM-standard-compliant way? [01:20:01.0000] <wodemaye__> getElementsByTagName("a")[0 [01:20:04.0000] <wodemaye__> oops [01:20:20.0000] <wodemaye__> is getElementsByTagName("a")[0].href.hostname standards-compliant? [01:21:15.0000] <odinho> wodemaye__: Don't think so, but it should be best to check spec. [01:21:38.0000] <wodemaye__> odinho, any specific place/doc u would look in? [01:21:50.0000] <annevk> wodemaye__: it.s [0].hostname [01:22:04.0000] <annevk> http://whatwg.org/C#htmlanchorelement [01:22:20.0000] <wodemaye__> annevk, is that guaranteed standards-compliant/cross-browser compatible? [01:22:29.0000] <odinho> wodemaye__: Write a test and test it :-) [01:22:37.0000] <annevk> well it's in the HTML standard [01:22:52.0000] <annevk> it probably works everywhere as it's ancient (iirc), but who knows [01:23:15.0000] <wodemaye__> annevk, but it's not deprecated? [01:23:30.0000] <annevk> no [01:24:07.0000] <wodemaye__> annevk, ur link fails. [01:24:25.0000] <odinho> wodemaye__: Works in Opera, Firefox, Chromium, and IE10. So that's them done. [01:24:26.0000] <annevk> DreamHost is having some issues I guess [01:24:49.0000] <wodemaye__> annevk, whatwg is hosted on dreamhost? [01:24:50.0000] <annevk> should work in a bit [01:25:01.0000] <annevk> yeah [01:25:03.0000] <wodemaye__> the goog couldn't spare a bit of hosting space, Hixie ? [01:26:05.0000] <annevk> they probably can, but we like to run the show without using company resources (other than time) [01:27:25.0000] <odinho> annevk: hmmz. the spec at w3, can't click stuff to get backlinks etc. Is that disabled for w3 compat mode? [01:28:15.0000] <annevk> odinho: W3C had issues with running scripts at some point [01:28:22.0000] <MikeSmith> no such issues now [01:28:30.0000] <odinho> wodemaye__: http://dev.w3.org/html5/spec/the-a-element.html#the-a-element http://dev.w3.org/html5/spec/the-a-element.html#dom-a-hostname [01:28:32.0000] <MikeSmith> the backlinks work in the author-view [01:29:06.0000] <MikeSmith> including the multi-page [01:29:34.0000] <MikeSmith> for the full spec, I thought I had them working, for the single-page version at leat [01:30:01.0000] <odinho> wodemaye__: Ah, look at author view instead. http://dev.w3.org/html5/spec-author-view/the-a-element.html#htmlanchorelement [01:30:09.0000] <odinho> MikeSmith: Yes, it did work in author view. :] [01:30:38.0000] <MikeSmith> I should fix it for the full spec if it's not working [01:31:11.0000] <MikeSmith> I guess I would notice these things if I actually used that when I needed to look up something [01:32:01.0000] <odinho> MikeSmith: lol, what do you mean, where do you look? at whatwg? at w3schools? (:P) [01:32:15.0000] <odinho> I gave you some heavy opposites there. [01:32:33.0000] <jgraham> MikeSmith is actually the secret villian behinf w3schools [01:32:45.0000] <odinho> babambmaaaaamm!!! :-o [01:33:02.0000] <MikeSmith> I put the "cool" in w3schools [01:33:05.0000] <jgraham> It is run from his HQ inside a Japanese volcano [01:33:16.0000] <wodemaye__> except there is no "cool" in w3schools. [01:33:25.0000] <wodemaye__> is w3s at all associated with w3c? [01:33:48.0000] <annevk> nope [01:33:53.0000] <MikeSmith> I only write my w3schools content when I'm loaded [01:33:58.0000] <odinho> lol [01:36:27.0000] <MikeSmith> the problem with sites like MDN is that they value accuracy too highly [01:36:34.0000] <MikeSmith> leaves no room for free improvisation [01:37:43.0000] <wodemaye__> loaded? MikeSmith [01:38:51.0000] <MikeSmith> wodemaye__: piped up [01:39:09.0000] <wodemaye__> MikeSmith, piped up? [01:39:11.0000] <odinho> MikeSmith: As a reader, there's also way too much relevant information on those pages, I find it much better when you can just get some incorrectly written super small non-helping description instead. [01:39:25.0000] <MikeSmith> odinho: now you're talking [01:39:26.0000] <wodemaye__> not a native english speaker so i don't know some slang, sorry. [01:39:47.0000] <wodemaye__> does it mean drunk? [01:39:55.0000] <MikeSmith> you'd understand w3schools much better if you used it while listening to Albert Ayler's "Love Cry" [01:40:20.0000] <MikeSmith> there are messages in there waiting to be heard [01:40:24.0000] <wodemaye__> but doesn'nt MDN specifically document the mozilla implementation of the standards? [01:40:32.0000] <Ms2ger> Not anymore [01:40:51.0000] <wodemaye__> Ms2ger, seriously? [01:40:58.0000] <Ms2ger> The documentation for web-exposed APIs should be vendor-neutral [01:42:11.0000] <benvie> they discuss vendor specifics but for all vendors [01:42:21.0000] <benvie> which is useful [01:42:33.0000] <Ms2ger> With help from paul_irish and other Googlers, too [01:43:13.0000] <odinho> Helpful list at the bottom too: https://developer.mozilla.org/en/http_access_control#Browser_compatibility [01:44:15.0000] <odinho> Although stuff does hang in there: Requires Gecko 2.0 (Firefox 4 / Thunderbird 3.3 / SeaMonkey 2.1) [01:44:35.0000] <wodemaye__> web-exposed APIs? [01:44:38.0000] <odinho> Well, Access-Control-Expose-Headers works in most all browsers now AFAIK, not just Gecko 2 :P [01:44:55.0000] <odinho> At least in Opera 12, which I know best. [01:45:14.0000] <Ms2ger> MDN also documents a lot of Gecko internals; not much point in trying to make those vendor-neutral [01:45:41.0000] <Ms2ger> /me glares at his inbox [01:46:35.0000] <wodemaye__> Ms2ger, i almost never used mdn cus i wanted to code to standards and not implementations, but holy shit there's some badass document'n in there. [01:46:42.0000] <wodemaye__> thanks dude. Ms2ger [01:46:50.0000] <Ms2ger> Np :) [01:47:21.0000] <benvie> it's often not clear when you've drifted out of the standard JS stuff into mozilla-specific land [01:47:36.0000] <odinho> benvie: Yeah, -- it would be nice to have them more seperated. [01:47:39.0000] <benvie> but it's usually noe just a section of an article, but like the whole section of the site [01:47:51.0000] <benvie> but they look the same and interlink [01:48:11.0000] <odinho> Although I can really see why Mozilla doesn't want to move it to a more neutral place :P [01:48:18.0000] <Ms2ger> There's been talk of separating them more, not sure how that's going ahead [01:48:20.0000] <benvie> hah yeah [01:48:51.0000] <MikeSmith> do we really not have any contributed test cases for postmsg? [01:49:16.0000] <odinho> I seem to remember Opera and Google wanting to part with their docs for a W3-hosted place like this. But Mozilla has the most thorough and best docs, so would be giving away most stuff by far. [01:49:17.0000] <benvie> well it seems like everyone's moving in the same general direction, which towards more js-centric and standardized in that fashion [01:49:45.0000] <MikeSmith> jgraham: btw and fwiw I agree it'd be better at this point to end the approved/submitted division [01:50:09.0000] <wodemaye__> benvie, what's a good/comprehensive/reliable/accurate documentation effort that aims primarily to document only standards with maybe some notes sprinkled through about implementation but explicitly made known as such. [01:50:19.0000] <MikeSmith> jgraham: I guess details about whether a test is approved or not could go into a manifest or something [01:50:21.0000] <benvie> well [01:50:39.0000] <benvie> I don't know that there is one [01:50:40.0000] <Ms2ger> jgraham, I approve of anything that makes the paths shorter, that would make life easier for thunderbird developers on windows :) [01:50:48.0000] <benvie> MDN does a good job as it is [01:51:07.0000] <benvie> so there's no push to redo its efforts [01:51:31.0000] <wodemaye__> what about the official w3/whatwg docs? those aren't the most readable... [01:51:41.0000] <wodemaye__> would u recommend referring to those often benvie ? [01:51:45.0000] <benvie> from my experience, the W3C specs combined with MDN are all that's eneded [01:51:49.0000] <odinho> MikeSmith: where is this email, if any? [01:51:57.0000] <Ms2ger> odinho, public-test-infra [01:52:08.0000] <benvie> let's see [01:52:13.0000] <odinho> All these lists I'm not on... *finding* [01:52:28.0000] <benvie> http://www.html5rocks.com/en/ is probably a big one for quality [01:52:43.0000] <odinho> benvie: Although often very WebKit-centric [01:52:59.0000] <benvie> yeah it's very much a google/webkit oriented thing [01:53:43.0000] <wodemaye__> benvie, but its fairly authoritative? [01:53:52.0000] <benvie> I guess it's kind of hard to find a party willing to put in the expertise resources to produce this content that ISN'T directly affiliated with one of the vendors [01:54:14.0000] <Ms2ger> Achievement unlocked: fantasai agrees with me on www-style [01:54:36.0000] <jgraham> Ms2ger: You sure? [01:54:47.0000] <benvie> and with how things are standardized, you can get away with doing a laege chunk of standardized...ish content that still favors your specific implementation's strengths [01:54:51.0000] <benvie> just because you know it best [01:54:58.0000] <MikeSmith> Ms2ger: watch out for lightening strikes [01:55:03.0000] <jgraham> Also, working backwards through time, lol windows [01:55:20.0000] <benvie> authoritative in this world is easy [01:55:25.0000] <Ms2ger> jgraham, yeah :/ [01:55:25.0000] <jgraham> MikeSmith: Yeah that seems much more sensible to me [01:55:27.0000] <benvie> does it exist in an implemtnation or not [01:55:41.0000] <benvie> it either can be used or can't be [01:56:01.0000] <Ms2ger> jgraham, in a meta element! ;) [01:56:36.0000] <jgraham> Ms2ger: You unlocked the trolling #whatwg achievement a long time agao. No need to do it again :p [01:57:03.0000] <Ms2ger> Why thank you, dear [01:57:33.0000] <benvie> the web api sphere is a finely balanced dance of anarchy. It's like a house of cards that is reinforced by its own frailty [02:10:16.0000] <MikeSmith> benvie: a dancing spherical house of cards? [02:12:17.0000] <Ms2ger> Yay, unprefixed border-image coming to Gecko [02:12:25.0000] <Ms2ger> /me bets Opera did it first [02:12:48.0000] <smaug____> :p [02:13:07.0000] <odinho> Never gets old, does it :P [02:13:50.0000] <Ms2ger> No :) [02:15:11.0000] <odinho> Webbrowser getting bought by a website, Opera did it first (?) : [02:18:13.0000] <smaug____> well, ok, that is something NS managed to do first [02:18:35.0000] <smaug____> /me hopes the rumors about FB buying Opera aren't true [02:19:54.0000] <odinho> smaug____: Ah yes. Then it's no danger, because it's already done. So no reason to do it. *phew [02:22:34.0000] <wodemaye__> /me hopes they are true! :) [02:24:22.0000] <smaug____> it would be very sad to lose Opera to an evil empire. [02:24:38.0000] <odinho> It would be very sad to suddenly work for an evil empire. [02:24:49.0000] <odinho> I would probably even have to get a facebook account. [02:26:47.0000] <AryehGregor> zcorpan, yes, the stack trace can be annoyingly large, I agree. Unfortunately, since it's not in a standard format, we can't really do anything to make it better except not print it. Which is probably a good idea for most tests, but not for my crazy complicated tests. :) [02:28:08.0000] <zcorpan> AryehGregor: we should make it a standard already :-P [02:28:36.0000] <AryehGregor> annevk, in the United States, I noticed that after I kept an HTTP connection open for more than a second or so, bandwidth dropped drastically. I suspect my ISP was prioritizing short-lived TCP connections to benefit typical HTTP over long downloads and such. [02:29:36.0000] <AryehGregor> (which is a good form of network non-neutrality, IMO, although it might do the wrong thing for videoconferencing) [02:30:06.0000] <annevk> this was ssh [02:30:09.0000] <AryehGregor> (it shouldn't hurt VoIP as long as they only throttle the bandwidth of long-lived connections) [02:30:34.0000] <AryehGregor> Right, but they'd want to hit BitTorrent and FTP too, so if I were them I'd just do it at the TCP level and ignore the protocol. [02:30:45.0000] <AryehGregor> That also means you don't have to actually inspect the traffic. [02:30:50.0000] <AryehGregor> Which is good because often you can't. [02:48:56.0000] <jgraham> AryehGregor: FWIW we don't actually have any local branches yet [02:49:39.0000] <jgraham> But it is a use case we have had, and addressed badly for non-W3C testsuites so it doesn't seem unlikely that we will have to address it for W3C ones [02:50:14.0000] <jgraham> (e.g. if there is some test that hardcodes domains due to testing document.domain, we would want to patch it to use different domains) [02:51:12.0000] <jgraham> Life would be much easier if Hixie didn't sleep. Does anyone understand navigation well enough to explain https://www.w3.org/Bugs/Public/show_bug.cgi?id=17245 to me? [02:54:47.0000] <annevk> didn't we have a copy of the Design Principles on the WHATWG Wiki at some point? [02:54:47.0000] <AryehGregor> jgraham, why not just fix it in the repo? You have commit access. [02:54:53.0000] <annevk> or was it a W3C Wiki page? [02:56:10.0000] <odinho> AryehGregor: Hmmm. Like it is now, it's not looked nicely upon to go into other vendors folders and just rewrite them :P [02:56:27.0000] <annevk> http://www.w3.org/html/wg/wiki/ProposedDesignPrinciples [02:56:32.0000] <AryehGregor> odinho, I think it's considered fine once they're formally submitted, at least if you talk to the submitter first, right? [02:56:36.0000] <AryehGregor> It should be, anyway . . . [02:56:44.0000] <annevk> seems to be have become quite the mess [02:57:40.0000] <jgraham> AryehGregor: Fix what? [02:57:51.0000] <AryehGregor> jgraham, if the test isn't usable from non-W3C-land. [02:57:54.0000] <odinho> AryehGregor: yea, -- but I don't feel like I should mess around in e.g. the submitted/Microsoft/ folder, I don't belong in there. [02:58:11.0000] <jgraham> It is quite possible that for some tests different servers will be needed on the public internet and on our internal network [02:58:41.0000] <AryehGregor> odinho, I think CSSWG policy, at least, is that you're allowed to. Anyway, you certainly can give it as feedback before it gets approved. [02:58:50.0000] <Ms2ger> odinho, I'm happy to mess around in the Opera folders ;) [02:58:51.0000] <AryehGregor> jgraham, sure, but there should be some way to do that without patching the tests. [02:59:18.0000] <MikeSmith> annevk: it seems like there was something on the WHATWG wiki [02:59:35.0000] <jgraham> AryehGregor: Perhaps, I guess it depends on the details of the tests [02:59:39.0000] <MikeSmith> um, what does "frictionless" mean in terms of Web development? [02:59:43.0000] <odinho> AryehGregor: Yea, I already have a support.js file where you can set some variables. Those should be possible (and most maybe want) to override at least. [03:00:09.0000] <odinho> Ms2ger: Yea, I'm okay with people messing around in the Opera folder. We're just that kind of company ^^ [03:01:07.0000] <annevk> MikeSmith: I cannot find it [03:01:32.0000] <jgraham> In any case I wouldn't like to design in the assumption that we will never have to make local patches [03:03:39.0000] <MikeSmith> annevk: yeah, I can't either [04:12:05.0000] <jgraham> There's nothing I like more than spec threads where someone says "we should change the spec to say X" and then others say "yeah we changed our implementation to do that already (but didn't tell anyone)" [04:12:36.0000] <odinho> Seems to have been a few lately. [05:00:42.0000] <MikeSmith> how is the word skägg pronounced? [05:02:38.0000] <jgraham> If you are a human? By configuring your vocal chords/tounge/lips in the correct way and modulating the flow of air through them as needed. If you are Ms2ger? By sending electrical signals that cause vibrations in the cone of a speaker. [05:03:21.0000] <jgraham> I imagine [05:03:33.0000] <jgraham> Not that I have ever examined his hardware of course [05:06:39.0000] <zcorpan> MikeSmith: http://sv.forvo.com/word/sk%C3%A4gg/#sv [05:08:52.0000] <MikeSmith> tack [05:10:04.0000] <Ms2ger> *takk [05:10:23.0000] <annevk> zcorpan: we should nail down that list then [05:12:42.0000] <MikeSmith> annevk: plh is going to shut down our working group if we don't get remaining issues resolved and publish a LCWD [05:12:58.0000] <MikeSmith> charter expires in 1 month [05:15:48.0000] <annevk> MikeSmith: for a guy that claims to care about patent commitments he seems to care