2009-06-01 [17:15:00.0000] hober: Sensible stuff [17:31:00.0000] hober: nice. (hi) [19:07:00.0000] Just making sure I'm not crazy: If I have

, the  node gets placed *outside* of, correct? 

[19:27:00.0000] Should active formatting elements list every contain a single element which is null? 

[19:27:01.0000] *ever 

[19:30:00.0000] Oh, that's their internal representation of Marker 

[19:48:00.0000] Where in the spec does it specify that, after I encounter the  tag, I'm supposed to somehow switch back into IN_FOREIGN_CONTENT? 

[19:50:00.0000] It looks like the spec makes a half-hearted attempt to keep the mode in "in foreign content", but as soon as we have some sort of complicated code inside that, we move to another mode and consequently lose the ability to "go back" 

[19:58:00.0000] Also, assuming that this is a spec error (I'd put it at an 80% chance it's not, but I'm hoping it is), PHP html5lib now has 100% tests passing 

[19:59:00.0000] erm, s/tests/specced tests/ 

[19:59:01.0000] And has, maybe, three nasty workarounds (one for empty doctypes, and two for namespaced attributes) 

[00:29:00.0000] ezyang: The table taint thing is the only case where the tests deviated from current spec intentionally because of me

[00:30:00.0000] hsivonen: you can flip a setting in ie8 to disable the xss filter

[00:30:01.0000] zcorpan_: where?

[00:33:00.0000] hsivonen: internet settings, security, custom level, scripting, enable xss filter

[00:34:00.0000] zcorpan_: you still around?

[00:34:01.0000] MikeSmith: yep

[00:35:00.0000] about "If the area element has no href attribute, then the area represented by the element cannot be selected, and the alt attribute must be omitted."

[00:35:01.0000] does v.nu not already handle that correctly?

[00:35:02.0000] /me wonders why the labels are "(o) Disable-ON ( ) Enable-OFF"

[00:36:00.0000] MikeSmith: dunno, that was re the spec bug you filed

[00:36:01.0000] oh

[00:36:02.0000] remind me which bug that was?

[00:37:00.0000] http://www.w3.org/Bugs/Public/show_bug.cgi?id=6976

[00:38:00.0000] ah

[00:38:01.0000] I didn't file that one, it just got auto-assigned to me

[00:38:02.0000] oh

[00:38:03.0000] because dude filed it against the h:tml draft

[00:39:00.0000] anyway, I'll reply to it and close it out

[00:39:01.0000] i didn't look carefully enough at the labels

[00:40:00.0000] np

[00:40:01.0000] it did make me think some about how v.nu error-reporting could be improved for this case

[00:41:00.0000] because right now, v.nu just says, "Error: Required attributes missing on element area." for cases like this

[00:41:01.0000] MikeSmith: hack jing to emit more useful messages? :)

[00:42:00.0000] well, I think hsivonen's already done some improvements that way

[00:42:01.0000] I was thinking about possibility adding it in the separate-pass assertions-checking

[00:42:02.0000] the assertions.sch Schematron file and the Assertions.java code

[00:43:00.0000] could have that specifically report, "An area element without an href attribute must not have an alt attribute."

[00:43:01.0000] or whatever

[00:43:02.0000] as another message?

[00:44:00.0000] yeah, it would need to be another message

[00:44:01.0000] if we did it that way

[00:44:02.0000] (as opposed to hacking jing)

[00:45:00.0000] i think it would be a better long-term solution to hack jing

[00:45:01.0000] zcorpan_: thanks. not the most discoverable thing

[00:45:02.0000] but then again i'm not the one who's hacking

[00:46:00.0000] hsivonen: indeed, ie's preferences are even worse than opera's :)

[00:46:01.0000] at least opera has opera:config

[00:46:02.0000] maybe I should put software.hixie.ch into my "intranet" and set the pref only for that

[00:47:00.0000] hmm. but then the compat mode thing could do intranet weirdness and confuse my live DOM testing

[00:48:00.0000] hsivonen: you could make the compat mode not depend on intranet, no?

[00:48:01.0000] yes

[00:49:00.0000]