little [05:16:00.0000] <annevk> MikeSmith: did he miss that Apple joined the Notifications WG? [05:16:20.0000] <MikeSmith> no, he knows that [05:16:50.0000] <MikeSmith> he would just kind of like to see some actual further progress made on the spec I guess [05:17:21.0000] <annevk> I did actually ping John again earlier today [05:17:34.0000] <annevk> apparently he's on the YouTube team so doesn't have much time [05:17:46.0000] <annevk> but I'll see what he says [05:18:31.0000] <annevk> I happen to not care much for WGs so threatening to close mine down seems counter-productive [05:18:39.0000] <MikeSmith> k [05:18:44.0000] <annevk> but I do care somewhat about Notifications [05:19:28.0000] <Ms2ger> Shut down the WG and move to the WHATCG? [05:19:52.0000] <MikeSmith> annevk: I wonder whether other John might be willing to work on the spec [05:19:57.0000] <MikeSmith> John Lee [05:19:58.0000] <MikeSmith> Apple [05:20:10.0000] <MikeSmith> he seems a bit more motivated right now [05:21:11.0000] <MikeSmith> annevk: hey there's a new CSS group for you to join [05:21:12.0000] <MikeSmith> http://www.w3.org/community/blog/2012/05/31/proposed-group-css-specifications-community-group/ [05:21:21.0000] <MikeSmith> no idea who floated that one [05:21:34.0000] <MikeSmith> "This group addresses and discusses proposed ideas for CSS specifications." [05:21:44.0000] <MikeSmith> both addresses and discusses [05:21:46.0000] <MikeSmith> in that order [05:22:25.0000] <annevk> MikeSmith: we could ask him if John doesn't do anything [05:22:33.0000] <annevk> MikeSmith: or I could do it; I have some time now [05:22:57.0000] <MikeSmith> whatever works for you [05:23:36.0000] <annevk> I have doubts John Lee has the time [05:23:53.0000] <MikeSmith> OK [05:25:06.0000] <zcorpan> hmm, seems nobody supports MutationNameEvent? [05:25:30.0000] <annevk> s/hmm, seems/yay/ [05:25:56.0000] <Ms2ger> Seems correct for Gecko [05:26:26.0000] <annevk> also Opera/WebKit checking window.MutationNameEvent [05:27:01.0000] <gsnedders> wilhelm: Given you keep on complaining about it, you might be interested in what the latest snapshot has a fix for. ;) [05:27:51.0000] <annevk> IE also doesn't have it [05:28:02.0000] <Ms2ger> IE10? [05:28:16.0000] <annevk> ooh [05:28:26.0000] <annevk> MutationNameEvent was a DOM Level 3 Events addition? [05:28:37.0000] <annevk> it's not in http://www.w3.org/TR/DOM-Level-2-Events/events.html anyway [05:29:17.0000] <Ms2ger> For document.renameNode(), apparently [05:29:42.0000] <zcorpan> filed https://www.w3.org/Bugs/Public/show_bug.cgi?id=17270 [05:29:49.0000] <annevk> fast zcorpan is fast [05:31:14.0000] <wilhelm> gsnedders: I have already been informed. Thanks, I'll try it out. (c: [05:32:38.0000] <annevk> I guess it would prudent to check Level 3 against Level 2 to see what else was added that's not actually useful [05:33:12.0000] <kennyluck> Can we publicly criticize Google on the WHATWG blog for not making enough effort in the standardization work? [05:33:25.0000] <MikeSmith> eh? [05:33:29.0000] <Ms2ger> And Apple? [05:33:33.0000] <Ms2ger> And Mozilla? [05:33:38.0000] <Ms2ger> And Microsoft? [05:33:39.0000] <zcorpan> and w3c? [05:33:42.0000] <MikeSmith> heh [05:33:44.0000] <Ms2ger> Not Opera, I guess [05:33:58.0000] <zcorpan> nope, we did it first [05:34:23.0000] <gsnedders> We shipped a Native HTML6 implementation last year. [05:34:26.0000] <Ms2ger> You were first not to make enough effort? [05:34:40.0000] <Ms2ger> Glad you got over that, then ;) [05:35:05.0000] <annevk> kennyluck: if that's the summary of the post I'm not sure how that's a useful post for the WHATWG blog [05:35:45.0000] <kennyluck> *shrug* [05:36:05.0000] <gsnedders> kennyluck: What's the justification for them not putting in enough effort? [05:36:09.0000] <MikeSmith> kennyluck: shop that idea out to due at .Net magazine. I bet he'd love it [05:37:06.0000] <kennyluck> gsnedders, why ask me? It seems pretty clear that we have very little or no expectation for Apple, but Google… hmm... [05:37:14.0000] <annevk> so why do FocusEvent and such still have init*Event()? [05:38:11.0000] <zcorpan> why does DOMActivate still exist in the spec? [05:38:22.0000] <zcorpan> why does DOM3 Events still suck? [05:38:59.0000] <zcorpan> so many questions :-) [05:39:09.0000] <kennyluck> why does CSS2.1 read like an advanced tutorial instead of a spec? [05:39:36.0000] <annevk> because it's CSS 2.0 patched instead of the rewrite it deserved [05:40:04.0000] <Ms2ger> I guess that applies equally to D3E [05:40:32.0000] <annevk> Media Queries too :( [05:42:06.0000] <Ms2ger> ALL THE SPECS [05:43:42.0000] <Ms2ger> Who's going to try and get CSS to define what "critical subresources" are? [05:44:14.0000] <annevk> /me frowns [05:46:16.0000] <MikeSmith> if berjon is working on updating the WebIDL stuff in respec, would be good to update it to use the same biblio stuff as anolis [06:57:46.0000] <annevk> I hope he succeeds [07:02:05.0000] <codacoder> anyone here? [07:06:35.0000] <jgraham> No [07:06:46.0000] <codacoder> I can tell ;) [07:07:06.0000] <codacoder> it's that vacant look [07:08:15.0000] <jgraham> We're so pretty [07:08:15.0000] <codacoder> Was hoping someone could explain something about the specs - specifically, about DEPRECATED/OBSOLETE [07:08:26.0000] <codacoder> lol [07:08:48.0000] <codacoder> a Brit I see. Prob as old as me too [07:08:52.0000] <Philip`> Which specs? [07:09:01.0000] <codacoder> html5 [07:09:18.0000] <Philip`> HTML5 doesn't do deprecated - things are either allowed, or not allowed [07:09:32.0000] <jgraham> Some are "obsolete but conforming" [07:09:33.0000] <Philip`> (which is totally independent of whether browsers must implement support for those things or not) [07:09:54.0000] <jgraham> Which sounds a lot like "deprecated with delicious figs" [07:10:02.0000] <codacoder> well... yes... but the spec I'm reading seems to be sitting painfully right on the fence re obsolete [07:10:18.0000] <Philip`> I think "deprecated" is more like "conforming but obsolete" [07:10:36.0000] <codacoder> Phil: yep [07:11:07.0000] <codacoder> I woundered why, then, there's a whole section on rendering obsolete crap [07:11:37.0000] <codacoder> so a definition of obsolete would be good [07:12:05.0000] <Philip`> Most people write non-conforming content (never mind obsolete-but-conforming content), so browsers are required to render that stuff correctly anyway, because that's necessary for real-world interoperability [07:12:20.0000] <codacoder> example: FRAMESET [07:12:35.0000] <codacoder> I agree - but the spec is not clear about what it means [07:13:05.0000] <codacoder> as a reader, it's not clear and can cause endless "going in circles" trying to tie it down [07:13:16.0000] <Philip`> "obsolete" means nothing more than "conformance checkers will warn about this" and "the spec writers would prefer you not to do this (but we know you're going to anyway)" [07:13:33.0000] <codacoder> right [07:13:43.0000] <Philip`> (Rather, "obsolete but conforming" means that) [07:14:16.0000] <Philip`> frameset is entirely non-conforming, which means "conformance checkers will give an error about this" and "the spec writers would prefer you not to do this (but we know you're going to anyway)" [07:14:25.0000] <codacoder> I guess it's an unfortunate flow I'm "carrying" that, deprecated -> obsolete -> gone [07:15:00.0000] <Philip`> The "obsolete -> gone" step never occurs in practice [07:15:08.0000] <codacoder> agreed [07:15:20.0000] <Philip`> (Browsers still support things that were deprecated in the first ever published HTML specs) [07:15:20.0000] <codacoder> but the spec does not make that clear, is my point [07:15:26.0000] <codacoder> I know [07:16:26.0000] <zcorpan> codacoder: the spec usually says something like "authors must not do X" and then "user agents must do Y when authors do X" [07:16:35.0000] <codacoder> context: I'm writing a test suite, I wanna make sure my wording is as correct as poss [07:16:55.0000] <zcorpan> test suite aimed at browsers? [07:17:15.0000] <Philip`> Would it be clearer if somewhere like http://www.whatwg.org/specs/web-apps/current-work/multipage/obsolete.html#obsolete repeated something like the text from http://www.whatwg.org/specs/web-apps/current-work/multipage/introduction.html#conformance-requirements-for-authors ("this specification defines in some detail the required processing for invalid documents as well as valid documents.")? [07:17:18.0000] <codacoder> and to "defend" my statements i need a "source" <- the spec aint helping (much) [07:17:45.0000] <codacoder> reading... [07:17:48.0000] <zcorpan> what are you testing? [07:17:59.0000] <codacoder> test-suite aimed at browser based app [07:20:21.0000] <zcorpan> if your test is testing the user agent (the browser), then you should ignore all requirements on authors/documents [07:20:46.0000] <codacoder> no, it's testing the content [07:22:30.0000] <codacoder> Philip: it needs to be clear what obsolete means in practice [07:23:58.0000] <codacoder> I think (IMO) it's less than clear and is missing the (old) "MUST" etc definitions. What I've read is wishy-washy (sorry) [07:26:35.0000] <Philip`> I think in pretty much every case there is a relevant series of "must"s somewhere in the spec, but it tries to only require things once and then uses wishy-washy language elsewhere to describe consequences of those requirements [07:26:47.0000] <kennyluck> Yeah, I guess I don't get the "obsolete but conforming" idea either. So here's a concrete question: If I were to write a book about HTML, should I include elements/syntax that are "obsolete but conforming"? [07:26:56.0000] <Philip`> and it's hard to find the right "must" if you don't know where to look [07:27:18.0000] <codacoder> kenny: Zackly! [07:27:35.0000] <codacoder> any of you here part of the authorship team? [07:27:44.0000] <jgraham> kennyluck: No [07:27:47.0000] <Philip`> kennyluck: If you're writing a reference book for people who have to maintain old pages, you should include everything that's ever used in practice, including non-conforming features [07:28:11.0000] <jgraham> (unless it is some very special case like Philip` says) [07:28:25.0000] <kennyluck> Philip`, yes, but what I am talking about new books. I don't exactly get what the middle class is for. [07:28:52.0000] <codacoder> Imagine - test fails. Reason: blah. Why -> link [07:28:55.0000] <jgraham> If you were writing a handbook for implementors with a catchy title like "HTML: Living Standard" then yes [07:28:55.0000] <Philip`> kennyluck: If you're writing a tutorial-style book for new content authors, you should just describe the features that you think they ought to use, which should be a subset of the non-obsolete conforming features [07:29:05.0000] <codacoder> It's almost impossible using the spec as it is [07:29:39.0000] <Philip`> (There's probably a lot of conforming features that aren't worth taking the effort to describe, so it depends on where you choose to focus as the author of the book) [07:30:08.0000] <codacoder> or in my case as the author of tests [07:30:38.0000] <webben> Philip`: "The "obsolete -> gone" step never occurs in practice". I wonder if that's really true. Didn't that happen with <layer>? [07:30:45.0000] <kennyluck> That I don't quite agree. I thought the whole point of making a subset of what browsers implement conforming is for educational purposes and yet you are telling me to build my own thing. [07:31:37.0000] <Philip`> webben: As far as I'm aware, that was never supported outside of Netscape [07:31:57.0000] <Philip`> (and never in a spec) [07:31:58.0000] <codacoder> layer was netscape - not a spec [07:32:20.0000] <webben> Philip`: True AFAIK, though a single browser supporting something is not necessarily a blocker to it becoming widespread in the corpus. [07:32:27.0000] <webben> Maybe <layer> never did. [07:33:19.0000] <codacoder> sticking to my point tho - this spec is too loose in the context i mentioned. It's not aiding me aid "them" [07:33:29.0000] <Philip`> http://philip.html5.org/data/tag-count-pages.txt says <layer> is on about 1% of pages, but I think it's almost always perfectly acceptable to treat it like <span>, which is what most browsers do now [07:34:00.0000] <webben> codacoder: I find it easiest to think of these things as: browsers need to support whatever features are needed to give users access to the corpus. authors should be encouraged to use whatever features give the best results for users [07:34:06.0000] <Philip`> codacoder: Could you give a more concrete example of the kind of information you can't find in the spec? [07:35:08.0000] <codacoder> said it earlier - may have scrolled past... imagine: test fails. Reason: blah. Why -> link [07:35:12.0000] <webben> codacoder: Yeah ... like what test are you unsure how to write? [07:35:30.0000] <codacoder> webben: I am not unsure [07:35:38.0000] <codacoder> I know exactly what I want to write [07:35:44.0000] <Philip`> kennyluck: I think there's a fundamental problem with different people having different ideas on what's a good subset for educational purposes; the spec just gives one idea (its notion of conformance) and a compromise with a slightly different idea (its notion of obsolete but conforming), under the belief that it's better to say something than nothing [07:36:00.0000] <webben> codacoder: Are you writing a document conformance checker of some sort? [07:36:31.0000] <codacoder> no - a test suite [07:36:32.0000] <Philip`> but that doesn't preclude anyone else from coming up with their own definition of 'good' HTML [07:36:37.0000] <codacoder> runs against an app [07:36:47.0000] <Philip`> codacoder: That didn't sound very concrete to me :-) [07:36:58.0000] <webben> codacoder: But testing for HTML conformance violations? [07:37:20.0000] <codacoder> Philip: test is (eg) FRAMESET or CENTER [07:37:59.0000] <webben> codacoder: You're looking for some text to link to that gives rationale for why FRAMESET (say) is not conforming HTML? [07:38:06.0000] <codacoder> more detail... $("center").length === 0 [07:38:22.0000] <codacoder> webben: yep [07:38:45.0000] <codacoder> it's too linky-linky-linky [07:38:53.0000] <webben> codacoder: It would be nice if the spec had rationale for all its design decisions, but I'm not sure that's realistics. [07:38:59.0000] <codacoder> like i said, it's not helping me help them [07:39:25.0000] <codacoder> and the terminology is way too wishywashy [07:39:30.0000] <webben> codacoder: For one thing, to a large extent the spec is an artefact of compromises. [07:39:50.0000] <codacoder> webben: you got it nailed there [07:39:51.0000] <webben> codacoder: So it's not necessarily reducible to fundamental design principles. [07:40:33.0000] <webben> codacoder: Worse still, this is true of pretty much all specifications where multiple parties need to agree, so the spec isn't really exceptional here. [07:40:55.0000] <webben> codacoder: Also, the rationale could easily be longer than the spec itself... [07:41:15.0000] <webben> codacoder: There are occasional efforts to write up some rationale outside the spec tho. [07:41:22.0000] <codacoder> I sort of agree. But in this case (dare I say it) the w3c older specs were better [07:42:02.0000] <webben> codacoder: http://wiki.whatwg.org/wiki/FAQ#Is_design_rationale_documented.3F [07:42:17.0000] <codacoder> in this regard, HTML5's definition of obsolete has no teeth and becomes pretty meaningless [07:42:24.0000] <codacoder> thanks [07:42:40.0000] <webben> codacoder: The older specs omitted a lot of rationale. [07:42:46.0000] <webben> They also omitted a lot of spec ;) [07:43:05.0000] <webben> codacoder: HTML5's definition of obsolete seems pretty straightforward. [07:43:07.0000] <kennyluck> Philip`, that's reasonable. But just like some people write specs because browser implementers don't bother reverse engineering each other. The idea of a conforming class for authors should be similar: not everyone would bother reading all rationale and going into the debates of why something is conforming and not and come up with his/her own class. The existence of the class in the middle means that the spec doesn't address this use [07:43:12.0000] <codacoder> understand I'm not here to "knock the specs" - here to improve them if I can [07:43:14.0000] <webben> Not sure what teeth you expect it to have. [07:43:19.0000] <webben> codacoder: Yep, understood. [07:44:08.0000] <codacoder> obsolete is absolutely (my baggage acknowledged) meaningless [07:44:34.0000] <webben> codacoder: It just means discouraged. That's it. [07:44:40.0000] <codacoder> and if I managed to link to the right spot in the spec, it would go against me, not for me [07:44:59.0000] <codacoder> but that was what deprecated means (meant) [07:45:29.0000] <webben> codacoder: No. [07:45:44.0000] <webben> codacoder: Deprecated implied future UAs could drop support. [07:46:45.0000] <codacoder> haha. see? it's a freakin nightmare! [07:47:12.0000] <codacoder> when i said "my baggage" this is what I meant! [07:47:40.0000] <webben> codacoder: I don't think these terms have ever been used or understood consistently. But you can just use the definition in the spec at hand. [07:47:43.0000] <codacoder> If i had a spec that was definitive (what else is a spec for at base values?) It would remove ALL baggage! [07:47:53.0000] <webben> codacoder: It is definitive. [07:48:05.0000] <codacoder> nope - it's "loose" [07:48:56.0000] <codacoder> I guess I should code as though I was writing a validator and take their suggestions for that [07:49:17.0000] <webben> codacoder: I don't get how http://www.whatwg.org/specs/web-apps/current-work/multipage/obsolete.html is loose. [07:49:45.0000] <webben> codacoder: It says authors must not use X, conformance checkers must warn about X, user agents must do Y with X. [07:50:20.0000] <webben> codacoder: In your test justification text, you could entirely ignore the language about "obsolete" if you think it's confusing. [07:50:28.0000] <codacoder> which is a toothless definition of obsolete [07:50:30.0000] <kennyluck> It says author *should not* use X. [07:50:40.0000] <codacoder> right [07:50:45.0000] <kennyluck> It really should be "must not" if that's the intention. [07:51:01.0000] <webben> kennyluck: Only for obsolete but conforming. [07:51:21.0000] <codacoder> I guess we're getting to it now... what in the hell does obsolete mean? I think they need a different term [07:51:47.0000] <kennyluck> I think it should probably be called "non-conforming but only triggers warning in conformance checker class" [07:52:10.0000] <webben> codacoder: Use the dictionary definition: http://www.merriam-webster.com/dictionary/obsolete [07:52:37.0000] <codacoder> webben: it's not about my ignoring something... it's about where I send a failing programmer to read something - what he reads should be clear, meaningful and "absolute" [07:52:51.0000] <codacoder> kenny: yep [07:52:55.0000] <webben> it is absolute [07:53:18.0000] <webben> /me shrugs [07:53:44.0000] <codacoder> webben: within its own confined context <- I say its chosen a bad one [07:53:51.0000] <webben> I'd prefer "discouraged" to "obsolete but conforming" and "forbidden" to "obsolete". [07:54:10.0000] <codacoder> RIGHT - much better [07:54:35.0000] <webben> In general, changing what goes into these categories let alone what they are called is a political minefield. [07:55:01.0000] <codacoder> oh yes. [07:55:19.0000] <codacoder> however, it should be better than it is [07:55:47.0000] <webben> Compromise often does result in things that are worse than they should be, but are better than not having them at all. [07:56:29.0000] <codacoder> agreed [07:56:36.0000] <webben> I often feel we'd have been better off just not bothering with author conformance requirements and leaving the whole business to linters. [07:57:04.0000] <webben> If nothing else it would have saved a lot of fairly pointless arguments. [07:57:28.0000] <codacoder> you got it [07:57:50.0000] <kennyluck> webben, that's true. [07:58:02.0000] <codacoder> in many ways, this entire conv is "evidence" [07:58:16.0000] <webben> I've been involved in a lot of those arguments, trying to make the conformance requirements make sense (from my perspective). [07:58:22.0000] <webben> It's very difficult to get people to agree. [07:58:40.0000] <webben> And I don't much like the set of conformance requirements we've got at the moment. [07:58:42.0000] <jgraham> No it's not! [07:58:47.0000] <webben> :) [07:58:52.0000] <codacoder> lol [08:00:01.0000] <kennyluck> But still, the problem here is quite obvious. "conforming" and "should not" are opposite words. [08:00:15.0000] <webben> kennyluck: They're not. [08:00:25.0000] <webben> kennyluck: That's standard IETF stuff. [08:00:38.0000] <webben> kennyluck: SHOULD NOT = don't do unless you've got good reason [08:00:52.0000] <webben> kennyluck: MUST NOT = non-conforming [08:01:25.0000] <kennyluck> webben, hmm.. ok I guess it makes some sense now. [08:03:15.0000] <codacoder> LEt's say I link to 15.2 (http://www.whatwg.org/specs/web-apps/current-work/multipage/obsolete.html) because someone has used (say) FRAMESET [08:03:35.0000] <codacoder> they read, then click on Frameset link... [08:04:17.0000] <codacoder> which takes them to a section talking about how frameset is meant to work (which is hardly my personal definition of obsolete) [08:04:19.0000] <kennyluck> webben, the IETF RFC for these keywords doesn't define what "conforming" means so I can't tell if that's true though. http://tools.ietf.org/html/rfc2119 [08:05:14.0000] <kennyluck> I think the normal practice is, though, when writing a test suite, is to include those "should"s and "should not"s. [08:05:50.0000] <kennyluck> In that sense, a violation of "should" seems to indicate that it is non-comforming… but I am not sure... [08:05:57.0000] <codacoder> kenny: agreed. I wanted to use the spec as definitve "why" - complaint is, it's not [08:08:31.0000] <codacoder> will you guys all sign wavers so I can store this entire thread in my tests? LOL [08:09:58.0000] <codacoder> anyway, I'm out of here. Thanks. It was informative and "confirming" (if not conforming) ;) [08:10:15.0000] <kennyluck> I have no problem and the channel log is public (see title) anyway. [08:10:15.0000] <Philip`> codacoder: "$("center").length === 0" - that makes it sound like you're implementing a conformance checker [08:10:25.0000] <Philip`> which would be an easier way to describe the problem :-) [08:11:37.0000] <codacoder> ok - just for Philip: that's only a tiny part of it, but yes, in this context, it pretty much is [08:12:02.0000] <codacoder> the test engine tests app functionality too [08:13:37.0000] <codacoder> ok going this time. Thanks again all. [08:13:38.0000] <annevk> smaug____: http://www.w3.org/TR/progress-events/#interface-progressevent [08:13:40.0000] <annevk> smaug____: btw [08:13:52.0000] <odinho> Seems I was under the false impression that Firefox nightly didn't allow sync XHR with CORS. But testing it, it does in fact do that. [08:13:59.0000] <annevk> smaug____: did Gecko disable cross-origin XMLHttpRequest? [08:14:01.0000] <annevk> oh [08:14:02.0000] <odinho> So time seems to be running out for that. [08:14:19.0000] <odinho> I don't see how they can wait so long with it if Moz really wants to do it. [08:15:39.0000] <odinho> annevk: A mozilla girl said it at the F2F. :-) [08:17:22.0000] <smaug____> annevk: disable CORS? [08:17:27.0000] <smaug____> oh, sync [08:17:36.0000] <smaug____> I don't think so [08:17:42.0000] <smaug____> sicking did suggest that [08:18:05.0000] <smaug____> annevk: yes, ProgressEvents spec has that, but implementations do have, IIRC, init*Event [08:18:20.0000] <smaug____> and createEvent("progressevent"); is supported, again, IIRC [08:18:29.0000] <smaug____> also some event related to storage handling... [08:19:01.0000] <smaug____> annevk: https://bugzilla.mozilla.org/show_bug.cgi?id=736058 [08:19:21.0000] <annevk> and bugs cannot be fixed? [08:19:45.0000] <annevk> aah [08:19:49.0000] <annevk> what nonsense is that [08:19:57.0000] <annevk> for a test you add that? [08:20:28.0000] <odinho> smaug____: So then I *don't* have to rewrite the CORS-tests to be async? Well, I'll make them more readable again then :P [08:21:07.0000] <smaug____> well, we haven't disabled CORS yet [08:21:14.0000] <smaug____> odinho: would be better to ask sicking [08:21:42.0000] <smaug____> I think disabling CORS could be quite risky [08:22:34.0000] <odinho> Yes, seems to be, esp. after CORS with sync XHR has worked for so long on the wild web. [08:22:59.0000] <odinho> But you have some release trains to test with. [08:25:02.0000] <smaug____> unfortunately surprisingly many sites are tested only with release builds [08:25:34.0000] <smaug____> but [08:25:36.0000] <smaug____> hmm [08:25:39.0000] <smaug____> perhaps a warning first [08:25:50.0000] <smaug____> "CORS with sync XHR will be disabled" [08:26:15.0000] <smaug____> warning are effective in some cases [08:26:18.0000] <smaug____> +s [08:30:16.0000] <odinho> smaug____: I meant for sitecompat, - I guess you'll have people reporting broken sites etc. [08:30:40.0000] <smaug____> sure [08:31:30.0000] <smaug____> but certain kinds of sites are used only with release builds [08:31:34.0000] <smaug____> like intranets [08:31:41.0000] <smaug____> some intranets [08:31:44.0000] <odinho> So if they're like "omg web iz broken!1" it might be hard, but if it's all like ... crickets ... it might be more okay. [08:31:51.0000] <odinho> smaug____: Yeah, I can see that. [08:58:11.0000] <dglazkov> good morning, Whatwg! [09:41:07.0000] <TabAtkins> Ms2ger: What spec talks about "critical subresources"? [09:41:15.0000] <Hixie> html [09:41:28.0000] <Ms2ger> CSS doesn't, that's the issue :) [09:41:38.0000] <TabAtkins> Ms2ger: Okay, then I don't get the context. [09:41:52.0000] <Ms2ger> https://www.w3.org/Bugs/Public/show_bug.cgi?id=17011 [09:42:33.0000] <TabAtkins> Ah, kk. Was wondering if you were just kvetching about a CSS problem without bringing it up with us again. ^_^ [09:43:33.0000] <TabAtkins> Hixie: If you need any help with Proxies for whatever reason, I can help. They're not hard. [09:44:06.0000] <Hixie> i want an object that just forwards all unknown method calls to another [09:44:12.0000] <TabAtkins> "critical subresources" sounds like it's potentially a useful term for CSS to define anyway. [09:44:15.0000] <Hixie> it's like 2 lines of code in perl [09:44:31.0000] <TabAtkins> Hixie: Slightly more than two lines in JS, but not much. [09:45:11.0000] <Hixie> it looked to be a lot more than two when i was looking at the proxy api [09:45:24.0000] <Hixie> and some long lines of unintuitive api calls at that [09:45:54.0000] <TabAtkins> You were looking at "proxy" instead of "direct proxy", weren't you. [09:46:08.0000] <TabAtkins> (It's the first google hit, I know.) [09:46:12.0000] <Hixie> no idea what the difference is [09:46:27.0000] <Hixie> i was looking at brendan' slide show, amongstother things [09:46:40.0000] <TabAtkins> The wiki page says, right at the top, that the "proxy" proposal is obsolete and superceded by direct proxies. [09:47:04.0000] <TabAtkins> iirc, all you need is: [09:47:59.0000] <Hixie> oh. if the stuff i was reading is dead then that's awesome. [09:48:04.0000] <Hixie> it was all claiming to be accepted and stuff [09:48:19.0000] <Hixie> is there a spec somewhere? [09:48:27.0000] <Hixie> that represents the latest thinking on js? [09:49:27.0000] <Hixie> [javascript direct proxy] doesn't help me [09:49:46.0000] <TabAtkins> var wrapped = Proxy(obj, {get: function(target, name, receiver) { if( methodIWantToCatch(name) ) { return doStuff(name); } else return target[name]; }); [09:50:00.0000] <TabAtkins> http://wiki.ecmascript.org/doku.php?id=harmony:direct_proxies [09:50:48.0000] <Hixie> yeah, that looks like what i was looking at [09:51:05.0000] <TabAtkins> Okay, that's not exactly hard. [09:51:15.0000] <Hixie> surely what you describe would result in the |this| pointing to the wrong object [09:51:22.0000] <Hixie> i want it pointing at the thing i'm proxying, not the proxy [09:51:46.0000] <TabAtkins> No, it works. [09:52:02.0000] <Hixie> how? [09:52:05.0000] <TabAtkins> The "return target[name]" part soft-binds this appropriately. [09:52:10.0000] <TabAtkins> For the unknown properties. [09:52:58.0000] <TabAtkins> var wrapped = Proxy(obj, {get: function(obj, name) { if( methodIWantToCatch(name) ) { return doStuff(name); } else return obj[name]; }); [09:53:01.0000] <TabAtkins> Easier to see now? [09:53:28.0000] <Hixie> oh sorry i think i misexplained [09:53:35.0000] <Hixie> let me be more clear [09:54:50.0000] <Hixie> i'm creating an object A at time t0. When A is created, it does a network request, and at time t1>t0, i'll use that data to create an object B. I want all calls to A to get proxied to B once A has created B. [09:55:10.0000] <TabAtkins> And before A has created B, what happens? [09:55:11.0000] <Hixie> does that make sense? [09:55:43.0000] <Hixie> before A has been created, some other code I write will do stuff, in particular, it will queue the requests to be run once I have B (all these calls are async) [09:56:05.0000] <Hixie> A will also have other methods [09:56:11.0000] <Hixie> e.g. to see how the network is doing [09:56:40.0000] <TabAtkins> Okay, the line I have above will work, with the obvious modifications inside the get trap. [09:56:42.0000] <Hixie> (an alternative would be for A to replace itself with B somehow... maybe i can do that with some prototype magic) [09:56:54.0000] <Hixie> your use of the word "obvious" is foreign to me :-P [09:56:54.0000] <TabAtkins> Check if the property being got is one that A "natively" has. If so, do it. [09:57:09.0000] <TabAtkins> Otherwise, check if B has been craeted. If not, queue it up. If so, forward it. [09:57:25.0000] <Hixie> what is "obj" in the first argument to Proxy? [09:57:25.0000] <TabAtkins> You can't swap things out. A has to be a Proxy the whole time. [09:57:37.0000] <TabAtkins> The object that you're wrapping with the proxy. [09:57:44.0000] <Hixie> A? [09:58:35.0000] <TabAtkins> No, it's B. [09:58:39.0000] <TabAtkins> A is the proxy. [09:58:40.0000] <Hixie> B doesn't exist yet [09:58:45.0000] <TabAtkins> Right. [09:58:47.0000] <TabAtkins> So. [09:59:11.0000] <TabAtkins> Hm. [09:59:40.0000] <Hixie> (what if i'm proxying to different objects depending on what the method is? or one of the arguments?) [10:00:10.0000] <Hixie> i just want to trap "method was called but not defined", i don't really want the JS system to know i'm proxying [10:00:40.0000] <TabAtkins> The JS *doesn't* know that you're proxying. That's the point of proxies - they're undetectable except by the isProxy method. [10:00:52.0000] <TabAtkins> So actually, I got it. You need some indirection because of your swap-out. [10:01:18.0000] <Hixie> i mean the JS compiler [10:01:31.0000] <TabAtkins> Then you have to invoke magic. [10:01:42.0000] <smaug____> ah, innerHTML serializes html:script and svg:script in a different way. [10:01:46.0000] <TabAtkins> Magic which is perfectly possible to do with proxies. [10:01:56.0000] <TabAtkins> Why do you care aout the J?S compiler? [10:02:14.0000] <Hixie> TabAtkins: oh i completely believe this is possible, my thesis is just that it's way more complicated than perl. [10:02:25.0000] <smaug____> hsivonen: ping [10:02:29.0000] <TabAtkins> I... somehow doubt that your actual use-case is two lines in perl. [10:02:31.0000] <Hixie> TabAtkins: i don't care about the compiler, i mean, if it needs to know it needs to know, i just don't know why i need to tell it [10:02:38.0000] <TabAtkins> A simple "forward everything I dont' have to thsi other object", sure. [10:03:00.0000] <TabAtkins> But you actually want something more complicated. [10:03:27.0000] <Hixie> use AutoLoader; sub AUTOLOAD { ...do whatever i want with $AUTOLOAD... } [10:03:37.0000] <Hixie> the code in that block will get run for any undefined method [10:04:05.0000] <TabAtkins> While not built-in, that's doable easily with the line I posted above. [10:04:07.0000] <Hixie> (it's almost identical to __noSuchMethod__, which would work fine too) [10:04:18.0000] <Hixie> TabAtkins: i believe that it's doable, i just don't see how [10:04:36.0000] <TabAtkins> Okay, let me send you an email with real code. [10:04:45.0000] <Hixie> k :-) [10:05:44.0000] <TabAtkins> Proxies is purposely a low-level API that allows any other proxy-style abstraction to be built on top, so it maps directly to ES's fundamental operations. [10:06:02.0000] <Hixie> yes [10:06:28.0000] <MikeSmith> TabAtkins: maybe post it as a gist and link to it here (along with e-mailing to Hixie). for people wandering across the logs [10:06:52.0000] <Hixie> TabAtkins: again, i'm not arguing that it's not powerful, or whatnot, just that it isn't simple. [10:07:08.0000] <Hixie> TabAtkins: i'm sure it is great for people who understand JS fundamentals [10:07:37.0000] <TabAtkins> Okay, I'll grant you that. I think it's simple because it's easy to think about in terms of those fundamentals, and easy to build abstractions on top of. [10:07:55.0000] <TabAtkins> MikeSmith: Sure, I'll do a blog post. [10:08:09.0000] <Hixie> when i look at the API definition and the first thing I see is "getOwnPropertyDescriptor", i wonder how many JS authors have any clue what that means [10:08:30.0000] <TabAtkins> The fundamental traps are confusing. ^_^ Most of the time you just need to worry about the derived traps like "get". [10:09:48.0000] <Hixie> i guess what i don't understand is the difference between "proxy" and "target". If I just want to have an object that traps all these operations, it's unclear to me whether that's the Proxy object, or the object I pass to the Proxy method. [10:09:54.0000] <Hixie> (also why is it a method and not a constructor?) [10:10:07.0000] <TabAtkins> It's a constructor. A lot of the fundamental constructors dont' require "new". [10:10:23.0000] <Hixie> that's confusing [10:10:39.0000] <gsnedders> [[Construct]] is mostly just a slightly magic [[Call]] [10:10:40.0000] <TabAtkins> The proxy stands between the author and the target. [10:11:29.0000] <Hixie> say i wanted to make an object that just logged all these traps [10:11:44.0000] <Hixie> so if i call foo.bar, it says "getting bar!" [10:11:59.0000] <Hixie> and if i call foo.quux = 2, it says "setting quux!" [10:12:06.0000] <Hixie> how do i do that? [10:12:20.0000] <Hixie> var proxy = Proxy(null, { get: ... }); ? [10:12:32.0000] <Hixie> var proxy = Proxy({}, { get: ... }); ? [10:12:42.0000] <TabAtkins> Proxy({}, {get: function(obj,name){ console.log("getting "+name+"!"); }, set: function(){...