element with logo but this outliner (http://gsnedders.html5.org/outliner/) doesn't show the alt text in the outline. Do you think it would be possible to make it reads the alt text and show it if no other text is present? [01:44:00.0000] Mrmil: zcorpan_ was trying to get me to implement that :P [01:45:00.0000] i think Hixie is wasting people's time with making , and do something other than being ignored [01:46:00.0000] gsnedders: Hehehe, did he succeed? :) [01:46:01.0000] Mrmil: I guess I probably will [01:46:02.0000] (get around to doing it) [01:47:00.0000] gsnedders: Ok. grins [01:48:00.0000] zcorpan_, he's right about whitespace roundtripping though [01:48:01.0000] zcorpan_, but supposedly editors have to deal differently with whitespace anyway [01:49:00.0000] zcorpan_, though that would indicate the current parsing algorithm is not suitable for editors [01:49:01.0000] i guess editors would want to preserve whitespace between attributes too [01:49:02.0000] yeah [01:49:03.0000] which means that the dom is not a suitable model for editors [01:50:00.0000] Hmm, Sims 3's first release date is tomorrow [01:50:01.0000] I think Dreamweaver preserves all sorts of crazy parse errors [01:50:02.0000] the suitability of the parsing algorithm depends on the kind of editor [01:50:03.0000] if you want to make a GUI-only editor, it's fine [01:51:00.0000] yeah i guess you'd want to preserve weird doctypes and bogus comments etc too [01:51:01.0000] if you want to make a dual GUI/text editor like Dreamweaver, you are in the world of hurt and complexity [01:55:00.0000] man, jd really has his job cut out for him when his customers go to lengths like these: http://www.reddit.com/r/programming/comments/8opdq/html_5_and_web_video_freeing_rich_media_from/c09yhr2 [02:02:00.0000] /me wonders when accessibility enters the Flash v. HTML5 comparisons with more force [02:02:01.0000] Hixie: his new story appears to be that HTML is just the next Silverlight [02:02:02.0000] How's Flash accessibility doing on Mac and Linux these days? [02:02:03.0000] othermaciej: wow, that's quite the compliment he's giving us [02:03:00.0000] hsivonen: that comparison is why I try to discourage people from citing Bespin as an HTML5 poster child... [02:03:01.0000] bespin is a poster child for us providing syntax highlighting :-) [02:03:02.0000] <Hixie> maybe in the next version... [02:03:03.0000] <Hixie> (i mean seriously, they must REALLY want that feature to do what they did in bespin) [02:04:00.0000] <Hixie> (it's a technological tour de force) [02:09:00.0000] <annevk42> Hixie, I've the feeling that's not enough [02:10:00.0000] <annevk42> Hixie, how would you e.g. implement code-auto-completion? [02:10:01.0000] <Hixie> we'd probably need that too, yes [02:11:00.0000] <Hixie> maybe textarea needs to be defined as an xbl2 binding to contenteditable content [02:13:00.0000] <hsivonen> there's also the issue that the DOM is rather heavy-weight as the model of a text editor view [02:13:01.0000] <othermaciej> so far I've had people tell me I should add more disclaimers to the design principles, I should either make them stated absolutely without disclaimers or delete them, or that I should state them without disclaimers but with an unstated expectation that they are not actually absolute [02:13:02.0000] <othermaciej> rarely have I seen a more clear example of the You Can't Please Everybody principle [02:14:00.0000] <hsivonen> maybe we should document that We Can't Please Everybody [02:14:01.0000] <annevk42> maybe we should have a principle: "Design Principles are no Magic Wand" [02:14:02.0000] <othermaciej> hsivonen: but what if there is a case where you can please everybody? [02:15:00.0000] <hsivonen> othermaciej: these absolutes are hard [02:16:00.0000] <othermaciej> I also can't believe that "Tables" was cited as an example of a good design principle [02:17:00.0000] <MikeSmith> I think the principle should be stated, For Any Given Proposal|Decision, We Will Have at Least One Jackass Formally Object to It, Verbosely [02:18:00.0000] <annevk42> /me wonders how bad the fallout will be from Hixie's latest email [02:19:00.0000] <othermaciej> MikeSmith: I think we may have actually had some decisions that occurred without a Formal Objection [02:19:01.0000] <othermaciej> can't remember though [02:20:00.0000] <MikeSmith> I think if we had a proposed decision to publish the statement, "It's very likely that the sun will rise tomorrow morning.", we'd have objections . [02:20:01.0000] <othermaciej> IE is down to 65.59% [02:21:00.0000] <othermaciej> that puts it at down 2.56% for the year to date [02:21:01.0000] <othermaciej> at this rate it will still take years to get under 50% [02:22:00.0000] <jgraham> othree: In any case "You can't please everybody" would be rejected on the grounds that it was just a thinly veiled excuse not to care about accessibility [02:23:00.0000] <jgraham> Or rather people would complain about it citing that as a reason [02:23:01.0000] <Hixie> i care about accessibility [02:23:02.0000] <Hixie> it's self-styled accessibility experts i don't care about :-) [02:23:03.0000] <jgraham> Hixie: I care about accessibility too [02:24:00.0000] <Hixie> of course many of them don't seem to be able to distinguish the two [02:25:00.0000] <othermaciej> I'm not generally in favor of self-styled experts of any stripe [02:25:01.0000] <jgraham> The point was that given a self-evident principle like "you can't please everybody", I predict that we would get complaints that the principle could be used as an excuse to preferentially exclude the disabled [02:25:02.0000] <othermaciej> I prefer the kind that can do without styling themselves [02:25:03.0000] <jgraham> Even though that would, of course, not be the case [02:25:04.0000] <Hixie> people already claim that [02:26:00.0000] <Hixie> so i don't think that would be a new problem [02:26:01.0000] <othermaciej> nowadays people seem more concerned with us excluding critical non-browser HTML implementations [02:26:02.0000] <othermaciej> such as libwww-perl and refrigerators [02:26:03.0000] <gavin_> I don't think jgraham was tying to claim that it would be anew problem :) [02:27:00.0000] <jgraham> Hixie: Sure. I was just pointing out how even the most obvious statement of compromise would be bent by those determined to see HTML5/HATWG/hatever in a bad light [02:27:01.0000] <Hixie> well sure [02:27:02.0000] <Hixie> but as you say [02:27:03.0000] <Hixie> even the most obvious statement would be taken that way [02:27:04.0000] <Hixie> so i wouldn't concern myself with that [02:28:00.0000] <jgraham> wwwww <- please insert any extra w's from this stash [02:29:00.0000] <othermaciej> jgraham: did you break your w? [02:31:00.0000] <abarth> i'm trying to think of some way to direct this discussion so that I can be quoted by Mr. Last Week [02:32:00.0000] <Hixie> you have such low ambitions :-P [02:32:01.0000] <Hixie> btw, thanks to plh's clarifications, i'm glad to say that html5 is now a standard (in development), rather than just a recommendation [02:32:02.0000] <abarth> Hixie: easy for you to say. you get quoted by him all the time :) [02:32:03.0000] <Hixie> at least on the whatwg site [02:32:04.0000] <othermaciej> abarth: your line above might be enough! [02:33:00.0000] <othermaciej> mpilgrim got cited so you can score without your name being on his enemies list apparently [02:33:01.0000] <Hixie> abarth: you just have to compliment easily offended people using long words that he won't understand [02:33:02.0000] <abarth> haha [02:34:00.0000] <Hixie> abarth: or, say things that are insults but are stated in ways that make the insult sound like they are the most obvious thing in the world to you [02:34:01.0000] <annevk42> othermaciej, I think mpilgrim made it to his list [02:34:02.0000] <Hixie> abarth: (e.g. implying that mr last week doesn't understand long words) [02:34:03.0000] <annevk42> considering what he said about him anyway [02:34:04.0000] <othermaciej> annevk42: he has now I guess [02:35:00.0000] <othermaciej> but he wasn't on the canonical list of Enemies of the People [02:35:01.0000] <Hixie> there's an enemies list? [02:35:02.0000] <abarth> I think Mr. Last Week is secretly a fan of Hixie and is trying to discredit the opposition by being rediculous. [02:35:03.0000] <annevk42> not an actual published list [02:36:00.0000] <othermaciej> ridiculous [02:36:01.0000] <gsnedders> abarth: I think we need to undermine the IETF HTTPbis WG by saying you shouldn't concat. headers with ", " :) [02:36:02.0000] <othermaciej> (spelling pet peeve) [02:36:03.0000] <othermaciej> annevk42: remember his TPAC checklist? [02:36:04.0000] <othermaciej> abarth: it's hard to tell! [02:36:05.0000] <gsnedders> /me notes annevk42 signed that and othermaciej didn't :P [02:36:06.0000] <othermaciej> abarth: his "satirical" posts sometimes make the opposition sounds dumb [02:36:07.0000] <annevk42> othermaciej wasn't there... [02:36:08.0000] <gsnedders> annevk42: I know. [02:37:00.0000] <gsnedders> http://www.flickr.com/photos/gsnedders/2979057018/in/set-72157608267037421/ [02:37:01.0000] <othermaciej> abarth: but the outright profane ones, or postings of the most unflattering photos of Hixie he could find on the internet, those seem sincerely mean-spirited [02:37:02.0000] <abarth> gsnedders: I don't really understand the HTTPbis WG. It's unclear to me how to help out there. [02:37:03.0000] <othermaciej> gsnedders: I'll get you at this year's TPAC if it ends up happening and people go [02:37:04.0000] <othermaciej> (I gather the fee is controversial...) [02:37:05.0000] <Hixie> in all fairness, i'm proud of that vampire photo [02:38:00.0000] <Hixie> it was taken at a party at hyatt's when i was an intern at netscape [02:38:01.0000] <gsnedders> othermaciej: I'm _probably_ not going. [02:38:02.0000] <gsnedders> /me wonders [02:38:03.0000] <gsnedders> Oh, that vampire photo [02:38:04.0000] <othermaciej> plenty o' cred, then [02:38:05.0000] <gsnedders> abarth: Nor do I understand them. [02:39:00.0000] <othermaciej> abarth: they seem more interested in spitting the RFC into a dozen pieces than in actually fixing anything [02:40:00.0000] <abarth> Hixie: how far back is the vampire picture? i've been scrolling through but don't see it yet [02:41:00.0000] <Hixie> i dunno, i don't actually read lastweek so i've no idea how many posts he's written [02:41:01.0000] <Hixie> (i have far better things to do with my life!) [02:41:02.0000] <Hixie> such as today, going to maker faire and then seeing wicked, which, by the way, is pretty good as a musical [02:41:03.0000] <Hixie> though i prefer into the woods [02:41:04.0000] <Hixie> (and avenue q, of course) [02:41:05.0000] <Hixie> (though anyone who's read the spec realises this) [02:42:00.0000] <gsnedders> Oh, I need to get ticket to go and see Avenue Q/ [02:42:01.0000] <gsnedders> /me is reminded :P [02:42:02.0000] <abarth> i missed avenue q, but i did make it to wicked [02:42:03.0000] <Hixie> man when i start writing examples later this year, i hope i don't run out of quotes from these musicals to put in the spec [02:42:04.0000] <Hixie> abarth: did you see the one in SF? [02:42:05.0000] <abarth> Hixie: yes [02:42:06.0000] <Hixie> abarth: i was _really_ impressed by the production value there [02:42:07.0000] <Hixie> holy crap [02:43:00.0000] <Hixie> that's some serious set dec [02:43:01.0000] <gsnedders> Hixie: Do you know anything about the London production? [02:43:02.0000] <othermaciej> I saw Up today [02:43:03.0000] <abarth> Hixie: Yeah, I always worry when I see a play outside of NY or London that I'm missing out, but the production was really good [02:43:04.0000] <othermaciej> I rarely look up MLW unless someone mentions it here [02:43:05.0000] <Hixie> up was pretty fantastic too [02:43:06.0000] <Hixie> though that's to be expected [02:44:00.0000] <othermaciej> but I am flattered that he adopted the name I assigned him [02:44:01.0000] <Hixie> what with it being the movie of the year and all [02:44:02.0000] <Hixie> though i may have to update the example in the spec [02:44:03.0000] <othermaciej> (I believe I coined the term "Mr. Last Week" back in the day) [02:44:04.0000] <Hixie> because up comes before a number of the other movies in the list [02:44:05.0000] <gsnedders> Hixie: what example? [02:44:06.0000] <Hixie> gsnedders: no idea (re london) [02:45:00.0000] <othermaciej> Up was surprisingly not-for-little-kids [02:45:01.0000] <gsnedders> Hixie: Damnit, I was counting on you! :P [02:45:02.0000] <gsnedders> Ah, Up doesn't come out here until October :( [02:45:03.0000] <gsnedders> /me always hates the fact that Pixar films don't come out here for months [02:45:04.0000] <Hixie> othermaciej: yeah, the trailers had me worried that it was going to be dreamworks-like, but woah was that a mistaken worry [02:46:00.0000] <othermaciej> Dreamworks is not in the same league as Pixar [02:46:01.0000] <Hixie> othermaciej: i guess i should have more faith in ye people of emeryville [02:46:02.0000] <gsnedders> Hixie: "The top 10 movies of all time"? [02:46:03.0000] <Hixie> gsnedders: no comment [02:46:04.0000] <gsnedders> :P [02:46:05.0000] <Hixie> and re the release date, i guess it's because they have to do all the subtitling and so on [02:46:06.0000] <Hixie> /me ducks [02:47:00.0000] <othermaciej> I say that having recently seen "Monsters vs. Aliens" and moderately enjoyed it [02:47:01.0000] <gsnedders> /me slaps Hixie [02:47:02.0000] <othermaciej> so I wonder if John Lasseter really has full creative control of all Disney animation now [02:47:03.0000] <othermaciej> because apparently they have a new cel-animated feature coming up [02:47:04.0000] <othermaciej> and I am wondering whether I should expect it to be any good [02:48:00.0000] <gsnedders> Hixie: Actually, Cars at least did in parts have a different audio track (and script, and hence animation) of the character's voices at parts here! [02:48:01.0000] <Hixie> wow [02:48:02.0000] <Hixie> i know monsters, inc did [02:48:03.0000] <Hixie> the non-us version had different graphics in a couple of places [02:49:00.0000] <Hixie> e.g. the "walk/don't walk" sign is symbols, iirc [02:49:01.0000] <Hixie> and the alert in the code 21 or whatever it's called was a different symbol [02:49:02.0000] <gsnedders> /me only got around see seeing Ratatouille in English over Christmas, having been bullied by French relatives into seeing it in Paris in French when it came out :P [02:49:03.0000] <Hixie> ew! ewewew! [02:49:04.0000] <Hixie> dubbing! [02:49:05.0000] <Hixie> ewwwww! [02:49:06.0000] <Hixie> /me backs away from the forever tainted gsnedders [02:50:00.0000] <gsnedders> Hixie: Did you see an undubbed version of Црна мачка, бели мачор the? [02:50:01.0000] <gsnedders> then [02:50:02.0000] <Hixie> yes [02:50:03.0000] <Hixie> i doubt there has ever been a dub for it [02:50:04.0000] <gsnedders> That is true. [02:50:05.0000] <jgraham> Please do not get me started on the insanity of Pixar release dates. The films are amazing. The delays of months and months and months are inexcusable. [02:51:00.0000] <gsnedders> /me wonders where he'll be in October… [02:51:01.0000] <Hixie> gsnedders: actually that movie is the first example of a movie that was so good that by the end i actually forgot i was reading subtitles [02:51:02.0000] <zcorpan_> /me marks all html5 alerts from the weekend as read [02:51:03.0000] <Hixie> lol [02:51:04.0000] <zcorpan_> if there was something interesting, let me know :) [02:52:00.0000] <Hixie> all the interesting stuff was mentioned in this channel i think [02:52:01.0000] <zcorpan_> k [02:52:02.0000] <Hixie> though if you haven't seen the wave keynote video, i recommend that [02:52:03.0000] <Hixie> http://code.google.com/events/io/ [02:53:00.0000] <zcorpan_> i saw that on friday [02:53:01.0000] <Hixie> and the first keynote is worth it just for the one slide [02:53:02.0000] <zcorpan_> or maybe i just saw part 2 [02:53:03.0000] <zcorpan_> s/part/day/ [02:53:04.0000] <gsnedders> Hixie: I must admit I was impressed by how well Ratatouille was done in French, though at parts it was obvious it wasn't the original language (though it was re-animated so the mouth movements all fit), as the language at times got too complex for the age-range trying to stay fairly close to the English [02:55:00.0000] <gsnedders> /me tends to think dubbed films aren't very good [02:56:00.0000] <annevk42> Spirited Away was not too bad I thought, but then I never watched the original [02:57:00.0000] <gsnedders> I didn't think the dubbing of that was overly good in that [03:18:00.0000] <zcorpan_> /me tries the custom search element on his site and finds that it is more to it than just copy-and-paste-and-be-done-with-it [03:24:00.0000] <zcorpan_> Hixie: the google web elements snippets aren't valid html5 [03:25:00.0000] <Hixie> i am shocked [03:25:01.0000] <Hixie> (if you really want me to tell the team, send me a mail to ianh⊙gc and i'll forward it appropriately) [03:27:00.0000] <Hixie> ok i should probably head to bed [03:27:01.0000] <Hixie> nn [03:35:00.0000] <zcorpan_> <feature name="API:example.org.geolocation" - http://www.w3.org/mid/4A1FF4AA.2030409⊙oc [03:35:01.0000] <zcorpan_> reversed dns labels++ [03:37:00.0000] <hsivonen> oh noes. Feature enumeration rides again. :-/ [03:37:01.0000] <Philip`> zcorpan_: Reversed DNS labels-- because they didn't write it in proper reverse order [03:38:00.0000] <zcorpan_> maybe they should be called Labels DNS Reversed [03:39:00.0000] <jgraham> Philip`: it is presumably using the .example tld [03:40:00.0000] <gavin_> instead of the example.org domain? [03:40:01.0000] <gavin_> seems unlikely [03:40:02.0000] <zcorpan_> although he never said it was a reversed dns label, he just minted the API: url scheme [03:40:03.0000] <jgraham> (Although that would be inconsistent with the rest of the message and willfully obscure, so I may not be being entirely serious) [03:41:00.0000] <zcorpan_> i guess my point is that people will misunderstand how to write reversed dns labels [03:42:00.0000] <hsivonen> I seriously doubt the utility of the <feature> feature [03:42:01.0000] <hsivonen> I guess I should send a comment, but I'm probably late considering BONDI's schedule [03:44:00.0000] <annevk42> "better late than never"? [03:45:00.0000] <hsivonen> yeah. [03:45:01.0000] <hsivonen> lunch first [03:56:00.0000] <Mrmil> May I ask for some light feedback? http://server.ebrana.cz/olda/_apps/html5/ [04:00:00.0000] <zcorpan_> Mrmil: looks pretty good [04:01:00.0000] <gsnedders> PHP html5lib is slow. Really slow. [04:01:01.0000] <gsnedders> /me wonders if it is in an infinite loop [04:03:00.0000] <Philip`> gsnedders: Wait until it finishes, and then you'll know it's not an infinite loop [04:03:01.0000] <Mrmil> zcorpan_: Yay, thanks :) Now I have a question - imagine that the 3 service float boxes didn't have the annotation paragraph(s), what code would you use? headers wouldn't make sense anymore because there would be no text associated for them. Would you use ul li instead? [04:03:02.0000] <gsnedders> /me adds output of current line/column [04:04:00.0000] <gsnedders> Over line 25k now… (86k lines, IIRC) [04:05:00.0000] <jgraham> gsnedders: php-html5lib alone is 86k lines? [04:05:01.0000] <gsnedders> jgraham: No, the spec [04:05:02.0000] <jgraham> Oh I see [04:06:00.0000] <gsnedders> line 38k [04:06:01.0000] <gsnedders> OK, I think it probably will finish… [04:07:00.0000] <gsnedders> 44k… [04:08:00.0000] <gsnedders> /me wonders how ever long it'll take with a profiler when it takes this long without [04:10:00.0000] <gsnedders> I guess we take the prize for slowest impl. of HTML 5 parsing algorithm, at least :P [04:11:00.0000] <jgraham> gsnedders: I bet I am slower [04:11:01.0000] <gsnedders> jgraham: Slowest automated impl. [04:11:02.0000] <gsnedders> /me would not be surprised if he was slower than jgraham [04:11:03.0000] <gsnedders> line 60k! [04:12:00.0000] <gsnedders> Uh, yeah, I'm bored now. [04:12:01.0000] <annevk42> empirical evidence suggests I'm faster than jgraham and Hixie at getting correct answers :p [04:13:00.0000] <jgraham> annevk42: Hey, that was a microbenchmark if ever I saw one [04:13:01.0000] <jgraham> (although I would bet that you are still faster than me) [04:14:00.0000] <gsnedders> Line 70k! [04:15:00.0000] <Philip`> gsnedders: Using the entire spec for profiling seems like a pretty stupid idea :-p [04:15:01.0000] <annevk42> you know, if I'm really bored one day I think I rather find something else to do than test my performance on the HTML5 parsing algorithm :p [04:17:00.0000] <gsnedders> Has any machine-run implementation taken over 15 minutes to parse the spec before? [04:17:01.0000] <Philip`> That's an easy record to beat [04:17:02.0000] <Philip`> Start parsing, press ctrl+Z, wait 16 minutes, type "fg" [04:17:03.0000] <gsnedders> (We're now under 10k lines to go!) [04:18:00.0000] <Philip`> gsnedders: Sounds like your performance is slightly sub-linear [04:18:01.0000] <gsnedders> Line 80k! [04:18:02.0000] <gsnedders> No, I think it's linear. [04:19:00.0000] <gsnedders> Just slow. [04:19:01.0000] <Philip`> Hmm, good point, I suppose linear performance would be within the error bars of my IRC timestamp precision [04:20:00.0000] <gsnedders> real 17m16.388s [04:20:01.0000] <gsnedders> user 15m43.416s [04:20:02.0000] <gsnedders> sys 0m12.038s [04:20:03.0000] <Philip`> Is the output correct? [04:20:04.0000] <gsnedders> Dunno. [04:20:05.0000] <gsnedders> It just created a DOM, then deleted it :P [04:21:00.0000] <gsnedders> ezyang: So, only around 16m of CPU time to parse the spec. I guess we're not that bad. [04:22:00.0000] <Philip`> Compared to, like, one second in Java? [04:23:00.0000] <hsivonen> 16 minutes to parse the HTML5 spec? [04:23:01.0000] <gsnedders> hsivonen: yes. [04:24:00.0000] <annevk42> whoa [04:24:01.0000] <hsivonen> /me goes back to manually inlining the read() method in the Java tokenizer and tweaking it not to write into the input buffer to allow speculative parsing in C++ [04:27:00.0000] <jgraham> gsnedders: May I suggest profiling on <10% of the spec [04:27:01.0000] <jgraham> Since that will probably contain enough interesting stuff to see where all the time is being taken [04:29:00.0000] <hsivonen> are PHP arrays slow when used as arrays of ints? [04:29:01.0000] <hsivonen> that is, in that case, do they perform like real arrays or like hashtables or like something else? [04:29:02.0000] <gsnedders> hsivonen: We don't use them. Hashtables. And I don't think they're slow, just memory hogs. [04:29:03.0000] <gsnedders> (We're talking around 70 bytes of overhead per value) [04:30:00.0000] <thomaslee_> Why does TextMetrics have a width but no height? [04:30:01.0000] <othermaciej> 70 bytes per value? [04:30:02.0000] <othermaciej> seriously? [04:30:03.0000] <gsnedders> othermaciej: Seriously. [04:30:04.0000] <hsivonen> gsnedders: ok. I was wondering how hard it would be to make the C++ translator to target PHP [04:31:00.0000] <gsnedders> hsivonen: You don't want to, basically. [04:31:01.0000] <Philip`> thomaslee_: Height of what? (Line height? Bounding box height? Em square height?) [04:31:02.0000] <hsivonen> as I understand it, GWT compiler Java char[] into a JS array [04:31:03.0000] <hsivonen> which seems reasonable performant compared to the DOM and stuff [04:32:00.0000] <hsivonen> gsnedders: is indexing into PHP strings O(1)? [04:32:01.0000] <gsnedders> hsivonen: A black box function (in an extension) which took a string and returned a DOMDocument (or something else if you created another object model) would probably be the only way to get sane perf [04:32:02.0000] <gsnedders> hsivonen: yes [04:32:03.0000] <gsnedders> Uh, doing it with a bit less debug stuff it did it in 1m44s [04:32:04.0000] <Philip`> thomaslee_: There are lots of metrics that could be returned (which is why there's a whole TextMetrics structure instead of a single value), but currently it's just intended to address use cases like drawing labels on graphs [04:32:05.0000] <thomaslee_> Philip`: line height I guess. I'm trying to center some text vertically within a known bounds. [04:33:00.0000] <Philip`> thomaslee_: ... and not to address e.g. complex multi-line text layout, and so it doesn't expose any of the metrics you need for that [04:33:01.0000] <gsnedders> hsivonen: Mainly because all arrays, no matter what type of the key, is just put in a hashtable, and AFIAK that has O(1) lookup [04:33:02.0000] <zcorpan_> /me sends an email to pf basically repeating the ToC of html5 when they ask if there are any other areas with a11y concerns in html [04:33:03.0000] <Philip`> thomaslee_: Can't you use ctx.textBaseline = 'middle' for that? [04:33:04.0000] <thomaslee_> Philip`: was hoping to be able to use something like: (y + (cellHeight / 2) - (metrics.height/2)) [04:34:00.0000] <thomaslee_> Philip`: maybe? reads docs* [04:34:01.0000] <othermaciej> hsivonen: JS arrays are arrays [04:34:02.0000] <gsnedders> I guess the main reason why it was so much slower was there were an extra couple million functions calls… [04:34:03.0000] <Philip`> thomaslee_: (Don't know how widely that's implemented, but metrics.height isn't implemented at all either :-) ) [04:34:04.0000] <hsivonen> othermaciej: OK. I thought JS arrays had some whacko hashtable dualism, too [04:35:00.0000] <othermaciej> they are also hashtables, but foo[n] where n is a smallish integer will be at an array access level of speed [04:35:01.0000] <Philip`> thomaslee_: http://www.whatwg.org/specs/web-apps/current-work/multipage/the-canvas-element.html#dom-context-2d-textbaseline [04:35:02.0000] <othermaciej> at least in decent implementations [04:35:03.0000] <othermaciej> but you can also set an arbitrary named property on an array [04:36:00.0000] <gsnedders> hsivonen: SPLFixedArray which is in a disablable extension, and new in PHP 5.3, only allows integer keys and is a really array is actually slower from a reading POV, but quicker writing and a lot lower memory usage [04:36:01.0000] <thomaslee_> Philip`: ah, so I can calculate the vertical centre and render the text using textBaseline middle [04:36:02.0000] <hsivonen> othermaciej: the GWT-compiled version turns DOMStrings into UTF-16 code unit arrays and runs the tokenizer on those [04:36:03.0000] <Philip`> thomaslee_: In theory, yes [04:36:04.0000] <thomaslee_> let's see how it goes :) [04:36:05.0000] <othermaciej> hsivonen: that I'm confused about [04:36:06.0000] <hsivonen> othermaciej: so if you put a very long string into the live DOM input field, you get a very long array internally [04:37:00.0000] <othermaciej> it's not in general a good idea to use JS arrays of code point values instead of JS strings [04:37:01.0000] <othermaciej> especially for stuff you got out of the DOM [04:37:02.0000] <jgraham> hsivonen: That seems like it could be bad. [04:37:03.0000] <othermaciej> strings will almost certainly have better perf [04:37:04.0000] <thomaslee_> Philip`: you've saved the day again :P [04:38:00.0000] <othermaciej> I wonder if Google actually uses GWT for any of their products [04:38:01.0000] <jgraham> I seem to remember that V8 does some magic for array indicies > about 1000 [04:38:02.0000] <hsivonen> ideally, the JS version would use strings after I've made the input buffer read-only [04:38:03.0000] <othermaciej> an array of code point values will be bigger and slower than a JS string [04:38:04.0000] <jgraham> othermaciej: Isn't wave supposed to be written entirely with GWT? [04:38:05.0000] <othermaciej> in near any implementation [04:38:06.0000] <hsivonen> but that would entail some kind of Java-to-Java transformation before showing the code to GWT [04:38:07.0000] <othermaciej> jgraham: dunno, I have not studied its internals [04:39:00.0000] <hsivonen> using arrays makes sense here for both Java and C++ and the JS version is a side show currently [04:39:01.0000] <othermaciej> hsivonen: I would think the sensible model is to use String on the Java side and strings on the JavaScript side [04:39:02.0000] <othermaciej> instead of char arrays in either place [04:39:03.0000] <othermaciej> oh [04:39:04.0000] <thomaslee_> Philip`: out of curiosity, how exactly are you involved in HTML 5? [04:40:00.0000] <othermaciej> if you actually need something mutable then yeah you can't use a JS string [04:40:01.0000] <hsivonen> othermaciej: given how java.io works and how SAX works, I want to have char[] in Java [04:40:02.0000] <hsivonen> othermaciej: also, the XML coercion magic will still write into the input buffer [04:40:03.0000] <hsivonen> but the XML coercion stuff is unnecessary in Gecko [04:40:04.0000] <hsivonen> where speculation will be necessary [04:41:00.0000] <gsnedders> hmm, DOMNode::appendChild is by far the most expensive method/function [04:42:00.0000] <Philip`> thomaslee_: I just hang out here and annoy people, and sometimes send feedback on the spec and write http://philip.html5.org/tests/canvas/suite/tests/ and stuff [04:42:01.0000] <hsivonen> fwiw, in Gecko, DOM appends show up on profiles independently of the actual CSS box creation for those appends [04:43:00.0000] <gsnedders> 46% of the time taken to parse the spec (which appears to be closer to 1m46s) is appendChild [04:43:01.0000] <othermaciej> Gecko's DOM is pretty slow all around [04:43:02.0000] <othermaciej> <http://www.hixie.ch/tests/adhoc/perf/dom/artificial/core/001.html> [04:43:03.0000] <othermaciej> (as far as core DOM operations at least) [04:44:00.0000] <gsnedders> (This is PHP's DOM, though) [04:45:00.0000] <Philip`> gsnedders: Maybe you should parse not into a DOM, but into an XML string, and then use the presumably-fast built-in XML parser to construct the DOM [04:45:01.0000] <othermaciej> (in retrospect though, that test can't distinguish slowness of the bindings glue from slowness of the core DOM op itself) [04:45:02.0000] <hsivonen> gsnedders: the DOM is C, right? [04:45:03.0000] <gsnedders> Philip`: That limits us to streamable output [04:45:04.0000] <gsnedders> hsivonen: yes [04:45:05.0000] <hsivonen> othermaciej: Gecko's DOM Core bindings glue is getting thinner [04:45:06.0000] <Philip`> gsnedders: Just do that for well-formable fragments, not for entire documents [04:46:00.0000] <jgraham> gsnedders: You could have a custom tree implementation that could serialize to XML [04:46:01.0000] <othermaciej> (WebKit nightly is nearly 7x as fast as recent Minefield on that test) [04:47:00.0000] <gsnedders> jgraham: That'll be more expensive as that means function calls. [04:47:01.0000] <hsivonen> aargh. I need to think about character reference tokenization again [04:47:02.0000] <gsnedders> The function call overhead _has_ to be avoided. [04:47:03.0000] <gsnedders> (Mainly by avoiding functions!) [04:48:00.0000] <zcorpan_> i wonder why accessibility people put so much weight on a small set of features (alt, summary, etc) while completely ignoring other features (drag and drop, details, etc) [04:48:01.0000] <hsivonen> zcorpan_: a no-regression policy, I think [04:49:00.0000] <Philip`> zcorpan_: Maybe the small set of features is all that there's any chance of getting implemented in the next decade [04:49:01.0000] <thomaslee_> Philip`: I see. Seems like you keep yourself busy :) [04:50:00.0000] <zcorpan_> Philip`: some features have already been implemented and not received accessibility expert review [04:51:00.0000] <othermaciej> it's not unreasonable to be more concerned about regressions, or what is perceived as such [04:51:01.0000] <gsnedders> jgraham: Function call overhead is big enough for it actually be to worthwhile to have an if statement for A–Z and only call strtolower on characters we know it'll have an effect on [04:52:00.0000] <zcorpan_> othermaciej: agreed, but i'm seeing something close to all-focus-on-regressions-and-completely-ignore-everything-else [04:52:01.0000] <Philip`> gsnedders: Is that function call overhead, or strtolower overhead? [04:52:02.0000] <gsnedders> Philip`: function call overhead more than strtolower overhead [04:53:00.0000] <jgraham> gsnedders: I don't see how it would add more function calls. It would just mean replacing calls to the DOM API to calls to your own code [04:53:01.0000] <jgraham> The big problem is that your own code might be even slower than the DOM API [04:53:02.0000] <othermaciej> zcorpan_: their approach may seem reasonable to them [04:53:03.0000] <Philip`> gsnedders: What if you say "if (char == 'A') char = 'a'; else if (char == 'B') ..."