}), ...}); [10:12:52.0000] <Hixie> ok [10:13:04.0000] <Hixie> i think it's ridiculous that you have to pass {} in that case. [10:13:35.0000] <Hixie> that's what i mean by "i don't want to have to tell the compiler what i'm proxying" [10:13:55.0000] <TabAtkins> It wasn't seen as worthwhile to provide a variant API that is in all ways identical to just passing an empty object as the first argument. [10:14:11.0000] <Hixie> i don't understand what the purpose of that argument is at all [10:14:15.0000] <TabAtkins> Though there is http://wiki.ecmascript.org/doku.php?id=harmony:virtual_object_api which is similar, but it requires you to define all the fundamental traps. [10:14:38.0000] <TabAtkins> Say you were not just throwing out console spam, but actually *doing* the operations (and also console spamming). [10:14:38.0000] <gsnedders> Hixie: [], {}, function(){} produce different objects. [10:14:47.0000] <gsnedders> Hixie: or host objects [10:14:49.0000] <TabAtkins> Then your call would look more like: [10:15:21.0000] <TabAtkins> Proxy(obj, {get: function(obj, name) { console.log("getting "+name+"!"); return obj[name]; }...}); [10:15:55.0000] <TabAtkins> The proxy is a wrapper around some other object, intercepting calls to that object. [10:16:06.0000] <TabAtkins> Thus the name - it's a proxy for the original object. [10:16:15.0000] <Hixie> i don't understand the use case (other than logging/debugging) for wrapping a single object that already existed when the proxy was made [10:16:22.0000] <TabAtkins> There's tons. [10:16:46.0000] <Hixie> i see lots of use case for proxying to multiple objects or objects that don't yet exist [10:16:56.0000] <Hixie> or for proxying to nothing at all but doing lots of magic [10:17:12.0000] <gsnedders> Hixie: Make array-like methods (filter, reduce) available on a NodeList, as an example [10:17:12.0000] <Hixie> but when do you just want to proxy to an existing object? [10:17:38.0000] <Hixie> for that you'd presumably just poke at the prototype, why would you make a proxy? [10:17:46.0000] <TabAtkins> Hixie: Before I start writing this code, what's your actual use-case? Making sure I capture the details correctly. [10:17:48.0000] <Hixie> in fact how would you even use a proxy to do that? [10:18:12.0000] <TabAtkins> ...that's precisely the main and most direct use-case for Proxies. I have no idea how you're missing it. [10:18:40.0000] <gsnedders> Hixie: A lot of libraries don't want to mutate built-in prototypes. Proxy get to return Array.prototype.reduce when "reduce" is got. [10:19:15.0000] <Hixie> TabAtkins: i want to gsnedders i don't understand why that is better, can you elaborate? [10:20:12.0000] <Hixie> TabAtkins: i have JS objects that represent objects in a server-side data structure, which is lazily loaded as needed. I don't know the types of the objects until I get them, but I need to instantiate the objects before I get them. The API to those objects is all async (callback-based). [10:20:13.0000] <gsnedders> Hixie: Consider the case of having mutliple libraries loaded. You don't want them defining slightly different NodeList.prototype.reduce with different APIs [10:21:34.0000] <Hixie> gsnedders: oh, you do in fact mean "Make array-like methods (filter, reduce) available on a NodeList" and not "Make array-like methods (filter, reduce) available on NodeList", i see [10:22:52.0000] <TabAtkins> Hixie: Okay, cool. You mentioned possibly having multiple objects to dispatch to, based on which method is called. Is that necessary, or was it just theorizing? [10:23:16.0000] <Hixie> TabAtkins: that's a different use case that i've had before, but not relevant to my immediate problems (though i'm curious how it would work too) [10:23:32.0000] <gsnedders> Hixie: Most libraries like jQuery return a wrapper around NodeList or so [10:23:34.0000] <TabAtkins> Okay. It's a trivial modification, but I'll ignore it for now. [10:32:06.0000] <ap> Hi Hixie! I spent some time refreshing my memory of appcache spec, but am still unsure of the answer: can an iframe use a different appcache than main frame? The use case is to cache main application in one cache, and also have a cache with localized data per language. Localized data would be loaded in an invisible iframe. [10:32:20.0000] <ap> Hixie: from what I see, that should work, right? [10:32:36.0000] <Hixie> yes, iframes are inependent [10:32:43.0000] <Hixie> d [10:32:49.0000] <ap> Hixie: thanks [10:42:15.0000] <Ms2ger> Heh [10:42:18.0000] <Ms2ger> css3-marquee [10:43:13.0000] <Hixie> i really should be working at the office rather than at home. bbl. :-) [10:43:32.0000] <Ms2ger> Why? :) [10:43:51.0000] <MikeSmith> whew [10:44:17.0000] <gsnedders> Work from offices? Pff! [10:44:25.0000] <MikeSmith> glad that all the tedious discussion about boring technical stuff is over for now [10:44:46.0000] <gsnedders> Though sometimes I think I should do so more… [10:44:49.0000] <MikeSmith> let's please get back to discussing responsible images [10:48:18.0000] <MikeSmith> oh man [10:48:20.0000] <MikeSmith> https://twitter.com/stevefaulkner/status/208252675110871040 [10:48:54.0000] <MikeSmith> but hey that's a very normal thing that happens very often [10:49:16.0000] <MikeSmith> where multi-year business agreements are extended for only a single month [10:49:29.0000] <MikeSmith> so that's probably not a sign of anything else going on at all [10:49:57.0000] <MikeSmith> they probably just plan to keep extending it for one month at a time for the next 5 years, right? [10:54:36.0000] <hober> could someone remind me who the Redundancy Activity Lead is? [ context: https://twitter.com/w3c/status/208170557169090560 ] [10:57:12.0000] <Ms2ger> You should tell glazou [11:02:35.0000] <hober> Ms2ger: there's already a Member-space thread about it [11:03:04.0000] <Ms2ger> You shouldn't tell me that in a public channel, dear :) [11:09:40.0000] <Ms2ger> hober++ [11:10:51.0000] <rniwa> /me wonders where ++ came from [11:10:59.0000] <rniwa> or attributed to [11:12:48.0000] <TabAtkins> Hm. Queueing method calls with proxies is kinda annoying. [11:13:05.0000] <rniwa> AryehGregor: yt? [11:13:10.0000] <AryehGregor> rniwa, briefly. [11:13:16.0000] <TabAtkins> Because the traps don't know it's a method call, they just know that a property is being requested. [11:13:19.0000] <rniwa> AryehGregor: hi, did you get what I mean for the pre bug? [11:13:19.0000] <AryehGregor> (usually I'm not around at this hour, though) [11:13:52.0000] <AryehGregor> rniwa, I replied on the bug asking for more details, didn't I? [11:13:58.0000] <TabAtkins> So you have to return a special object that'll intercept the call and queue the info for you. [11:13:58.0000] <rniwa> AryehGregor: maybe I'll go visit you in Israel one of these days :) unless you're coming to the TPAC in France this year [11:14:04.0000] <rniwa> AryehGregor: yeah and I replied [11:14:18.0000] <AryehGregor> rniwa, didn't see that yet. I'll look tomorrow. Nope, not coming to TPAC. [11:14:21.0000] <rniwa> AryehGregor: with an example [11:14:24.0000] <rniwa> AryehGregor: okay. [11:14:38.0000] <rniwa> fwiw, i've been to google's tel-aviv office for bidi work :) [11:14:42.0000] <AryehGregor> :) [11:14:52.0000] <rniwa> AryehGregor: it'll be nice to catch up in person every now and then [11:15:18.0000] <rniwa> latency is killing our lively conversatino sometimes :\ [11:15:23.0000] <AryehGregor> Yes, true. [11:19:06.0000] <Philip`> MikeSmith: I guess that's a hint that Google's going to announce next month that it's finished organising the world's information and will henceforth shut down, and they don't want any outstanding contracts at that time [11:19:27.0000] <MikeSmith> heh [11:19:53.0000] <rniwa> AryehGregor: maybe I can go to Israel after/before TPAC [11:20:05.0000] <MikeSmith> Philip`: they should hire me for that job [11:20:10.0000] <rniwa> AryehGregor: I'd rather go fly to Israel from/to France [11:20:14.0000] <rniwa> AryehGregor: than from bay area. [11:20:25.0000] <MikeSmith> Philip`: the shutting down part, I mean [11:20:25.0000] <rniwa> since i got really sick both ways the last time I flew :( [11:20:33.0000] <AryehGregor> rniwa, make sure you coordinate with me well in advance -- I don't work full-time, so I'd have to make sure it's at a time that works for me. [11:20:37.0000] <rniwa> AryehGregor: of course, only if you have a free time around that time [11:20:46.0000] <AryehGregor> rniwa, want to send me an e-mail? [11:20:51.0000] <rniwa> AryehGregor: yeah, will do that :) [11:20:55.0000] <AryehGregor> Thanks! [11:20:56.0000] <rniwa> AryehGregor: but TPAC is in October [11:21:02.0000] <rniwa> AryehGregor: so will probably contact you around August or so [11:21:10.0000] <AryehGregor> rniwa, okay. [11:21:17.0000] <rniwa> it's waaay too early to book a plane ticket, etc... :) [11:21:18.0000] <MikeSmith> Philip`: I like shutting stuff down. I wish I could shut down some more things. Failed social experiments...  [11:21:44.0000] <MikeSmith> failed organizations [11:22:52.0000] <MikeSmith> refund us finally our confiscated gods [11:22:54.0000] <MikeSmith> and all that [11:23:24.0000] <Ms2ger> Failed standards organizations? [11:24:15.0000] <divya> :)) [11:24:44.0000] <rniwa> MikeSmith: since you're here... [11:24:56.0000] <rniwa> MikeSmith: is there anyway to codify instructions for manual tests? [11:25:00.0000] <rniwa> MikeSmith: in HTML or JavaScript? [11:25:24.0000] <rniwa> MikeSmith: the thing is WebKit has a bunch of test automation extensions like window.layoutTestController in our test runner [11:25:38.0000] <rniwa> MikeSmith: and we can emulate keyboard, mouse, etc... events with that. [11:26:06.0000] <rniwa> MikeSmith: so if we could codify instructions in W3C manual tests, we could automate those tests as well. [11:28:16.0000] <MikeSmith> um [11:28:37.0000] <MikeSmith> I don't understand what it means to automate manual tests [11:28:45.0000] <AryehGregor> rniwa, definitely coordinate with someone at Mozilla, because we have the same issue. [11:28:52.0000] <MikeSmith> wait [11:28:55.0000] <MikeSmith> what issue? [11:29:06.0000] <MikeSmith> I think I'm missing something here [11:29:12.0000] <rniwa> oops sorry [11:29:17.0000] <rniwa> MikeSmith: this is about W3C test suites [11:29:25.0000] <MikeSmith> /me nods [11:29:27.0000] <rniwa> MikeSmith: there are a bunch of manual tests in there [11:29:34.0000] <MikeSmith> yeah [11:29:42.0000] <rniwa> MikeSmith: which says something like "click X and do Y" [11:29:47.0000] <MikeSmith> ah [11:29:54.0000] <MikeSmith> WebDriver? [11:30:04.0000] <MikeSmith> emulating user actions? [11:30:20.0000] <rniwa> MikeSmith: right. we have WebDriver-like extension in our test runner [11:30:31.0000] <rniwa> MikeSmith: except that they're exposed as a javascript object. [11:30:37.0000] <MikeSmith> ok [11:30:38.0000] <rniwa> MikeSmith: so instead of you have an external program that drives a test [11:30:44.0000] <MikeSmith> yeah [11:30:47.0000] <rniwa> MikeSmith: the test itself can control what test runner does. [11:31:04.0000] <rniwa> MikeSmith: I suspect Mozilla has a similar mechanism from what AryehGregor just said. [11:31:14.0000] <AryehGregor> Yes. [11:31:49.0000] <MikeSmith> unless I'm missing something, this is exactly the scope of the standard Web driver API work [11:32:03.0000] <MikeSmith> Simon Stewart et all [11:32:21.0000] <MikeSmith> minus one l [11:33:04.0000] <MikeSmith> http://dvcs.w3.org/hg/webdriver/raw-file/default/webdriver-spec.html [11:33:28.0000] <MikeSmith> if not, it should be the scope [11:33:40.0000] <MikeSmith> the ultimate goal being, automate everything [11:33:55.0000] <MikeSmith> or "anything that can be automated must be automated" [11:33:58.0000] <rniwa> MikeSmith: hm... the last time i checked, WebDriver API was supposed to be used externally? [11:34:25.0000] <MikeSmith> sorry, what do you mean by "externally"? [11:34:35.0000] <MikeSmith> this is an API that's exposed to Web applications [11:34:44.0000] <rniwa> MikeSmith: oh is it? [11:34:53.0000] <MikeSmith> yeah man [11:34:57.0000] <MikeSmith> that's the whole idea [11:35:03.0000] <rniwa> MikeSmith: so if you have test.html [11:35:06.0000] <MikeSmith> unless I'm deeply confused [11:35:15.0000] <rniwa> MikeSmith: we can invole methods, etc... of WebDriver in test.html itself? [11:35:23.0000] <MikeSmith> yes [11:35:36.0000] <MikeSmith> once it's actually implemented of course [11:36:12.0000] <MikeSmith> but man you are talking to the village idiot here [11:36:15.0000] <MikeSmith> me I mean [11:36:29.0000] <MikeSmith> I recommend pinging Simon Stewart about it [11:36:47.0000] <MikeSmith> dude can make it much more clear than me [11:37:06.0000] <rniwa> MikeSmith: ok. is he n #testing at irc.w3.org? [11:37:22.0000] <MikeSmith> sometimes [11:37:27.0000] <rniwa> MikeSmith: the last time i talked with someone wokring on webdriver [11:37:34.0000] <MikeSmith> he may also be on #chromium [11:37:44.0000] <MikeSmith> David Burns is also working on this [11:38:10.0000] <MikeSmith> and Eran Messeri [11:38:17.0000] <rniwa> MikeSmith: oh, Simon Stewart is a googler :\ [11:38:17.0000] <MikeSmith> rniwa: yeah? [11:38:26.0000] <MikeSmith> yeah man [11:38:28.0000] <Ms2ger> Boo, Googlers :) [11:38:30.0000] <MikeSmith> jesus [11:38:37.0000] <rniwa> MikeSmith: okay, I'll go talk with him. thanks! [11:38:43.0000] <MikeSmith> do you guys actually talk to each other? [11:39:03.0000] <MikeSmith> rniwa: your company is too big, chief [11:39:03.0000] <rniwa> hm... there are at least 3 Simon Stewart at Google :( [11:39:08.0000] <MikeSmith> wow [11:39:27.0000] <MikeSmith> that's a clear sign it's time to move on to somewhere smaller [11:39:42.0000] <Ms2ger> Consider Mozilla, for example... [11:39:55.0000] <Ms2ger> We're always looking for people who know things about editing :) [11:40:11.0000] <MikeSmith> we too [11:40:19.0000] <MikeSmith> and we have only 60 people on staff [11:40:47.0000] <MikeSmith> maybe 30 people if you count the ones that are actually doing productive work [11:40:51.0000] <MikeSmith> maybe [11:41:02.0000] <MikeSmith> hard to trump that [11:42:20.0000] <rniwa> Ms2ger: nah... you g!uys have AryehGregor & ehsan [11:42:34.0000] <Ms2ger> Do you know what our code looks like? [11:42:36.0000] <rniwa> oops s/g!uys/guys/ & s/ehsan/ehsan!/ [11:42:59.0000] <Ms2ger> Hmm, ehsan factorial would be nice... [11:43:00.0000] <rniwa> MikeSmith: i think google has something like 30,000 employees now... [11:43:07.0000] <rniwa> LOL [11:43:15.0000] <rniwa> Ms2ger: that'll be a lot of ehsan indeed. [11:44:34.0000] <MikeSmith> rniwa: you should test all of those people to determine how well the can estimate how many basketballs can fit in the room [11:45:22.0000] <Ms2ger> MikeSmith, surely HR already did that? [11:45:36.0000] <TabAtkins> That's the first question you answer to get in the door. [11:45:47.0000] <rniwa> Ms2ger: to make you feel better, take a look at http://trac.webkit.org/browser/trunk/Source/WebCore/editing/ReplaceSelectionCommand.cpp [11:45:56.0000] <MikeSmith> OK, well at least make the draw the organizational structure of Google on a whiteboard [11:46:11.0000] <rniwa> Ms2ger: which is mutually recursive (we bail out when the actual recursion happens) with DeleteSelectionCommand (another class) & mergeParagraphs [11:46:48.0000] <rniwa> with moveParagraphs [11:46:54.0000] <MikeSmith> TabAtkins: my answer is, it's a fucking stupid question that reveals asshattedness in the person assking it [11:47:19.0000] <MikeSmith> but hey that's just me [11:48:10.0000] <Ms2ger> rniwa, hmm, that code seems to be calling RefPtr::get() a lot [11:48:22.0000] <TabAtkins> MikeSmith: Yes, I was joking. ^_^ [11:48:28.0000] <rniwa> MikeSmith: someone said the joke that we should ask every SWE in our company to see if they can understand the recent "proof" of P != NP [11:48:53.0000] <rniwa> Ms2ger: that's okay. RefPtr::get() is an inline function that just obtains a raw pointer [11:49:04.0000] <rniwa> Ms2ger: it's basically zero-cost. [11:49:15.0000] <Ms2ger> rniwa, yeah, but our refptrs have an operator T* [11:49:29.0000] <MikeSmith> TabAtkins: problem is I guess I can't tell where the joke starts. the truth is actually stranger than the fiction [11:49:35.0000] <rniwa> Ms2ger: the reason we don't have that is due to PassRefPtr I believe [11:49:42.0000] <Ms2ger> Oh, how silly [11:50:22.0000] <rniwa> Ms2ger: it actually saves us a lot of CPU cycles :) [11:50:29.0000] <rniwa> s/ us// [11:50:38.0000] <rniwa> Ms2ger: not to mention it improves the cache locality [11:51:27.0000] <rniwa> although Darin (Adler) recently told me a depressing story about how editing was the reason we introduced RefPtr... :'( [11:51:55.0000] <Ms2ger> Heh [11:52:26.0000] <MikeSmith> anyway, building an organization where the product-dev work in largely driven by Assperghers-syndome engineers combined with the revenue side being driven by completely unethical moneygrubbers is clearly the recipe for long-term success [11:52:57.0000] <TabAtkins> Worked for us! (You gotta make sure the revenue-type people don't infiltrate management, is all.) [11:53:27.0000] <MikeSmith> TabAtkins: then you have the cases like Marius Milner who combine the best of both worlds [11:53:35.0000] <rniwa> TabAtkins: "revenue-type people don't infiltrate management" indeed is very important for almost all companies [11:54:51.0000] <MikeSmith> whoever decided to hire Marius and give him free reign should get a super-big gold star [11:55:43.0000] <TabAtkins> OMG, people are still talking about the wifi bullshit? [11:55:54.0000] <MikeSmith> yeah [11:56:23.0000] <MikeSmith> because I guess people sorta are surprised when other people blow smoke up their asses [11:56:26.0000] <TabAtkins> You send data to unencrypted websites over unencrypted wifi, everyone in your vicinity has access to it. [11:56:35.0000] <MikeSmith> yeah [11:56:40.0000] <MikeSmith> and then you lie about it [11:56:47.0000] <MikeSmith> nothing wrong with that [11:57:07.0000] <MikeSmith> you claim you had no clue what was going on [11:57:21.0000] <MikeSmith> was just one crazy dude out there doing stuff on his own [11:57:44.0000] <MikeSmith> nothing evil about that [11:58:05.0000] <MikeSmith> anyway [11:58:21.0000] <MikeSmith> we all work for really wonderful organizations [11:58:30.0000] <TabAtkins> It was random bits of data grabbed across a few seconds as a car drove by. [11:58:42.0000] <MikeSmith> yep [11:58:47.0000] <MikeSmith> exactly [11:58:52.0000] <MikeSmith> that's all it was [11:59:21.0000] <MikeSmith> so no reason to not be totally transparent publicly about the fact that's what you were doing [12:01:25.0000] <MikeSmith> on the bright side there's no record of Marius actively working to prevent people who love each other from having the legal right to get married [12:01:30.0000] <MikeSmith> so that's a plus [12:01:59.0000] <MikeSmith> like I said, we all work for really wonderful organizations [12:02:47.0000] <MikeSmith> with really exemplary leadership [12:03:49.0000] <Ms2ger> /me wonders what got MikeSmith on his horse [12:05:16.0000] <MikeSmith> Ms2ger: nothing but love, an [12:24:37.0000] <jgraham> Please stop writing so much when I'm not looking [12:24:53.0000] <jgraham> /me tries to figure out if people eventually got the right story about WebDriver [12:28:42.0000] <jgraham> rniwa: So the story with webdriver is that it's typically external; you have some script outside the browser that drives the interaction [12:28:48.0000] <TabAtkins> Hm. Hixie: After finishing the impl, I'm pretty sure you don't need proxies at all. [12:29:03.0000] <jgraham> But they have made a js API for it. I don't know if that could be self-driving [12:29:05.0000] <rniwa> jgraham: can it be inside? [12:29:30.0000] <rniwa> jgraham: e.g. it'll be useful to be able to do something like [12:29:41.0000] <rniwa> div.sendKeyDown(); [12:29:47.0000] <rniwa> assertSomethingHappened() [12:30:02.0000] <rniwa> jgraham: if the API is only available to external programs [12:30:12.0000] <rniwa> jgraham: then asserting conditions will be much more complicated [12:30:18.0000] <rniwa> jgraham: because of concurrency, etc... [12:30:31.0000] <jgraham> Right. So the way we do that is to use watir which is a particularly crappy ruby wrapper around WebDriver [12:30:58.0000] <jgraham> Designed by the kind of people who says things like "tests can be specifications" [12:31:18.0000] <jgraham> And yes, I think concurrency is a problem [12:31:39.0000] <jgraham> Because it sort of tries to pretend that the API is sync, but it obviously isn't really [12:31:44.0000] <rniwa> jgraham: oh no :( cucumber... [12:32:09.0000] <rniwa> jgraham: yeah, so I'd vote for exposing it via JavaScript. [12:32:17.0000] <rniwa> jgraham: we certainly don't want to expose it all the time. [12:32:22.0000] <TabAtkins> Hixie: Ah, nm, I see the flaw in what I was going to write. You either need (a) proxies, (b) private names, or (c) to add the methods to every instance instead of them being on the prototype. [12:32:29.0000] <jgraham> Well it isn't really "exposed" in any sense [12:32:33.0000] <rniwa> but it's not an issue for browser vendors... [12:32:46.0000] <jgraham> I mean, you have to connect externally to the browser [12:33:06.0000] <rniwa> jgraham: right... so what I'm advocating is to expose it via DOM [12:33:28.0000] <jgraham> So, I don't really know how well that would work [12:33:30.0000] <rniwa> jgraham: webElementDiv = window.webDriver(div); [12:33:44.0000] <jgraham> It at the least sounds *different* to WebDriver [12:34:01.0000] <jgraham> But it was someone at Google who was working on the JS bindings [12:34:13.0000] <jgraham> I don't remember his name though [12:34:17.0000] <rniwa> jgraham: and then you can do something like webElementDiv.sendKeys(~~) [12:34:32.0000] <jgraham> ... I don't know the API details [12:34:33.0000] <benvie> if you're referring to what I think you are, I did that here using code generation from IDL and prototye accessors https://github.com/Benvie/svgstuff/blob/master/lib/defs.js [12:35:34.0000] <jgraham> I know that Mozilla are using the python bindings and I think Google/FB use it to test their websites (probably with Java/PHP bindings, respectively) [12:35:52.0000] <jgraham> But about the JS bindings I know nothing [12:35:58.0000] <rniwa> Simon Stewart? [12:36:06.0000] <jgraham> Well he will know who knows [12:36:16.0000] <jgraham> But it wasn't him that wrote them [12:36:25.0000] <jgraham> But he is project lead or something [12:37:03.0000] <jgraham> None of the Opera|ex-Opera people who would know more are here at the moment [12:37:25.0000] <rniwa> benvie: i don't understand. how are you emulating sendKeys, etc... [12:37:29.0000] <rniwa> benvie: from javascript? [12:37:53.0000] <benvie> well I did that using node and ffi but not really finished [12:38:06.0000] <rniwa> benvie: i don't think that's what we're looking for. [12:38:14.0000] <benvie> yeah [12:38:17.0000] <benvie> similar [12:38:21.0000] <benvie> but no =D [12:38:31.0000] <rniwa> benvie: yeah, the idea is similar. [12:38:51.0000] <rniwa> jgraham: i think there's a real value in exposing these APIs through javascript [12:38:55.0000] <rniwa> jgraham: so that tests can be self-contained [12:39:18.0000] <rniwa> jgraham: one drawback, however, is that we probably won't be shipping this API in production [12:39:31.0000] <gsnedders> rniwa: And that makes it impossible to test IE/Opera as a third-party. [12:39:32.0000] <rniwa> jgraham: so ordinary web developers can't use them to run tests :\ [12:39:42.0000] <gsnedders> rniwa: Or even release builds of Safari, say [12:39:45.0000] <rniwa> gsnedders: right. [12:39:48.0000] <rniwa> gsnedders: but that's already true. [12:39:55.0000] <rniwa> gsnedders: we can leave a manual test instruction [12:40:01.0000] <rniwa> gsnedders: at least that's what we do in webkit [12:40:13.0000] <rniwa> gsnedders: we have window.eventSender which lets us emulate keyboard/mouse events [12:40:18.0000] <rniwa> gsnedders: but we also leave manaul test instruction [12:40:27.0000] <rniwa> gsnedders: so that we can run those tests in firefox, etc... [12:40:44.0000] <gsnedders> rniwa: WebDriver being external allows automated testing of the browsers, which is better than that [12:40:59.0000] <rniwa> gsnedders: I don't think everyone is on the page, however. [12:41:20.0000] <rniwa> gsnedders: I remember Microsoft explicitly said they won't be implementing it for example. [12:41:28.0000] <rniwa> s/said/saying/ [12:41:56.0000] <rniwa> gsnedders: and I'm skeptical that we can use WebDriver API as is in our test harness [12:42:17.0000] <rniwa> gsnedders: what's the point of a test suite if browser vendors can't run them? [12:42:33.0000] <rniwa> or rather "don't" [12:43:19.0000] <jgraham> rniwa: Why are you skeptical? [12:43:30.0000] <rniwa> jgraham: that we can support WebDriver in our test harness. [12:43:46.0000] <jgraham> Hmm, well I guess I don't know how you run tests [12:43:50.0000] <rniwa> jgraham: fwiw, we don't use a full-blown web browser to run tests. [12:44:05.0000] <jgraham> Well we soert-of don't [12:44:09.0000] <rniwa> jgraham: we have a special test runner called DumpRenderTree or WebKitTestRunner (for webkit2) [12:44:39.0000] <jgraham> How unlike a real browser is it? [12:44:40.0000] <rniwa> jgraham: that exoses special objects such as layoutTestController, eventSender, textInputController in the global scope (i.e. on window object) [12:44:53.0000] <rniwa> jgraham: it's so unlike a real browser that it doesn't even have a window. [12:45:13.0000] <rniwa> jgraham: you can't see anything until the test completes [12:45:19.0000] <jgraham> Hmm, not having a window does sound like it could be a problem [12:45:20.0000] <rniwa> jgraham: and you can't interact with it. [12:45:23.0000] <gsnedders> jgraham: It literally just makes a render tree and (sometimes) a screenshot available. That's it. [12:45:35.0000] <gsnedders> jgraham: There is no interaction, you start it, it does that, over. [12:45:39.0000] <rniwa> gsnedders: right. [12:45:55.0000] <jgraham> Oh well that seems like a kind of broken way of testing [12:46:05.0000] <jgraham> In that it is very unlike anything you will ship [12:46:19.0000] <jgraham> (very different codepaths) [12:46:26.0000] <rniwa> jgraham: in practice, it can test things pretty well. [12:46:45.0000] <rniwa> jgraham: we do all the paining, layout, etc... normally [12:46:45.0000] <rniwa> it [12:46:52.0000] <rniwa> jgraham: it's just that it doesn't have any real UI to it. [12:46:57.0000] <rniwa> s/ to it// [12:47:06.0000] <jamesr_> we also do some testing with a fuller browser, of course, but we don't run every test that way [12:47:19.0000] <rniwa> jgraham: and doesn't let user interact because the whole point of the test runner is to automate testing [12:47:24.0000] <rniwa> jgraham: and run them as fast as possible. [12:47:37.0000] <jgraham> So it is rendering to an offscreen buffer, effectively? [12:47:40.0000] <rniwa> jgraham: if we were to start the entire web browser for each test case, it would be impratically slow. [12:47:43.0000] <jamesr_> correct [12:47:46.0000] <rniwa> jgraham: yes. [12:47:50.0000] <jamesr_> or in many cases not rendering at all [12:48:08.0000] <jgraham> rniwa: Sure, obviosuly you have to not do that [12:48:11.0000] <rniwa> jgraham: things like GPU acceleartion, etc... need a special treatment because of that. [12:48:11.0000] <gsnedders> Do you test stuff like GPU painting with it? [12:48:15.0000] <jamesr_> if the test doesn't depend on pixels [12:48:47.0000] <jamesr_> for tests that need it, we test that part of the GPU pipeline (normally using osmesa so it can run on VMs easily) [12:48:59.0000] <jgraham> We can use our remote-debugging support to load pages, wait for the browser to become idle (or for a result to be recieved) and load the next test, without restarting [12:49:23.0000] <jgraham> Although WebDriver/Watir tests are an exception, sadly [12:49:25.0000] <gsnedders> So you could, theoretically, have GPU pipeline bugs in layout tests that don't get found? [12:49:54.0000] <rniwa> gsnedders: our regular tests can't find them. but as jamesr said, we have special tests for those. [12:49:59.0000] <jamesr_> we could have bugs anywhere that don't get found, but exercise most of the GPU pipeline in layout tests (the parts that aren't covered are OS-integration things like IOSurface swapping, etc) [12:50:08.0000] <jamesr_> some of our regular layout tests run through the GPU path [12:50:51.0000] <rniwa> with GPU acceleration, though, testing everything is virtually impossible because you then have to try all permutations of OS, GPU chip/board, driver, etc... [12:50:52.0000] <jamesr_> but not all - most of our tests don't render at all, they just make JS assertions or dump the render tree without painting it [12:51:21.0000] <jgraham> (and we have a lightweight platform layer and render to a virtual buffer) [12:52:38.0000] <gsnedders> jgraham: That's not true, we render to Xvfb — our platform layer we run on is complete, just the platform beneath it isn't so standard. [12:53:35.0000] <gsnedders> jamesr_: Was mainly asking because I know we do have some general layout bugs with the GPU pipeline (though we equally don't normally run all tests through it) [12:53:48.0000] <jgraham> gsnedders: I meant "lightweight" as in "not the full desktop browser (except when we test that specifically)" [12:54:01.0000] <jgraham> Possibly that wasn't sufficiently obvious [12:54:14.0000] <jamesr_> i think it's a terminology thing but we would normally call that a rendering or paint bug, not layout [12:54:42.0000] <jamesr_> layout figures out parameters on a bunch of C++ objects representing what we call the render tree. painting goes through those data structures and generates pixels [12:54:55.0000] <gsnedders> jamesr_: Right, I wouldn't call that a layout bug per-se, but it's a bug affecting layout which is caught by the layout tests, even if the bug lies elsewhere [12:54:57.0000] <jamesr_> if the GPU pipeline is broken it'd tend to break the latter of those two steps but it couldn't really break the former [12:55:20.0000] <gsnedders> affecting layout insofar as what the user sees [12:55:32.0000] <jamesr_> yes definitely. we have tests to cover that [12:55:42.0000] <rniwa> gsnedders: we