? :-) [04:53:04.0000] <gsnedders> jgraham: Well, instead of appendChild in C we have more function calls within our implementation of it [04:53:05.0000] <Philip`> zcorpan_: Some people talked about accessibility of <canvas> too, which is a new thing and not a regression [04:54:00.0000] <gsnedders> jgraham: Hence in total we end up with more function calls [04:54:01.0000] <zcorpan_> Philip`: yes that's basically the exception to the rule [04:54:02.0000] <zcorpan_> though i haven't seen that much discussion about <canvas> [04:55:00.0000] <gsnedders> Philip`: I doubt that's quicker, haven't tried though [04:56:00.0000] <Philip`> zcorpan_: I think most of the discussion has been saying "<canvas> must be accessible", and the only concrete proposal is to make the spec say "authors must provide an accessible alternative to any <canvas> content" (which the spec already requires) [04:57:00.0000] <gsnedders> /me installs kdesdk4 on his old Power Mac G5 as that has more RAM than this laptop, hence runs KCachegrind with large profiles a lot better. [04:57:01.0000] <zcorpan_> Philip`: indeed [04:57:02.0000] <Dashiva> Philip`: And the proposal to make not doing so illegal :) [04:58:00.0000] <Philip`> zcorpan_: so there hasn't been any interesting technical discussion on accessibility there [04:58:01.0000] <annevk42> zcorpan_, there's been discussion about <video> and <audio> too [04:58:02.0000] <zcorpan_> annevk2: true [04:58:03.0000] <hsivonen> should mv_embed be considered ready for production use for enabling Ogg <video> in IE using Cortado or VLC plugin? [04:58:04.0000] <Dashiva> There was some talk about making <canvas> contain a tree of elements that could be used to extract structure info [04:58:05.0000] <Dashiva> Apparently some OSes do it like that [04:59:00.0000] <Dashiva> But as I recall, that just reduced to authors who cared doing it, and those who didn't care still wouldn't do anything [04:59:01.0000] <jgraham> I think in generaly the a12y people have benn concered with things that fit cleanly into familar WCAG categories [05:00:00.0000] <zcorpan_> /me wonders what a12y stands for [05:00:01.0000] <zcorpan_> accessibillity? [05:01:00.0000] <jgraham> Can I not spell/count/something [05:01:01.0000] <jgraham> What a surprise [05:01:02.0000] <Dashiva> Just remember that a11y is a name [05:01:03.0000] <Dashiva> And l10n is an animal :P [05:02:00.0000] <jgraham> What about l18n? [05:02:01.0000] <jgraham> or whatever it is [05:02:02.0000] <Dashiva> i18n? [05:02:03.0000] <zcorpan_> i18n [05:02:04.0000] <Dashiva> That's just weird [05:03:00.0000] <jgraham> Oh yeah, I. That makes more sense [05:04:00.0000] <jgraham> /me can only remember simple facts like "it has 18 in it and is something about making software for funny foreign types" [05:05:00.0000] <zcorpan_> like swedes? [05:05:01.0000] <Dashiva> And then there's that m12n or whatnot which I never remember what is [05:05:02.0000] <beowulf> /me wonders what the real title of the film se7en is [05:06:00.0000] <annevk42> Dashiva, l12n? [05:06:01.0000] <Dashiva> No, mNumberLetter [05:06:02.0000] <zcorpan_> m12n is modularization [05:06:03.0000] <zcorpan_> xhtml2 wg term :) [05:06:04.0000] <Dashiva> Aha [05:07:00.0000] <Philip`> beowulf: Semicitizen? [05:07:01.0000] <Philip`> beowulf: or Sempergreen? [05:07:02.0000] <Dashiva> 7 would be the total number of letters, wouldn't it? [05:07:03.0000] <zcorpan_> /me wonders whether Philip` has an English dictionary which he can query with regexps [05:07:04.0000] <zcorpan_> Dashiva: no [05:08:00.0000] <zcorpan_> Dashiva: the numbers of letters that are replaced with the number [05:08:01.0000] <Philip`> zcorpan_: /usr/share/dict/words [05:08:02.0000] <Philip`> (on any sensible OS) [05:09:00.0000] <jgraham> http://www.morewords.com/ does what you want (probably) [05:09:01.0000] <jgraham> (although it didn't find Philip`'s matches so it obviously sucks) [05:11:00.0000] <Philip`> /me first heard the word "l10n" when Ultima Online had a "Public L10n Test" server, but it was drawn in a handwritingy font so it looked very much like "Lion" [05:12:00.0000] <Philip`> (UO's concept of "l10n" was to remove all the nice grammar in English, like adjectives in front of nouns for the names of magical weapons, and replace everything with boring-but-easy-to-localise phrases instead) [05:13:00.0000] <annevk42> btw, for future reference: http://en.wikipedia.org/wiki/Numeronym [05:32:00.0000] <hsivonen> yay for ecj: [05:32:01.0000] <hsivonen> 9392: goto 0 [05:32:02.0000] <hsivonen> 9395: goto 9277 [05:32:03.0000] <hsivonen> 9398: goto 0 [05:33:00.0000] <hsivonen> seems like a trivial optimization not to generate gotos like that [05:35:00.0000] <Dashiva> zcorpan_: That doesn't seem to match most of the ones listed here? [05:52:00.0000] <zcorpan_> Dashiva: it does afaict [05:52:01.0000] <zcorpan_> l10n - "ocalizatio".length == 10 [05:53:00.0000] <Dashiva> Hmm, indeed. I must be sleepy still. [06:09:00.0000] <Dashiva> I wish lastweek would link some good music for once [06:25:00.0000] <Philip`> Dashiva: Maybe he'll take requests if you have something specific to recommend [06:37:00.0000] <hsivonen> I wonder if anyone makes good old-fashioned Java bytecode optimizers anymore [06:37:01.0000] <hsivonen> that'd flatten multilevel gotos and stuff [06:43:00.0000] <roc> probably [06:43:01.0000] <roc> I doubt they matter [06:43:02.0000] <roc> the JIT takes all that apart anyway [06:44:00.0000] <hsivonen> roc: the thing is that the JIT doesn't kick in at all if a method grows over 8000 byte codes in size [06:45:00.0000] <roc> well, that's dumb [06:45:01.0000] <hsivonen> roc: Or I could document that if anyone uses this library, they have to configure their JIT in a special way [06:45:02.0000] <roc> I guess you should avoid making such methods :-) [06:45:03.0000] <hsivonen> roc: but most likely people would just ignore that and say that HTML5 is slow [06:46:00.0000] <hsivonen> roc: as long as the JIT kicks in, one huge method of gotos JITs to faster code than a bunch of methods [06:49:00.0000] <jgraham> hsivonen: Break it into to so something? [06:50:00.0000] <jgraham> s/to/two/ [06:50:01.0000] <jgraham> s/so/or/ [06:50:02.0000] <hsivonen> jgraham: then I have to transfer local variables over [06:50:03.0000] <hsivonen> jgraham: and using fields instead of local variable is very bad if you are trying to minimize the # of bytecodes [06:51:00.0000] <hsivonen> But I may end up splitting some NCR stuff out [06:51:01.0000] <hsivonen> Of course, if Hixie adds a new state for --!>, I'll have to make room for some more bytecodes... [06:52:00.0000] <jgraham> Yeah, we really need at least --!> and preferably -- > to work [06:52:01.0000] <jgraham> So I would start clearing room :) [06:54:00.0000] <hsivonen> hmm. I guess at some point there's no point in counting byte codes and I should just document that HotSpot has a lame magic limit and how to disable the limit [06:55:00.0000] <hsivonen> limits like that feel really applet-ish [06:55:01.0000] <hsivonen> who'd want the JIT not to do its thing on a server? [07:00:00.0000] <Philip`> Maybe someone who doesn't want their server to crash because it's trying to JIT a million-bytecode method all at once? [07:04:00.0000] <hsivonen> so many constrains: validation needs, Java library needs, speculative parsing in C++, JIT limits... [07:06:00.0000] <Philip`> I guess there's also the constraint of being a single codebase, rather than one for each different use [07:10:00.0000] <hsivonen> the 8000 bytecode limit is really the most annoying constraint [07:10:01.0000] <hsivonen> I'm very close to giving up and fixing it in documentation [07:10:02.0000] <hsivonen> "use this incantation to get a 10x speedup" [07:11:00.0000] <roc> JVM methods can't be more than 64k bytes of bytecodes [07:11:01.0000] <hsivonen> roc: the main tokenization loop is currently at 8701 bytecodes, so it's not too near that limit [07:12:00.0000] <hsivonen> oh, yeah, CRLF is hugely annoying, too. a lot of the extra bytecode fat is due to CRLF handling [07:12:01.0000] <roc> I once wrote a bytecode instrumentation tool that had the ability to take apart methods grew to over 64K bytes and automatically partition them into smaller methods [07:12:02.0000] <roc> without worrying too much about performance of course [07:13:00.0000] <heycam> roc, can you commit that to rhino's codebase please? :) [07:13:01.0000] <heycam> i run in to it from time to time [07:14:00.0000] <heycam> it'd be nice not to just tell it to use interpreted mode unconditionally in advance, in those situations [07:14:01.0000] <heycam> oops, please ignore the smiley, i'm trying to cut back [07:14:02.0000] <hsivonen> (it seems that major design decisions in Antlr have been made due to the 8000 bytecode limit) [07:18:00.0000] <hsivonen> I wonder if dalvik and whatever AppEngine uses have this kind of magic limits [07:19:00.0000] <heycam> while trying to optimise some of rhino's codegen once, i noticed it was generating a 4000 byte method that was a massive switch that just returned false from every branch [07:19:01.0000] <heycam> (true, that wasn't compilation of js to java, and just a generated method to support compiled js, but still... seemd a bit excessive) [07:26:00.0000] <ezyang> gsnedders: What was the next most expensive function call after appendChild()? [07:27:00.0000] <gsnedders> uh, dunno. [07:27:01.0000] <ezyang> Wait, you didn't profile it? [07:27:02.0000] <gsnedders> Yes, I did. [07:27:03.0000] <ezyang> Ok [07:27:04.0000] <gsnedders> I can't remember what the profile said :P [07:28:00.0000] <ezyang> You should post that up somewhere public [07:28:01.0000] <gsnedders> (It kinda crashes my laptop opening it in KCachegrind though as it eats so much memory.) [07:29:00.0000] <gsnedders> (So I'm installing KCachegrind on my old Power Mac G5 which has more memory.) [07:29:01.0000] <ezyang> Yeah, swapping can be pretty painful [07:29:02.0000] <ezyang> How big is the profile file? [07:30:00.0000] <gsnedders> It got to 600MB before I killed php stopping it becoming bigger. [07:30:01.0000] <ezyang> Hmm. [07:30:02.0000] <ezyang> My suggestion is to take a smaller HTML file and use that [07:30:03.0000] <gsnedders> (And KCachegrind tends to use double the size of the input when loading.) [07:30:04.0000] <gsnedders> That's boring! :P [07:31:00.0000] <ezyang> If html5lib is not pathological, it should still be representative [07:31:01.0000] <ezyang> Anyway, my current ordering of priorities is: 1. Clean up the APIs, 2. Kick out a release, 3. Optimize the hell out of it [07:32:00.0000] <ezyang> Also, I start working today, so I won't have as much time to work on html5lib [07:33:00.0000] <gsnedders> on my blog the profile is totally different [07:33:01.0000] <ezyang> Oh? [07:33:02.0000] <gsnedders> HTML5_Tokenizer::parse is slowest [07:33:03.0000] <ezyang> Is this total runtime or runtime within? [07:33:04.0000] <gsnedders> Just within it. [07:34:00.0000] <gsnedders> TreeConstructor::emitToken is next slowest [07:34:01.0000] <ezyang> Well, duh [07:34:02.0000] <gsnedders> then consumeCharacterReference [07:34:03.0000] <ezyang> Those are our two big honking switch cases [07:34:04.0000] <gsnedders> Then Tokenizer::emitToken [07:34:05.0000] <ezyang> consumeCharacterReference isn't surprising either [07:34:06.0000] <gsnedders> then InputStream::charsUntil [07:35:00.0000] <ezyang> These are all fairly obvious, and non-trivial to fix, unfortunately [07:35:01.0000] <ezyang> I'll run the profiler myself (I tend to favor Wikipedia pages) and see what I can do, after (1) and (2) are done [07:35:02.0000] <gsnedders> /me wonders what a decent medium sized page is [07:36:00.0000] <gsnedders> /me does from a non-bias POV http://en.wikipedia.org/wiki/United_Kingdom [07:38:00.0000] <gsnedders> appendChild is 8th most expensive there [07:39:00.0000] <gsnedders> I think we could remove our own appendChild and just use DOMNode::normalize [07:40:00.0000] <Philip`> gsnedders: The mean page size is something like 20KB [07:40:01.0000] <Philip`> (in my samples) [07:47:00.0000] <gsnedders> ezyang: Where do you want the call for parsing to actually go? [07:47:01.0000] <zcorpan> hsivonen: Start: Go to "X-UA-Compatible <meta>?". - you've forgotten to curlify the quotes there [07:48:00.0000] <hsivonen> zcorpan: thanks [07:49:00.0000] <hsivonen> zcorpan: fixed [07:52:00.0000] <gsnedders> ezyang: What are these "godawful libxml bugs"? [07:52:01.0000] <zcorpan> hsivonen: IE=6 is the same as IE=5 [07:52:02.0000] <zcorpan> i.e. quirks mode [07:52:03.0000] <hsivonen> zcorpan: thanks. (that'll take a bit more time to update) [07:53:00.0000] <annevk42> really? that's somewhat unexpected [07:54:00.0000] <jgraham> /me wonders how long annevk42 will be known as "Anna" for in the html5lib wiki [07:54:01.0000] <annevk42> jgraham, considering how often that wiki is updated, pretty long [07:54:02.0000] <jgraham> That page was only created yesterday or so though [07:55:00.0000] <zcorpan> annevk2: well i just tested it. it wasn't what i expected, either [07:55:01.0000] <jgraham> /me is not sure why it is so important to know who the project is done by rather than ho to use it [07:56:00.0000] <annevk42> ah, i guess that must be from the person I added to the project [07:56:01.0000] <annevk42> he said he was going to work on documentation [07:56:02.0000] <beowulf> jgraham: babes might be interested [07:59:00.0000] <gsnedders> /me laughs, and shakes his head [08:03:00.0000] <annevk42> "formal W3C Note" hmm [08:17:00.0000] <zcorpan> re http://www.w3.org/mid/4A23E22C.2010502⊙gd - i thought there were more issues, like what to do with whitespace and lone % [08:21:00.0000] <zcorpan> hsivonen: forgetting a ">" of an end tag results in lots of messages [08:21:01.0000] <zcorpan> hsivonen: see e.g. http://www.webforum.nu/showthread.php?t=174764 [08:21:02.0000] <zcorpan> <a href="http://www.djurssommerland.dk/">Djurs sommarland</a är något man inte får missa om man ... [08:22:00.0000] <zcorpan> on a start tag it's even worse [08:28:00.0000] <zcorpan> hsivonen: maybe if a tag has an illegal attribute, silence all other illegal attribute messages and duplicate attributes messages for that tag [08:28:01.0000] <zcorpan> hsivonen: also, it seems pointless to emit any other messages than "attributes on end tag" for end tags [08:32:00.0000] <zcorpan> http://bugzilla.validator.nu/show_bug.cgi?id=579 [08:37:00.0000] <hsivonen> zcorpan: thanks [08:53:00.0000] <hsivonen> I wonder why Flickr slideshows are Flash [08:53:01.0000] <hsivonen> is it all because of the ability to go full-screen? [09:05:00.0000] <jgraham> Gah, the sanitizer in html5lib.filters.sanitizer seems to be broken [09:22:00.0000] <jgraham> hsivonen: It works with the video, of course [09:23:00.0000] <jgraham> But it would be a neat project to do a flickr slideshow viewer in html5 [09:23:01.0000] <hsivonen> jgraham: they had Flash slideshows long before video [09:24:00.0000] <hsivonen> but it would be cool if Flickr went Flashless [09:25:00.0000] <jgraham> hsivonen: I know. But at the moment it seems like a good reason not to use HTML [09:26:00.0000] <jgraham> The nice thing about a flashless slideshow is that I would get better colours in browsers that support embedded colour profiles [09:26:01.0000] <zcorpan> hmm, the <script src>//documentation</script> abnf only supports LF and multiline comments can't contain stars or slashes at all if i'm reading it correctly [09:26:02.0000] <hsivonen> support for color profiles doesn't necessarily mean better color :-/ [09:27:00.0000] <hsivonen> a lot of color management is GIGO, unfortunately [09:27:01.0000] <jgraham> hsivonen: For photography it generally means better colour [09:28:00.0000] <hsivonen> jgraham: if you actually have right profiles [09:28:01.0000] <hsivonen> throughout the workflow [09:28:02.0000] <jgraham> Since a large fraction of the people have reasonably well calibrated monitors and are using profile-aware applications like Lightroom [09:28:03.0000] <jgraham> (and photoshop) [09:32:00.0000] <jgraham> e.g. compare http://www.flickr.com/photos/jgraham/3567837536/sizes/l/ in various browsers [09:32:01.0000] <gsnedders> /me wonders what 15 minute walk is [09:33:00.0000] <jgraham> (although I actually think that I made some mistakes with the processing of that in terms of aesthetic judgement, but that's not really the point) [09:34:00.0000] <gsnedders> jgraham: if Bibbi said 15 minutes walk, what sorta distance do you think she'd mean? [09:35:00.0000] <jgraham> gsnedders: I guess the sort of distance that you could walk in 15 minutes [09:35:01.0000] <jgraham> gsnedders: Do you have an address or anything? [09:35:02.0000] <gsnedders> jgraham: No [09:35:03.0000] <gsnedders> Just, "Close to the office (about 15 minutes walk)" [09:35:04.0000] <jgraham> Well that could be anywhere but I guess within 1km or so [09:36:00.0000] <zcorpan> i have about 15 minutes walk to the office [09:36:01.0000] <jgraham> me too [09:36:02.0000] <gsnedders> 15 minutes at what speed for a start :P [09:36:03.0000] <zcorpan> walking speed of course [09:36:04.0000] <jgraham> but we live at pi/2 radians from each other [09:37:00.0000] <jgraham> gsnedders: I guess normal walking speed is about 2.5 miles/hour [09:37:01.0000] <jgraham> so 15minutes walk is about 5/8 of a mile. Or 1km [09:37:02.0000] <jgraham> Which is what I said before [09:39:00.0000] <gsnedders> /me works out basically where 1km away from the centre is [09:39:01.0000] <gsnedders> (Or rather, from where the office is) [09:40:00.0000] <gsnedders> /me notes Google Maps puts the office right in the middle of Centrum Ring, which I kinda doubt :P [09:40:01.0000] <zcorpan> if (window.widget.harharhar) [09:42:00.0000] <hsivonen> /me notes that Chrome doesn't show the video thumbnails in the YouTube demo [09:45:00.0000] <zcorpan> /me notes that the youtube demo uses <div><h1> instead of <details><legend> [09:46:00.0000] <zcorpan> the document outline will be shorter if they migrate to <details> [09:46:01.0000] <jgraham> gsnedders: According to http://maps.forum.nu/gm_sensitive_circle2.html I live about 1.5km from the office. But I guess it takes a little more than 15 minutes for me to get in [09:52:00.0000] <gsnedders> /me needs to work out when he'll leave Linköping [09:52:01.0000] <jgraham> gsnedders: It is like Hotel California [09:52:02.0000] <gsnedders> jgraham: ? [09:53:00.0000] <gsnedders> I think not the day after I finish, as the only flight that day is a 6:30am [09:53:01.0000] <jgraham> gsnedders: Just a throwaway pop-culture reference [09:53:02.0000] <gsnedders> jgraham: Ah. Pop-culture. That's the problem :P [09:55:00.0000] <gsnedders> Wait, what… [09:55:01.0000] <gsnedders> Flights from Stockholm (Skavasta) to Berlin (Schönefeld) for 55SEK!? [09:56:00.0000] <jgraham> gsnedders: Ryanair? [09:56:01.0000] <gsnedders> jgraham: Yeah [09:56:02.0000] <jgraham> gsnedders: Check the "Taxes" and "Fees" [09:56:03.0000] <gsnedders> Taxes and Fees it claims to be 0. [09:56:04.0000] <gsnedders> That's what is surprising me so much [09:57:00.0000] <jgraham> gsnedders: Oh you found one of those flights. In that case you will only have to pay for luggage, booking fee, etc. [09:57:01.0000] <gsnedders> Every day apart from Friday and Sunday [09:57:02.0000] <gsnedders> (On the Sunday it is 314.78 SEK) [09:58:00.0000] <gsnedders> /me probably won't get a plane ticket from Berlin to Britain yet, as he doesn't know whether he'll be going to London or to home… [10:04:00.0000] <gsnedders> WTF? [10:04:01.0000] <gsnedders> Flights for £0.00 (excluding taxes/fees) [10:05:00.0000] <gsnedders> At 6am, though :P [10:10:00.0000] <jgraham> gsnedders: 0 excluding taxes/fees == about 50GBP all told, usually [10:12:00.0000] <gsnedders> It was more than that [10:12:01.0000] <gsnedders> But I don't care. [10:12:02.0000] <gsnedders> /me is going for a flight at a more reasonable time of day :P [10:25:00.0000] <jgraham> /me has paid an extrodinary amount of money for the last few sets of Ryanair flights he has booked [10:26:00.0000] <jgraham> Because I don't tend to book early enough and because they are not nearly as cheap as they make out [13:27:00.0000] <annevk42> how does http://software.hixie.ch/utilities/js/live-dom-viewer/saved/124 demonstrate margin collapsing? [13:27:01.0000] <annevk42> oh, zcorpan is not here [13:35:00.0000] <Hixie> aaah, the epitomy of process over progress: http://www.w3.org/mid/4A2428A2.6080205⊙gd [13:38:00.0000] <Dashiva> But what if about:blank totally changes meaning during the process? ;) [13:41:00.0000] <gsnedders> Hixie: s/y/e/ [13:42:00.0000] <Hixie> so it seems [13:42:01.0000] <Hixie> thanks [13:42:02.0000] <Hixie> Dashiva: exactly [13:43:00.0000] <gsnedders> /me fixes a test case in php html5lib by fixing the test handler to actually follow the test format :P [13:43:01.0000] <annevk42> Hixie, yeah :/ [13:43:02.0000] <annevk42> Hixie, though if they think it really matters maybe they should leave the value space up to some other RFC... [13:44:00.0000] <annevk42> /me wonders who'll write all of those [13:44:01.0000] <Hixie> i have faith that adam will do the right thing [13:44:02.0000] <annevk42> the right thing was just made harder [13:45:00.0000] <annevk42> /me kind of like about:noreferrer [13:45:01.0000] <annevk42> likes* [14:08:00.0000] <jgraham> /me wonders who the people are that go "oh no that's just a draft standard I can't possibly implement the well defined thing that I need for interoperability" [14:11:00.0000] <tantek> jgraham, Perhaps those that then must maintain back/bug-ward compatibility with their implementation of the draft standard for 3+ years for enterprise users and others that don't upgrade implementations very often. [14:14:00.0000] <jgraham> tantek: In the case where the draft is really unstable that seems reasonable. But in the case of e.g. about: it's fixed by legacy anyway [14:15:00.0000] <Hixie> i like larry's observation that we don't understand the ietf process given how he doesn't understand the whatwg process :-) [14:15:01.0000] <tantek> jgraham, agreed. when a draft standard is simply documenting existing (hopefully widely known) interoperable behavior, the excuse of "draft" does not hold much weight. [14:21:00.0000] <annevk42> afaict Larry is wrong too [14:21:01.0000] <annevk42> tools.ietf.org is generally considered to be the space of official IETF documents [14:21:02.0000] <annevk42> moreover, I believe there's no documented guarantee that the links there won't break either [14:21:03.0000] <annevk42> is not generally* duh [14:37:00.0000] <expensivelesbian> Hello all, just a random question - if one wanted to write some sort of "visualiser" (VU meters, graphical EQ bars and the like), taking the output of an <audio> element, would this be possible currently, with HTML5 + JavaScript? If this question is off topic, someone please say nicely and I'll scurry off elsewhere [14:38:00.0000] <annevk42> http://lists.w3.org/Archives/Public/public-html/2009Jun/0036.html really, what's the obsession with introducing versioning? is it some kind of nirvana? [14:40:00.0000] <annevk42> expensivelesbian, it's not off-topic, but I don't think it's possible either currently [14:40:01.0000] <roc> it's not possible [14:40:02.0000] <annevk42> expensivelesbian, unless you skip the <audio> element and parse the stream directly in ECMAScript, but besides that being hard you might run into some limitations as well (e.g. no real byte stream representation) [14:43:00.0000] <expensivelesbian> ok, thanks for the answers. I'll have a ponder on this [14:45:00.0000] <annevk42> I guess due to <canvas> you can do a lot more with <video> [14:45:01.0000] <annevk42> but then audio manipulation is prolly much harder [14:46:00.0000] <expensivelesbian> well, it was that bounding boxes demo got me thinking, "there must be something _like_ that for poor old neglected <audio>" [14:50:00.0000] <annevk42> expensivelesbian, sorry [14:50:01.0000] <annevk42> expensivelesbian, there have been discussions, but nothing concrete yet [14:59:00.0000] <annevk42> quotes in http://edward.oconnor.cx/2009/BarCamp-San-Diego-5/ are nice -- navigation within that presentation is a bitch though, what's back? [15:26:00.0000] <hober> I don't think I got around to implementing back... [15:27:00.0000] <hober> I should really make a print stylesheet too.. [16:22:00.0000] <ezyang> gsnedders: (a very late response) Put it in HTMLParser. That should be the "public interface" [16:23:00.0000] <ezyang> gsnedders: Godawful libxml bug == "Namespace Error" when you try to set a namespaced attribute. [16:24:00.0000] <jwalden> so, looking at the websocket protocol, what constraints are enforced on the server regarding processing of the input? is it an error if tab-delimited "header" lines are received? what if "multi-line headers" are received? basically, websockets looks like HTTP, but at what point does accepting HTTP input that's more lax than the websocket client input constitute a violation of the protocol... [16:24:01.0000] <jwalden> ...requirements? [16:26:00.0000] <Hixie> what do you mean by "enforced"? [16:26:01.0000] <Hixie> the protocol spec says what you're expected to suport [16:26:02.0000] <Hixie> support [16:26:03.0000] <Hixie> as a server-side author [16:27:00.0000] <jwalden> s/enforced/required/, then [16:27:01.0000] <jwalden> the client is required to follow the exact steps given in the draft [16:27:02.0000] <Hixie> http://tools.ietf.org/html/draft-hixie-thewebsocketprotocol-12#page-15 [16:28:00.0000] <jwalden> the server, it seems, has latitude to accept more than that [16:28:01.0000] <jwalden> is the server required to accept input from a client that doesn't actually follow the strictures of the client-processing section? [16:28:02.0000] <jwalden> yes, I've been looking at that [16:29:00.0000] <ezyang> gsnedders: My idea is to implement SimpleTree and watch our performance problems melt away ^_^ [16:29:01.0000] <Hixie> jwalden: the server doesn't have to do any error handling, if that's what you mean [16:29:02.0000] <jwalden> okay [16:30:00.0000] <jwalden> that seems...dangerous [16:30:01.0000] <jwalden> not in any tangible sense [16:30:02.0000] <Hixie> the ability to control authors is basically nil [16:30:03.0000] <Hixie> so getting them to implement particular error handling behavior... forget it [16:30:04.0000] <jwalden> but in the sense of I-can-plausibly-believe-this-will-bite-us-at-some-point [16:30:05.0000] <jwalden> no, this is error handling server-side [16:31:00.0000] <Hixie> right [16:31:01.0000] <Hixie> authors write the server side [16:31:02.0000] <jwalden> oh, there are authors on both sides [16:31:03.0000] <jwalden> hum, this might work then [16:32:00.0000] <Hixie> i could write a new section [16:32:01.0000] <Hixie> which is "full requirements for server-side implementors that actually follow specs" [16:32:02.0000] <Hixie> if you like [16:32:03.0000] <Hixie> might not be a bad idea [16:32:04.0000] <jwalden> dunno, yet [16:32:05.0000] <Hixie> if you do think that would be useful, send a mail to whatwg, and i'll add it [16:33:00.0000] <jwalden> I'm just looking at this from the POV of the Mozilla bug, in which our HTTP server is being lightly hijacked to support all this, worried about being more generous in accepting input than it should be [16:33:01.0000] <Hixie> is this for testing or for production? [16:34:00.0000] <Hixie> for regression testing, it doesn't matter what the server does, it just has to check that the client sends exactly the right bytes [16:34:01.0000] <jwalden> I try not to differentiate the two when I write servers (or rather, when I write the one server I've written so far) [16:34:02.0000] <Hixie> a regression test suite server and a production server are completely different [16:34:03.0000] <Hixie> you can't, e.g., use apache to do serious regression testing of client-side http behaviour [16:34:04.0000] <jwalden> yes, but that doesn't mean one shouldn't make an effort not to make it safe-by-default [16:35:00.0000] <Hixie> i don't really see how a regression-testing server framework can be unsafe -- it has to accept any input whatsoever [16:36:00.0000] <jwalden> hm, I think we're talking about different meanings; "production" to me is use in which inputs are unconstrained, "regression-testing" in the sense of inputs are constrained but may intentionally be random junk [16:36:01.0000] <jwalden> constrained by whatever's in the test suite using it [16:40:00.0000] <jwalden> on an unrelated note, reading RFC-style formatting rather than whatwg-style formatting is no fun :-( [16:40:01.0000] <jwalden> especially since things like "find" don't work due to ASCII text-decorations [16:42:00.0000] <Hixie> jwalden: an "in production" server is one serving end-user traffic [16:42:01.0000] <jwalden> we agree on that [16:42:02.0000] <Hixie> jwalden: a regression testing server is one used to test another tool's implementation [16:43:00.0000] <jwalden> sure, just say that I like my regression testing servers to have as rigorous an implementation as a production server must have [16:43:01.0000] <Hixie> so an in-production server can do things like disconnect if the input is bogus [16:43:02.0000] <Hixie> but a regression testing server needs to report pass/fail for all inputs [16:55:00.0000] <gsnedders> http://larry.masinter.net/versioning-html.html [16:55:01.0000] <gsnedders> (from www-tag) [16:56:00.0000] <ezyang> 'lo gsnedders [16:56:01.0000] <ezyang> Did you see my late late late msgs? [16:56:02.0000] <gsnedders> ezyang: yes [16:57:00.0000] <gsnedders> ezyang: I've fixed another test case (actually, I've fixed the test runner). [16:57:01.0000] <ezyang> Makes sense [16:57:02.0000] <ezyang> The strDom() function is hack on hack on hack 2009-06-02 [17:00:00.0000] <ezyang> Hixie: ping? [17:16:00.0000] <Hixie> ezyang: here [17:17:00.0000] <ezyang> Hixie: Can you comment on what IN_FOREIGN_CONTENT is supposed to do when we are processing an HTML element as "in body", and that element wants us to change the mode? [17:18:00.0000] <Hixie> um [17:18:01.0000] <Hixie> nothing beyond what the spec says as far as i can tell [17:18:02.0000] <Hixie> how do you mean? [17:19:00.0000] <ezyang> As far as I can tell, the spec doesn't record that you were originally in "in foreign content" when it hops to another mode, so when you pop out of that, you stay in "in body", and don't go back to "in foreign content" [17:19:01.0000] <ezyang> I just wanted to verify that this was true before sending mail [17:32:00.0000] <theMadness> Augh, damn user agents. [17:32:01.0000] <theMadness> They create a plethora of <a> nodes when I wrap some block item inside an <a> tag. [17:33:00.0000] <theMadness> I guess I have overestimated the current support of html5. [17:34:00.0000] <Hixie> ezyang: minimal testcase? do you mean like <svg><table> ? [17:41:00.0000] <ezyang> <svg><foreignObject><table></table></foreignObject></svg> [17:42:00.0000] <ezyang> Uhh, make that <svg><foreignObject><table></table></foreignObject><g/></svg> [17:43:00.0000] <Hixie> ezyang: wouldn't you switch back to foreign content when you reset after the </table>? [17:44:00.0000] <ezyang> Possibly. Let me check that part of the spec again [17:45:00.0000] <ezyang> Yep, looks like it. [17:49:00.0000] <Hixie> anyone here understand RDF/XML? [17:53:00.0000] <Dashiva> I figured RDF/XML would be mostly self-explanatory, but if you're asking that question, I guess not [18:08:00.0000] <ezyang> Is there any good way of just getting messages about changes to the html5lib spec that are w.r.t. tokenization/tree-building? [18:08:01.0000] <ezyang> erm, s/html5lib/html5/ [18:09:00.0000] <Hixie> yeah, look for 't' in the first field [18:09:01.0000] <ezyang> Hm... so I still have