normally categorize those as painting bugs :) [12:55:54.0000] <rniwa> /me hates technical jargons [12:56:07.0000] <rniwa> why can't we stop inventing obnoxious jargons [12:56:20.0000] <gsnedders> jamesr_: I guess my point is more painting bugs can turn up (and hence regress!) in layout tests as well as painting tests. [12:56:29.0000] <jgraham> Hmm a painting bug sounds like you have the right information in layout but the gfx layer did something wrong (and it is often fixed by forcing a repaint) [12:56:40.0000] <jamesr_> jgraham, depends on the bug [12:57:00.0000] <jamesr_> gsnedders, certainly, and that's why we have the capability of going through the painting path (including GPU where applicable) in our layout test harness [12:57:08.0000] <jgraham> jamesr_: Sure, I think I'm just saying that I would draw a distinction along those lines [12:57:14.0000] <jamesr_> and then checking that those pixels either match a golden PNG or the rendering of a reference file [12:57:22.0000] <jgraham> Which might be an Opera thing or might be something I made up [12:57:54.0000] <Hixie> TabAtkins: still think it's trivial? :-) [12:58:00.0000] <jamesr_> we also have repaint tests which paint once, change something, then paint again and make sure we actually update all the pixels that are supposed to look different [12:58:01.0000] <jgraham> "then checking that those pixels either match a golden PNG" - I'm sorry :) [12:58:24.0000] <jamesr_> we have a lot of golden PNGs (as in golden file testing) in WebKit [12:58:43.0000] <jgraham> We still have too much of it in Opera [12:58:49.0000] <jgraham> It is evil and must die [12:59:00.0000] <gsnedders> jamesr_: And regressions could slip through if it's not done often. [12:59:29.0000] <jamesr_> we have bots that run it as fast as they can cycle (it's not every checkin, but it's every hour or two on the slowest configs) [12:59:58.0000] <gsnedders> Right, so it's part of the normal testing cycle, even if it isn't done on every run. [13:00:18.0000] <gsnedders> It just sounded as if it was exceptional and not normally done at all from what you said before. [13:00:18.0000] <Philip`> jgraham: I guess the problem is exacerbated by Opera supporting too many platforms? [13:00:40.0000] <gsnedders> Philip`: And subtle differences between products. [13:01:24.0000] <rniwa> anyways, [13:01:27.0000] <jamesr_> ah no, it's part of the configuration of the test so not all tests are configured to hit every path, but we run all the tests continuously in many configurations [13:01:43.0000] <rniwa> to re-iterate my point, it would be really nice if we could expose WebDriver API in tests themselves [13:01:58.0000] <rniwa> so that we can import W3C tests without havingt to manually modify them [13:02:14.0000] <jgraham> Philip`: > 1 you mean? :) [13:02:31.0000] <jgraham> Or really > 0 [13:02:42.0000] <gsnedders> It'd bad enough with one platform, having to manually verify things :P [13:02:58.0000] <jgraham> Since even on one platform people do annoying things like change the font engine every so often [13:04:03.0000] <gsnedders> (Font rendering was a big reason why we tested on Windows 2000 for a long time, not wanting to go through n thousand screenshots and relabel them) [13:07:29.0000] <rniwa> gsnedders: yeah... font anti-aliasing kills us :( [13:08:13.0000] <rniwa> gsnedders: we have to generate thousands of very similar PNGs just to cope up with font-aliasing differences between different versions of Mac/Win/etc... [13:08:32.0000] <rniwa> (or maybe we've disabled anti-aliasing on Windows; /me doesn't remember) [13:14:41.0000] <rniwa> jgraham, gsnedders: so... do you think exposing it via JavaScript would be an option at all? [13:14:48.0000] <rniwa> with manual instructions? [13:15:17.0000] <rniwa> jgraham, gsnedders: alternatively, i would be fine with having some external file that instructs WebDriver what to do. [13:15:45.0000] <rniwa> jgraham, gsnedders: as long as it's machine-readable so that we can manipulate it to work with our test harness. [13:16:28.0000] <gsnedders> I'd much rather be able to test release browsers. [13:16:49.0000] <rniwa> gsnedders: that could be done manually, right? [13:16:55.0000] <rniwa> gsnedders: we already do that manually. [13:17:20.0000] <rniwa> gsnedders: but i'm open to options that allows automated testing of browsers in production. [13:17:26.0000] <rniwa> gsnedders: that sounds like a very valuable goal to have. [13:17:52.0000] <rniwa> gsnedders: as long as the configuration is such that we can also use it in our test harness [13:20:17.0000] <TabAtkins> Hixie: More details are needed. These async methods, are they part of a stable, closed set? Or are they determined by the server-side object? [13:21:57.0000] <rniwa> gsnedders: would that sound good with you? [13:22:06.0000] <rniwa> s/would/does/ [13:22:15.0000] <rniwa> ugh... s/with/to/ [13:25:53.0000] <TabAtkins> If they're a closed set, I can do this without proxies at all. [13:26:38.0000] <TabAtkins> Hixie: The difficulty here that makes it harder than other languages is that Javascript doesn't actually have any concept of "methods". It just has properties, whose values might be callable. [13:32:05.0000] <Hixie> TabAtkins: there is a finite number of known types that will be instantiated, each with a finite number of methods. [13:32:21.0000] <Hixie> TabAtkins: sure. s/method/field/ is fine. [13:32:24.0000] <TabAtkins> Okay, that's the "closed set" option. Cool. [13:32:48.0000] <TabAtkins> Hixie: Renaming doesn't help. ^_^ It just means that I need to be a little fancy to intercept method calls. [13:32:49.0000] <Hixie> (i actually do have an open-ended set of methods but one of the methods from the closed set is what i'll use to invoke the open set) [13:33:01.0000] <Hixie> TabAtkins: ah [13:41:20.0000] <TabAtkins> Unless I'm crazy, this turned out to be *really* easy once I stopped trying to use Proxies, because they're not necessary. [13:41:34.0000] <TabAtkins> Though you likely need Private Names to avoid exposing some of the data here. [13:43:44.0000] <jgraham> So what are you doing? Swapping prototypes from one that queues the results to one that doesn't. That's pretty evil so presumably not... [13:43:59.0000] <jgraham> s/./?/ [13:44:53.0000] <jgraham> Or I guess you can just rewrite the properties at runtime in the object itself [13:45:09.0000] <jgraham> None of this sounds healthy though [13:45:18.0000] <TabAtkins> Hixie: http://www.xanthir.com/blog/b4JB0 [13:45:48.0000] <TabAtkins> jgraham: Tell me if I'm doing anything wrong here. [13:47:20.0000] <Hixie> TabAtkins: swapping prototypes is what i suggested a few hours ago when you were saying to use proxies :-) [13:47:30.0000] <TabAtkins> Hixie: I'm not swapping protos... [13:47:42.0000] <Hixie> oh, i misread what jgraham said, my bad [13:47:52.0000] <TabAtkins> I thought proxies were necessary when I thought that the set of async methods was open-ended. [13:48:45.0000] <Hixie> wait so i have to list every function that every possible class of object might implement? [13:48:58.0000] <TabAtkins> No, you have different FarObject classes. [13:49:12.0000] <Hixie> i don't know the type of the FarObject until i get back the data from the server [13:49:21.0000] <TabAtkins> Oh! [13:49:31.0000] <TabAtkins> Hm, then. Let me think about this. [13:49:39.0000] <Hixie> if i knew the type of the object it would be trivial, i just wouldn't need to create a proxy at all [13:49:43.0000] <TabAtkins> We'll probably be back to needing a proxy if we want to be efficient. [13:49:49.0000] <Hixie> i could just create the object and have the object itself do the queuing [13:49:58.0000] <TabAtkins> Yeah, that's what I ended up doing. ^_^ [13:50:21.0000] <Hixie> (which btw is what i think i will probably end up doing, by having every object ID actually include its type as well) [13:50:40.0000] <Hixie> (but i still wish it was easy to do it without that) [13:50:50.0000] <TabAtkins> If Type1 has asyncFoo and type2 has asyncBar, and you call asyncBar on the object before it's loaded but it ends up being Type1 when you get the data, what happens? [13:51:46.0000] <Hixie> in practice, you don't do that, because you'll have some idea of what the superclass of the object is and won't call things that that superclass doesn't support [13:52:13.0000] <Hixie> and only when you get information back from the methods of that superclass would you then call the methods of the subclass (the class it actually is) [13:52:18.0000] <Hixie> anyway i think my original point, which is that this is not as trivial as in perl, stands proven [13:52:18.0000] <TabAtkins> I think the "won't call" is more theory than practice. ^_^ [13:52:30.0000] <benvie> I implemented synchronizing across remote DOM implementations by using unique IDs for every new object and recording all the inputs and outputs of every action [13:52:31.0000] <TabAtkins> I defy you to actually write this in two lines of perl. [13:52:56.0000] <Hixie> i gave you the two lines of perl to define a catchall method [13:53:08.0000] <benvie> so document.createElement becomes #1 GET #2, apply #2 'div' creates #3 [13:53:09.0000] <benvie> and so on [13:53:14.0000] <Hixie> actually doing the dispatch would be a few more, but sure, hold on [13:53:15.0000] <TabAtkins> Yeah, the catchall part is easier because Perl has methods. [13:53:45.0000] <Hixie> Mozilla has (had?) __noSuchMethod__ [13:53:48.0000] <Hixie> in JS [13:53:52.0000] <Hixie> which would make this easy too [13:54:12.0000] <TabAtkins> That doesn't obey the raw JS semantics. ^_^ [13:54:20.0000] <Hixie> but it's easy :-) [13:54:31.0000] <TabAtkins> Of course it is! [13:56:43.0000] <creis_> Hixie: Do you know anything about the rel=external link type (http://blog.whatwg.org/the-road-to-html-5-link-relations#rel-external)? [13:56:56.0000] <creis_> Looks useful for something I'm considering, but it doesn't appear to have made it into the spec. [13:57:19.0000] <creis_> Just wondering if it's a dead proposal or still being considered. [13:57:52.0000] <Hixie> creis_: it's defined here: http://microformats.org/wiki/rel-external [13:58:01.0000] <Hixie> creis_: and registered here: http://microformats.org/wiki/existing-rel-values#HTML5_link_type_extensions [13:58:23.0000] <Hixie> creis_: the w3c had us remove it from the spec text itself for some reason or other, but i expect it'll be back in the spec sooner or later [13:58:38.0000] <Hixie> creis_: i'm planning on revamping how registration of things like rel values is done [13:58:45.0000] <Hixie> creis_: now that we're on a truly "living standard" model [13:59:14.0000] <tantek> Hixie - have there been any issues with using the microformats wiki? [13:59:23.0000] <tantek> (for rel registration) [13:59:50.0000] <creis_> Hixie: Cool, thanks. Slightly related, do you know if there's any consideration about applying similar link types to window.open, now that the "features" argument is supposed to be ignored? [14:00:43.0000] <Hixie> tantek: not especially, it's just not that great that the values that are "standard" aren't in the spec [14:00:50.0000] <Hixie> tantek: i don't expect use of the wiki to stop [14:01:09.0000] <Hixie> tantek: just that once things are "accepted" they be put in the spec so that people can find them more easily [14:01:34.0000] <tantek> ok cool. so the wiki helps with standardizing, then when things are stable, they can be incorporated into the spec. I'm fine with that. [14:01:36.0000] <Hixie> creis_: i don't think i've heard of any problem that that would solve [14:01:45.0000] <tantek> I'd like to avoid having redefining though - which is what happened with rel-tag [14:01:52.0000] <Hixie> tantek: indeed [14:02:10.0000] <Hixie> tantek: (well, rel=tag was more about matching existing practice than redefining) [14:02:15.0000] <Hixie> tantek: (but i agree in principle) [14:02:29.0000] <tantek> hixie - your opinion of existing practice, not without dispute [14:02:46.0000] <creis_> Hixie: One example is rel=noreferrer, which suppresses window.opener if used in combination with target=_blank. It would be nice to have a way to call window.open where opener was not set (e.g., with rel=external or something similar). [14:02:52.0000] <tantek> (document vs. blog post granularity) [14:03:24.0000] <tantek> Hixie, I'd say if you think the definition of a rel value on the wiki (and respective spec page there) is insufficiently detailed, it's probably not mature enough to make it into the standard [14:04:19.0000] <Hixie> creis_: makes sense. i don't recall if we have a solution for that offhand, but either way, please don't hesitate to file a bug or send e-mail about it to get it on the list. [14:04:35.0000] <Hixie> tantek: agreed, again in principle [14:04:37.0000] <creis_> Hixie: Thanks, I'll follow up. [14:04:56.0000] <tantek> Hixie, great. we'll cross further bridges when we get to them then. [14:07:02.0000] <jgraham> Hixie: It occurs to me that there is something I was going to ask you when you weren't around. Maybe I should work out what it was while you are around... [14:08:17.0000] <jgraham> Ah, maybe you already answered in a bug [14:09:34.0000] <Hixie> jgraham: yeah i think i answered in teh bug [14:10:16.0000] <jgraham> I wonder what happens if you don't use the location interface but set the src attribute [14:10:24.0000] <Hixie> on iframe? [14:10:34.0000] <Hixie> setting src on iframe iirc kills the browsing context entirely [14:10:36.0000] <jgraham> (this might also be covered already, the spec is still ladaing) [14:10:39.0000] <Hixie> and recreates it [14:11:13.0000] <jgraham> *loading [14:11:47.0000] <jgraham> But it is still possible that you would get one or two history positions in the joint session history [14:13:38.0000] <Hixie> when you set src="" the load happens with replacement enabled, iirc [14:13:40.0000] <Hixie> which means you get 1 [14:13:43.0000] <Hixie> the about:blank is nuked [14:13:52.0000] <Hixie> but i don't recall exactly [14:14:18.0000] <jgraham> I don't see that in http://www.whatwg.org/specs/web-apps/current-work/#attr-iframe-src [14:14:52.0000] <Hixie> TabAtkins: http://damowmow.com/playground/perl-forwarder.pm (not tested, includes all the boilerplate as well except for the actual loading code) [14:14:59.0000] <jgraham> But it is quite possible I am missing something because navigation feels quite like "you are in a maze of twisty algorithms, all alike" [14:15:01.0000] <TabAtkins> Hixie: Take my last scenario, with Type1 and Type2. What's the desired behavior difference, before the data is loaded, between calling asyncBar() (valid function, but not for the type you eventually figure out for it) and calling asyncUnknown() (no type has this function)? [14:16:27.0000] <Hixie> jgraham: search for "Furthermore, if the browsing context's session history contained only one Document when the process the iframe attributes algorithm" [14:16:36.0000] <Hixie> jgraham: just below the "process the iframe attributes" algorithm [14:17:34.0000] <Hixie> TabAtkins: i guess while you're loading, both should just do nothing, and once you're loaded, both should throw an exception (so in particular, the former shoudl throw an exception when the queue of methods is flushed) [14:17:43.0000] <Hixie> TabAtkins: at least, that's what the perl impl i threw together did [14:18:15.0000] <Hixie> TabAtkins: the perl implementation of AUTOLOAD ended up being 11 lines, of which 6 are trivial [14:18:31.0000] <jgraham> Hixie: But is that right? I can have multiple entries in the iframe history e.g. I initially lood 001.html and while that is loading switch the location.href to 002.html and while that is loading set the src to 003.html [14:19:13.0000] <jgraham> /me thinks the autoload thing would be just as easy in python but with less line noise [14:19:18.0000] <Hixie> jgraham: since navigation cancels existing navigations, i don't think you can have more than one [14:19:52.0000] <Hixie> jgraham: oh uh [14:20:01.0000] <jgraham> What if one ran to completion? [14:20:03.0000] <Hixie> jgraham: i misunderstood (because i was assuming setting src blows away the browsing context) [14:20:36.0000] <Hixie> jgraham: actually what i said still stands [14:21:01.0000] <Hixie> jgraham: you'd have a race condition in the example you give [14:21:20.0000] <Hixie> jgraham: "while that is loading" could mean either before the session history is updated, or after [14:21:28.0000] <Hixie> jgraham: if it's before, then it's as if it never happened [14:21:38.0000] <Hixie> jgraham: if it's after, then the condition doesn't apply [14:23:28.0000] <jgraham> Hixie: It is the latter case I wonder about, I think [14:24:29.0000] <Hixie> jgraham: so we'd be talking about a case where an iframe is loaded with a file that, while it is loading, replaces the src, and the question is do you get one entry or two. [14:24:37.0000] <Hixie> jgraham: should be easy enough to test :-) [14:24:57.0000] <jgraham> Yeah I guess I will do that tommorrow [14:25:06.0000] <Hixie> jgraham: (i'd also check to see if you get 2 or 3 entries in the same case but with a page being loaded completely first) [14:25:22.0000] <Hixie> jgraham: (in case setting src="" does indeed blow away the session history) [14:25:32.0000] <jgraham> I am pretty sure the answer will be that you get N+1 entries rather than N+2 given that you start with N [14:25:32.0000] <Hixie> jgraham: (which i could have sworn it did, i dunno why the spec doesn't say that) [14:25:52.0000] <jgraham> But I will see [14:25:57.0000] <Hixie> i am pretty sure you end either with N+2 entries or 1 :-) [14:26:03.0000] <Hixie> but not N+1 :-) [14:26:15.0000] <jgraham> Heh [14:26:22.0000] <Hixie> N+1 will be hard to spec, so i hope i'm right :-P [14:26:31.0000] <jgraham> Let's go with ">0" [14:30:28.0000] <smaug____> jgraham: whenever doing something with session history, test the implementations (which all do different things) and pick up the behavior which you like the best :) [14:34:32.0000] <jgraham> smaug____: That's what I am doing. I am also doing step 2) Avoid hypocritsy and make the world a better place by giving feedback on the spec, especially where we want to diverge from it [14:34:59.0000] <jgraham> *hypocrisy [14:35:33.0000] <smaug____> jgraham: wasn't someone going to spec session history [14:35:46.0000] <smaug____> I mean the parts which aren't spec'ed [14:36:00.0000] <smaug____> jgraham: someone from Opera, not you [14:36:31.0000] <TabAtkins> Hixie: http://www.xanthir.com/blog/b4JB0 [14:36:33.0000] <TabAtkins> 32 lines. [14:37:28.0000] <smaug____> jgraham: session history becomes very interesting when you start defining what should happen when DOM is modified [14:37:52.0000] <smaug____> (remove/add iframes, move them, etc) [14:38:29.0000] <jgraham> smaug____: When does DOM modification affect history? Apart from document.open() of course. Which is more "demolition" than "modification" [14:38:51.0000] <TabAtkins> As written the object never advertises that it has the async methods, even after it's loaded. That'll take another line to add them to 'near' and make it distinguish them properly. [14:39:16.0000] <jgraham> Well when I tested that last moving iframes always caused them to reload, so I would always expect an extra position in session history [14:39:31.0000] <smaug____> /me tries to find some tests [14:39:51.0000] <jgraham> The only exception was chrome when you added/removed a node into the same place in a single step [14:40:21.0000] <jgraham> When it badly optimised away the action and missed the expected side effects [14:41:54.0000] <smaug____> where are my tests.. [14:44:40.0000] <smaug____> jgraham: http://mozilla.pettay.fi/moztests/history2/Start.html [14:44:55.0000] <smaug____> that is a case which behaves quite differently in different browsers [14:45:04.0000] <smaug____> note, the comments about browser engines aren't valid anymore [14:45:24.0000] <smaug____> Gecko follows IE behavior nowadays, at least in most cases [14:45:45.0000] <smaug____> since IE has traditionally been the least broken [14:46:21.0000] <jgraham> /me frames that sentence and puts it on the wall [14:49:26.0000] <jgraham> smaug____: Interesting test, thanks [15:09:14.0000] <Hixie> TabAtkins: what is a Name? [15:10:04.0000] <TabAtkins> The syntax isn't decided yet, but the functionality is http://wiki.ecmascript.org/doku.php?id=strawman:private_names [15:10:08.0000] <TabAtkins> it's a gensym, basically. [15:10:24.0000] <TabAtkins> The way that JS finally gains private properties. [15:10:50.0000] <TabAtkins> Alternately, http://www.xanthir.com/blog/b4FJ0 [15:12:09.0000] <Hixie> ah [15:12:10.0000] <Hixie> weird [15:12:46.0000] <Hixie> and "..." is an in-place array expansion operator? [15:13:07.0000] <TabAtkins> Yeah. [15:13:28.0000] <Hixie> well this is certainly simpler than i feared [15:13:35.0000] <Hixie> but not as simple as i'd like [15:13:36.0000] <Hixie> :-) [15:14:18.0000] <Hixie> whenever i look at the new JS stuff I feel like it's invented by computer science researchers [15:14:29.0000] <Hixie> rather than being copied from mainstream programming languages [15:14:52.0000] <Hixie> i can't help but wonder how regular authors will deal with it [15:15:04.0000] <TabAtkins> This is largely because it is. Also because the ESWG prefers to add new primitives, and then come along later and see what easy-to-use APIs to standardize on top of them. [15:15:35.0000] <Hixie> seems kinda like the opposite of what we do for the rest of the web [15:15:43.0000] <TabAtkins> Kinda, yeah. [15:15:43.0000] <Hixie> where we try desperately not to innovate [15:15:47.0000] <Hixie> anyway [15:15:52.0000] <TabAtkins> Or, wait. [15:16:12.0000] <TabAtkins> The point is that libraries get to innovate on top of the new primitives, and then we come through and bless the best solution. [15:17:11.0000] <Hixie> what i don't understand is why the primitives and libraries have to be in JS rather than in other languages [15:17:20.0000] <Hixie> as in [15:17:32.0000] <Hixie> why don't we look at existing libraries and languages and copy the already known best solutions [15:17:54.0000] <Hixie> rather than making JS have arcane computer science theory primitives, as well as, later, higher-level solutions to some problems [15:18:20.0000] <TabAtkins> Because without the primitives, a lot of problems *can't* be solved in a reasonable way? [15:19:33.0000] <Hixie> in the abstract, i find that hard to refute [15:20:53.0000] <Hixie> but for example, why do we need to provide a way for a property to be sealed dynamically, rather than just having statically defined classes with properties that can't be changed dynamically anyway? [15:21:49.0000] <TabAtkins> I can't comment much on the desire for sealing properties. [15:22:10.0000] <TabAtkins> But I suspect that at least some of it is that getting WG agreement on a class syntax is much harder than WG agreement on the ability to seal. [15:22:22.0000] <TabAtkins> (Which is why JS *still* doesn't have a class syntax.) [15:23:05.0000] <Hixie> fricking committee-driven design [15:25:52.0000] <Hixie> imageSmoothing... should affect patterns, right? [15:26:09.0000] <TabAtkins> I think so. [15:27:10.0000] <TabAtkins> Patterns can be resized and whatnot, right? [15:27:12.0000] <TabAtkins> If so, yes. [15:27:39.0000] <Hixie> yeah [15:32:49.0000] <rafaelw_> hixie: reference for H4X? [15:33:07.0000] <Hixie> dunno offhand [15:33:17.0000] <Hixie> it's basically just e4x but for html [15:33:33.0000] <Hixie> i don't think anything formal has been written, since it never gets any traction [15:34:04.0000] <zewt> heh, the worst features of all languages seem to be where they get "creative" [15:34:58.0000] <rafaelw_> ok. that makes it hard for me to respond =-) [15:36:02.0000] <Hixie> just imagine e4x but without namespaces and without comment support. [15:36:18.0000] <Hixie> (and with some of the <svg>/<math> namespace magic, i guess) [15:36:59.0000] <Hixie> if it's something you think we can get traction on amongst implementors of js engines, i'm sure we can come up with an actual proposal that a js person could turn into a real spec [15:37:24.0000] <TabAtkins> Unrelated: all of a sudden, today all the IE people started sending mail in large blue text. [15:37:32.0000] <TabAtkins> Maybe they all got a corporate Outlook upgrade? [15:38:05.0000] <Hixie> rafaelw_: (oh, another change from e4x would be that the elements created using h4x actually be DOM nodes, not a parallel data structure like in e4x) [15:38:44.0000] <smaug____> I don't think that was the idea with e4x. Brendan just never implemented the DOM binding [15:39:49.0000] <rafaelw_> I guess my question is how would it avoid the context element problem? [15:41:16.0000] <Hixie> smaug____: e4x had a kind of implied casting, but there were definitely two different types of objects in play [15:41:22.0000] <Hixie> smaug____: even with that mapping that never got implemented [15:41:32.0000] <tantek> objects are overrated [15:41:34.0000] <Hixie> rafaelw_: just always require end tags and don't do any optional tags [15:41:46.0000] <Hixie> rafaelw_: so there's only one parse mode [15:42:03.0000] <Hixie> rafaelw_: (it doesn't have to match the HTML syntax, it just has to be at least as terse for most purposes) [15:42:15.0000] <rafaelw_> And it wouldn't do any fixup? [15:42:27.0000] <Hixie> right, any syntax errors would be JS syntax errors [15:42:28.0000] <rafaelw_> So "<option>Foo<option>Bar" throws an error? [15:42:30.0000] <Hixie> and wouldn't compile [15:42:35.0000] <Hixie> it doesn't throw, it just doesn't compile [15:42:46.0000] <Hixie> like "var = foo" [15:43:29.0000] <Hixie> rafaelw_: i think maybe i misrememberd the proposal name. try googling for e4h? [15:43:56.0000] <rafaelw_> The main problem I see with this is it means required page authors to learn another HTML which is different from the HTML they know. [15:44:26.0000] <Hixie> i don't think that's a huge problem, authors dealt fine with JSON e.g. [15:44:31.0000] <Hixie> which isn't quite the same as JS [15:44:31.0000] <smaug____> it wouldn't be that different, at least in most cases [15:45:03.0000] <Hixie> rafaelw_: http://lists.w3.org/Archives/Public/public-script-coord/2011OctDec/thread.html#msg65 [15:45:09.0000] <Hixie> E4H [15:46:09.0000] <Hixie> quasi-literals is the other thing people have proposed in this space [15:46:15.0000] <Hixie> which i think has a bit more traction over in JS land [15:46:47.0000] <Hixie> which isn't quite as cool imho but is more generic and, at least in some proposals, still gets you the syntax checking [15:48:14.0000] <wodemay> what's with these? why are y'all putting presentational elements back into html5? https://en.wikipedia.org/wiki/HTML_tag#Presentation What's the rationale? [15:48:38.0000] <Hixie> wodemay: ? [15:49:07.0000] <Hixie> wodemay: the premise of the question is incorrect, unless you mean <canvas>, i guess [15:49:36.0000] <wodemay> Hixie, <b>/<i>/<u> (i understand u, there's no semantic equivalent. but the other two...?) [15:49:53.0000] <Hixie> wodemay: please read the HTML spec, you'll find none of those are presentational [15:49:58.0000] <TabAtkins> The wikipedia page explains the semantic meaning of those elements. [15:50:01.0000] <TabAtkins> As does the HTML spec. [15:50:48.0000] <wodemay> TabAtkins, I have read both. But I find it a hard sell to believe that those are not presentational. [15:51:06.0000] <Hixie> what's your definition of "presentational"? [15:51:24.0000] <TabAtkins> Okay. Shrug, then. If the part where they say "these have meaning" doesn't mean "not presentational" to you, then I'm not sure what to say. [15:51:35.0000] <wodemay> as with <hr>, since it's explicitly horizontal. [15:51:44.0000] <TabAtkins> Huh? [15:51:46.0000] <wodemay> /me rereads the wiki and spec. [15:52:06.0000] <TabAtkins> <hr> will generate a vertical rule in vertical text. [15:52:37.0000] <rafaelw_> hixie: re: E4X & js quasi literals [15:52:42.0000] <Hixie> the definition of <hr> doesn't even mention the word "horizontal" [15:52:56.0000] <rafaelw_> Conceptually what E4X proposes are two complementary things: [15:53:01.0000] <Hixie> wodemay: make sure you're reading the latest spec, maybe you're looking at an out of date one. http://whatwg.org/html [15:53:10.0000] <rafaelw_> a) HTML literals in JS (so you get parse errors at load time) [15:53:15.0000] <wodemay> Hixie, okay, thanks. [15:53:18.0000] <rafaelw_> b) A new HTML parser [15:53:41.0000] <rafaelw_> Maybe it's worth separating these issues. [15:53:51.0000] <Hixie> rafaelw_: well it's not a new HTML parser, it's just a new set of literals in the JS parser [15:54:27.0000] <rafaelw_> So it wouldn't use any parser? [15:54:35.0000] <Hixie> it's JS [15:54:45.0000] <Hixie> it uses the same parser as parsing "function () { }" or "-35.23" [15:54:50.0000] <Hixie> same as E4X [15:55:05.0000] <rafaelw_> Would "<div><tr><td>Foo</td><tr></div>" be valid? [15:55:11.0000] <rafaelw_> Would "<div><tr><td>Foo</td></tr></div>" be valid? [15:55:33.0000] <Hixie> what do you mean by "valid" in this context? [15:55:42.0000] <rafaelw_> would it throw a parse error. [15:56:04.0000] <rafaelw_> is it "statically valid" [15:56:05.0000] <Hixie> the script var foo = <div><tr><td>Foo</td></tr></div>; would not throw a parse error, but would generate a DOM tree that is non-conforming to HTML. [15:56:16.0000] <Hixie> and assign it to the variable "foo" [15:56:19.0000] <wodemay> TabAtkins, sorry... I was a little confused and remembered somethiing different as the HTML5 subsections of those elements' sections on the wiki and was basing my memory of the official spec's explanations on that. But still, don't the tagnames derive their etymology from an earlier presentational incarnation? Shouldn't they get new names that are semantic to the core? [15:56:25.0000] <Hixie> or i should say, |foo| [15:56:29.0000] <rafaelw_> i see. [15:56:33.0000] <rafaelw_> what about [15:56:43.0000] <rafaelw_> var foo = <div>Foo</div><div>Bar</div> [15:57:03.0000] <TabAtkins> wodemay: Yes, they were presentational in old versions. But they were often used in a way that had a somewhat consistent meaning, which HTML5 now codifies. [15:57:29.0000] <Hixie> wodemay: the tag names are just opaque strings, where possible we reuse those that are already implemented because it means we get things for free (e.g. the default rendering of the "instance of term" element <i> is already italics in old browsers because those browsers think they mean "italics") [15:58:24.0000] <Hixie> rafaelw_: there's no concrete proposal, so it's hard to say. one could say that two HTML literals side by side result in an implied documentFragment, but personally I would prefer that we say that if you want a DF you have to wrap the elements in <>...