to sign up for html5-commits, and just filter out things w/o t in the first field? [18:10:00.0000] <Hixie> or you can use the tracker [18:10:01.0000] <Hixie> it has little check boxes that filter out everything else [18:10:02.0000] <Hixie> note that i often screw up [18:10:03.0000] <Hixie> so don't rely on it too much :-) [18:10:04.0000] <ezyang> So, "Affects Tools" is what I want? [18:11:00.0000] <ezyang> Man, I wish this had an RSS feed [18:11:01.0000] <ezyang> (and an easier way of unchecking everything) [18:12:00.0000] <Hixie> yeah [18:12:01.0000] <Hixie> you can provide patches, i believe [18:13:00.0000] <Hixie> the code is in code.google.com/html5 [18:13:01.0000] <ezyang> Sweet. [18:13:02.0000] <ezyang> Assuming it's not too grody, I'll take a stab [18:13:03.0000] <ezyang> ftr, it's http://code.google.com/p/html5/ [18:13:04.0000] <ezyang> Aw, it's still svn [18:14:00.0000] <ezyang> Can I get commit access? [18:16:00.0000] <Hixie> sure [18:17:00.0000] <Hixie> what's your account? [18:17:01.0000] <ezyang> edwardzyang⊙tc [18:17:02.0000] <ezyang> (same as html5lib) [18:18:00.0000] <Hixie> i'm not an html5lib admin :-) [18:19:00.0000] <Hixie> done [18:19:01.0000] <Hixie> don't break anything :-) [18:19:02.0000] <ezyang> Aye-aye, captain! [18:28:00.0000] <ezyang> Uh oh... html5 is using a svn:externals of testdata, which means it's old. [18:29:00.0000] <Hixie> i think most of the stuff in there is stale [18:29:01.0000] <ezyang> Heh. [18:29:02.0000] <Hixie> it's just a place for us to throw stuff related to html5 [18:29:03.0000] <ezyang> Ok. [18:29:04.0000] <Hixie> if it's useful it usually moves on to live elsewhere [18:38:00.0000] <ezyang> Meh, after looking at the Python code I decided I don't care enough [18:39:00.0000] <Hixie> hah [18:45:00.0000] <jwalden> Hixie: seems to me that "Basic d2FsbGU6ZXZl" is missing some caps and a hyphen, no? :-) [18:45:01.0000] <jwalden> /me pat-pats window.atob [18:46:00.0000] <jwalden> /me runs the other authstrs through atob for entertainment [18:46:01.0000] <jwalden> guess that's just stylistic :-) [18:49:00.0000] <ezyang> Hmm... should I setup some require_once's, or setup the entire HTML Purifier style autoloader shish-kebab [19:07:00.0000] <ezyang> Barring any objections, I'd like to release 0.1.0. [19:07:01.0000] <ezyang> (of the php html5lib) [19:10:00.0000] <Hixie> jwalden: :-) [19:16:00.0000] <ezyang> Uploaded [19:16:01.0000] <ezyang> I can't figure out how to update the project summary [20:45:00.0000] <MikeSmith> irc.mozilla.org down? [20:46:00.0000] <Hixie> wfm [20:49:00.0000] <boblet> Hey all, I’d like to ask about the id attribute ‘opaque strings’ comment (3.3.3.1); what is meant by ‘Particular meanings should not be derived from the value of the id attribute’? Is this intended for browser makers? [20:49:01.0000] <Hixie> yes [20:49:02.0000] <Hixie> it means that id="header" doesn't mean the element is a header [20:50:00.0000] <Hixie> it could just as easily be a footer, or a link, or a duck [20:51:00.0000] <boblet> aah, that makes sense. So as authors we’d be encouraged to use meaningful names (as the class attribute spec specifically states), but that meaning shouldn’t affect how the browser treats the element with the id huh [20:52:00.0000] <Hixie> yeah [20:52:01.0000] <Hixie> if you can come up with a better way for the spec to say that, drop me a line :-) [20:54:00.0000] <boblet> How would this sit with eg microformats? eg if the text specifically said “User agents should not derive particular meanings from the value of the id attribute” (which would remove the ambiguity for me), wouldn’t this also be problematic for eg Operator, Optimus et al? [20:55:00.0000] <boblet> Hrm, perhaps adding your example “eg id="header" should not be interpreted as <header>” is clearest? [20:58:00.0000] <Hixie> do they use id="" attribute values? [20:58:01.0000] <Hixie> i'd rather not define things by examples, though i could certainly _add_ examples if you want [20:59:00.0000] <Hixie> drop me a mail if you do [21:01:00.0000] <boblet> Microformats plugins/bookmarklets just look for the matching value, which could be class="vcard" or id="vcard". Using class="" is most common, but it’s possible an author might use id="". I’ll email you these comments. Thanks for the clarification [21:06:00.0000] <Hixie> k [21:13:00.0000] <othermaciej> I think the justification I just sent for the Design Principles may be longer than the actual Design Principles [21:14:00.0000] <othermaciej> not sure how to feel about this [21:16:00.0000] <ezyang> /me turns on obnoxious flag [21:16:01.0000] <ezyang> Anyone want to update the html5lib status page? [21:16:02.0000] <ezyang> /me turns off obnoxious flag [21:26:00.0000] <othermaciej> aaaand I self-replied twice [21:26:01.0000] <othermaciej> should have planned more before sending [23:29:00.0000] <Hixie> othermaciej_: you should just put your e-mail into the design principles [23:30:00.0000] <othermaciej_> all of it? [23:30:01.0000] <Hixie> more or less [23:30:02.0000] <Hixie> i mean it'll need editing [23:30:03.0000] <othermaciej> yeah [23:30:04.0000] <Hixie> but it would give the context people are asking for [23:30:05.0000] <Hixie> even if they don't agree with it [23:30:06.0000] <othermaciej> obviously the part where I say which principles are better or worse shouldn't be pasted in [23:30:07.0000] <Hixie> sure [23:30:08.0000] <Hixie> i'd just drop the ones that didn't work [23:31:00.0000] <Hixie> or fix them so they can [23:31:01.0000] <othermaciej> sections 1 and 2 seem like the best fits [23:31:02.0000] <othermaciej> though they are awfully long [23:32:00.0000] <othermaciej> anyway, I guess that seems like a good idea [23:32:01.0000] <othermaciej> or at the very least I could write an introductory section that addresses the same points [23:34:00.0000] <othermaciej> by the way, the prediction that "you can't please everyone" would draw objections was fulfilled [23:34:01.0000] <othermaciej> though I was surprised that the objection came from Sam [00:37:00.0000] <heycam> if i read the spec correctly, the load event dispatched once a document is loaded does not bubble. (same for most (all?) other "fire a simple event" things.) [00:37:01.0000] <hsivonen> the HTML5 version indictor is charset=utf-8, mmkay? [00:39:00.0000] <heycam> but event handler attributes (including onload="") are listeners for bubbling [00:39:01.0000] <heycam> unless: bubbling listeners get AT_TARGET phase events, and events that don't bubble are still dispatched AT_TARGET [00:39:02.0000] <heycam> which i think is true [00:40:00.0000] <heycam> so never mind [00:52:00.0000] <othermaciej> seems so quiet here tonight [01:00:00.0000] <hsivonen> could a html5lib admin add my hsivonen⊙if account to the project in addition to or in place of the @gmail.com account? [01:16:00.0000] <annevk42> heycam, that's true [01:17:00.0000] <annevk42> hsivonen, done [01:17:01.0000] <hsivonen> annevk42: thank you [01:22:00.0000] <hsivonen> $ hg outgoingabort: Python support for SSL and HTTPS is not installed [01:22:01.0000] <hsivonen> what macports incantation do I need? [01:24:00.0000] <hsivonen> how do I make hg tell me which python interpreter it is using? [01:25:00.0000] <Hixie> use lsof :-) [01:29:00.0000] <hsivonen> is there a way to leave lsof waiting so that I can first start lsof and then start hg? [01:33:00.0000] <jgraham> hsivonen: Does less `which hg` followed by reading the line at the top not answer your question? [01:34:00.0000] <hsivonen> jgraham: yes. thanks [01:34:01.0000] <hsivonen> for the log: the incantation is sudo port install py25-socket-ssl [01:35:00.0000] <hsivonen> but now I get a 404 for hg outgoing [01:35:01.0000] <hsivonen> hg outgoing https://hsivonen⊙if@html5lib.googlecode.com/hg/abort: HTTP Error 404: Not Found [01:36:00.0000] <hsivonen> /me wonders if Google's hg support has a bug with non-Gmail accounts [01:36:01.0000] <hsivonen> with the two @ signs [01:37:00.0000] <zcorpan> Hixie: "CSS properties respectively" is marked up as a dfn [01:37:01.0000] <jgraham> hsivonen: s/@/%40/ [01:37:02.0000] <jgraham> hsivonen: They do have a bug [01:37:03.0000] <hsivonen> jgraham: thanks [01:38:00.0000] <jgraham> (only the first @) [01:38:01.0000] <Hixie> zcorpan: thanks [01:39:00.0000] <hsivonen> jgraham: how should I escape % in hgrc? Python interprets % as something magic [01:39:01.0000] <hsivonen> %%? [01:39:02.0000] <hsivonen> yes, %% [01:40:00.0000] <hsivonen> yay for overlapping escapes [01:49:00.0000] <Hixie> holy crap, shelley managed to write an entire blog post about html5 without saying i was wrong about anything [01:49:01.0000] <othermaciej> no wai [01:50:00.0000] <Hixie> i really don't understand why people have so much trouble with the design principles [01:50:01.0000] <Hixie> are people really that desperate for black-and-white rules? [01:50:02.0000] <Hixie> language design isn't something you do by following a flowchart [01:50:03.0000] <Hixie> you have to weigh all the various goals, etc [01:51:00.0000] <othermaciej> haha, she hates the Design Principles because they are against RDFa [01:51:01.0000] <Hixie> i just don't get why every time we find the need to create new technologies, someone says "you're ignoring the don't reinvent the wheel principle", especially when we documented why we were doing so for that feature [01:52:00.0000] <Hixie> or why every time any one principle is overriden by some other concern, people assume that means we should always ignore that principle [01:52:01.0000] <Hixie> e.g. i mention that the design principles are merely guidelines, and shelley's response is to tell hsivonen that he's wrong about caring about dom consistency [01:52:02.0000] <Hixie> but that's just stupid [01:52:03.0000] <Hixie> because you have to care about dom consistency [01:53:00.0000] <Hixie> just like you have to care about everything else [01:53:01.0000] <Hixie> why do these people find this so challenging to grasp? [01:54:00.0000] <othermaciej> I am not sure [01:56:00.0000] <ojan> where is this thread going on? [01:57:00.0000] <othermaciej> I don't know how to explain it any further without getting repetitive [01:57:01.0000] <othermaciej> but yeah [01:57:02.0000] <Hixie> ojan: public-html is where most of it is, shelley's blog was what most recently triggered by confusion though [01:57:03.0000] <othermaciej> there could be valid reasons to override the DOM Consistency principle [01:58:00.0000] <Hixie> sure [01:58:01.0000] <Hixie> just look at the xml:lang disaster [01:58:02.0000] <othermaciej> for example, the need to Support Existing Content leads to some differences already [01:58:03.0000] <Hixie> we had to override it for that [01:58:04.0000] <othermaciej> right, like xml:lang vs lang [01:58:05.0000] <Hixie> xml:lang is the main reason we HAVE the principle, in fact :-P [01:58:06.0000] <Hixie> so we never have to repeat that disaster [01:58:07.0000] <zcorpan> would be nice to drop xml:lang and make lang equivalent to xml:lang for all elements [01:59:00.0000] <zcorpan> likewise with id [01:59:01.0000] <othermaciej> but just because you feel like it is not a good reason to override it [01:59:02.0000] <Hixie> holy crap! windows 7 just launched in a vm in the background and played its startup sound at full volume in stereo, sounded like a spaceship had landed in my hall or something! [01:59:03.0000] <othermaciej> and I would even say your counter-argument doesn't have to be based on any Design Principle [01:59:04.0000] <othermaciej> but you have more odds of being convincing if it is [01:59:05.0000] <Philip`> /me tried to explain HTML5's lang vs xml:lang rules to someone implementing an RDFa parser, and they didn't seem to entirely think it was a lovely elegant idea [02:00:00.0000] <Hixie> well people have clearly failed to notice that i never refer to the design principles explicitly [02:00:01.0000] <Hixie> and that i've been applying them completely independent of whether we have the design principles doc or not [02:00:02.0000] <Philip`> (I think the spec is probably less confusing now than it was a few days ago, though) [02:01:00.0000] <Hixie> the spec doesn't mention the string "xml:lang" any more (with a couple of exceptions) [02:01:01.0000] <Hixie> which might make it clearer [02:01:02.0000] <othermaciej> I think the Design Principles are more useful for those who disagree with the direction of the spec than those who largely agree [02:02:00.0000] <zcorpan> Hixie: i guess now you can remove the part in terminology starting "When an XML name, such as an attribute or element name, is referred to in the form prefix:localName" [02:02:01.0000] <Hixie> zcorpan: really? there's none left? [02:02:02.0000] <Hixie> othermaciej: yeah, but i don't think they see it that way :-) [02:02:03.0000] <zcorpan> Hixie: i think xml:lang was all there ever was (ignoring examples in that section) but i might be wrong [02:03:00.0000] <Hixie> i'm sure there are others [02:03:01.0000] <zcorpan> xml:base maybe [02:06:00.0000] <zcorpan> "Similarly, if the newly created element has an xmlns:xlink attribute in the XMLNS namespace whose value is not the XLink Namespace, that is a parse error." [02:06:01.0000] <zcorpan> is "xmlns:xlink" the local name or qualified name? [02:06:02.0000] <Hixie> qalified [02:07:00.0000] <Hixie> qualified, even [02:07:01.0000] <Hixie> it gets complicated with xmlns: attributes [02:07:02.0000] <Hixie> e.g. xmlns="" is in the XMLNS namespace iirc [02:07:03.0000] <annevk42> yes it is [02:07:04.0000] <hsivonen> Hixie: WCAG, unlike design principles, have litigable precision for exceptions [02:08:00.0000] <hsivonen> WCAG 2.0 that is [02:08:01.0000] <Hixie> WCAG 2.0 has precision? [02:08:02.0000] <Hixie> of any kind? [02:08:03.0000] <hsivonen> except where WCAG is incomprehensible without informative supporting materials [02:08:04.0000] <Hixie> good to know [02:08:05.0000] <annevk42> you can make things somewhat confusing too [02:08:06.0000] <annevk42> data:text/xml,<test x:xmlns="x" xmlns:x="x"/> [02:08:07.0000] <Hixie> that's only confusing because the order of the attributes is backwards :-P [02:09:00.0000] <zcorpan> "PS: Interestingly, this behavior could be used to send xhtml (and get it shown on IE) without sending different headers to each browser: Content-Type: text/html Content-Type: application/xhtml+xml" [02:09:01.0000] <zcorpan> why hasn't anyone thought of that before? [02:10:00.0000] <annevk42> violates HTTP? [02:10:01.0000] <Philip`> Because it's insane? [02:10:02.0000] <Hixie> does apache even let you do that? [02:10:03.0000] <othermaciej> I was thinking if there continues to be debate on whether the principles should be absolute, I could cite some examples from other domains [02:10:04.0000] <zcorpan> the PHP content negotiation scripts also violate HTTP and are insane [02:10:05.0000] <hsivonen> zcorpan: I see a new meme for all the cool people coming up [02:10:06.0000] <Philip`> and perhaps unlikely to survive proxies [02:10:07.0000] <othermaciej> Math: Every natural number is either prime or composite - except 1. [02:11:00.0000] <zcorpan> Hixie: i guess not in a simple way [02:11:01.0000] <othermaciej> Perl: There's more than one way to do it. -- but sometimes there is only one way, does this undermine the very foundations of Perl? [02:11:02.0000] <Hixie> when is there only one way?! [02:11:03.0000] <Hixie> that i want to see [02:11:04.0000] <othermaciej> Python: none of the principles are absolute, all are extremely vague: http://www.python.org/dev/peps/pep-0020/ [02:12:00.0000] <othermaciej> Hixie: I guess it would depend on what the definition of "one" is [02:12:01.0000] <Philip`> If there's one way to do it, call that way X; then you can do eval("X") and now there's two ways to do it [02:12:02.0000] <Hixie> even by a reasonable definition such that Philip`'s two things are considered the same [02:12:03.0000] <othermaciej> I think there's only one way to add two numbers, short of adding no-op layers of wrapping [02:12:04.0000] <jgraham> I think the difference between the python "design principles" and the HTML ones is rather telling [02:12:05.0000] <Hixie> adding two numbers... hmm [02:12:06.0000] <philipj> Hixie, where do keep that list of mails you have on your backlog where one can vote for important issues (but doesn't)? Google fails me. [02:13:00.0000] <Philip`> You can add two numbers with a lookup table [02:13:01.0000] <Hixie> philipj: whatwg.org/issues [02:13:02.0000] <othermaciej> jgraham: telling in what way? [02:13:03.0000] <philipj> thanks [02:13:04.0000] <Hixie> philipj: but if you have a specific request, just let me know :-) [02:13:05.0000] <Hixie> philipj: it'll be quicker for you :-) [02:13:06.0000] <othermaciej> jgraham: I wish we had enough community consensus that our design principles could be one-liners [02:13:07.0000] <philipj> just wondering if you forgot about the media resource selection algorithm feedback [02:13:08.0000] <othermaciej> but with a higher level of dissent, it's more useful to add some explanation [02:14:00.0000] <jgraham> othermaciej: e.g. the line "Although that way may not be obvious at first unless you're Dutch. [02:14:01.0000] <Hixie> philipj: "In the search loop, step 6 can never be reached because step 5 either goes to 1 or aborts the task. Simply removing "Otherwise, abort the task." should [02:14:02.0000] <Hixie> do the trick." [02:14:03.0000] <Hixie> philipj: ? [02:14:04.0000] <Hixie> philipj: or something else? [02:14:05.0000] <hsivonen> Hixie: example of WCAG 2.0 precision: zooming has a precise percent requirement: 200% [02:14:06.0000] <othermaciej> jgraham: you mean the fact that readers are expected to have a sense of humor? [02:15:00.0000] <philipj> not only that, but the part about async/sync inconsistencies with when src attr or <source> element is selected [02:15:01.0000] <annevk42> "Although that way may not be obvious at first unless you're Dutch." [02:15:02.0000] <hsivonen> Hixie: example of WCAG 2.0 imprecision: you can't tell without supporting materials if browser-native zoom satisfies the req [02:15:03.0000] <annevk42> Python is nice [02:15:04.0000] <Hixie> philipj: i thought i fixed that [02:15:05.0000] <philipj> let me have a look [02:15:06.0000] <Hixie> philipj: oh you mean src vs <source> [02:15:07.0000] <Hixie> philipj: i don't know of a way to fix that [02:15:08.0000] <othermaciej> annevk2: do you find it obvious? [02:16:00.0000] <Hixie> philipj: i have an e-mail from you starting "This is an improvement in predictability" [02:16:01.0000] <philipj> Hixie, is there a problem with the solution I suggested? [02:16:02.0000] <jgraham> othermaciej: The fact the community accepts humour and the fact that one of the design principles points out that perfect design is ulikely to happen [02:16:03.0000] <hsivonen> Hixie: my favorite is "Interruptions can be postponed or suppressed by the user, except interruptions involving an emergency." [02:16:04.0000] <annevk42> That I always make the right design decisions when writing Python because I'm Dutch seems very obvious and logical to me :) [02:16:05.0000] <othermaciej> heh [02:17:00.0000] <Hixie> philipj: i guess it might work, i'll have to look closer [02:18:00.0000] <philipj> I've implemented basically by ignoring step 5 "let src equal the first value that was assigned" and some minor changes [02:18:01.0000] <Hixie> philipj: k. i expect we can make that work. i'll get to it probably before the end of the month (unless you need the spec updated before then) [02:19:00.0000] <philipj> well as I said it's all in my mail, I just wanted to check if you had it in your backlog [02:19:01.0000] <philipj> there's no big rush as far as I'm concerned [02:20:00.0000] <Hixie> othermaciej: lol "Namespaces are one honking great idea -- let's do more of those!" [02:20:01.0000] <Hixie> philipj: yup, definitely got it [02:21:00.0000] <othermaciej> Hixie: not applicable in our context! [02:23:00.0000] <jgraham> With that exception (and a few others), I actually think that a lot of the Python design principles are things that we should be striving for [02:25:00.0000] <annevk42> just due to sheer complexity of the Web platform we sort of fail on "If the implementation is hard to explain, it's a bad idea." [02:25:01.0000] <jgraham> annevk42: But we should aim for that in new features [02:25:02.0000] <jgraham> At least at a high level [02:26:00.0000] <jgraham> I mean details of web security and so on will always be hard to explain so the details of new features will never be obvious [02:26:01.0000] <Philip`> /me wonders if there is a principle like "ignore principles that seem inconveniently contrary to our view of the world (e.g. ones that promote namespaces)" [02:27:00.0000] <annevk42> Philip`, I guess that principle is that not all programming principles apply to those of markup languages [02:28:00.0000] <Philip`> How does one decide which ones do apply? [02:29:00.0000] <jgraham> Philip`: Given an arbitary set of possibe principles? [02:30:00.0000] <jgraham> The ones that make sense based on experience, etc. If you get it wrong then it just means that the result will likely be bad [02:30:01.0000] <annevk42> Philip`, somewhat informed judgment? [02:30:02.0000] <othermaciej> jgraham: many of the new features fail [02:30:03.0000] <othermaciej> jgraham: for instance Workers - no way any implementation is going to be easy to explain [02:31:00.0000] <othermaciej> but hopefully the interface can remain easy to explain [02:32:00.0000] <annevk42> the difference is that Python is just a programming language, not so much an API to all kinds of features [02:32:01.0000] <Philip`> It's an API to an interpreter [02:33:00.0000] <annevk42> (and Python is also going to violate that principle once they get further with JIT'ing) [02:33:01.0000] <jgraham> othermaciej: Possibly that means that workers is too complex. Possibly it just means that the complexity is inevitable if you want that feature. [02:33:02.0000] <jgraham> But if we had a principle like that we would at least look to minimise internal complexity whereever possible [02:33:03.0000] <jgraham> (DOM consistency is rather similar) [02:34:00.0000] <jgraham> (but obviously more specific) [02:34:01.0000] <Hixie> i wonder if julian realises that complaining that i haven't replied to his e-mail is delaying my replying to his e-mail because i end up reading and responding to his e-mails complaining about it... [02:38:00.0000] <othermaciej> jgraham: I think there is inherent implementation to concurrency [02:39:00.0000] <othermaciej> but also, the message-passing model should have the least implementation complexity of various concurrency models [02:40:00.0000] <othermaciej> however I think Workers / MessagePorts may have more complexity than expected, as we are learning in the course of implementing in WebKit [02:40:01.0000] <othermaciej> jgraham: there is Avoid Needless Complexity [02:41:00.0000] <othermaciej> jgraham: but in general, I think it's more important to avoid interface complexity than implementation complexity [02:41:01.0000] <Hixie> i should go to bed [02:41:02.0000] <othermaciej> unless the implementation complexity required is truly excessive, and beyond what is reasonable for the feature in question [02:43:00.0000] <hsivonen> I should try to invoke Avoid Needless Complexity and Solve Real Problems against table taint [02:44:00.0000] <hsivonen> having a writable datum in stack nodes means speculative parsing needs to copy the nodes instead of holding references to the nodes from different speculations [02:45:00.0000] <hsivonen> woohoo! <datagrid> content model got sane [02:45:01.0000] <zcorpan> hsivonen: hmm i can't reopen your bug, but now it has more info [02:46:00.0000] <hsivonen> zcorpan: which bug? [02:46:01.0000] <zcorpan> http://www.w3.org/Bugs/Public/show_bug.cgi?id=6978 [02:47:00.0000] <zcorpan> maybe it should be a per-document flag so that when it gets set, all further </script> tags are omitted in the serialization [02:48:00.0000] <zcorpan> but you could still make it run by inserting <style></script></style> [02:48:01.0000] <zcorpan> so not worth it i guess [02:48:02.0000] <zcorpan> maybe the whole script should be omitted instead [02:49:00.0000] <zcorpan> i.e. instead of serializing <script>contents, serialize <script></script> (or nothing) [02:49:01.0000] <zcorpan> i guess nothing is better considering <script src> [02:50:00.0000] <zcorpan> maybe script elements shouldn't be inserted in the dom until the end tag is seen? [02:50:01.0000] <zcorpan> during parsing [02:52:00.0000] <hsivonen> zcorpan: Deferring insertion would be annoying [02:52:01.0000] <zcorpan> yeah [02:54:00.0000] <Hixie> what is this solving, anyway [02:56:00.0000] <zcorpan> does IE do it? [02:59:00.0000] <zcorpan> nope [02:59:01.0000] <zcorpan> that opera does it might be a side-effect of wanting to be ie-compatible [03:00:00.0000] <zcorpan> though we emit more tags than ie does at the moment [03:01:00.0000] <Hixie> ok well i'm going to bed. [03:01:01.0000] <Hixie> nn. [03:08:00.0000] <hsivonen> Hixie: nn [03:08:01.0000] <othermaciej> gnight Hixie [03:21:00.0000] <annevk42> /me thought Hixie was going to bed [03:21:01.0000] <annevk42> LIES! [03:23:00.0000] <othermaciej> /me wonders if chaals will eventually conclude that he's not saving any time by repeatedly making his same argument that getting rid of the design principles would save time [03:24:00.0000] <annevk42> time will tell -- harharhar [03:35:00.0000] <othermaciej> did my explanation/justification of the document make sense btw? [03:35:01.0000] <othermaciej> I know it ended up a bit close to TL;DR territory [03:40:00.0000] <annevk42> ok, let that be the last thing I read before I start packing for the CSS meeting [03:40:01.0000] <annevk42> /me has a flight in 3 hours [03:44:00.0000] <othermaciej> ah, wouldn't want to miss that [03:45:00.0000] <zcorpan> Hixie: <a href="©="> is equivalent to <a href="©="> [03:46:00.0000] <annevk42> othermaciej, looks good [03:46:01.0000] <annevk42> othermaciej, I agree with Hixie that some of that should probably go into the introduction [03:46:02.0000] <othermaciej> annevk42: yeah, that suggestion made sense to me [03:46:03.0000] <annevk42> othermaciej, I also I agree with removing and adding principles as suggested in the end [03:46:04.0000] <othermaciej> annevk42: I will see what happens in the email discussion for a day or two before I do that [03:47:00.0000] <othermaciej> (either of those things) [03:47:01.0000] <annevk42> cool [05:10:00.0000] <zcorpan> my browsers have less spacing for <isindex> than what the spec gives (with implying ) [05:16:00.0000] <zcorpan> <isindex type=checkbox> [05:21:00.0000] <zcorpan> the spec matches ie at least [05:22:00.0000] <zcorpan> except for the p and the label [05:25:00.0000] <Mrmil> mmm coffee... [05:51:00.0000] <gsnedders> ezyang: http://stuff.gsnedders.com/phphtml5lib/ [06:22:00.0000] <Philip`> gsnedders: Why JPEG? [06:24:00.0000] <Philip`> Also, how did it spend 101.70% of its time inside {main}? [06:31:00.0000] <gsnedders> Philip`: Because it's sanely small :P [06:31:01.0000] <gsnedders> Philip`: Because it's not 100% reliable :P [06:40:00.0000] <zcorpan> is it 101.70% reliable? [06:40:01.0000] <gsnedders> zcorpan: Totally. [06:44:00.0000] <Dashiva> Well, ponder that statement. Is something that is 101.70% reliable a) more than 100% reliable, b) the same as 100% reliable, it makes no sense to speak of more, or c) less than 100% reliable, because it makes no sense to speak of more [06:48:00.0000] <Lachy> Dashiva, it means it performed better than expected or possible [06:56:00.0000] <Lachy> I'm a little confused about the changes regarding conformance of unescaped ampersands that are being discussed on the list. AFAICT, the spec still says that unescaped ampersands in attribute values are still ambiguous [06:56:01.0000] <Lachy> is it only the parsing section that has been changed to not emit a parse error? [06:57:00.0000] <zcorpan> Lachy: "or, if the character is in an attribute value, one or more characters in the ranges U+0030 DIGIT ZERO .. U+0039 DIGIT NINE, U+0041 LATIN CAPITAL LETTER A .. U+005A LATIN CAPITAL LETTER Z, and U+0061 LATIN SMALL LETTER A .. U+007A LATIN SMALL LETTER Z, followed by a U+003D EQUALS SIGN character (=)." [06:57:01.0000] <hsivonen> now I'm confused whether "RT" on twitter means re-tweet or reply to [06:57:02.0000] <hsivonen> I thought it was the former, but I see usage that looks like the latter [06:57:03.0000] <Lachy> hsivonen, it's supposed to mean re-tweet [06:58:00.0000] <hsivonen> ok. good [06:58:01.0000] <Lachy> some people must be misusing it [07:00:00.0000] <Lachy> that part zcorpan just quoted is what confused me. It's quite difficult to comprehend [07:02:00.0000] <Philip`> /me has non-fun trying to get Perl and Python implementations of YAML to interoperate [07:02:01.0000] <Lachy> and, after rereading it, I think it means that href="foo&bar=" not ambiguous, but href="foo&bar" is because it explicitly says it needs to be followed by an equals sign [07:02:02.0000] <Philip`> (particularly for handling Unicode characters) [07:02:03.0000] <Lachy> (though, I would'nt be surprised if I still got that wrong, since the paragraph is so badly written) [07:12:00.0000] <gsnedders> /me blames Hixie [07:16:00.0000] <Philip`> Doesn't it just mean an ampersand is ambiguous if it doesn't match /&(\s|<|&|[0-9A-Za-z]+=)/? [07:16:01.0000] <Philip`> (which would be odd because & would be defined as ambiguous, I guess) [07:19:00.0000] <jgraham> Philip`: Do they interoperate on non-Unicode characters? [07:20:00.0000] <Philip`> jgraham: Not quite [07:20:01.0000] <gsnedders> /me wonders how long `git blame -L '/>Tokeni/,/>Tree construct/' source` will take to run [07:20:02.0000] <Philip`> One library outputs strings like "foo \ [07:20:03.0000] <Philip`> bar" [07:20:04.0000] <Philip`> (using the slash to escape the newline because it's just there for pretty wrapping) [07:20:05.0000] <Philip`> and a different library parses that as a literal \ in the string [07:21:00.0000] <Philip`> I should probably use a different interchange format, like JSON [07:21:01.0000] <Philip`> but I think it works now so I'm not going to change it [07:27:00.0000] <jgraham> http://blog.digitalbazaar.com/2009/05/31/admitting-that-javascript-was-a-mistake/ makes the exciting claim that Javascript was a mistake because it is a functional programming language [07:36:00.0000] <Philip`> jgraham: and that we should use a procedural language like Python instead [07:37:00.0000] <Lachy> "Debugging in general (which would be impossible without Firebug)." -- clearly, that's not true. Venkman was a great debugger well before firebug [07:44:00.0000] <Lachy> using python in browsers instead of JavaScript would be an even bigger mistake, cause then we'd have to deal with the insane, unintuitive design of python [07:54:00.0000] <Philip`> Lachy: Why is that worse than the insane, unintuitive design of JavaScript? [07:54:01.0000] <jgraham> Lachy: There is really nothing wrong with Python [07:54:02.0000] <jgraham> It's not really designed to go in browsers though [07:54:03.0000] <Philip`> I imagine one serious problem would be either the language would fork, or the language in one environment (in-browser vs out-of-browser) would have to make huge sacrifices for the other [07:55:00.0000] <Philip`> where nobody sensible uses JS outside of browsers, so there aren't widely-diverging demands on it [07:55:01.0000] <jgraham> Philip`: Indeed. JS-as-a-language has basically taken success in the browser at the expense of no success outside the browser [07:56:00.0000] <Lachy> Philip`, the design of javascript is more intuitive than python, though I admit it does have some concepts like closures and prototype chains that can be difficult to grasp [07:57:00.0000] <jgraham> Lachy: I would toatlly love evidence backing up that statement [07:57:01.0000] <Lachy> jgraham, which part of the statement? The complexity of closures and prototype chains, or that it's more intuitive than python? [07:58:00.0000] <jgraham> Because I think basically exactly the opposite; python has a few tough features but is basically sane whereas js is basically complex all the way through [07:58:01.0000] <jgraham> Lachy: The intuitivemenss part [07:59:00.0000] <Philip`> I think Perl is sane, so I don't have problems with either JS or Python [08:01:00.0000] <jgraham> As a random example javascript will allow you to do Constructor() (without "new") to execute a constructor function and not only will you not get the Object you were expecting back you will probably mess up the global namespace [08:02:00.0000] <Lachy> calling Constructor() without "new" is just a regular function call [08:02:01.0000] <jgraham> Lachy_: That's the point [08:02:02.0000] <jgraham> function Construct() {this.a = 2}; b = Construct() [08:02:03.0000] <jgraham> Sets a global variable a to 2 [08:04:00.0000] <Lachy> as an example of a completely backwards and unintuitive feature in python, to join an array of strings with some delimiter, instead of calling array.join(delim), you have to call "delim".join(array) [08:05:00.0000] <Lachy> and they use weird names like lists and dictionaries instead of array [08:06:00.0000] <jgraham> Lachy: I thought that was crazy untill I realised that doing an operation on a string and getting a string back is more consistent than doing an operation on an array and getting a string back [08:07:00.0000] <Lachy> no, doing an operation on the object itself (an array in this case) to obtain some output makes much more sense from an object oriented programming perspective [08:07:01.0000] <jgraham> Lachy: They are actually rather good names. "array" implies things that aren't necessarily true [08:07:02.0000] <Lachy> like what? [08:07:03.0000] <jgraham> (moreso in js than in python even) [08:08:00.0000] <Philip`> Like O(1) access time [08:08:01.0000] <jgraham> Lachy: a = new Array(); a[Math.pow(2,32)-1] = 1; [08:09:00.0000] <jgraham> Lachy: Ho much memory does that use and how fast would it be enumerated? [08:09:01.0000] <jgraham> s/Ho/How/ [08:09:02.0000] <Dashiva> -2, surely [08:10:00.0000] <jgraham> Dashiva: Oh yeah, sorry [08:11:00.0000] <Philip`> Why -2? [08:11:01.0000] <Dashiva> Otherwise .length can overflow [08:12:00.0000] <Philip`> Ah [08:12:01.0000] <jgraham> Philip`: Array index properties have to be < 2^32 -1 [08:12:02.0000] <Lachy> indeed, it does overflow. a.length returns 0 in the first case [08:12:03.0000] <jgraham> Lachy: That's a bug [08:13:00.0000] <jgraham> Or maybe I'm wrong. [08:13:01.0000] <Lachy> Both Opera and Firefox do the same thing [08:14:00.0000] <jgraham> Yeah it is common behaviour. I thought it was a bug specwise but maybe something that has been adopted as common behaviour [08:15:00.0000] <Dashiva> It's a bug specwise, yes. But what does that matter :P [08:15:01.0000] <Lachy> jgraham, anyway, in answer to your question, if you created a list of the same length in python, how much memory would it use and how fast would it be enumerated? [08:16:00.0000] <Philip`> You can't assign to random indexes, without allocating the space first [08:16:01.0000] <Philip`> (using append, extend, etc) [08:19:00.0000] <Lachy> Philip`, sure, I'm not concerned about the specific method used to create the list of that length. My question was just pointing that the the performance question also applies to python under the same circumstances [08:20:00.0000] <Dashiva> I think his point is that in Python some of the things implicit in JS are made explicit [08:20:01.0000] <Lachy> the implicit features in JS are a nice feature to have [08:21:00.0000] <Philip`> Lachy: The same circumstances don't exist in Python [08:21:01.0000] <Lachy> Philip`, surely you can create a list of that length in some way though [08:22:00.0000] <Philip`> Lachy: Only if you fill in all the intermediate values [08:22:01.0000] <Philip`> Lachy: which is different to what the JS example is doing [08:23:00.0000] <Lachy> well, than that's a point to JS then cause being able to create arrays of that length without needing to fill in all the intermediate values explicitly, at least until they need a non-undefined value, is a feature [08:25:00.0000] <Philip`> I think jgraham's point was that that feature in JS clearly can't be implemented with the data structure that's typically called an array [08:26:00.0000] <jgraham> Lachy: It is not at all obvious that something called an "Array" would actually be a hashtable, or something similar under the hood [08:26:01.0000] <Philip`> and so it's misleading to call it "Array" in JS [08:28:00.0000] <Lachy> the underlying technical implementation is really irrelevant to whether or not it's an array [08:30:00.0000] <Lachy> the fact that it can be used as a stack, queue, hash table or array may be a reason to not call it just an array though, but that's a very minor issue [08:30:01.0000] <Philip`> The underlying technical implementation is critical to the distinction between arrays and lists and hashtables and hybrids [08:31:00.0000] <Philip`> and people should understand the distinctions, so they don't write silly code that has terrible performance [08:31:01.0000] <jgraham> Lachy: Also there are plenty of thother things; var a = {}; a[1]="a"; a["1"] = "b"; //oops I overwrote my key [08:32:00.0000] <jgraham> The fact the assignment creates globals [08:32:01.0000] <jgraham> The with statement [08:33:00.0000] <Philip`> Optional semicolons [08:33:01.0000] <Lachy> yes, the with statement is a mistake, just like the goto and break statements in other languages [08:33:02.0000] <jgraham> eval [08:33:03.0000] <jgraham> (Yes I know Python has eval but it is almost never used) [08:34:00.0000] <jgraham> (and it is much worse in a web context) [08:34:01.0000] <Lachy> people who use eval in javascript are usually clueless programmers (or they're using it for JSON parsing, despite the security problems) [08:36:00.0000] <Philip`> I know someone who used eval("a."+b) because they weren't aware that a[b] was equivalent [08:36:01.0000] <jgraham> the way that enumeration works [08:36:02.0000] <jgraham> A bunch of stuff that should get better in ES5 wrt defining object properties [08:37:00.0000] <Rik|work> Philip`: I believe it's one of the most reason people use eval [08:37:01.0000] <jgraham> the fact that strings, booleans and numbers have two different representations [08:37:02.0000] <Philip`> The use of '+' for both string concatenation and addition [08:37:03.0000] <jgraham> One of which is totally useless except for spec-internal algorithms [08:37:04.0000] <gsnedders> How many parse errors should "<!DOCTYPE" produce? I, and the PHP impl., think 1 (as of the change in spec of 2009-05-26). The test case says 2. [08:37:05.0000] <gsnedders> (Likely answer: test case is unchanged.) [08:38:00.0000] <gsnedders> s/26/24/ [08:38:01.0000] <jgraham> Philip`: Never had a problem with that but the implicit type shifiting is odd [08:39:00.0000] <jgraham> also == vs === [08:39:01.0000] <jgraham> Object equality [08:39:02.0000] <jgraham> [1,2,3] != [1,2,3] [08:42:00.0000] <gsnedders> What should "<a a=a&>" output? [08:42:01.0000] <gsnedders> (From the tokenizer) [08:47:00.0000] <gsnedders> "If no match can be made, then no characters are consumed, and nothing is returned. If the current state is not the character reference in attribute value state, or the U+0026 AMPERSAND character is not followed by one or more characters in the ranges U+0030 DIGIT ZERO to U+0039 DIGIT NINE, U+0041 LATIN CAPITAL LETTER A to U+005A LATIN CAPITAL LETTER Z, and U+0061 LATIN SMALL LETTER A to U+007A LATIN SMALL LETTER Z, fo [08:47:01.0000] <gsnedders> by a U+003D EQUALS SIGN character (=), then this is also a parse error." [08:47:02.0000] <gsnedders> That really doesn't read well. [08:49:00.0000] <gsnedders> So in the character reference in attribute value state then "&foo=" is a parse error, but "&foo" is not, right? [08:49:01.0000] <gsnedders> Wrong way round. [08:50:00.0000] <gsnedders> "&foo=" is not a parse error, "&foo>" is. [08:52:00.0000] <gsnedders> "<a a=a&>" is right in the test, my impl. is not. [08:54:00.0000] <Lachy> http://www.molly.com/2009/06/02/why-xhtml2/ [08:56:00.0000] <beowulf> the listed advantages are not familiar to me as an html author [08:57:00.0000] <beowulf> actually, i want to smash things now [08:57:01.0000] <Lachy> http://twitter.com/mollydotcom/status/2004173388 [08:57:02.0000] <Lachy> apparently, there's a new version of that article coming [08:58:00.0000] <Lachy> "Search engines are ‘user agents’ that don’t support frames!" - I've never once seen any evidence to support that claim [08:59:00.0000] <Lachy> in fact, the first result returned is a framed page: http://www.google.com/search?q=Java+docs [08:59:01.0000] <Lachy> so it's a bogus claim [09:00:00.0000] <beowulf> an advantage of xhtml2 to an html author is "CSS gives more presentational possibilities than plain HTML" [09:00:01.0000] <beowulf> i know you all think html authors are stupid, but that's insulting [09:01:00.0000] <Lachy> yeah, I read that and wondered why a general advantage of using CSS was listed as one specifically benefiting XHTML2 [09:02:00.0000] <beowulf> "Visible on more devices" # gah [09:03:00.0000] <Lachy> of course, that claim is based in a theoretical future world in which support for XHTML2 has been added to devices and the number outweighs the number of devices that support HTML today [09:04:00.0000] <jgraham> The claim is based on the erronous assumption that it is not possible to parse HTML on small devices [09:05:00.0000] <jgraham> WWhich is just crazy given that there are small devices shipping with HTML parsers today and even smaller devices that render html content using e.g. mini [09:05:01.0000] <beowulf> that claim is based on never having written anything for a small device [09:06:00.0000] <beowulf> and most of those claims are based on never having worked on html more complicated than a w3c page [09:06:01.0000] <Lachy> later tonight, after Molly posts the proper version of that post, I will blog a detailed rebuttal of all his bogus points [09:09:00.0000] <beowulf> i find that whole 'less presentation, more structure' section insulting [09:11:00.0000] <Lachy> LOL. I missed this one the first time through: "Advantages: All the ones that you can imagine, because XML is a Good Thing" [09:12:00.0000] <beowulf> fair enough, i don't mind if it's xml or text, but some of those points are written by someone who's guessing what most people do, or so it would seem to my dumb html author mind [09:15:00.0000] <Lachy> beowulf, it's written by Steven Pemberton [09:16:00.0000] <beowulf> Lachy: who's he? [09:17:00.0000] <Lachy> XHTML2 WG chair [09:17:01.0000] <Lachy> or, at least member, if he's not the chair, but I'm pretty sure he is [09:19:00.0000] <Lachy> http://www.molly.com/2009/06/02/the-real-why-xhtml-discussion/ [09:21:00.0000] <Lachy> "Now, if HTML had never had a lax parser, but had always been strict, there wouldn’t be an incorrect (syntax-wise) HTML page in the planet, because everyone uses a ’suck it and see’ approach" [09:22:00.0000] <Lachy> I wonder how he explains the presense of non-well-formed XHTML and SVG on the web [09:29:00.0000] <takkaria> "HTML is just not addressing the use cases of the real world any more." hmm [09:31:00.0000] <Philip`> Lachy_: That viewpoint seems to be limited to the case where people write a small number of static pages and look at them all to see that they're correct [09:31:01.0000] <Philip`> which was perhaps a valid viewpoint in 1995 [09:32:00.0000] <Philip`> but now everything is dynamic and mashed up, so your site's visitor is the first person in the world to see a particular page and encounter any errors on it [09:34:00.0000] <Philip`> /me reads the blog post [09:35:00.0000] <jgraham> Philip`: Despite which there are a lot of conversations around HTML which revolve around the 1995 model being a good approximation to the truth [09:35:01.0000] <Philip`> "2. Look at it in the browser, if there is a problem, fix it, and look again." - ah, that seems to be where the problem is, because most sites contain an infinite number of pages and you can't look at them all in your browser when developing the site [09:36:00.0000] <Philip`> "I want the browsers to tell me when my pages are wrong" - sounds like a job for a validator [09:37:00.0000] <jgraham> What's the URL to the number of outstanding emails that Hixie has? [09:37:01.0000] <Philip`> (possibly a validator that's built in to the browser as an extension) [09:37:02.0000] <jgraham> (the really slow canvas graph one) [09:40:00.0000] <Philip`> jgraham: http://www.whatwg.org/issues/data.html ? [09:40:01.0000] <jgraham> Philip`: Yes [09:40:02.0000] <jgraham> Thanks [09:58:00.0000] <gsnedders> And we're back in 2010 for the due date [10:09:00.0000] <gsnedders> Hixie: Can you add an id to the NCR table? [10:30:00.0000] <gsnedders> "<!DOCTYPE" should return one parse error now, right? [10:33:00.0000] <gsnedders> /me commits it seeming nobody argues otherwise :P [10:33:01.0000] <gsnedders> (And it was the specific aim of the spec change to make it only return one parse error) [10:34:00.0000] <shepazu> Lachy_: there is little or no non-well-formed SVG on the web, fwiw [10:35:00.0000] <shepazu> there is some SVG that is invalid because Adobe's viewer allowed magic namespaces [10:36:00.0000] <jgraham> shepazu: Ho much SVG is created by automatic tools? [10:36:01.0000] <jgraham> (Other than e.g. inkscape) [10:36:02.0000] <shepazu> as far as XHTML, I seriously doubt there is much non-well-formed XHTML that is served exclusively as XHTML, and not text/html [10:37:00.0000] <shepazu> jgraham: I have no way of knowing that... [10:37:01.0000] <shepazu> but there was a lot of hand-generated SVG back in the early days, the majority, perhaps [10:37:02.0000] <shepazu> before Inkscape, definitely the vast majority [10:38:00.0000] <shepazu> there was also script-generated SVG, which will of course be non-WF [10:38:01.0000] <shepazu> unless it's generated as strings, I suppose [10:39:00.0000] <shepazu> but it didn't make it to the web, because it would have been corrected when it didn't work [10:41:00.0000] <shepazu> so, Lachy_'s premise doesn't seem to be borne out by reality [10:41:01.0000] <gsnedders> How does <x x=1 x=2 X=3> give two parse errors? [10:41:02.0000] <jgraham> gsnedders: How many did you expect? [10:41:03.0000] <gsnedders> jgraham: One. [10:42:00.0000] <gsnedders> One for each duplicated attribute? [10:42:01.0000] <shepazu> no, one could argue that this prevented or slowed the uptake of SVG, but that's another matter (and I think it's far less critical than that it was not supported by browsers and people had to download the Adobe viewer) [10:42:02.0000] <jgraham> gsnedders: Don't you lowercase the attribute name in the tokenizer? [10:42:03.0000] <gsnedders> jgraham: Yes. [10:42:04.0000] <shepazu> bbl [10:43:00.0000] <jgraham> gsnedders: So you end up with 3 x attributes, giving two parse errors, no? [10:43:01.0000] <gsnedders> jgraham: I didn't know whether the spec gave one parse error when it hit the first dupe or for all [10:44:00.0000] <jgraham> gsnedders: I'm not reading the spec of anything. But I would expect 2 in some naive way [10:44:01.0000] <gsnedders> The spec says 2 [10:44:02.0000] <gsnedders> I was being naïve and assuming the other logical option :P [10:47:00.0000] <Philip`> shepazu: "I seriously doubt there is much non-well-formed XHTML that is served exclusively as XHTML, and not text/html" - I agree with that doubt, but that's because I doubt there is much XHTML (regardless of well-formedness) that is served exclusively as XHTML [11:46:00.0000] <gsnedders> "<x x x" should output two parse errors, right? One for dupe attr, and one for EOF? [12:31:00.0000] <jgraham> gsnedders: It seems bad to point to RFC2119 and say "oh we can't have a must requirement about that because browsers will interoperate" [12:33:00.0000] <jgraham> Because RFC2119 is clearly not written to deal with the situation that HTML is in where most authors are unaware of the specs and all implementations have to be error correcting [12:34:00.0000] <jgraham> /me wonders what the bad side effects of saying "HTML documents must be UTF8 encoded" would be [12:34:01.0000] <jgraham> Apart from the fact that it wouldn't match the default [12:35:00.0000] <jgraham> I gues you could say "HTML documents must be UTF8 encoded and must have <meta charset=utf8>" [12:36:00.0000] <jgraham> I can't see people going fo this though [12:36:01.0000] <tantek> jgraham you would simply have to define the processing model for when an HTML processor encountered an HTML document that was not encoded in UTF8. [12:36:02.0000] <jgraham> tantek: Of course the actual processing would have to work like it does today [12:37:00.0000] <jgraham> This would purely be an author requirement [12:38:00.0000] <jgraham> At the very least it seems sensible to remove the ASCII loophole for declaring encodings [12:38:01.0000] <tantek> by "remove" do you mean "must ignore"? [12:39:00.0000] <tantek> you can't really "remove" anything from the author's array of options, you can only make it "do nothing" or have some other minimal effect. [12:39:01.0000] <jgraham> tantek: I mean make it mandatory to declare ascii documents wither in <meta> or through headers [12:40:00.0000] <jgraham> at the moment non-ascii has to be declared to be conforming, ascii does not [12:40:01.0000] <tantek> I'm not sure what that means unless you define what happens when authors don't do it. [12:40:02.0000] <jgraham> tantek: In terms of implementation requirements it doesn't mean anything [12:41:00.0000] <jgraham> Implementations still have to look at HTTP headers, <meta> elements, use heuristics or fall back to win-1252 [12:41:01.0000] <jwalden> Lachy_: that's not overflow; 2**32-1 is not an array index property, so setting it has no effect on the length property [12:41:02.0000] <jgraham> jwalden: You would have thought so [12:42:00.0000] <jgraham> If you, say, read the spec. But the real world disagrees [12:42:01.0000] <jwalden> jgraham: no, that's exactly how the browsers work [12:43:00.0000] <jwalden> arrays-as-objects with all the concomitant baggage was a flaw, but the real world does pretty well with them anyway [12:43:01.0000] <jgraham> Oh wait a second my test earlier was totally wrong [12:44:00.0000] <jgraham> jwalden: Yes you are quite right. Apolologies for doubting you :) [12:45:00.0000] <jgraham> s/lo// [12:45:01.0000] <jwalden> no worries :-) [12:45:02.0000] <jwalden> I've stared long and hard at the array-index-property parts of es3 [12:46:00.0000] <jwalden> frankly, arrays-as-objects is probably a bigger problem for implementations than for developers, since the convert-to-sparse determination is unspecified but necessary for performance and constraining memory use [12:49:00.0000] <jwalden> it's also a compatibility concern: SpiderMonkey has specifically tweaked its thresholds to perform better on certain JS benchmarks promulgated by the authors of engines which have different thresholds that don't fall over on said tests [12:49:01.0000] <jwalden> I do so love benchmarketing [12:51:00.0000] <jwalden> (threshold for sparsifying, to be clear, although I would be surprised if others haven't been similarly tweaked) [12:51:01.0000] <Hixie> heycam: yt? i was wondering if you'd done anything on http://mcc.id.au/2009/01/string-handling/string-handling recently or had any plans to do more [12:52:00.0000] <jgraham> jwalden: It is pretty surprising for authors if things that are instantaneous in one browser take an excessive amount of time in others [12:53:00.0000] <jwalden> this wasn't "excessive" amounts of time, though, it was time measured in dozens of milliseconds [12:53:01.0000] <jwalden> efficiently storing array properties is faster than doing a generic property-hash-storage system, but it's not order-of-magnitude slower, not enough for authors to get confused about it [12:54:00.0000] <jgraham> jwalden: Yeah I agree about a few milliseconds on typical microbenchmarks not being so important [12:57:00.0000] <jwalden> this is what benchmarketing drives you to do :-\ [13:09:00.0000] <gsnedders> /me wonders if there is any nice way to fix the horridness of PHP html5lib without making it slower [13:10:00.0000] <ezyang> Reimplement it in C [13:11:00.0000] <Hixie> jwalden: just make a benchmark that needs the optimisation you think is better for real world pages [13:12:00.0000] <gsnedders> ezyang: yeah, without doing that? :P [13:12:01.0000] <ezyang> I mean, to a certain extent it would be smart [13:12:02.0000] <ezyang> At the very least, tokenization is something that probably should be in C [13:13:00.0000] <gsnedders> ezyang: I think any any extent it would be smart :P [13:13:01.0000] <gsnedders> s/any/to/ [13:13:02.0000] <gsnedders> /me has the tokenizer working as an iterator now [13:13:03.0000] <gsnedders> we're talking 50% slower. [13:13:04.0000] <ezyang> Not surprising [13:14:00.0000] <ezyang> Turning it into an iterator means you get all of those function calls we oh so cleverly removed back [13:14:01.0000] <ezyang> And more [13:14:02.0000] <gsnedders> Yup [13:15:00.0000] <gsnedders> It's interesting that it is actually slower than having all the function calls within the tokenizer [13:15:01.0000] <gsnedders> I mean, there are probably around the same number as I have it now. [13:15:02.0000] <ezyang> That's because iterator generates a bunch more calls [13:15:03.0000] <ezyang> Like, it checks if you're at the end() on every round [13:15:04.0000] <gsnedders> ezyang: I know, but I mean including those [13:15:05.0000] <ezyang> Hmm, unclear. [13:16:00.0000] <gsnedders> /me wonders if it the calls to array_shift [13:16:01.0000] <gsnedders> (as that is O(n) [13:16:02.0000] <gsnedders> /me profiles [13:16:03.0000] <ezyang> array_shift() shouldn't be O(n)... [13:17:00.0000] <ezyang> But I suppose PHP's array implementation would make it so. [13:17:01.0000] <gsnedders> It is. It re-indexes everything. [13:17:02.0000] <gsnedders> There's an SPL queue in 5.3… [13:17:03.0000] <gsnedders> /me sighs [13:18:00.0000] <ezyang> If you want, you can branch and work on PHP5.3 only [13:18:01.0000] <gsnedders> No point [13:18:02.0000] <gsnedders> What we could do if create a class which implements the queue like bit of the weird SPL queue structure and use arrays internally there [13:18:03.0000] <gsnedders> array_shift isn't that bad actually [13:19:00.0000] <ezyang> Don't boher with that, then [13:19:01.0000] <gsnedders> It's the same culprits as before [13:19:02.0000] <gsnedders> It's all function call overhead more or less… [13:19:03.0000] <gsnedders> Did I say how much I love PHP yet? [13:19:04.0000] <ezyang> Maybe you're interpreting the results wrong? [13:20:00.0000] <gsnedders> I reckon 15% of the cost to be the function calls foreach creates. [13:22:00.0000] <ezyang> /me puts html5lib performance on his queue of things to do after work [13:23:00.0000] <Philip`> Please feel free to optimise the Python html5lib too [13:23:01.0000] <gsnedders> /me was really trying to do this to get the final few test cases to pass [13:24:00.0000] <gsnedders> I guess we could do the not-so-nice way the Python impl does it [13:24:01.0000] <Philip`> What's not-so-nice about it? [13:24:02.0000] <gsnedders> Philip`: Just think about who wrote it :P [13:24:03.0000] <Philip`> Uh... annevk2? [13:25:00.0000] <ezyang> What really needs to happen is the core string manipulation stuff needs to be written out in C [13:25:01.0000] <ezyang> And then we make wrappers around it in all the appropriate languages [13:25:02.0000] <Philip`> or jgraham? [13:25:03.0000] <gsnedders> ezyang: That's not the problem. The most expensive thing is the tree construction. [13:25:04.0000] <gsnedders> ezyang: http://stuff.gsnedders.com/phphtml5lib/ [13:25:05.0000] <ezyang> Oh, well, then you should factor out the tree-specific stuff, and then make the SimpleTree impl. [13:25:06.0000] <gsnedders> Well, that isn't always true [13:26:00.0000] <gsnedders> That's never true :P [13:26:01.0000] <gsnedders> Oh well. [13:26:02.0000] <gsnedders> Ignore me. [13:26:03.0000] <gsnedders> :P [13:26:04.0000] <gsnedders> With the spec the most expensive thing is DOMNode->appendChild :P [13:26:05.0000] <ezyang> gsnedders: Screens don't do anything for me. The great thing about CacheGrind is you can reframe the data and look further [13:26:06.0000] <gsnedders> ezyang: Indeed. But some people wanted to see [13:27:00.0000] <jgraham> gsnedders: Did you have an actual point about python html5lib or are you just casting aspersions? [13:27:01.0000] <annevk2> gsnedders, hey, I read that! [13:27:02.0000] <gsnedders> jgraham: Just insofar as it diverges from the spec. [13:27:03.0000] <gsnedders> annevk2: :D [13:27:04.0000] <gsnedders> /me buys annevk2 a beer… next year. [13:27:05.0000] <Philip`> gsnedders: The spec is meant to be diverged from [13:27:06.0000] <jgraham> gsnedders: In what way? [13:28:00.0000] <Philip`> gsnedders: (That's why it says that any algorithm which produces equivalent output to the specified one is acceptable) [13:28:01.0000] <jgraham> Apart from known issues with failing tests? [13:28:02.0000] <ezyang> Philip`: You have to admit, though; it's a lot easier (w.r.t. to maintainability) to do that once the spec is stable. [13:28:03.0000] <gsnedders> Philip`: I know :P [13:29:00.0000] <gsnedders> /me would quite like to write an impl. that uses something like Parsec [13:29:01.0000] <Philip`> ezyang: Indeed, but hopefully there are enough test cases and enough independent implementations that any problems caused by spec changes can be found relatively easily [13:29:02.0000] <jgraham> Somehow the headline "Microsoft unveils new controller" sounds rather sinister [13:30:00.0000] <roc> One Microsoft Way! [13:31:00.0000] <ezyang> It takes me on average about thirty minutes to resolve a spec-code impedance. Maybe towards the end it was 15 minutes, after I got good at figuring out what was different. [13:34:00.0000] <gsnedders> ezyang: Practice! :P [13:35:00.0000] <ezyang> My point, though, is that even when the code is ~ the spec, it's nontrivial to resolve bugs. [13:35:01.0000] <ezyang> And it can onl get worse [13:35:02.0000] <gsnedders> ezyang: Solution: read all diffs with spec :P [13:36:00.0000] <gsnedders> ezyang: I brought the tokenizer up to date with the spec, FWIW [13:36:01.0000] <jgraham> gsnedders: That's not a very practical solution [13:36:02.0000] <ezyang> gsnedders: Awesome [13:37:00.0000] <gsnedders> ezyang: (and had to update a few test cases which means probably everything else fails now) [13:37:01.0000] <gsnedders> ezyang: I rewrote the entire NCR handling so we don't consume so much [13:38:00.0000] <ezyang> About 80% I thought I had to update a test-case, it was actually my fault [13:38:01.0000] <ezyang> You should probably double check. [13:39:00.0000] <gsnedders> ezyang: The aim of the commit in the spec was to reduce to one parse error from two. I just did that in the test :P [13:41:00.0000] <ezyang> Oh, that was the only change [13:41:01.0000] <ezyang> Yeah, that sounds legit [13:42:00.0000] <gsnedders> The other thing I did to test suite was add one test that I broke in uncommitted code locally [13:48:00.0000] <gsnedders> /me wonders about getting hsivonen's parser working as a PHP ext [13:48:01.0000] <ezyang> Nonono! [13:48:02.0000] <ezyang> It has to be C. [13:48:03.0000] <gsnedders> ezyang: Uh, surely we can get C++ to work? [13:48:04.0000] <ezyang> C++ is not acceptable [13:48:05.0000] <ezyang> It has a horrid time integrating back to C [13:49:00.0000] <ezyang> And most bindings work better with C code than C++ [13:49:01.0000] <jgraham> gsnedders: Write a parser in brainfuck and then write a brainfuck to C compiler [13:49:02.0000] <jgraham> No need to write C [13:51:00.0000] <gsnedders> ezyang: That's more work :P [13:54:00.0000] <jgraham> ezyang: Attributes like xlink:href in <body> should have local name "xlink:href" and be in no namespace, unless that changed recently [13:54:01.0000] <jgraham> /me sees no evidence for such a change [13:55:00.0000] <ezyang> jgraham: Oh? [13:55:01.0000] <ezyang> Where in the spec does it say that? [13:56:00.0000] <jgraham> ezyang: Where in the spec does it say otherwise :) [13:56:01.0000] <jgraham> The only time that you insert attributes in a namespace is when you adjust foreign attributes for a token [13:56:02.0000] <jgraham> Which only happens for foreigncontent [13:56:03.0000] <jgraham> afaik [13:57:00.0000] <gsnedders> Hmm… "When an end tag token is emitted with attributes, that is a parse error." is hard to implement when once you've hit attributes you still don't know it'll actually be emitted [13:57:01.0000] <ezyang> Ah. That's probably right. [13:57:02.0000] <gsnedders> ezyang: Also, why are there things related to foreign content commented out in tokenizer? [13:57:03.0000] <ezyang> You should revert those changes (or I can revert them) [13:57:04.0000] <ezyang> gsnedders: Not sure what you're talking about [13:57:05.0000] <gsnedders> ezyang: seee // XXX not implemented [13:57:06.0000] <ezyang> jgraham: I know exactly why I thought that was otherwise the case [13:58:00.0000] <jgraham> ezyang: I can revert them if you like [13:58:01.0000] <ezyang> gsnedders: Oh, I never implemented that [13:59:00.0000] <ezyang> Please do so. [13:59:01.0000] <gsnedders> /me wonders why test cases don't test that [13:59:02.0000] <ezyang> jgraham: That would be loverly. Also, update the wiki page to mention the difference between "xlink href" and "xlink:href" [13:59:03.0000] <ezyang> It's very subtle [14:00:00.0000] <jgraham> ezyang: If you have the wiki page can you update it? I would need to find it... [14:01:00.0000] <ezyang> http://wiki.whatwg.org/wiki/Parser_tests [14:01:01.0000] <ezyang> I lost my account info [14:22:00.0000] <jwalden> Hixie: it's not that I don't think the optimization is bad, it's that the dividing point between when and when not to optimize is not being determined by real-world scripts but rather by artificial use in benchmarks [14:22:01.0000] <jwalden> heycam: fyi, http://lists.w3.org/Archives/Public/public-webapps/2009AprJun/0720.html [14:22:02.0000] <gsnedders> /me never ceases to be amazed by how bad RSS/Atom are created [14:23:00.0000] <Dashiva> hehe [14:24:00.0000] <Dashiva> JF linking to his blog reminded me about "small and narrow minds" [14:24:01.0000] <Dashiva> We haven't really come far from back there... [15:35:00.0000] <zcorpan> "Wow, cool! There’s been a rip in the space/time continuum and this article has slipped through from 1998!" - http://www.molly.com/2009/06/02/the-real-why-xhtml-discussion/ [15:40:00.0000] <arun> zcorpan, hah :) [15:41:00.0000] <arun> He had me at "books are dead." [16:35:00.0000] <Dashiva> "XML parsing is really easy. It is rather trivial to write an XML parser." 