</> [15:58:47.0000] <rafaelw_> i see. [15:59:09.0000] <rafaelw_> so it literally becomes a short-hand for equivalent calls to createElement, appendChild, setAttribute, etc... [15:59:15.0000] <Hixie> yeah [15:59:18.0000] <TabAtkins> Yeah, that's the easiest way to think of it. [16:00:18.0000] <smaug____> not so fun part would be that workers shouldn't allow that syntax [16:00:26.0000] <wodemay> Hixie, but isn't part of the html5 concept that webpages coded in the 90s will still work on browser versions released in 2020 anyway? Doesn't said concept cover your scenario? Can't HTML5-thespec name them something diff? [16:00:44.0000] <TabAtkins> wodemay: What's the value in naming them differently, then? [16:00:56.0000] <zewt> (<i> does mean italics, because that's how it renders and that's what everyone uses it for; you can wave your arms around fiercely trying to say it means something else, but that's what it means :) [16:01:07.0000] <rafaelw_> well. i can tell you that the ship has sailed on new features for ES6. [16:01:24.0000] <rafaelw_> JS quasi's are approved in proposal. [16:01:54.0000] <rafaelw_> presumably you could define that HTML quasis parse this way, but i don't think you'd get load-time parse errors like you want. [16:02:16.0000] <rafaelw_> in any case, you could look at what you're describing as an alternate proposal for the behavior of Document.parse(). [16:02:18.0000] <rafaelw_> no? [16:02:49.0000] <wodemay> TabAtkins, Let's say someone is born 2015 and begins learning to code HTML5 docs in the year 2030. There shouldn't be any remnants of the stone ages of mixing pres and content. They shouldn't ever *think* of tags as having presentational meaning. Naming them differently will prevent them from getting misused. [16:03:00.0000] <rafaelw_> who is parsing doesn't matter to authors. [16:03:14.0000] <Hixie> smaug____: yeah, we'd have to define some solution to that [16:03:26.0000] <TabAtkins> wodemay: That seems gratuitous. There's nothing wrong with the name, and in a lot of cases, the legacy uses of it are consistent with the new semantics. [16:03:48.0000] <Hixie> smaug____: long term we need some solution to the DOM in workers anyway, even if it's just some nerfed versions of the objects [16:03:48.0000] <wodemay> its like the debate about @hidden vs. @irrelevant. [16:04:01.0000] <Hixie> wodemay: well, <i> still works, right? :-) [16:04:12.0000] <wodemay> Hixie, your point? [16:04:14.0000] <smaug____> I do think DOM will just work in workers [16:04:55.0000] <Hixie> wodemay: nameing them differently doesn't prevent them from being misused... just look at people using <blockquote> for indentation :-) [16:05:52.0000] <wodemay> Hixie, but accurate naming does have value (again, i'll cite the @hidden vs. @irrelevant debate) [16:07:16.0000] <TabAtkins> Hixie: Did you say somewhere that <intent> is okay to be a self-closing tag in the head? [16:07:28.0000] <Hixie> wodemay: certainly, but that value must be balanced against other concerns, e.g. backwards compatibility [16:07:45.0000] <Hixie> TabAtkins: my latest thinking on web intents is in a recent e-mail to the whatwg list, i do not recall my precise statements therein [16:07:55.0000] <Hixie> TabAtkins: though i intend to look at that topic again soon [16:07:57.0000] <TabAtkins> Darn, that means I"ll have to look it up. [16:08:11.0000] <Hixie> heh sorry :-) [16:09:35.0000] <wodemay> Anyone know when ECMA6 is meant to become widely implemented? [16:10:12.0000] <TabAtkins> Like HTML, it's implemented in bits and pieces. [16:10:14.0000] <ojan> Hixie: i think at some level the rub is what you svg/mathml magic you do [16:10:34.0000] <wodemay> How about `multi [16:10:36.0000] <wodemay> line [16:10:42.0000] <ojan> Hixie: i think also that lots of people will want <p>, <li>, <td> without a close tag to work [16:10:42.0000] <wodemay> string-literals`? [16:11:04.0000] <ojan> Hixie: other than that, i'm ok with your proposal as an alternative to implied context parsing [16:11:41.0000] <ojan> Hixie: wait are you saying also that things like <img> will require a close tag? [16:12:14.0000] <Hixie> ojan: i would just have the syntax support <foo/> [16:12:19.0000] <gsnedders> wodemay: I'd guess sometime in the next year and a half. [16:12:29.0000] <ojan> Hixie: ok...so it's basically xml? [16:12:30.0000] <Hixie> ojan: we could hardcode some specific tag names if people think that's necessary [16:12:35.0000] <Hixie> ojan: yeah [16:12:36.0000] <wodemay> gsnedders, but it's already in chromium/v8... right? [16:12:42.0000] <Hixie> ojan: like i said, it's very close to e4x [16:13:01.0000] <ojan> Hixie: would you be opposed to <p>, <li>, etc without a close tag? [16:13:11.0000] <Hixie> ojan: not especially, if it helped sell the feature :-) [16:13:16.0000] <gsnedders> wodemay: Not really. Both SM and V8 implement relatively small parts of ES6 [16:13:17.0000] <ojan> lol, ok [16:13:21.0000] <Hixie> ojan: wouldn't be my first choice though :-) [16:13:29.0000] <ojan> Hixie: just trying to understand what you're proposing... [16:13:44.0000] <gsnedders> IE10 now supports at least Flash in Metro :( [16:13:51.0000] <wodemay> gsnedders, i was referring specifically to multiline string literal backtick syntax. [16:15:02.0000] <gsnedders> wodemay: Oh, that I don't know off-hand. [16:15:03.0000] <Hixie> ojan: making omitting an end tag work would be... interesting, because you end up having to hard-code into the JS parser the list of elements that imply the </p> [16:15:13.0000] <gsnedders> Please don't do that. [16:15:14.0000] <roc_> I'm a wee bit frustrated to post to the list saying "we're changing our behavior in Gecko and the spec needs to be updated" and get replies from other implementors saying "oh, we already changed that ages ago" [16:15:52.0000] <Hixie> roc_: agreed (and my apologies if the reason is that i'm not keeping the spec up to date fast enough) [16:16:10.0000] <roc_> I don't think so in this case :-) [16:16:50.0000] <gsnedders> Nor did jgraham when he mentioned it earlier :( [16:17:41.0000] <wodemay> rafaelw_, BTW, the "ship" isn't ever intended to "sail," so to speak, for new additions or features to HTML5, correct? [16:18:18.0000] <Hixie> wodemay: the ship sails when implementations ship [16:18:35.0000] <Hixie> boy that sentence could be clearer [16:18:36.0000] <smaug____> roc_: which case is this? [16:18:49.0000] <wodemay> Hixie, but that's a very regular sailing. It isn't a singular ship that sails and then never comes back. It's more like a ferry, right? [16:19:04.0000] <Hixie> wodemay: not sure a shipping metaphor is very apt to be honest :-) [16:19:11.0000] <smaug____> ah, some video thing [16:19:23.0000] <Hixie> wodemay: to put it in non-metaphorical terms, we can change the spec in any way that is compatible with the web [16:19:26.0000] <wodemay> hahah. but u follow me? Hixie [16:19:32.0000] <Hixie> wodemay: at any time [16:19:34.0000] <wodemay> Hixie, what does that mean? [16:19:53.0000] <Hixie> wodemay: e.g. we can't rename <p> to <b>, because that would cause zillions of pages to render differnetly than expected [16:20:06.0000] <jamesr_> they'd be upside down [16:20:18.0000] <Hixie> wodemay: but we can introduce a new element <askfjya78e6ads> because there is no page in the world that depends on it being implemented or not [16:20:30.0000] <Hixie> (i assume) [16:20:33.0000] <Hixie> (we'd have to check) [16:20:34.0000] <wodemay> so 20 years from now the current spec will still be some iteration of html5, but it might have some tags deprecated and some added, correct? and then the same will be true as compared to then 30 years from now? [16:21:00.0000] <Hixie> wodemay: well we don't call it "html5", we call it "html", but yes, essentially. until someone comes along with something that makes html obsolete once and for all. [16:21:08.0000] <wodemay> Hixie, hahah, talk about being thorough. ;) somehow i feel like there aren't! [16:21:13.0000] <Hixie> which i'm sure everyone here is hoping will happen as soon as possible [16:21:30.0000] <wodemay> Hixie, hahahahaha. me included! [16:21:55.0000] <wodemay> There should be something like JSON that makes it less redundant. or sthg that draws on HAML syntax. [16:22:07.0000] <wodemay> but why not call it "html5", Hixie ? [16:22:15.0000] <roc_> the thing is, people keep trying to kill HTML and we keep trying to stop them [16:23:28.0000] <Hixie> roc_: whose trying to stop them? [16:23:30.0000] <roc_> generally because the replacement is worse in some important dimension [16:23:45.0000] <roc_> Mozilla, other browser vendors, Web standards people [16:23:50.0000] <Hixie> roc_: people keep trying to kill html and they keep failing, i don't think we're stopping them [16:24:09.0000] <roc_> we're killing Flash [16:24:15.0000] <Hixie> adobe is killing flash [16:24:37.0000] <kennyluck> Who are trying to kill HTML? [16:24:49.0000] <roc_> Microsoft and Adobe tried [16:24:50.0000] <Hixie> wodemay: the question is not "why not", the question is "why" :-) [16:25:07.0000] <smaug____> W3C tried [16:25:11.0000] <wodemay> the answer in that case is "b/c it's more specific," Hixie [16:25:24.0000] <Hixie> wodemay: specific how? [16:25:26.0000] <roc_> hehe [16:26:54.0000] <wodemay> Hixie, it specifies the version number ..? [16:27:16.0000] <Hixie> roc_: i think the metaphor that applies here is something like "several people are building towers out of blocks. the people building the tower not labeled 'html' keep knocking down their towers by mistake or poor design. the tower labeled 'html' is a big heap and so doesn't fall." [16:27:23.0000] <wodemay> smaug____, what's the difference between WHATWG and W3C? [16:27:32.0000] <Hixie> roc_: it's not like the people building the html tower keep going over and knocking down the other towers [16:27:34.0000] <roc_> that is a good analogy [16:27:35.0000] <gsnedders> wodemay: But browsers never implement all of one version — they just implement bits and pieces. [16:27:54.0000] <gsnedders> wodemay: Version numbers only make sense if the spec gets implemented in full, but it never does. [16:28:10.0000] <Hixie> wodemay: there is no version. the spec changes daily, the browsers implement different slices of it at any one time and also change daily. [16:28:39.0000] <wodemay> is that why the html5 doctype dec is html and not html5? [16:28:50.0000] <roc_> that's not really what "kill" means in this context. No-one ever kills a competitor's product directly, only by out-competing it. And that's what we've done with HTML. [16:30:43.0000] <ojan> Hixie: what would you say the advantage of E4H is compared to implied context parsing? [16:30:54.0000] <ojan> Hixie: is it just that it's simpler for web developers? faster? [16:31:39.0000] <gsnedders> ojan: Harder to screw up. People notice syntax errors in JS. [16:31:50.0000] <Hixie> ojan: by "implied context parsing" you mean "runtime parsing of a string"? if so, the main advantage is compile-time syntax checking. [16:32:56.0000] <Hixie> roc_: i think there's a qualitative difference between out-competing by intentionally addressing a competitor's weaknesses and explicitly having as a goal the competitor's downfall, and out-competing by ignoring the competition and having them fall by the wayside [16:33:04.0000] <Hixie> roc_: and i would strongly suggest that HTML has done only the latter. [16:33:34.0000] <Hixie> ojan, rafaelw_: here's a strawman proposal: http://www.hixie.ch/specs/e4h/strawman [16:33:36.0000] <ojan> Hixie: oic...you wouldn't get syntax erros with the current quasis proposal either :( [16:33:42.0000] <Hixie> ojan, rafaelw_: (doesn't handle SVG yet) [16:33:44.0000] <astearns> <video> completely ignored everyone else, I'm sure :) [16:33:58.0000] <Hixie> ojan: oh, really? that's sad. i thought brendan said we would, based on the thread i cited earlier. [16:34:38.0000] <ojan> Hixie: i might be misreading http://wiki.ecmascript.org/doku.php?id=harmony:quasis [16:34:51.0000] <Hixie> astearns: <video> on HTML has basically failed, so i don't think it's a good example if you're trying to support roc's argument :-) [16:35:15.0000] <Hixie> astearns: (failed due to the codec gridlock; i don't think it's a permanent failure) [16:35:25.0000] <astearns> I hope it's not permanent [16:35:38.0000] <TabAtkins> It won't be. [16:35:41.0000] <Hixie> well worst case it'll resolve itself when the relevant patents expire [16:36:16.0000] <astearns> but whether something has succeeded or failed isn't relevant to whether it's a counterexample to "only the latter" [16:36:20.0000] <roc_> it'll resolve itself at the latest when improvements in video compression reach diminishing returns [16:36:30.0000] <roc_> but who knows? That could be 50 years away [16:36:37.0000] <rafaelw_> hixie: looking [16:36:49.0000] <zewt> roc: don't forget to add twenty years for patents to expire [16:37:28.0000] <Hixie> astearns: granted, but at least for me, adding <video> was just about filling in HTML's weaknesses, not trying to do something to hurt another platform [16:38:09.0000] <Hixie> astearns: e.g. if we wanted to hurt Flash we could have done things like make it click-to-play, or made it harder to sandbox, or made it not be able to see the DOM, or any number of things [16:38:26.0000] <Hixie> astearns: instead, alongside work on HTML there has been much work from browser vendors in actually making Flash work better [16:38:49.0000] <Hixie> hardly the sign of trying to damage a competing platform [16:39:12.0000] <Hixie> even Apple, who arguably did try to damage Flash on iOS, seem to be working to help Flash on MacOS with their sandboxing [16:39:21.0000] <astearns> I agree, but you're moving the goalposts from "ignoring" to "not intending to harm" [16:39:36.0000] <wodemay> Hixie, what does the term "HTML5" specify, then? [16:39:38.0000] <roc_> I think it's fair to say that a number of people have had a goal of enriching HTML to make Flash unnecessary, with the hope it will go away eventually [16:40:00.0000] <Hixie> astearns: "ignoring" was a poor choice of words; true. we do look at existing practices in other platforms. [16:40:08.0000] <roc_> on the grounds that we want authors to target multi-vendor royalty-free standards instead of single-vendor platforms. [16:40:22.0000] <Hixie> roc_: yeah, that's probably fair [16:41:14.0000] <Hixie> roc_: but i don't think that's the main reason flash has done poorly [16:42:29.0000] <zewt> bryan of the screaming caps name posting to a thread about mailing list practices is rather ironic [16:43:44.0000] <Hixie> ojan, rafaelw_: ok, added a strawman way of supporting svg and mathml too (requires that the element names be prefixed if they're in those namespaces) [16:44:45.0000] <Hixie> ojan:, rafaelw_: an alternative would be to only require that for the element that crosses to another namespace, and default the rest to the parent element's namespace [16:45:02.0000] <Hixie> ojan, rafaelw_: but that's a minor detail that could be figured out if this was to go anywhere [16:48:32.0000] <ojan> Hixie: i'm torn here. on the one hand, compile-time errors are nice. on the other, there's a cost in people needing to learn new things and there will perpetually be confusion about parsing that works in E4H, but not HTML or vice versa. [16:48:45.0000] <ojan> Hixie: at some level, how is this different from pushing for XHTML? [16:49:45.0000] <ojan> Hixie: also, I really like the API for quasis as a way of generating a DOM using javascript variables in an XSS-safe way [16:50:01.0000] <ojan> Hixie: see the safehtml function in the quasi's proposal to see what i mean [16:50:12.0000] <ojan> Hixie: and you don't get that with E4H, unless i'm missing something [16:51:29.0000] <zewt> ojan: compile-time errors aren't very nice; it makes code backwards-compat/polyfill/etc very hard [16:52:51.0000] <zewt> i don't want to have to serve entirely different scripts for different browser capabilities [16:53:50.0000] <gsnedders> zewt: That's true of any langauge extention, though [16:53:55.0000] <gsnedders> *language [16:54:04.0000] <Hixie> ojan: the substitution thing lets you sub in the value of expressions in e4h [16:54:25.0000] <zewt> gsnedders: sure, which is why language extensions shouldn't be used unless the justification is very strong [16:54:51.0000] <Hixie> ojan: but in general my argument isn't that we should do e4h, my argument is that hte use case of "make a dom tree" should be solved using a better solution than "parse a string at runtime" and a better solution than "use an excessively verbose api" [16:55:25.0000] <ojan> Hixie: well...the proposal with backing so far has been quasis + safehtml [16:55:35.0000] <Hixie> zewt: no, it's why the language should be designed to make syntax errors block-scopable, imho :-) [16:55:42.0000] <ojan> Hixie: i suppose you could make safehtml do roughly what e4h does [16:55:49.0000] <ojan> Hixie: while still using quasis [16:56:20.0000] <Hixie> ojan: i'm not very familiar with it, do you have a link? the only proposal i was aware of was "make the html parser even more complicated and add DocumentFragment.innerHTML" (or document.parse(), same idea) [16:56:25.0000] <ojan> Hixie: but you would still get a runtime erro instead of compile error [16:56:29.0000] <ojan> Hixie: i linked above [16:56:37.0000] <Hixie> got it [16:58:25.0000] <Hixie> ojan: so this just gets implemented as a JS lib? not in the browsers? [16:58:31.0000] <Hixie> ojan: the html-specific part, i mean [16:58:53.0000] <ojan> Hixie: it can be, but in the case of the html-specific part, the browser would also provide it builtin [16:59:24.0000] <ojan> Hixie: so, pages could polyfill w a JS function [16:59:38.0000] <Hixie> (btw, i don't really understand how the filtering is done, seems like it'd be a huge rathole of possible ways things should be escaped) [16:59:45.0000] <Hixie> so how does the HTML parsing happen here? [16:59:48.0000] <ojan> Hixie: for browsers that support quasies, but not the safehtml functio that is.