2009-06-03 [17:05:00.0000] <gsnedders> /me finds http://lists.w3.org/Archives/Public/spec-prod/2009AprJun/0010.html [17:26:00.0000] <gsnedders> /me thinks from his query that there are 2245 TRs [17:41:00.0000] <heycam> Hixie, i do plan to finish that testing [17:41:01.0000] <heycam> (and fix the tests up) [17:42:00.0000] <heycam> i'll have a small amount of time after the svg wg's f2f next week that i'll be working on webidl stuff [18:08:00.0000] <gsnedders> /me laughs at HTML 5 now calling itself a "Draft Standard" [18:08:01.0000] <Hixie> i was so happy to hear that html5 was a standard [18:08:02.0000] <Hixie> i always thought that we were supposed to not call w3c things standards [18:08:03.0000] <gsnedders> Me too. [18:09:00.0000] <gsnedders> /me still agrees with you, and just thinks it is more W3C fucked-up-ness [18:09:01.0000] <gsnedders> Uh, the spec is loading really slowly here. [18:09:02.0000] <gsnedders> Like 4KB/s [18:10:00.0000] <gsnedders> Do we have any idea about implementations for iframe@sandbox? [18:10:01.0000] <Hixie> wfm [18:10:02.0000] <Hixie> no [18:27:00.0000] <gsnedders> Oh, just my local net is busy [19:23:00.0000] <heycam> hi, what in the spec determines how external javascript encoding is determined? [19:35:00.0000] <Hixie> heycam: html5 has some stuff for how to do it for <script> [19:35:01.0000] <Dashiva> heycam: I believe "Running a script" in 4.3.1 the script element [19:37:00.0000] <heycam> ok thanks [19:38:00.0000] <heycam> so i'm wondering if/how the processing of scripts labelled as text/javascript differs from what rfc 4329 says to do with them [19:40:00.0000] <heycam> this is the question i'm trying to answer: "Why is HTML5 making the default be text/javascript rather than application/ecmascript?" [19:40:01.0000] <Dashiva> There's a bug for that [19:41:00.0000] <heycam> yeah, but i don't remember a clear reason being given [19:41:01.0000] <heycam> i'll look it up again [19:42:00.0000] <Hixie> the type doesn't matter to the implementations [19:42:01.0000] <Hixie> the implementations completely ignore the type part of the content-type [19:42:02.0000] <Hixie> you could have image/png and it'd work as well [19:42:03.0000] <Hixie> at least over the wire [19:42:04.0000] <Hixie> the type="" attribute is a different matter [19:42:05.0000] <Hixie> but all this is academic [19:42:06.0000] <heycam> ah, i'm asking about the type="" attribute [19:42:07.0000] <Hixie> the only reason that i use text/javascript instead of application/ecmascript is that application/ecmascript is stupid [19:43:00.0000] <Hixie> everyone calls it text/javascript [19:45:00.0000] <heycam> i guess what i want to know is: how would things break if you normatively referenced rfc 4329 for the processing of <script type=text/javascript> and <script type=application/ecmascript>, and for if you changed the default type of <script> with no type="" to be application/ecmascript [19:45:01.0000] <Hixie> what does rfc4329 say that can be referenced? [19:45:02.0000] <Hixie> i didn't realise it defined anything useful other than the type names [19:46:00.0000] <heycam> it gives some requirements on encoding detection / error reporting, i think [19:46:01.0000] <Hixie> s/useful/usable/, i guess, given my stance on the names :-) [19:46:02.0000] <Hixie> oh? [19:46:03.0000] <heycam> i thought so [19:46:04.0000] <Hixie> /me looks [19:46:05.0000] <heycam> but i could be wrong, i just skimmed [19:46:06.0000] <heycam> i remember implementing distinct codepaths for text/javascript and application/ecmascript in batik, a while ago, based on what rfc 4329 said [19:47:00.0000] <Hixie> rfc4329 is misleading then [19:48:00.0000] <heycam> i.e., doesn't describe reality? [19:48:01.0000] <Hixie> uh yeah, having now looked at this seems highly in variance with implemented practice [19:48:02.0000] <heycam> ok [19:48:03.0000] <Dashiva> "Use of the "text" top-level type for this kind of content is known to be problematic." [19:49:00.0000] <Dashiva> What is the reason here? I'm not in the know. [19:49:01.0000] <heycam> default us-ascii encoding? [19:49:02.0000] <Hixie> or iso-8859-1 depending on which spec you read [19:49:03.0000] <Dashiva> Is that it? [19:49:04.0000] <Hixie> (it's nonsense) [19:49:05.0000] <Hixie> i believe so [19:49:06.0000] <Hixie> there is no real problem [19:50:00.0000] <heycam> well i see that html5 maps us-ascii to windows-1252 [19:50:01.0000] <Hixie> the problem is a persistent myth promulgated by a variety of spec writers [19:51:00.0000] <Dashiva> Hmm, the RFC claims charset defaults to UTF-8 [19:51:01.0000] <Dashiva> That's not really true, is it? [19:51:02.0000] <Hixie> no [19:51:03.0000] <Hixie> i would basically ignore 4329 altogether [19:52:00.0000] <heycam> i'd like concrete things that rfc 4329 conflicts with to answer the question about why it uses text/javascript and doesn't reference that rfc, if you could [19:52:01.0000] <heycam> s/why it/why html5/ [19:52:02.0000] <Hixie> search for 4329 in html5 [19:53:00.0000] <heycam> yep i did that [19:53:01.0000] <heycam> there's just that one note [19:53:02.0000] <Hixie> right [19:53:03.0000] <Hixie> that's the reason for why it uses text/javascript [19:53:04.0000] <Hixie> the reason it doesn't refer to the charset algorithm is the charset algorithm is wrong [19:53:05.0000] <Hixie> e.g. it ignore charset="" on <script> [19:54:00.0000] <Hixie> doesn't take into account the refering document's charset [19:54:01.0000] <Hixie> does support UTF-32 [19:54:02.0000] <Hixie> and suggests that the type is relevant [19:54:03.0000] <Dashiva> Hmm, it defers to XML Media Types on the processing. That's why text/ and application/ are different [19:55:00.0000] <Hixie> 3023 [19:55:01.0000] <Hixie> that's where a lot of this nonsense started [19:55:02.0000] <Hixie> regarding text/ vs application/ [19:55:03.0000] <heycam> that text/javascript is more commonly used might be a reason to refer to the language as "text/javascript", but that doesn't have anything to do with requirements on how to process script, right? [19:55:04.0000] <Hixie> right [19:55:05.0000] <heycam> the encoding/charset stuff sounds like the kind of answer i want [19:55:06.0000] <MikeSmith> wakaba: you around? [19:56:00.0000] <heycam> "RFC 4329 requires <script type=text/javascript> to be processed like A, but in reality browsers need to be processed like B." [19:56:01.0000] <Hixie> /me should point out that he's not a big fan of tiny specs that over-solve non-problems, like 4329 [19:56:02.0000] <heycam> s/browsers need to be/browsers need to process it/ [19:56:03.0000] <MikeSmith> wakaba: if/when you have time, please try a traceroute to svn8.cvsdude.com and let me know what you get [19:57:00.0000] <MikeSmith> wakaba: I'm on my emobile connection and can't get any route to it [19:57:01.0000] <Hixie> heycam: did this come up in the svgwg? [19:57:02.0000] <heycam> the quoted question at the top did, yes [19:58:00.0000] <Hixie> the only reason i'm making the default text/ instead of application/ is that there's no reason to introduce application/ and everyone calls it text/ [19:58:01.0000] <Hixie> i'm not doing it because of charsets [19:58:02.0000] <heycam> hmm ok [19:58:03.0000] <Hixie> (and that is already mentioned in the spec) [19:58:04.0000] <heycam> i guess the more pertinent question then is why you don't normatively reference rfc 4329 [19:59:00.0000] <heycam> to which the answer is the charset/encoding stuff? [19:59:01.0000] <Hixie> yeah [19:59:02.0000] <Hixie> though nobody other than you has asked that :-) [19:59:03.0000] <Hixie> asked me that, anyway [19:59:04.0000] <heycam> didn't jreschke? [19:59:05.0000] <Hixie> did he? [19:59:06.0000] <heycam> i thought that's what that bugzilla bug was [19:59:07.0000] <Hixie> i thought he spoke of the type name only [19:59:08.0000] <heycam> but i couldn't find it [19:59:09.0000] <heycam> ah ok [20:00:00.0000] <heycam> processing for all the many different media types to identify javascript are handled the same in html5, yes? [20:00:01.0000] <Hixie> for type="", language="", or Content-Type: ? [20:01:00.0000] <heycam> type="" [20:01:01.0000] <heycam> (you've said Content-Type is ignored, right (except for its charset parameter)?) [20:02:00.0000] <Hixie> let's see... (for type...) [20:04:00.0000] <Hixie> heycam: the only difference appears to be the one listed here: http://www.whatwg.org/specs/web-apps/current-work/#support-the-scripting-language [20:05:00.0000] <heycam> the e4x bit? [20:10:00.0000] <Hixie> yeah [20:10:01.0000] <heycam> is this right?: <script src=""> without a type="" attribute that references a resource served as whatever media type, without a charset parameter, will be decoded using the referencing html document's encoding [20:11:00.0000] <heycam> (also if the resource doesn't have a BOM) [20:11:01.0000] <Dashiva> Yes [20:11:02.0000] <Dashiva> Although you probably meant without a charset attribute? [20:12:00.0000] <heycam> oh i meant charset parameter on the Content-Type [20:12:01.0000] <heycam> (but also without a charset="" on the <script>) [20:12:02.0000] <heycam> or does that charset parameter on the Content-Type not do anything either? [20:12:03.0000] <Dashiva> Yeah, because type doesn't really enter into it [20:12:04.0000] <Dashiva> It does [20:13:00.0000] <heycam> ah i see what you mean [20:13:01.0000] <heycam> the charset parameter in the type="" is ignored, it's the charset="" attribute that makes a difference [20:14:00.0000] <heycam> ok so that's a good concrete answer to the question of why rfc 4329 isn't normatively referenced [20:14:01.0000] <heycam> thanks [20:15:00.0000] <heycam> this is the kind of thing that it's hard to get a sense of quickly looking at the algorithms [20:16:00.0000] <heycam> informative, declarative text that summarises how encoding of an external script is determined would have been handy to answer my question, i think [20:17:00.0000] <Dashiva> It would be nice [20:17:01.0000] <jruderman> does "unknown parameters" in http://www.whatwg.org/specs/web-apps/current-work/#support-the-scripting-language refer to things like ;e4x=1 and ;version= ? [20:17:02.0000] <Dashiva> But when Hixie writes an algorithm it usually means he wasn't able to define things neatly declaratively [20:18:00.0000] <heycam> try harder, Hixie, gosh! [20:18:01.0000] <heycam> i don't mind even if the declarative informative parts aren't completely accurate [20:18:02.0000] <ezyang> /me finally got around to running kcachegrind on profiler output of html5lib php [20:18:03.0000] <heycam> and defer to the algorithm for details [20:19:00.0000] <heycam> but i guess that could be misleading for people, and might encourage them not to look into the details in the algorithm [20:22:00.0000] <Hixie> heycam: it's not clear such text is a good idea, since because it would be simpler than the real text, people would use it instead of the real text [20:23:00.0000] <heycam> yes i could understand that [20:23:01.0000] <Hixie> jruderman: it means parameters that the browser doesn't know [20:23:02.0000] <Hixie> heycam: oh, you just said that. nevermind :-) [20:23:03.0000] <Hixie> heycam: the other problem is that i don't know which parts of the spec people will want these summaries for [20:24:00.0000] <heycam> right [20:24:01.0000] <Hixie> heycam: and doing it for every little thing would double the size of the spec [20:24:02.0000] <jruderman> Hixie: i wasn't sure what "parameters" meant in that context [20:24:03.0000] <Hixie> jruderman: MIME parameters [20:24:04.0000] <jruderman> i was afraid that would be the answer ;) [20:24:05.0000] <Hixie> jruderman: see the MIME spec for details :-) [20:38:00.0000] <ezyang> Bottom line is, 4 seconds to parse a 2MB document is not bad. [20:39:00.0000] <ezyang> I don't think I'm going to waste too much time optimizing the hell out of it. [20:39:01.0000] <ezyang> A C implementation is generally useful, but lots of work. [20:39:02.0000] <ezyang> (this is to gsnedders) [20:44:00.0000] <Hixie> "I think, in fact, that if I may pose myself as not belonging to the hegemony, then I am one of the least conspirational persons of that lot - too little, perhaps." [20:44:01.0000] <Hixie> i am going to have to frame that and print it [20:44:02.0000] <Hixie> er, print it then frame it [20:50:00.0000] <gavin_> paranoid about not being paranoid enough? [20:56:00.0000] <Hixie> and using the word "hegemony" in the same sentence as claiming to not be conspirational [21:08:00.0000] <MikeSmith> I wonder if some people read their mailing-list postings aloud to themselves before sending them [21:08:01.0000] <MikeSmith> while looking in the mirror [21:09:00.0000] <MikeSmith> I mean people other than me [21:09:01.0000] <MikeSmith> or if they wink at themselves while doing so [21:10:00.0000] <ezyang> /me wink wink [21:10:01.0000] <ezyang> /me wink wink wink if you know what I mean [21:10:02.0000] <ezyang> wink wink [22:03:00.0000] <Hixie> haha, awesome [22:03:01.0000] <Hixie> unrelated people are referring to html5 as a "draft standard" now: http://www.cre8asiteforums.com/forums/index.php?showtopic=71518&mode=linearplus [22:05:00.0000] <hobo> well thats good news [22:06:00.0000] <othermaciej> Hixie: that's what I've always called it when I talk about it... [22:06:01.0000] <othermaciej> but it's nice that I can be pedantically correct now when I say that [22:28:00.0000] <shepazu> Hixie: did it escape your attention that Google was calling HTML5 a standard at Google I/O? [22:28:01.0000] <Hixie> no, but google are hardly "unrelated" :-) [22:28:02.0000] <othermaciej> shepazu: calling it a "draft standard" is definitely an improvement [22:28:03.0000] <shepazu> people have always used the term "standard" pretty loosely... [22:28:04.0000] <othermaciej> that is what I called it in my Google I/O talk [22:29:00.0000] <Hixie> it was in fact google's calling it a standard that cause plh to complain, which caused me to update the spec's title to use the word "standard" [22:29:01.0000] <othermaciej> Google also claimed many things are part of HTML5 which are actually now separate specs [22:29:02.0000] <shepazu> and that were never part of HTML5, like Geolocation API [22:29:03.0000] <othermaciej> I've advised potential Apple presenters to try to be more precise [22:29:04.0000] <shepazu> Hixie: you may have misunderstood... PLH wasn't complaining about the word, just noting that it's not yet done [22:30:00.0000] <othermaciej> calling it a standard without being clear about the draft status is incorrect [22:30:01.0000] <othermaciej> and potentially misleading [22:30:02.0000] <shepazu> the word really only matters for the pedantic, or when it actually matters to make a distinction [22:30:03.0000] <Hixie> shepazu: yes, i understand that (see the comments on the blog post for details) [22:31:00.0000] <othermaciej> PLH had a point, although if I were him, I would focus less on that and more on the extremely positive PR being generated for a W3C draft [22:31:01.0000] <shepazu> othermaciej: yes, I think that was what PLH was pointing out... that it's still a draft [22:31:02.0000] <othermaciej> like if you look at this: http://www.w3.org/QA/2009/05/_watching_the_google_io.html [22:32:00.0000] <shepazu> I think that PLH's post was optimistic on the whole... it told people how to get involved, said that it was still ongoing, and talked about the various related activities [22:32:01.0000] <othermaciej> there's about 1.5 sentences in the whole thing that seem happy about HTML5 getting a positive reception [22:32:02.0000] <Hixie> othermaciej: the reason google people refer to all this as "html5" is that there's no other catchphrase for it, btw, an issue that i brought up months ago in this channel [22:32:03.0000] <Hixie> othermaciej: not sure what to do about it [22:32:04.0000] <Hixie> othermaciej: until we have another marketable term, google'll probably continue calling it all "html5" [22:32:05.0000] <Hixie> othermaciej: because people respond to that term [22:32:06.0000] <othermaciej> Hixie: I think Apple will probably end up riding the catchphrase wave, though maybe be clear that it includes more than just the HTML 5.0 draft standard [22:33:00.0000] <shepazu> there's nothing wrong with marketing terminology, and there's nothing wrong with being more precise, depending on the audience [22:33:01.0000] <othermaciej> though maybe with the synchronicity of HTML 5 and ES5, we could call the whole current generation of Web standards "Web 5.0" [22:33:02.0000] <othermaciej> (see, this is why I'm not in marketing) [22:33:03.0000] <shepazu> othermaciej: you, sir, are correct :) [22:33:04.0000] <Hixie> shepazu: i think there _is_ something wrong with calling, e.g. the geolocation spec "HTML5" [22:34:00.0000] <Hixie> othermaciej: yes, i agree that "Web 5.0" is what we should use [22:34:01.0000] <shepazu> Hixie: and that's why you weren't up on stage :) [22:34:02.0000] <shepazu> I think that cat's out of the bag [22:34:03.0000] <othermaciej> Hixie: if I actually thought that was a good term I could probably put it in the buzz-vocabulary [22:34:04.0000] <Hixie> no, i wasn't on stage because i refused because i didn't want to appear vendor-biased :-) [22:34:05.0000] <othermaciej> all it would take is one WebKit blog post plus get some key apple folks to talk about it [22:35:00.0000] <othermaciej> but I'm not sure "Web 5.0" is actually good [22:35:01.0000] <shepazu> fwiw, I think "Web 5.0" is lame [22:35:02.0000] <othermaciej> it's not bad in itself, but it will remind people of "Web 2.0" and "web 3.0" [22:35:03.0000] <shepazu> it's too Web 2.0 [22:35:04.0000] <shepazu> yes [22:35:05.0000] <Hixie> well, i'm fine with whatever term people want, but until we have one, expect google's marketing to continue saying "html5" :-) [22:35:06.0000] <shepazu> Open Web is much better [22:35:07.0000] <Hixie> at least the chrome guys now say "OWP" [22:36:00.0000] <Hixie> but "OWP" is apparently not a term that people respond to [22:36:01.0000] <shepazu> OWP is also terrible [22:36:02.0000] <othermaciej> for Open Web Platform? [22:36:03.0000] <Hixie> othermaciej: yeah [22:36:04.0000] <shepazu> Open Web. [22:36:05.0000] <othermaciej> the problem with that is, it's not specific to the current generation of improvements [22:36:06.0000] <othermaciej> (or Open Web) [22:36:07.0000] <Hixie> (chrome use OWP for the concept in general, not this generation) [22:36:08.0000] <othermaciej> yeah [22:37:00.0000] <othermaciej> we say similar things - "open web", "web platform", "web technology", "standards-based web" [22:37:01.0000] <othermaciej> I like the 5 as a signifier of the current generation [22:37:02.0000] <zcorpan> W3C 5 [22:37:03.0000] <othermaciej> it's just too bad Web x.0 is somewhat poisoned by past meaningless buzzwords [22:38:00.0000] <othermaciej> zcorpan: that sounds like a model number for a european luxury car [22:38:01.0000] <othermaciej> "Web 5" could avoid the .0 taint, but it's not as euphonious [22:39:00.0000] <arun> I wasn't too thrilled with 3D being associated with HTML5 in past catchphrasing at Google I/O, but whatever :) [22:39:01.0000] <othermaciej> I think of the things they referred to as HTML5, only AppCache, <canvas> and <video> are actually in HTML5 [22:41:00.0000] <olliej> and all of them have been shipping for more than a year [22:41:01.0000] <arun> The 5 signifier is good, but unfortunately ECMAScript 5 is merely a renaming of the ersthwile ECMAScript Edition 3.1, so it's almost 5 because of standards-setting org rigidity :) [22:42:00.0000] <arun> Open Web, 5th Generation [22:43:00.0000] <heycam> Open Web: The Next Generation [22:43:01.0000] <othermaciej> 5th Generation Web? [22:43:02.0000] <heycam> Hixie is Picard, rubys is Riker (because of the beard, obviously), ... [22:43:03.0000] <arun> heh :) Or, just to futz with O'Reilly, maybe Web 5.0? [22:44:00.0000] <othermaciej> arun: we just talked about Web 5.0, and how it may bring up bad associations with Web 2.0... [22:44:01.0000] <othermaciej> but then again, for the target audience, the associations may be positive [22:44:02.0000] <arun> :) [22:44:03.0000] <tantek> a whole nother meaning to "5 > 2" [22:45:00.0000] <othermaciej> I've almost talked myself into Web 5.0 [22:46:00.0000] <arun> I'm definitely warming up to Web 5.0. 5.0 is attractive numerology; Fx 3.5, etc. There's fives in sufficiently needle-moving things. [22:46:01.0000] <zcorpan> maybe Opera 10 should be renamed to Opera 5 [22:47:00.0000] <shepazu> AppWeb? [22:48:00.0000] <shepazu> zcorpan: no, it should be Opera X [22:48:01.0000] <zcorpan> shepazu: too apple-y [22:48:02.0000] <shepazu> then the next version should be Opera X: beagle [22:49:00.0000] <shepazu> followed by pug, chihuahua, and weinerdog [22:50:00.0000] <zcorpan> what would the UA string be? Opera/X.beagle ...? [22:50:01.0000] <othermaciej> I asked Apple marketing for advice on a buzzword [22:53:00.0000] <shepazu> I suspect it might be too late... HTML5 is probably going to stick [22:56:00.0000] <heycam> Hixie, minor nit: s/WebIDL/Web IDL/ throughout the spec (except when it's a [WebIDL] reference) [22:56:01.0000] <Hixie> heycam: can you file a bug or drop me a mail? [22:57:00.0000] <heycam> yep [22:57:01.0000] <Hixie> thanks [23:46:00.0000] <zcorpan> /me splits off the Ogg controversy to a separate page on wikipedia [23:57:00.0000] <zcorpan> i think that's enough wiki editing for today [00:22:00.0000] <arun> I'd like to replace FileList as defined here: http://dev.w3.org/2006/webapi/FileUpload/publish/FileUpload.html#FileList-if with something that uses sequence<File> which is much more accurate to real world usage. heycam, that's merely typedef FileList sequence<File> right? [00:23:00.0000] <heycam> yep, apart from the fact that sequences aren't defined properly at the moment [00:23:01.0000] <heycam> how do you want the sequences to behave? [00:23:02.0000] <arun> OK, but what if FileList were to also* have a unsigned int for length? [00:23:03.0000] <arun> Frankly, I want sequences to behave like ECMAScript 3.0 Arrays :) [00:24:00.0000] <heycam> arun, ok [00:24:01.0000] <heycam> i think there a different use cases for different sequence-ish/array-ish behaviour [00:24:02.0000] <heycam> so you just want to expose an array as an idl attribute [00:24:03.0000] <heycam> is the array mutable? [00:24:04.0000] <arun> heycam, hmm... I may have simplified. [00:25:00.0000] <arun> The array represents the number of File objects selected by the user. [00:25:01.0000] <arun> And I was thinking that FileList would expose an unsigned int for length. [00:25:02.0000] <heycam> guess i should look at the spec [00:26:00.0000] <arun> http://dev.w3.org/2006/webapi/FileUpload/publish/FileUpload.html#FileList-if is what I want to replace . [00:26:01.0000] <heycam> ok [00:27:00.0000] <heycam> so it's a read only sequence [00:27:01.0000] <arun> So you'd say myFile = inputElement.files[0]; yourFile = inputElement.files[1]; (where inputElement.length = 2) [00:27:02.0000] <heycam> can't change the length, or what File objects are stored in there [00:27:03.0000] <heycam> you can certainly do that sort of thing at the moment with an interface and use of [IndexGetter] [00:27:04.0000] <arun> Yep. [00:27:05.0000] <heycam> /me has a telcon now, so you'll have to excuse me for a bit [00:27:06.0000] <arun> Yeah, I wonder if [IndexGetter] is better than sequence here. [00:28:00.0000] <Hixie> why would sequence not work? [00:28:01.0000] <Hixie> /me hopes sequence<> will work soon because HTML5 uses it :-) [00:29:00.0000] <heycam> coz i've commented most of it out [00:29:01.0000] <arun> I'm ok with using sequence<> and like it, actually. But in heycam's draft, it seems commented out. [00:29:02.0000] <heycam> i'll take a look at all the specs currently using sequence and hash something out in a couple of weeks [00:29:03.0000] <arun> I'm warming up to [IndexGetter] now actually. [00:30:00.0000] <arun> /me goes back to reading Web IDL... [00:33:00.0000] <Hixie> it being commented out is no reason not to use it if you ask me :-) [00:33:01.0000] <Hixie> it just means we don't know exactly what it'll be yet :-) [00:34:00.0000] <arun> Hixie, unless [IndexGetter] is invalid for this use case (is there a mutable/immutable thing that I should know?), it seems to do the trick admirably. [00:34:01.0000] <arun__> I like sequence<> but this seems quick and dirty for my use case. [00:36:00.0000] <Mrmil> Hi, I'll take the advantage that there's so many of you in the morning (well, my morning that is) and would like you to have a look at http://server.ebrana.cz/olda/_apps/html5/ and tell me what I can fix again. Thanks. :) [00:36:01.0000] <Hixie> arun: [IndexGetter] works fine here too [00:37:00.0000] <Hixie> /me uses [IndexGetter] a lot in his specs [00:37:01.0000] <Hixie> it's a lot more work :-P [00:37:02.0000] <arun> FWIW I want sequence<> used in the emerging 3DWeb spec. [00:37:03.0000] <arun> /me now has to go back to setting an agenda for the upcoming meeting sigh [00:38:00.0000] <Hixie> oh is there a mailing list for that yet? [00:38:01.0000] <arun> Mrmil, at first blush, the site seems ok. [00:39:00.0000] <arun> Hixie, there is* a closed mailing list presently. [00:40:00.0000] <zcorpan> Mrmil: you could remove class="hidden" and replace the selector with nav h2 [00:41:00.0000] <Hixie> arun: closed? [00:41:01.0000] <Hixie> arun: how do we join it? (and when will it be open?) [00:41:02.0000] <arun> Hixie, I continue to exert gentle pressure for more openness (I'm the Chair of the WG), but by "closed" I mean that it is for members only. [00:41:03.0000] <Mrmil> zcorpan: that's true but I have an experience that the more complex the site is, the more hidden elements it has. So it's more space-wise to have short hidden classes do the trick than long selector lines in the css sheet. [00:42:00.0000] <Hixie> arun: how do i become a member? [00:42:01.0000] <arun> Hixie, your company is on it, and if you want to be on it, that's pretty straightforward. [00:42:02.0000] <Mrmil> zcorpan: And I don't want to hide every nav h2. [00:42:03.0000] <zcorpan> Mrmil: fair enough [00:43:00.0000] <Mrmil> zcorpan: Hookey. [00:43:01.0000] <zcorpan> Mrmil: <q>appealing</q> - i think this is not quite proper use of <q> per spec [00:43:02.0000] <Mrmil> arun: Thanks :) [00:43:03.0000] <arun> Hixie, ask Fette et al, but basically you sign up with your company credentials on khronos. [00:43:04.0000] <Mrmil> zcorpan: Hehe, I know, I'll rewrite the text today. [00:44:00.0000] <Hixie> arun: hm, ok, will ask, thanks [00:44:01.0000] <Hixie> arun: how does a random person become a member? [00:44:02.0000] <Hixie> arun: we should post a blog post like the whatwg one telling people how to join [00:44:03.0000] <zcorpan> Mrmil: the <address> only "applies" to the <aside> section per spec. although one could argue that that the spec is being silly [00:45:00.0000] <zcorpan> Mrmil: <div class="clear"></div> - you could use :after instead [00:45:01.0000] <Mrmil> zcorpan: could do, no problem [00:45:02.0000] <arun> Hixie, right now, the random person can only become a member by invitation based on merit as judged by other members (unlike how public-html⊙wo works -- but note I'm not implying anything about merit on that list ;-) ) OR if they pay. [00:46:00.0000] <Hixie> hm, that sucks [00:46:01.0000] <Hixie> we should fix that [00:46:02.0000] <arun> Hixie, essentially this is a closed group; and I'm trying :-) [00:46:03.0000] <Hixie> well, i guess i'll join and help you :-) [00:46:04.0000] <Hixie> who's against it? [00:46:05.0000] <annevk2> are documents public? [00:46:06.0000] <Mrmil> zcorpan: done. Address in aside only sounds a little limitating (hope this word exitsts) [00:47:00.0000] <arun> annevk2, I'd like to make 'em so by being efficient in releasing specs. [00:47:01.0000] <arun> annevk2, but we've barely started still. [00:47:02.0000] <zcorpan> Mrmil: you could send feedback to the list regarding <address> in <aside> [00:48:00.0000] <arun> Hixie, what do you mean by [IndexGetter] is more work to use than sequence<>? You mean, instead of a typedef statement, you have to deal with the whole interface stuff? [00:48:01.0000] <Mrmil> zcorpan: Ok, will do [00:50:00.0000] <Hixie> arun: you have to define what all the indices are and so on [00:50:01.0000] <arun> Hixie, blah. I see your point. [00:50:02.0000] <arun> sequence<> would have been easier. [00:50:03.0000] <annevk2> just use sequence [00:50:04.0000] <annevk2> it will be defined in the end [00:51:00.0000] <annevk2> IndexGetter and all is only needed when you need special members on the interface [00:52:00.0000] <arun> annevk2, hmm... I'm defining an interface which exposes an array and an unsigned long. Seems like a single typdef won't do the trick; I need an interface statement. [00:52:01.0000] <annevk2> what's the unsigned long? [00:53:00.0000] <annevk2> got to run, be back in a bit [00:55:00.0000] <Mrmil> /me wonders what to do next. Another page or polish the one? [00:58:00.0000] <zcorpan> Mrmil: you could rewrite the page in xhtml2 [00:59:00.0000] <Mrmil> zcorpan: Or I can as well jump out of window. [01:04:00.0000] <MikeSmith> anybody know of a way to tell svn to ignore ssl certificate errors? [01:06:00.0000] <Hixie> arun: if sequence<> defines an Array, then that's got your unsigned long [01:07:00.0000] <arun> Hixie, eh? I'm talking about readonly unsigned long length; [01:07:01.0000] <Hixie> right, depending on how heycam defines sequence<>, that might come for free [01:08:00.0000] <arun> heheh, that's a little imprecise. I'll roll with a classic interface with [IndexGetter] I think :) [01:08:01.0000] <arun> I like sequence<> tho'. [01:09:00.0000] <arun> I think we should get it spec'd to allow length attributes to come for free (modulo mutability) [01:11:00.0000] <arun> /me heads out for a brief pause [01:12:00.0000] <Hixie> yeah might be good to have a |sequence<T>| and a |readonly sequence<T>| or |[Readonly] sequence<T>| [01:13:00.0000] <Hixie> man i never have any idea what kristof is talking about [01:16:00.0000] <Mrmil> Hixie: Is it just me or does sound a little arrogant? [01:17:00.0000] <Hixie> who, kristof? [01:17:01.0000] <Mrmil> yes [01:18:00.0000] <Hixie> not compared to a lot of the traffic on the public-html list! [01:18:01.0000] <Hixie> (including my own!) [01:18:02.0000] <Hixie> i just wish he'd include context so i had any idea what he was talking about [01:18:03.0000] <Hixie> he's dramatically improved his tone since the early days [01:19:00.0000] <Hixie> which is good [01:19:01.0000] <othermaciej> bonus points to Sam for jumping on the bandwagon of "disagree == ignore" [01:19:02.0000] <othermaciej> which is kind of an arrogant way to think about things [01:19:03.0000] <zcorpan> i always typo <frameset> as <fieldset> [01:19:04.0000] <Mrmil> Hixie: Interesting :) [01:20:00.0000] <othermaciej> Kristof has sensible things to say at times but other times I get lost reading his messages [01:24:00.0000] <hsivonen> Hixie: might not hurt to put "authoring tools should default to using UTF-8 for newly-created documents" or something of that nature in the spec [01:25:00.0000] <Hixie> file a bug, suggesting where to put it [01:25:01.0000] <hsivonen> Hixie: OK. [01:35:00.0000] <ojan> Hixie: is http://dev.w3.org/html5/websockets/ the most up to date websockets draft? [01:36:00.0000] <ojan> i couldn't find a version on whatwg.org [01:36:01.0000] <annevk42> yes [01:36:02.0000] <Hixie> yes [01:36:03.0000] <ojan> thx [01:36:04.0000] <annevk42> but there's also an I-D that goes along with it [01:36:05.0000] <ojan> a wot? [01:36:06.0000] <Hixie> http://tools.ietf.org/html/draft-hixie-thewebsocketprotocol [01:37:00.0000] <annevk42> thot /\ [01:37:01.0000] <ojan> :) [01:38:00.0000] <Hixie> you implementing? [01:38:01.0000] <ojan> Hixie: no. i have a meeting tomorrow with people who'd like to implement though. [01:38:02.0000] <Hixie> ah, cool [01:39:00.0000] <ojan> figured i should read the spec first :) [01:39:01.0000] <Hixie> :-) [01:39:02.0000] <Hixie> the text thing describes the wire protocol [01:39:03.0000] <Hixie> the w3c one defines the dom api [01:39:04.0000] <ojan> i see, i was just about to ask that [01:39:05.0000] <Hixie> (the ietf got upset that we were defining wire protocols in the w3c) [01:40:00.0000] <ojan> /me is already sick of standards politics [01:43:00.0000] <hsivonen> ojan: you probably haven't seen the hybi list yet... [01:43:01.0000] <Hixie> it's very odd having my boss post to the whatwg list [01:43:02.0000] <Hixie> not to mention dannyb, who's also on my team [01:44:00.0000] <ojan> hsivonen: no, i haven't. what is that? [01:44:01.0000] <hsivonen> ojan: I mean, reading the hybi list is an interesting dose of politics [01:44:02.0000] <ojan> ah [01:44:03.0000] <hsivonen> ojan: it's the IETF list for web socket [01:44:04.0000] <hsivonen> ojan: and for things that got lumped into the same list [01:44:05.0000] <Hixie> technically it's the IETF list for bidirectional HTTP [01:44:06.0000] <Hixie> and websocket somehow got dragged into it [01:45:00.0000] <Hixie> though most of the people there don't want it [01:45:01.0000] <Hixie> in practice if websocket is implemented, it'll become a moot issue [01:45:02.0000] <MikeSmith> Hixie: who's your boss? [01:46:00.0000] <othermaciej> I'm not sure what "bidirectional HTTP" even means [01:46:01.0000] <Hixie> chris dibona [01:46:02.0000] <ojan> Hixie: we have a number of teams asking for it [01:46:03.0000] <Hixie> othermaciej: that's part of the problem [01:46:04.0000] <Hixie> ojan: you should have seen the wave team salivating when i told them about it :-P [01:46:05.0000] <othermaciej> people seem to have all sorts of different and weird ideas about it [01:46:06.0000] <othermaciej> WebSocket is on deck for us Apple folks as one of the next WebKit features to tackle, once Safari 4 final ships [01:47:00.0000] <ojan> othermaciej: there are chromium people who plan to start working on it soon as well. i have a meeting with them tomorrow. [01:48:00.0000] <othermaciej> ojan: maybe there should be some coordination so we don't end up with duplicated efforts [01:48:01.0000] <ojan> othermaciej: totally. [01:49:00.0000] <othermaciej> (and I am remiss in not really sharing our engine feature plans, but we'll probably let webkit-dev know soonish about some of the big things we are interested in working on) [01:49:01.0000] <ojan> othermaciej: that would definitely help in our roadmap planning. [01:50:00.0000] <Hixie> should usemap="#x" require a matching id="x" or a matching name="x"? hsivonen, zcorpan, any opinions? spec says name="", but i think that's a typo [01:50:01.0000] <othermaciej> would help to hear the same from Google [01:50:02.0000] <ojan> othermaciej: are there specific webkit folk they should contact outside of webkit-dev? [01:50:03.0000] <hsivonen> Hixie: isn't name what actually works everywhere except Gecko's XML mode? [01:50:04.0000] <othermaciej> webkit-dev is probably better than contacting specific folks in any case [01:51:00.0000] <hsivonen> Hixie: I think I sent feedback when I last tested this, but now I can't remember what I said [01:51:01.0000] <Hixie> oh, wait, i was looking at <img name> [01:51:02.0000] <Hixie> but it should be <map name>, right? [01:51:03.0000] <Hixie> and that exists [01:51:04.0000] <Hixie> so nevermind [01:51:05.0000] <Hixie> i'm an idiot [02:21:00.0000] <ojan> Hixie: what happens with websockets if the server sends the client some data and the client closes the socket before receiving that data? [02:22:00.0000] <ojan> or, more importantly, the other way around...client sends the server data, but the server closes the socket before receiving the data [02:22:01.0000] <ojan> is there a way for the client to detect that the server never got the transmission? [02:26:00.0000] <ojan> hm...i guess TCP/IP deals with this for you so that situation isn't actually possible? namely, the packet transmission would fail [02:27:00.0000] <ojan> nevermind...it is possible. client calls postMessage, then gets disconnected from the internet. is there anyway for them to see that the server never got the postMessage data? [02:28:00.0000] <Philip`> What if the server got the message but then got disconnected from the internet itself, so it couldn't send any acknowledgement back [02:28:01.0000] <Philip`> s/$/?/ [02:28:02.0000] <Hixie> ojan: if they care, they have to have the server ack everything [02:29:00.0000] <ojan> Hixie: i see [02:55:00.0000] <hsivonen> http://www.w3.org/mid/E34714E1E6587741AD32C8E445B6F4AB01637C9D⊙dwsc is interesting [03:04:00.0000] <MikeSmith> hsivonen: I very much hope that vcs for validator.nu moves to something other than subversion eventually [03:06:00.0000] <jgraham> /me starts doing a little mercurial dance in the background [03:08:00.0000] <Philip`> /me averts his gaze in embarrassment [03:09:00.0000] <jgraham> Philip`: Just be glad I asn't doing the Git dance [03:10:00.0000] <Philip`> YAML really is a rubbish output format [03:10:01.0000] <Philip`> (probably because it's incredibly complicated, and therefore nobody implements it properly) [03:11:00.0000] <Philip`> JSON seems to work a lot better, i.e. it actually works [03:11:01.0000] <hsivonen> MikeSmith: ok. if now is the time when svn is actually a problem, I guess exploring a move to hg is now justified. [03:12:00.0000] <othermaciej> Hixie, ojan: depending on the particular case, there might be simpler ways to check than having the server ack everything [03:12:01.0000] <othermaciej> for example, the client could include a sequence number in every message, and when making a new connection to the server, the server could report the last sequence number it got [03:12:02.0000] <Hixie> indeed [03:12:03.0000] <othermaciej> TCP guarantees you won't lose just middle messages [03:13:00.0000] <hsivonen> MikeSmith: CVSDude doesn't have hg yet, so switching to hg means either bitbucket or Google. [03:13:01.0000] <othermaciej> and as long as you remain connected there's no need to ack [03:13:02.0000] <othermaciej> it's only after losing connectivity that it's possible some tail of the stream never arrived [03:13:03.0000] <ojan> yup [03:13:04.0000] <hsivonen> MikeSmith: I'll think about this and will ask fantasai after I've thought about this [03:13:05.0000] <MikeSmith> hsivonen: OK [03:15:00.0000] <MikeSmith> hsivonen: it's not an acute problem for many at the moment, since I finally managed to solve the particular problem I was having (but only after spending a good part of the afternoon trying to troubleshoot it) [03:15:01.0000] <MikeSmith> but I've found subversion to be a chronic time-waster [03:16:00.0000] <hsivonen> yeah, svn sucks once you've seen hg [03:16:01.0000] <hsivonen> svn seemed great when you had only seen cvs [03:16:02.0000] <zcorpan> contenteditable is such a mess [03:16:03.0000] <hsivonen> also, gotta check on the hg Eclipse plug-in situation [03:17:00.0000] <jgraham> zcorpan: and in other news... [03:17:01.0000] <MikeSmith> hsivonen: I find the worst thing about the svn client and library is that when I do run into problems with it, it is very time-consuming to try to troubleshoot them -- mostly because of the almost total lack of any useful error messages about what's causing the problem [03:18:00.0000] <MikeSmith> in the case of the problem I ran into today, it was because of libneon [03:18:01.0000] <hsivonen> /me has no idea what libneon is [03:18:02.0000] <MikeSmith> but there was no way I could have ever figured that out from the error output that svn was giving [03:18:03.0000] <MikeSmith> hsivonen: I didn't know what libneon was until today either [03:19:00.0000] <MikeSmith> apparently it is what subversion relies on for https [03:40:00.0000] <MikeSmith> is there some parameter I can pass to http://html5.org/tools/web-apps-tracker to make it show more than the default number of changes that it shows? [03:40:01.0000] <MikeSmith> annevk42: ↑ [03:40:02.0000] <MikeSmith> (assuming it was annevk42 who wrote the interface) [03:41:00.0000] <jgraham> MikeSmith: There is but I don't know what [03:42:00.0000] <MikeSmith> k [03:42:01.0000] <Philip`> MikeSmith: ?limit=123 [03:43:00.0000] <MikeSmith> Philip`: thanks [03:43:01.0000] <Philip`> according to http://code.google.com/p/html5/source/browse/trunk/web-apps-tracker/trackerlib.py [03:43:02.0000] <annevk42> limit [03:43:03.0000] <annevk42> you can even get infinite by using -1 [03:46:00.0000] <MikeSmith> sweet [03:52:00.0000] <MikeSmith> hsivonen: anyway, I'm very happy to help in any way I can with migration to mercurial hosting, if/when you want to migrate [03:54:00.0000] <Hixie> ok bed time [03:54:01.0000] <Hixie> nn [04:18:00.0000] <hsivonen> no <video> in Opera 10 beta, right? [04:21:00.0000] <annevk42> I don't think we'll do <video> for 10 [04:21:01.0000] <hsivonen> ok [04:25:00.0000] <hsivonen> hmm. Opera just autoupdated itself to Opera/10.00 (Macintosh; Intel Mac OS X; U; en) Presto/2.2.1 build 6343 calling itself alpha with the old appearance [04:40:00.0000] <hendry> hmm, was hoping Opera 10 would support border-radius [04:41:00.0000] <hsivonen> I guess Alpha really means Alpha at Opera, since the engine features didn't get a trunk update between Alpha and Beta [04:42:00.0000] <hsivonen> /me wonders how many concurrent engine branches Opera has [04:43:00.0000] <zcorpan> Opera 10 alpha and Opera 10 beta both use core 2.2 which is pretty old now [04:48:00.0000] <hsivonen> I was hoping to get to a point where I could write that three out of five support Theora natively, one supports it with XiphQT and the last one has issues but has Cortado :-/ [04:48:01.0000] <hsivonen> I guess I'll have to wait until Opera 10.next [04:49:00.0000] <hsivonen> /me wonders how <video> is doing in Epiphany-WebKit [05:49:00.0000] <hsivonen> ezyang: the test case for <!DOCTYPE html><body xlink:href=foo><math xlink:href=foo></math> is now wrong [05:49:01.0000] <hsivonen> xlink:href doesn't go into the XLink namespace on HTML elements [05:50:00.0000] <hsivonen> (I'll fix) [05:50:01.0000] <jgraham> /me thought he had fixed that [05:51:00.0000] <hsivonen> /me does hg pull [05:51:01.0000] <hsivonen> ok. I had a stale local repo [05:54:00.0000] <hsivonen> hmm. the frameset tests have changed to be less WebKit-compatible [05:54:01.0000] <hsivonen> /me tests IE [05:56:00.0000] <hsivonen> I'm unable to figure out what the actual rendered effect of IE's crazy DOM would be [05:56:01.0000] <hsivonen> but the tests are now less aligned with Opera, too [06:08:00.0000] <hsivonen> why are </body> and </html> differnt in 'in head' and 'in head noscript'? [06:32:00.0000] <Mrmil> /me wonders if anyone is alive here. He added one little thing to his test page and might need a little feedback again. [06:33:00.0000] <jgraham> Everyone's dead, Dave [06:33:01.0000] <Mrmil> Auuuuu I'll eat you my little fishie... [06:40:00.0000] <Mrmil> never mind then [06:40:01.0000] <annevk42> it's easier to just post your question and later check the minutes if someone picked up on it [06:40:02.0000] <annevk42> this IRC channel works very well asynchronously from time to time [06:41:00.0000] <Mrmil> I know, I just posted it here 5 million times that I'd feel like spamming [06:42:00.0000] <Mrmil> It's here http://server.ebrana.cz/olda/_apps/html5/ and I added the little searchform in the head [06:44:00.0000] <Philip`> I don't particularly like how the search box expands when I click on it - that seems like unusual and unexpected behaviour [06:44:01.0000] <Mrmil> better now? [06:45:00.0000] <Mrmil> right, if it should do that, it should probably expand on the other side, hm hm [06:46:00.0000] <Philip`> I like it now that it doesn't surprisingly expand for no reason :-) [06:46:01.0000] <Mrmil> Ok :) [06:47:00.0000] <Mrmil> the html of it is pretty simple, there is no problem I think [06:47:01.0000] <Philip`> Hmm, does HTML 5 not have <input type=search>? [06:48:00.0000] <Philip`> Oh, it does [06:48:01.0000] <Philip`> http://www.whatwg.org/specs/web-apps/current-work/multipage/forms.html#text-state-and-search-state [06:48:02.0000] <Philip`> Maybe that would be appropriate for the search box? [06:49:00.0000] <Mrmil> damn, there you go, I am so used to this html that I forgot there could be even better solution. Thanks Philip, I'll fix that [06:49:01.0000] <Mrmil> I haven't really looked into the html5 forms, but it's on my list [06:50:00.0000] <Philip`> Hmm, I think <time datetime="2009-06-03T13:20:34"> is invalid [06:51:00.0000] <Philip`> If it's a date-and-time then it needs to be a global date-and-time, i.e. include timezone information [06:51:01.0000] <Mrmil> I checked that, it's valid [06:52:00.0000] <Philip`> According to the spec, or according to a possibly-buggy validator? [06:52:01.0000] <Mrmil> according to Hixie [06:52:02.0000] <Mrmil> he said it was optional [06:53:00.0000] <Philip`> http://www.whatwg.org/specs/web-apps/current-work/multipage/text-level-semantics.html#the-time-element goes to http://www.whatwg.org/specs/web-apps/current-work/multipage/infrastructure.html#valid-date-or-time-string which goes to http://www.whatwg.org/specs/web-apps/current-work/multipage/infrastructure.html#valid-global-date-and-time-string [06:53:01.0000] <Philip`> which says you need either Z or [+-]\d\d:\d\d [06:53:02.0000] <Philip`> Am I missing something here? [07:04:00.0000] <Mrmil> 2009-06-03T13:20:34+02:00 so this would be CET time right [07:05:00.0000] <Philip`> CET with daylight saving would be +02:00, I believe [07:06:00.0000] <Mrmil> right [07:18:00.0000] <wakaba_> MikeSmith: http://pastebin.com/f3a189e59 [07:18:01.0000] <wakaba_> my emobile connection don't have route, either... [07:22:00.0000] <Mrmil> Philip`: Whew, we have old php here, had to do it manually. Good now? <time datetime="2009-06-03T15:02:40+02:00"> [07:23:00.0000] <Philip`> Mrmil: Looks good to me, though I'm not sure why validator.nu accepts it without the timezone [07:25:00.0000] <Mrmil> Philip`: well, me neither, though it does sound weird to have the stamp without a timezone [07:30:00.0000] <ezyang> hsivonen: (stark) So... is this fine distinction between having an actual namespace and being a literal foo:bar actually what happens in browsers? [07:32:00.0000] <hsivonen> ezyang: yes [07:32:01.0000] <ezyang> Good grief [07:33:00.0000] <ezyang> It's almost as if they ignored all of the good bits of XML. [07:34:00.0000] <Philip`> The implementations existed before XML did, and then it became too late to change because of legacy content [07:34:01.0000] <hsivonen> ezyang: what would be the good bits in this case? [07:34:02.0000] <ezyang> The good bits would be consistency and namespacing. [07:35:00.0000] <Philip`> Controversial :-o [07:35:01.0000] <hsivonen> I agree about consistency. I disagree on namespacing. [07:35:02.0000] <ezyang> As a tool author, namespaces are uniformly a good thing. It's unclear about that on the browser side [07:35:03.0000] <Philip`> HTML is consistent - you parse a sequence of non-whitespace characters and use that as the attribute name, and that's it (and you don't put the attribute into any particular namespace) [07:37:00.0000] <annevk42> Where is the Google Wave desktop drag & drop API defined? [07:38:00.0000] <annevk42> I couldn't find it in the Gears documentation [07:39:00.0000] <jgraham> annevk42: Is there work going on anywhere to make that stuff work without gears? [07:41:00.0000] <Philip`> Does Hg make it possible to import an external library into my repository, apply some local patches, and then easily keep the external library up-to-date with external revisions? (I presume it's just some kind of simple merging, but I don't know enough about Hg to know if there's a specific way to make it work properly) [07:42:00.0000] <jgraham> annevk42: (e.g. in webapps since I don't really follow that) [07:45:00.0000] <jgraham> Philip`: It seems like the answer is "Maybe". There are a couple of adddsons that claim to do things like svn:externals but it's not clear if they work well or not [07:47:00.0000] <annevk42> jgraham, about as much work as work on video streaming and audio manipulation APIs [07:47:01.0000] <jgraham> annevk42: Is that a long way of saying "no but people would like it to"? [07:48:00.0000] <annevk42> jgraham, though in the Wave video Google claims they want to work on this soon [07:48:01.0000] <annevk42> jgraham, a long way of saying no work at all [07:48:02.0000] <annevk42> jgraham, but also that I think there should :) [07:48:03.0000] <jgraham> annevk42: Agreed :) [07:48:04.0000] <Philip`> jgraham: Ah, I don't want it to be automatic - I just want it to be a straightforward manual process (e.g. copy the external library's latest files, then commit it to Hg somehow, then hope that a normal Hg checkout will incorporate those latest files plus any custom patches) [07:49:00.0000] <Philip`> (presumably with some merging happening somewhere but I don't really know where) [07:49:01.0000] <jgraham> Philip`: Presumably you can just clone the original then make your changes and then pull the original changes every so often and remerge [07:50:00.0000] <Philip`> Hmm [07:50:01.0000] <Philip`> Where would the original changes be pulled from? [07:51:00.0000] <jgraham> Philip`: The VCS of the external library [07:51:01.0000] <jgraham> I guess [07:52:00.0000] <jgraham> /me feels he is missing something [07:52:01.0000] <Philip`> Can you pull from it if it's using SVN? [07:53:00.0000] <jgraham> Philip`: Yeah if you use {some random tool that lets you do that} [07:53:01.0000] <Philip`> /me doesn't really know much about this at all [07:53:02.0000] <jgraham> hgsvn maybe? [07:53:03.0000] <Philip`> Ah, okay [07:53:04.0000] <jgraham> http://www.selenic.com/mercurial/wiki/WorkingWithSubversion [07:57:00.0000] <Philip`> Sounds like it's all kind of straightforward, which is nice [08:04:00.0000] <MikeSmith> wakaba_: thanks. cause of my problem turned out to not be a routing issue but instead a debian packaging flub combined with fundamental borkedness in svn [08:46:00.0000] <Hish> hi. is there an official way how to implement svg in html5? I only know that xhtml/svg is working... [08:47:00.0000] <annevk42> you mean as an author? [08:48:00.0000] <annevk42> at some point <div> <svg> <circle /> </svg> </div> (with appropriate attributes) will work [08:49:00.0000] <Hish> without namespaces? [08:51:00.0000] <Philip`> Hish: Yes [08:52:00.0000] <Hish> ok. thx a lot. [08:52:01.0000] <Philip`> Hish: (though you'll have to wait years until you can use it reliably) [08:52:02.0000] <Hish> sure. i stick with xhtml/svg for the time being [08:55:00.0000] <gsnedders> I guess in recent browsers you can do something like changeNSElements = document.getElementsByTagName("svg").getElementsByTagName("") and change the namespace of each element [08:55:01.0000] <gsnedders> I guess you'll have to do it for the svg element itself [08:59:00.0000] <jgraham> gsnedders: If you were a bit more clever than that you could probably implement something not too dissimmilar to the foreign content mode of the parser using js working on the post-parse tree [08:59:01.0000] <jgraham> With some restrictions like no self-closing elements [09:00:00.0000] <gsnedders> jgraham: Probably. But that's more work. :) [09:00:01.0000] <beowulf> /me subscribes to gsnedders newsletter [09:01:00.0000] <gsnedders> beowulf: Pff. Writing one is too much effort. [09:02:00.0000] <jgraham> gsnedders: It would be nice to support MathML and HTML in SVG in HTML. So you would at least want to special-case <foreignContent> [09:04:00.0000] <gsnedders> hmm… I'm not getting Opera 10 Beta as an auto-update :\ [10:25:00.0000] <ezyang> gsnedder: ping [10:28:00.0000] <jgraham> /me wishes we could just drop all the week numbering stuff from HTML5 [10:50:00.0000] <gsnedders> ezyang: Including my full nick in pings helps :P [10:52:00.0000] <ezyang> heh [11:10:00.0000] <gsnedders> I could do with some way of raising warnings in Anolis [11:13:00.0000] <gsnedders> And then do things like warn about referencing obsolete docs [12:00:00.0000] <takkaria> wow [12:00:01.0000] <takkaria> http://lists.w3.org/Archives/Public/www-tag/2009Jun/0000.html [12:00:02.0000] <takkaria> and more specifically, http://larry.masinter.net/versioning-html.html [12:04:00.0000] <tantek> /me notes that the versioning-html.html document itself lacks traditional W3C document versioning, i.e. as described/shown here: http://www.w3.org/2001/06/manual/#Errata [12:04:01.0000] <zcorpan> /me fails to resist the temptation to discuss bytes-to-DOM of XML and error handling in XML with the XML Core WG [12:05:00.0000] <zcorpan> http://lists.w3.org/Archives/Public/public-xml-core-wg/2009Jun/0009.html [12:05:01.0000] <gsnedders> zcorpan: There is fairly decent error handling dictated in the XML spec at least :P [12:15:00.0000] <zcorpan> gsnedders: but not error recovery [12:30:00.0000] <gsnedders> zcorpan: No, but it is handling :P [13:09:00.0000] <zcorpan> yay my edits on wikipedia haven't been reverted yet [13:12:00.0000] <gsnedders> /me wonders what edits [13:12:01.0000] <zcorpan> to the html5 page [13:13:00.0000] <atwilson> I'm somewhat confused about the language surrounding task sources in section 6.5.4 - in particular, if an object is no longer used as a task source for an event loop (for example, a MessagePort that has been closed, does that imply that all queued (but not yet dispatched) events for that task source are removed from the queue? [13:13:01.0000] <atwilson> Anyone know what behavior is intended? [13:16:00.0000] <Hixie> atwilson: why would that cause it to no longer be used? [13:17:00.0000] <atwilson> Ah. [13:17:01.0000] <atwilson> I read this: When a port's port message queue is open, the event loop must use it as one of its task sources. and assumed that causing close() would also close the message queue, but apparently that is not the case. [13:18:00.0000] <Hixie> nope [13:18:01.0000] <Hixie> i should probably make that clearer [13:18:02.0000] <atwilson> Close just prevents the ports from communicating with one another, but doesn't affect the open status of the queues. [13:18:03.0000] <Hixie> maybe use different words than open/close [13:20:00.0000] <atwilson> Yeah, maybe. The spec pretty clearly doesn't say that the message queue is closed - that was entirely my misreading, although I guess explicitly stating that the port message queue should be left open wouldn't be bad. [13:20:01.0000] <zcorpan> "W3C: Web Storage draft standard" - http://en.wikipedia.org/wiki/DOM_storage [13:22:00.0000] <Hixie> atwilson: i've reowrded it to say "enabled" vs" disabled" and to mention that it never goes back to disabled [13:22:01.0000] <atwilson> Hixie: OK. Note that it does* go back to disabled if the port is cloned. [13:22:02.0000] <Hixie> actually no, it just gets emptied [13:23:00.0000] <Hixie> this has much the same effect, of course [13:23:01.0000] <atwilson> Ah, you are right again. [13:23:02.0000] <Hixie> i have an unfair advantage in that i wrote this nonsense :-) [13:23:03.0000] <Hixie> sadly i still sometimes find i wrote something that i didn't think i wrote :-( [13:24:00.0000] <atwilson> Anyhow, I like having the port never go back to disabled. Thanks for clarifying. [13:25:00.0000] <Hixie> only 850 e-mails to go to hit my OKR [13:26:00.0000] <Hixie> that's about 30 e-mails a day until the end of the month [13:26:01.0000] <Hixie> should be feasible... [14:21:00.0000] <jgraham> gsnedders: You could use the warnings module with anolis to generate warnings, maybe [14:46:00.0000] <gsnedders> jgraham: Are they not just exceptions, which would make them hard to use? [14:50:00.0000] <gsnedders> jgraham: What could PMS do with them anyway? [14:50:01.0000] <gsnedders> jgraham: I guess it could put them in a comment before the root element [14:57:00.0000] <jgraham> gsnedders: They are basically exceptions yes [14:57:01.0000] <jgraham> But higher level code can choose which ones to react to and how to deal with them [14:57:02.0000] <jgraham> Dunno if it's a good idea or not [14:59:00.0000] <jgraham> gsnedders: In particular it might be more sutiable for language/api level stuff than application level stuff [15:00:00.0000] <jgraham> For your use case it is probably a bad idea [15:02:00.0000] <gsnedders> /me orders Lightroom [15:02:01.0000] <gsnedders> /me hopes it is all right to use his father's name in the billing address (as that must be as on credit card) while doing it at the educational store in my name… [15:10:00.0000] <gsnedders> Now, phase two of spending my father's money: plane tickets to get to/from Sweden. [15:12:00.0000] <gsnedders> noooo… [15:12:01.0000] <gsnedders> flight sold out since yesterday! [15:12:02.0000] <gsnedders> Sux. [15:12:03.0000] <gsnedders> Oh, no. [15:12:04.0000] <gsnedders> Just Ryanair website maybe broken… [15:13:00.0000] <gsnedders> I should probably be using a released web browser… [15:22:00.0000] <gsnedders> Yay. Ryanair is awesome. Buying single plane tickets you have to buy them one at a time. [15:34:00.0000] <jgraham> gsnedders: Their website is intentionally crap to make you think that they are cheap [15:34:01.0000] <gsnedders> jgraham: They're cheaper than Flyglobespan at least :D [15:35:00.0000] <gsnedders> (Which is a similar low-cost airline) [15:35:01.0000] <jgraham> hsivonen: Does the change to the test with <body><frame></frame></frame><frameset> reflect some anticipated change in the spec or can I not read? [15:35:02.0000] <gsnedders> Oh, fun fact: last October I flew from Edinburgh to Paris by Air France because it was cheaper than flying by EasyJet [15:36:00.0000] <jgraham> AFAICT the only way to set the frameset-OK flag to true is by not having an explicit <body> [15:36:01.0000] <jgraham> Although it is too late and I am too tired for me to try reading the spec properly, so I could be wrong [15:37:00.0000] <jgraham> gsnedders: I bet Ryanair are much much worse. They are much worse than EasyJet [15:37:01.0000] <gsnedders> jgraham: Indeed. [15:38:00.0000] <gsnedders> jgraham: My flight from Nice to Edinburgh on Flyglobespan, a apparently low-cost airline cost something like £120 [15:42:00.0000] <gsnedders> Sadly, however, Ryanair is the only airline that flies the routes I need this summer. [15:42:01.0000] <gsnedders> STN–NYO: 73.26 GBP [15:42:02.0000] <gsnedders> NYO–SXF: 220 SEK [15:42:03.0000] <gsnedders> SXF–STN: 25 EUR [16:00:00.0000] <hober> Mr. Last Week must be disappointed that http://www.google.com/squared/search?q=WHATWG+teamsters doesn't result in anything. [16:02:00.0000] <Philip`> http://www.google.com/squared/search?q=html5+elements - hmm 2009-06-04 [18:20:00.0000] <heycam> what's the best way to specify fallback for mathml content in an html 5 document (say, a png of the math)? [20:06:00.0000] <jcranmer> Hixie: can you convice the Google Video/YouTube people to use <video> ? [20:09:00.0000] <jcranmer> tell them they're evil bastards until they do so [20:09:01.0000] <jcgregorio> jcranmer: http://www.youtube.com/html5 [20:10:00.0000] <jcgregorio> that was demoed at Google IO [20:10:01.0000] <jcranmer> can it load ALL videos in there? [20:12:00.0000] <jcranmer> that seems to break in FF 3.6a1pre [20:15:00.0000] <kinetik> jcranmer: it's serving h.264 [20:44:00.0000] <shepazu> http://www.google.com/squared/search?q=races [21:53:00.0000] <roc> that's cool, but check this out: [21:53:01.0000] <roc> http://www.google.com/squared/search?q=programming+languages [21:53:02.0000] <roc> why is there a "death" column? [21:54:00.0000] <roc> with very ... strange ... values [22:09:00.0000] <MikeSmith> <cite style="font-style: normal">Lincoln</cite> [22:09:01.0000] <MikeSmith> nice [23:03:00.0000] <hsivonen> jgraham: whoa. sorry. I just noticed that the test case didn't match Safari and Opera and checked that <frame> and </frame> don't set frameset-o to not ok [23:12:00.0000] <hobo> What is going to happen with the onbeforeunload and onunload events, Opera doesn't acknowledge them [23:12:01.0000] <hobo> does whatwg deal with that? [23:14:00.0000] <roc> I find it hard to believe Opera doesn't support them [23:16:00.0000] <hobo> i dunno my test keeps failing [23:16:01.0000] <Hixie> the html5 spec does define those events, yes [23:17:00.0000] <hobo> The previous specs included them as well right? [23:17:01.0000] <hobo> Opera just does whatever? [23:21:00.0000] <Hixie> i don't think they've ever been defined [23:21:01.0000] <Hixie> some spec might have mentioned them though i doubt it [01:10:00.0000] <annevk42> Opera doesn't support those events [01:10:01.0000] <annevk42> (yet) [01:12:00.0000] <annevk42> heycam, no fallback is needed for MathML; it just needs to be supported [01:12:01.0000] <heycam> annevk42, ? [01:12:02.0000] <heycam> what if i want to target browsers that do support mathml in text/html as well as those that don't? [01:13:00.0000] <hsivonen> heycam: tough luck [01:14:00.0000] <hsivonen> heycam: actually, there's a way [01:15:00.0000] <hsivonen> heycam: 1) make all your MathML text nodes into CDATA sections [01:16:00.0000] <hsivonen> heycam: 2) stick and <img> with the image of the formula into annotation-xml [01:16:01.0000] <hsivonen> heycam: 3) hide annotation-xml using a selector that has an explicit namespace so it only matches if the elements really got assigned to the MathML namespace [01:17:00.0000] <heycam> that's so crazy it just might work [01:17:01.0000] <Philip`> What if a browser implements the HTML5 parsing algorithm but not MathML rendering? [01:18:00.0000] <heycam> in reality, pubrules won't let me use mathml in text/html anyway so the question is moot (for my purpose) [01:18:01.0000] <annevk42> Philip`, browsers have bugs [01:18:02.0000] <heycam> annevk42, likes [01:18:03.0000] <heycam> erm, lies! [01:18:04.0000] <hsivonen> Philip`: oh, right. are there MathML elements that in the normal course of MathML rendering hide <annotation> or <annotation-xml>? [01:18:05.0000] <annevk42> film at eleven [01:18:06.0000] <hsivonen> Philip`: at least <annotation-xml> as child of <semantics> isn't hidden automatically, IIRC [01:19:00.0000] <annevk42> heycam, you wish :p [01:19:01.0000] <hsivonen> heycam: anyway, I recommend playing with <annotation> and <annotation-xml> [01:19:02.0000] <heycam> hsivonen, ok [01:20:00.0000] <heycam> are there any browsers that support mathml in text/html yet? [01:20:01.0000] <hsivonen> heycam: there are my tryserver builds [01:20:02.0000] <heycam> k [01:20:03.0000] <hsivonen> heycam: SVG is easier, because you can count on <desc> getting hidden by the UA when SVG is supported [01:21:00.0000] <heycam> put fallback content inside <desc>? [01:21:01.0000] <hsivonen> heycam: right [01:21:02.0000] <heycam> it is a kind of description i suppose [01:42:00.0000] <annevk42> http://sandfly.net.nz/blog/2009/05/the-html5-video-tags-fatal-flaw/#comment-463 -- very typical indeed :p [01:46:00.0000] <hsivonen> sigh. the next commenter blames Mozilla instead of blaming MPEG-LA [01:46:01.0000] <heycam> oh my goodness is that blog the shape of things to come (re @font-face usage)? [01:47:00.0000] <annevk42> isn't it awesome? [01:47:01.0000] <hsivonen> /me notes that howcome promoted Theora before Mozilla was visibly behind Theora [01:47:02.0000] <annevk42> @font-face is the new <marquee> + animated GIF [01:47:03.0000] <hsivonen> (not to suggest blaming howcome. MPEG-LA should be blamed.) [01:48:00.0000] <hsivonen> annevk42: I haven't seen any animated @font-face yet [01:48:01.0000] <annevk42> lol [01:50:00.0000] <heycam> and of course you know that svg fonts can have videos and animated paths and whatnot inside glyphs [01:51:00.0000] <annevk42> not really, but I do now... [01:53:00.0000] <Philip`> It should only take a few years for things to settle down, so that people only use @font-face when it's a sensible design decision rather than using it simply to be cutting-edge [01:54:00.0000] <Philip`> and in the meantime I hope Opera adds a 'disable font download' option [01:55:00.0000] <heycam> /me imagines an svg font whose glyphs have child foreignObjects with marquees in them [01:55:01.0000] <jwalden> /me ponders self-referential SVG fonts [01:56:00.0000] <Philip`> The advantage of SVG fonts is that they're weird and complicated and nobody knows how to use them, so they won't get abused [01:57:00.0000] <Philip`> whereas anyone can download a TTF file from some random 1001 FREE FONTS! site and stick it on their own page [01:58:00.0000] <jwalden> frankly, I don't understand the use case for generic SVG font embedding, outside of use within SVG files themselves [01:58:01.0000] <jwalden> a large amount of complexity for not that much gain over font files themselves [01:59:00.0000] <jwalden> "I've always wanted to animate my font's glyphs!" [02:22:00.0000] <othermaciej> jwalden: SVG fonts do have some capabilities that are not available in typical outlined font formats [02:22:01.0000] <othermaciej> jwalden: such as multicolored characters [02:22:02.0000] <othermaciej> whether anyone should use such capabilities is beyond me [02:22:03.0000] <jwalden> wait, do fonts have intrinsic colors? [02:22:04.0000] <othermaciej> they don't have to [02:22:05.0000] <othermaciej> but they can [02:22:06.0000] <jwalden> what does that even mean in conjunction with CSS? [02:23:00.0000] <jwalden> castles in the sky [02:23:01.0000] <othermaciej> the outline definition doesn't by default define a color, in which case the CSS color is followed [02:23:02.0000] <othermaciej> but you can draw part or all of the character with an explicit color [02:23:03.0000] <jwalden> gag gag gag gag gag [02:24:00.0000] <jwalden> /me wonders what other absurd bells and whistles exist [02:24:01.0000] <othermaciej> also, for very one-off custom letterforms for headings, it may be easier to draw them with SVG than to create a throwaway TTF [02:24:02.0000] <othermaciej> but still I think TrueType Web fonts are much more useful [02:24:03.0000] <jwalden> it still seems unlikely that conversion tools are so difficult to procure or so impossible to use to validate the concept [02:25:00.0000] <Philip`> For custom headings you might as well just use an SVG file, rather than create a font and use it in @font-face [02:25:01.0000] <jwalden> although I suppose you do need SVG for your glyphs to be able to have realtime socket access to change in response to external network inputs [02:26:00.0000] <othermaciej> I don't think that SVG fonts were necessarily created with well thought out use cases and as a meaningful alternative to TrueType [02:26:01.0000] <othermaciej> it just seemed like something that an output format for Adobe Illustrator should have [02:27:00.0000] <jwalden> ah, back to the famed Macromedia hijinks again [02:28:00.0000] <Philip`> Macromedia? [02:28:01.0000] <Philip`> /me thought it was always an Adobe thing [02:30:00.0000] <othermaciej> SVG was Adobe's pet project before they formed Macrodobe and decided SVG should die [02:41:00.0000] <zcorpan> Philip`: you can't select the text if you just use an svg file, though [02:42:00.0000] <zcorpan> although i probably wouldn't bother creating an svg font as opposed to just use an svg file [02:47:00.0000] <hsivonen> othermaciej: my impression of the history is that having Illustrator export normal hinted OpenType fonts along with SVG would have been a horrible pirate vector [02:47:01.0000] <hsivonen> othermaciej: so SVG needed something that you can't drop in your system font folder [02:48:00.0000] <hsivonen> othermaciej: and then it just ballooned from there [02:48:01.0000] <othermaciej> hsivonen: I did not at the time follow nor have I since studied the actual history [02:48:02.0000] <othermaciej> I assumed it was part of SVG's desire to do everything in SVG [02:48:03.0000] <hsivonen> othermaciej: the above is speculation from few data points [02:49:00.0000] <hsivonen> I haven't done actual mailing list archeology [02:49:01.0000] <roc> does anyone actually support the full SVG font madness with arbitrary SVG content in glyphs? [02:49:02.0000] <roc> with animations and all? [02:49:03.0000] <roc> and looking at that blog, is it too late to drop support for font-face? [02:49:04.0000] <othermaciej> WebKit supports arbitrary content in glyphs, though I don't know if declarative animations will work [02:50:00.0000] <hsivonen> othermaciej: I do note, however, that SVG (1.1 IIRC) spec has an unusual field of use prohibition for fonts [02:50:01.0000] <hsivonen> the most interesting Process question is if the PP applies if you violate the field of use prohibition [02:50:02.0000] <othermaciej> hsivonen: field of use prohibition seems like a bizzarre thing for a W3C spec to contain [02:51:00.0000] <roc> with the complete <use>-alike style inheritance, it seems like an implementation nightmare [02:52:00.0000] <hsivonen> roc: do you plan on implementing it all anyway? [02:52:01.0000] <jgraham> Hey the fonts on that blog are useful. They tell me that I have better things to do than read the blog without even having to look at the words. Such is the amazing ability of typography [02:52:02.0000] <roc> I'm no sure [02:53:00.0000] <roc> I think to do it "right" we'd have to clone the SVG content into anonymous children of the text [02:53:01.0000] <roc> or something equally crazy [02:54:00.0000] <roc> for non-HTML text this is really crazy [03:09:00.0000] <othermaciej> I think we only allow use of SVG fonts from SVG - not 100% sure though [03:09:01.0000] <jgraham> Philip`: BTW hg rebase looks like it might be useful for your source control problem [03:09:02.0000] <othermaciej> using SVG fonts also makes you take a slower and more memory-intensive code path [03:09:03.0000] <jgraham> /me thinks he will want to solave a similar problem so is quite interested in solutions [03:12:00.0000] <Philip`> "hg: unknown command 'rebase'" [03:12:01.0000] <jgraham> Philip`: It hould be packaged with hg > 1.1 I think, but you may have to enable it [03:13:00.0000] <Philip`> Ah, looks like an extension thingy [03:14:00.0000] <Philip`> I found that pulling from an SVN-converted-to-Hg repository didn't work anyway, because it pulled everything into the root directory, whereas I wanted it moved into some subdirectory instead [03:14:01.0000] <Philip`> so I just gave up and copied the files with no history [03:14:02.0000] <othermaciej> good god, that font is truly fucking terrible [03:14:03.0000] <Philip`> and if I want to patch the originals then I guess I can figure out some way to do that in the future [03:16:00.0000] <Philip`> http://sandfly.net.nz/blog/2009/02/a-better-handwriting-font-with-yourfontscom/ - apparently he is "very happy with the result" [03:17:00.0000] <Philip`> /me doesn't think it really aids in readability much [03:17:01.0000] <othermaciej> I am curious what is really up with Google's H.264 license by the way; if they can really legitimately ship ffmpeg with H.264 support, it seems like Mozilla could too if it coughed up the money [03:17:02.0000] <othermaciej> I'm not really sure if their explanation makes sense or not though [03:18:00.0000] <hsivonen> othermaciej: it wouldn't work for downstream distributors [03:19:00.0000] <othermaciej> is Chrome (as opposed to Chromium) redistributable? [03:20:00.0000] <annevk42> I think so [03:20:01.0000] <hsivonen> othermaciej: I meant not only redistributable as binary but redistributable as in compilining your own build and shipping that [03:20:02.0000] <hsivonen> compiling [03:20:03.0000] <annevk42> I think the point is that the patent license is for Chrome as a whole, not for usage of FFmpeg [03:21:00.0000] <othermaciej> I don't know the exact nature of Google's patent licenses, but it's unclear to me if source vs. binary would be a critical legal issue [03:21:01.0000] <roc> I'm waiting for our lawyers to reach some conclusions before I comment further [03:23:00.0000] <othermaciej> I don't think you can build your own Chrome (as opposed to Chromium) but that's mainly due to trademarks [03:23:01.0000] <roc> but even if Google's interpretation is completely right, I think it subverts the intent of the LGPL [03:23:02.0000] <othermaciej> I was surprised by their interpretation, such a thing had not occurred to me [03:23:03.0000] <othermaciej> but the legal issues are too subtle for me to have a useful opinion on whether is right [03:25:00.0000] <roc> We've fought pretty hard to keep Web standards royalty free [03:25:01.0000] <roc> I'm not in a hurry to abandon that fight now [03:25:02.0000] <Philip`> One email from a Google person made it sound like it was a loophole that was closed in the LGPL3 [03:25:03.0000] <othermaciej> I wish video standards had the same IP norms as Web standards [03:26:00.0000] <hsivonen> it seems it's that it's possible to be open-source in terms of copyright while still being freedom-deprived for downstream (e.g. by using BSD on code that infringes on a 3rd party patent) [03:26:01.0000] <othermaciej> even more, I wish algorithms and algorithmic techniques could not be patented [03:26:02.0000] <roc> absolutely [03:27:00.0000] <othermaciej> ffmpeg itself seems to be a case of open source software that is freedom-deprived upstream [03:27:01.0000] <othermaciej> since it infringes on third-party patents and doesn't grant you a license [03:27:02.0000] <othermaciej> er, downstream [03:27:03.0000] <roc> yeah but at least if you download ffmpeg from them, you get the same rights they have [03:27:04.0000] <othermaciej> and, in the US and other countries where software patents are valid [03:27:05.0000] <roc> when you download it from Google, you don't [03:28:00.0000] <roc> The open source community was up in arms when Novell appeared to have struck a patent license deal with Microsoft [03:29:00.0000] <othermaciej> GPL certainly seems less ambiguous about this than LGPL [03:29:01.0000] <othermaciej> since you can't link with non-GPL code which has a hypothetical grant of rights [03:30:00.0000] <othermaciej> I also believe the Novell move was seen as conceding Microsoft had legitimate patent claims against Linux, which was something that hadn't been convincingly demonstratd [03:30:01.0000] <othermaciej> afaik no one disputes that MPEG-LA has legitimate patent claims against ffmpeg (at least in countries where it holds patents) [03:30:02.0000] <roc> it's true that the GPL and the LGPL are different here [03:31:00.0000] <roc> I don't know why that is [03:31:01.0000] <roc> but I always thought they were in the same spirit [03:31:02.0000] <roc> modulo the virality [03:31:03.0000] <othermaciej> patents in Software and especially on open standards are kind of bullshit [03:31:04.0000] <hsivonen> /me wonders if the FSF is going to comment on this whole thing [03:32:00.0000] <othermaciej> I would be curious what they think, or a group like the SFLC [03:32:01.0000] <roc> does anyone really think that the Linux kernel does not infringe Microsoft's patents? [03:32:02.0000] <roc> the kernel is big and complicated [03:32:03.0000] <roc> Microsoft's patents are many and broad [03:32:04.0000] <roc> of course they intersect [03:33:00.0000] <othermaciej> it does make one wonder why Microsoft won't cite specifics in public [03:39:00.0000] <roc> I've never understood why Microsoft hasn't gone on an all-out patent attack on free software [03:39:01.0000] <roc> I would have [03:39:02.0000] <Philip`> Maybe they're not evil? [03:40:00.0000] <othermaciej> maybe they really don't have any patents that are all that good [03:40:01.0000] <roc> neither of those options seem credible [03:40:02.0000] <jgraham> maybe they don't want IBM to sue them? [03:42:00.0000] <roc> seems like a risk worth taking, they can always settle with IBM for the right amount of money [03:43:00.0000] <othermaciej> maybe they fear antitrust litigation [03:43:01.0000] <jgraham> Maybe they fear precipitating change in the whole patent system [03:43:02.0000] <othermaciej> though, I heard a recent story to the effect that Microsoft was threatening and pressuring WinMobile handset vendors who bundled a non-Pocket IE browser [03:44:00.0000] <roc> antitrust maybe... [03:44:01.0000] <roc> for a while I thought they might be holding fire so that the EU didn't get spooked away from formalizing software patents [03:44:02.0000] <roc> but that receded [03:45:00.0000] <roc> my brother is a patent attourney, a partner in the law firm that handles Microsoft's NZ work [03:45:01.0000] <roc> you can imagine what our family gatherings are like [03:45:02.0000] <othermaciej> that must be... interesting [05:35:00.0000] <hsivonen> hmm. it seems we may be lacking test case coverage for EOF inside doctype in each possible position [05:36:00.0000] <jgraham> hsivonen: In tree construction or in tokenizer? [05:36:01.0000] <jgraham> In tokenizer we will lack coverage for things that changed since Philip` made all those tests [05:36:02.0000] <hsivonen> jgraham: tokenizer [05:37:00.0000] <jgraham> In tree-construction we lack tests everywhere [05:37:01.0000] <jgraham> e.g. html5lib failed 0 tests due to not implementing the framesetOK flag at all [05:38:00.0000] <Philip`> /me really ought to update his tokeniser / test generator [05:39:00.0000] <jgraham> Philip`: Where is the code? I have no intentions to make changes I just want to know if it makes more sense to me than last time I looked at it... [05:40:00.0000] <Philip`> jgraham: There's some version at http://canvex.lazyilluminati.com/svn/tokeniser/ which might (or might not) be close to the latest version that exists [05:41:00.0000] <Philip`> test_gen.ml is the bit that generates tests, based on the interpreter in tokeniser_interp.ml which executes the algorithm in tokeniser_spec.ml [05:45:00.0000] <Philip`> Hmm, I'm not entirely certain that any of it makes sense to me [05:46:00.0000] <hsivonen> the tokenizer alone seems to have over 50 different errors and warnings [05:47:00.0000] <Philip`> I think test_gen basically has a list of test strings (initially [""]), then for each one it looks at the final interpreter state (just before it would consume an EOF) and finds all the 'interesting' characters for that state, and then constructs a new set of strings, and then trims out ones that end up in duplicate states of others, and then repeats a few times [05:48:00.0000] <Philip`> hsivonen: What kinds of errors? [05:49:00.0000] <hsivonen> Philip`: I mean the number of different error messages the V.nu tokenizer can emit [05:50:00.0000] <Philip`> hsivonen: Oh, right [06:00:00.0000] <jgraham> Is the live dom viewer broken in IE8-in-windows 7 for anyone else or is it just me? [06:00:01.0000] <Philip`> "this is clearly going to be a long discussion" - surely not! [06:00:02.0000] <Philip`> jgraham: Broken in what ways? [06:00:03.0000] <jgraham> Specifically using w() to write to the output [06:00:04.0000] <Philip`> jgraham: (e.g. some way other than the XSS filter breaking it?) [06:00:05.0000] <Philip`> Oh [06:00:06.0000] <jgraham> (it orks in compat view but that is rather useless) [06:01:00.0000] <Philip`> /me doesn't remember whether w() worked in IE8-in-Vista [06:39:00.0000] <hsivonen> maikmerten: does Cortado provide enough scripting surface to JScript to be able to support the HTML5 video API in IE? [06:40:00.0000] <maikmerten> hsivonen, actually right now I'm not 100% sure the IE-Java scripting bridge works at all ;-) [06:40:01.0000] <maikmerten> but anyway, right now Cortado is just exposing bare essentials somewhat HTML5-compliant [06:41:00.0000] <maikmerten> no events, no fancy stuff, mostly just play and pause and a bit of broken seeking [06:42:00.0000] <maikmerten> does IE still support ActiveX? If it does I guess the best idea may be to develop an ActiveX-thingie 100% targeted at doing HTML5-video in IE [06:42:01.0000] <hsivonen> maikmerten: is the IE-Java bridge a Cortado thing or something in IE or the Sun plug-in? [06:42:02.0000] <hsivonen> maikmerten: any plans to expose scripting API in cortado to support HTML5 emulation shims? [06:43:00.0000] <hsivonen> maikmerten: can sites easily autoinstall ActiveX anymore? [06:44:00.0000] <maikmerten> hsivonen, I assume in worst case it would be the usual "click yes to install this security hole" thingie [06:44:01.0000] <maikmerten> hsivonen, currently Cortado just declares some methods/fields public in its main applet class [06:45:00.0000] <maikmerten> there's no bridge whatsoever to e.g. do events [06:45:01.0000] <maikmerten> and I've found Java <-> browser communication to be rather unreliable [06:45:02.0000] <hsivonen> I wonder if having a JS shim doing polling from timeout in order to emulate events would kill perf [06:45:03.0000] <maikmerten> (browsers/applets hanging and stuff) [06:46:00.0000] <hsivonen> Well, I hope someone writes a Theora ActiveX control for IE with an accompanying JS shim [06:46:01.0000] <maikmerten> hsivonen, here's a demo demonstrating one use case for the (limited) HTML5 media API support in Cortado: http://people.xiph.org/~oggk/elephants_dream/elephantsdream.html [06:47:00.0000] <maikmerten> hmmm... actually, no, doesn't demonstrate that really [06:48:00.0000] <maikmerten> (I supposed it would read the time to display a fitting subtitle with DOM-manipulation, but it's just "all Java subtitles") [06:49:00.0000] <maikmerten> hsivonen, another possibility would perhaps be to do a Vorbis/Theora player in Flash, but I have no idea how performant that would be [06:50:00.0000] <othermaciej> probably not very [06:50:01.0000] <othermaciej> the Flash H.264 codec is native code, and even that uses shocking amounts of CPU [06:50:02.0000] <maikmerten> right. The interesting question would be "how much slower than Java is it" ;-) [06:51:00.0000] <othermaciej> I'd expect ActionScript is a lot slower than Java [06:51:01.0000] <maikmerten> (after all Java is* fast enough to decode video, audio and do all the scaling for ~ SD video on common rigs) [06:51:02.0000] <othermaciej> it's actually slower than the fastest current JS implementations, for many kinds of code [06:51:03.0000] <maikmerten> well, it depends on how effective their JIT is - but getting that fast is easier on statically typed languages... [06:52:00.0000] <hsivonen> cool. the applet loads reasonably fast on the elephants dream demo [06:52:01.0000] <hsivonen> much faster than a random JavaFX demo [06:53:00.0000] <maikmerten> http://people.xiph.org/~maikmerten/demos/bigbuckbunny-applet.html <-- another demo with a higher resolution video [06:53:01.0000] <maikmerten> (no fancy subtitles, though) [06:53:02.0000] <MikeSmith> I think a significant blocking issue with codecs will always remain what is implemented in hardware level on mobile devices [06:54:00.0000] <maikmerten> that indeed is an issue - and to make matters worse this is a moving target (quite literally, too) [06:54:01.0000] <jgraham> MikeSmith: Alays is a long time [06:54:02.0000] <jgraham> Always [06:54:03.0000] <MikeSmith> well, in general, implementing XYZ in flash does not solve problems for mobile handsets [06:54:04.0000] <hsivonen> maikmerten: your demo showed an unverifiable signature warning in IE8/Java6 [06:55:00.0000] <MikeSmith> for browsing from mobile devices, I mean [06:55:01.0000] <maikmerten> hsivonen, aye, it's a signed applet with a Xiph.org certificate without official blessing [06:55:02.0000] <maikmerten> downside: User nag, upside "does play remote video" [06:56:00.0000] <MikeSmith> jgraham: "always" meaning until CPU and RAM resources on mobile handsets approach same level as PCs [06:56:01.0000] <hsivonen> maikmerten: yeah, the actual video performs well [06:57:00.0000] <maikmerten> MikeSmith, one would have to have a close look on what mobile devices are relevant, what hardware capabilities they have (hardwired decoders like in 1997 or programmable DSP stuff) and (in case DSPs are common) get DSP implementations done [06:57:01.0000] <hsivonen> this stuff really needs to be packaged for dummies so that one can use one script in head and then <video src=foo.ogg> in body and have things Just Work [06:57:02.0000] <maikmerten> hsivonen, that would be http://people.xiph.org/~maikmerten/demos/bigbuckbunny.html ;-) [06:58:00.0000] <maikmerten> hsivonen, (advanced version of this at http://www.celt-codec.org/presentations/ ) [07:01:00.0000] <maikmerten> anyway, as for making-IE-do-HTML5-media I guess something slammed together using liboggplay and and an additional DirectShow hook (for additional codec coverage) may be best [07:02:00.0000] <maikmerten> I develop Java software every day, I'm pretty fond of Java - but applets* are something always giving me a "don't use unless necessary"-feeling [07:03:00.0000] <maikmerten> after all the Java plugin used to be a great annoyance to basically everyone (slow, crashy) and the from-scratch-new plugin isn't thaaat common, I'd guess [07:03:01.0000] <hsivonen> maikmerten: ooh. excellent. re: celt-codec video [07:04:00.0000] <hsivonen> I like Java, too, but I browse the Web without applets [07:04:01.0000] <hsivonen> maikmerten: you should do more marketing for the script [07:05:00.0000] <hsivonen> maikmerten: thanks. I need to study the script some more [07:05:01.0000] <maikmerten> hsivonen, I hope we can put author's instructions of how to use that sort of scripting into the support resources of HTML5-media supporting browser vendors [07:06:00.0000] <maikmerten> hsivonen, be warned, I initially wrote that script, so parts are crude and DOM-API usage at worst [07:07:00.0000] <maikmerten> hsivonen, (gmaxwell from Wikimedia cleaned it somewhat, so it still may give pain but the bleeding is gone) [07:11:00.0000] <maikmerten> hmmm... anybody from Chrome/Chromium here? [07:11:01.0000] <maikmerten> apparently Chrome (dev channel build) exactly duplicates this Firefox bug: https://bugzilla.mozilla.org/show_bug.cgi?id=480058 [07:14:00.0000] <gsnedders> Yay! Swedish accommodation contracts that mean nothing to me! [07:14:01.0000] <jgraham> gsnedders: Me neither. And I signed two [07:15:00.0000] <gsnedders> "Since I don't know how much swedish you know…" [07:15:01.0000] <hsivonen> duplicating a bug exactly is called interop :-) [07:17:00.0000] <gsnedders> jgraham: one of you and zcorpan lives west from the office, right? [07:17:01.0000] <MikeSmith> bugwards compatibility [07:19:00.0000] <gsnedders> 1.6km more or less [07:20:00.0000] <MikeSmith> hmm, is "bugwards compatibility" not a familiar term? [07:20:01.0000] <MikeSmith> (google seems to suggest it's not) [07:20:02.0000] <maikmerten> hsivonen, okay, I just installed an ActiveX control in IE (Bitdefender online virus scanner) to see what the experience is like (the wonders of having Windows caged in a virtual machine) [07:21:00.0000] <gsnedders> Also, apparently the flat comes with a bunch of things with Swedish names :P [07:21:01.0000] <maikmerten> hsivonen, it involves clicking "yup, install that thing" twice, that's it [07:21:02.0000] <maikmerten> (first is a yellow "web page wants to install bla bla bla" in the web-browser - not unlike the Mozilla pop-up blocker) [07:22:00.0000] <hsivonen> maikmerten: ok [07:22:01.0000] <maikmerten> (when allowing this it'll ask again on installation) [07:23:00.0000] <maikmerten> the program keeps installed even after reboots, so I guess this would be a two-click solution for IE to install basically anything [07:26:00.0000] <maikmerten> however, no applet (that for sure) or ActiveX control (well, that one could exchange the rendering engine ;-) ) will be able to e.g. properly layer HTML content on top of the video, do transparency and whatnot [07:26:01.0000] <maikmerten> but then again there's a chance most people want "just video, thank you" [07:29:00.0000] <Philip`> I think most people really do want their videos to be translucent and embedded in a rotating SVG triangle [07:30:00.0000] <hsivonen> can laying an iframe on top of an ActiveX control punch a rectangular hole in the rendering of the ActiveX control? [07:30:01.0000] <Philip`> otherwise they'd still be happy with Flash [07:30:02.0000] <maikmerten> (hmmm... this whole pimp-your-browser talk makes me think writing a (signed) applet downloading and installing XiphQT on Safari may be something workable) [07:33:00.0000] <hsivonen> maikmerten: except Mac users run with Java off [07:33:01.0000] <maikmerten> hsivonen, by default? [07:33:02.0000] <hsivonen> maikmerten: by experience of instability and slowness and now due to unpatched security holes [07:33:03.0000] <maikmerten> hsivonen, I don't believe in the casual Mac user noticing how badly their Java is maintained ;-) [07:35:00.0000] <maikmerten> (this may be a vastly inaccurate opinion based on no numbers whatsoever) [08:00:00.0000] <gsnedders> ezyang: ping [08:00:01.0000] <ezyang> pong [08:01:00.0000] <gsnedders> ezyang: Trying to get a single emitToken in Tokenizer, what do you think of adding a callback function and calling call_user_func in Tok? [08:01:01.0000] <ezyang> That'll be slower. Why are you trying to get a single emitToken? [08:02:00.0000] <gsnedders> ezyang: So we don't need handling like setting PCDATA on end tag in every emitToken we have, nor removing duplicate attributes, etc. [08:04:00.0000] <ezyang> I was under the impression that's why it was a function call [08:04:01.0000] <gsnedders> I mean for the tokenizer tests we have an entirely separate emitToken, quite possibly with different bugs to the real tokenizer. [08:04:02.0000] <gsnedders> Which is kinda bad. [08:05:00.0000] <ezyang> Oh, I see. [08:06:00.0000] <gsnedders> I want to have a single emitToken in Tokenizer which calls another function, normally the emitToken method in TreeConstructor, but could be something else for the tests etc. [08:06:01.0000] <ezyang> So, with SimpleTest, you don't have to write any extra code: you just partially mock emitToken in TreeConstructer and do the wraparound [08:07:00.0000] <gsnedders> But that impls TreeConstructor class which makes it impossible to run the whole test suite as then TreeConstructor is already defined. [08:07:01.0000] <ezyang> No, you make a new TreeConstructer, and then you use our $builder parameter to inject the mock [08:08:00.0000] <Lachy> sigh. The same old arguments are being rehashed again about the summary attribute on public-html. :-( [08:09:00.0000] <gsnedders> /me wonders how that avoids defining two classes of the same name [08:10:00.0000] <ezyang> SimpleTest does all the hairy stuff for you. [08:10:01.0000] <ezyang> It's just a line or two of code [08:11:00.0000] <Philip`> Lachy: Maybe somebody should write them all down for easy reference when the discussion comes up again in another few months [08:14:00.0000] <hsivonen> aaargh! Astral non-characters! [08:14:01.0000] <ezyang> Hooray! [08:14:02.0000] <hsivonen> UTF-16 sucks [08:15:00.0000] <hsivonen> /me goes resurrect code from SVN [08:16:00.0000] <jgraham> Philip`: You would have to find someone who understood all the issues but was not percieved as biased [08:41:00.0000] <hsivonen> yay. down to one tokenizer test failure and that's only because I get a parse error between characters [08:49:00.0000] <hsivonen> sigh. it's diffecult to get the right character and error ordering when CR causes an early flush [09:05:00.0000] <mpilgrim> http://www.google.com/squared/search?q=html+5+elements [09:06:00.0000] <annevk42> http://www.google.com/squared/search?q=w3c [09:07:00.0000] <Philip`> mpilgrim: http://krijnhoetmer.nl/irc-logs/whatwg/20090604#l-8 - old news :-p [09:07:01.0000] <annevk42> hmm, http://www.google.com/squared/search?q=WHATWG and http://www.google.com/squared/search?q=whatwg are different [09:07:02.0000] <mpilgrim> damn it [09:10:00.0000] <Philip`> http://www.google.com/squared/search?q=polygons [09:10:01.0000] <Philip`> Apparently the internal angle of a circle is 20 degrees [09:10:02.0000] <hsivonen> LOL. add column "Status" to squaring "W3C" [09:12:00.0000] <hsivonen> (XForms is "unsupported") [09:12:01.0000] <hsivonen> (but then HTML is "closed") [09:12:02.0000] <hsivonen> (Semantic Web is "resolved: (list of) uris") [09:13:00.0000] <Philip`> /me tries adding the column "color" [09:13:01.0000] <Philip`> Apparently HTML is green, and CSS is red [09:14:00.0000] <hsivonen> I get "purple" for HTML [09:14:01.0000] <mpilgrim> http://www.google.com/squared/search?q=w3c+standards [09:15:00.0000] <hsivonen> and CSS is "black;" for me [09:15:01.0000] <mpilgrim> google has recreated the Celestial Emporium of Benevolent Knowledge! [09:16:00.0000] <mpilgrim> ( cf http://www.themodernword.com/borges/borges_quotes.html ) [09:17:00.0000] <Philip`> It's like Wolfram Alpha, except instead of giving you either useful answers or "I don't know", it always gives you complete rubbish [09:18:00.0000] <mpilgrim> it's powered by FACTS! from the INTERNET! [09:18:01.0000] <mpilgrim> (bugzilla bug #1: there are no facts on the internet) [09:19:00.0000] <annevk42> HTML is '#000"/>' for me [09:19:01.0000] <annevk42> in the 'color' column [09:19:02.0000] <annevk42> lol, DOM's Type is a "String" [09:19:03.0000] <hsivonen> annevk42: apparently the facts are different for everyone [09:19:04.0000] <annevk42> and for HTML it's "Security" [09:20:00.0000] <Philip`> That's the great thing about facts [09:20:01.0000] <hsivonen> perhaps google adapts them to the profile of your past searches [09:20:02.0000] <annevk42> I'm not signed in [09:20:03.0000] <mpilgrim> http://www.google.com/squared/search?q=natural+language+processing [09:20:04.0000] <hsivonen> annevk42: Google knows you even when you aren't [09:21:00.0000] <hsivonen> maybe Google Squared could be more accurate if it mined RDF [09:21:01.0000] <annevk42> when I search for uppercase planets the Pluto image is from Disney :) [09:22:00.0000] <mpilgrim> hsivonen: http://www.google.com/squared/search?q=rdf [09:22:01.0000] <Philip`> http://www.google.com/squared/search?q=programming+languages - apparently Pascal died in Paris, France [09:22:02.0000] <hsivonen> Philip`: and Prolog died in 2000 [09:23:00.0000] <Philip`> hsivonen: Hmm, I get "no value found" there [09:23:01.0000] <Philip`> but JavaScript died in "Nissan 300zx z32 ConZult Diagnostics OBDII Fault Codes Engine ECU Computer Software" [09:23:02.0000] <hsivonen> Philip`: it's the 'Year of Death' column [09:23:03.0000] <mpilgrim> http://www.google.com/squared/search?q=operating+systems <-- won't linus be surprised that O'Reilly Media wrote Linux [09:24:00.0000] <jgraham> mpilgrim: I was just looking at that. I this MS will be surprised windos is Free [09:24:01.0000] <mpilgrim> "Author: DOS" [09:25:00.0000] <mpilgrim> there's a kernel of truth in that somewhere [09:25:01.0000] <Philip`> I think one might conclude that parsing unstructured data is very much an unsolved problem [09:28:00.0000] <mpilgrim> http://www.google.com/squared/search?q=things+to+do+on+a+saturday yields exactly 1 result: "The Circus Starring Britney Spears" [09:28:01.0000] <mpilgrim> well, i know what I'M doing this saturday [09:30:00.0000] <annevk42> $128 [09:30:01.0000] <annevk42> /me added a column cost that came back with some random value from a seemingly unrelated site [09:31:00.0000] <annevk42> squared is a lot of fun [09:31:01.0000] <annevk42> cost of HTML is apparently $49.95 [09:31:02.0000] <hsivonen> more fun that the pre-scripted fun on Wolfram|Alpha [09:32:00.0000] <annevk42> "css properties" is sort of good [09:33:00.0000] <annevk42> it copies the applies to/inherited data [09:33:01.0000] <annevk42> to compare, the cost of <frame> is $9,881,575 [09:33:02.0000] <annevk42> and class is 137,802,000 JPY [09:33:03.0000] <annevk42> table is "FREE!" [09:34:00.0000] <mpilgrim> http://www.google.com/squared/search?q=microformats says VoteLinks is $1250 [09:34:01.0000] <mpilgrim> also, rel-nofollow is depicted by scantily clad women [09:35:00.0000] <annevk42> all the good things are forbidden :p [09:36:00.0000] <mpilgrim> it's not even like the quality degrades linearly [09:37:00.0000] <mpilgrim> http://www.google.com/squared/search?q=dog+breeds is one of the front-page suggestions [09:37:01.0000] <mpilgrim> but i change it to http://www.google.com/squared/search?q=cat+breeds and it tells me that President George Bush is Persian [09:37:02.0000] <tantek> mpilgrim - wow that is wacky [09:37:03.0000] <tantek> (the microformats "square") [09:37:04.0000] <mpilgrim> hey tantek [09:37:05.0000] <mpilgrim> i'm guessing your irc client flashed on "microformats" [09:38:00.0000] <tantek> actually on VoteLinks [09:38:01.0000] <mpilgrim> ha [09:38:02.0000] <mpilgrim> you should make them less expensive if you want people to adopt them [09:38:03.0000] <tantek> LOL [09:38:04.0000] <tantek> as this channel knows, using the "rev" attribute is quite expensive [09:39:00.0000] <tantek> what's the going pay-for-play rate on HTML5 features again? [09:39:01.0000] <annevk42> tens of millions? [09:40:00.0000] <mpilgrim> apparently $600: http://www.google.com/squared/search?q=html+5+features [09:40:01.0000] <annevk42> /me forgot [09:40:02.0000] <annevk42> regardless, you can buy HTML for 49.95USD [09:42:00.0000] <Philip`> http://googleblog.blogspot.com/2009/06/square-your-search-results-with-google.html - "This technology is by no means perfect." - gosh [09:44:00.0000] <mpilgrim> i submitted a bunch of bugs when it was strictly internal [09:44:01.0000] <jgraham> Philip`: Well at least they haven't gone all Wolfram on us proclaiming that they have invented world-changing technology only for it to be revealed as something rather less [09:45:00.0000] <mpilgrim> they don't seem to be fixed, but it's hard to tell [09:45:01.0000] <Philip`> jgraham: Indeed, but it does still seem surprisingly useless [10:46:00.0000] <rubys1> Hixie: ping? [11:50:00.0000] <mpilgrim> shelleyp: did you want a serious answer to your question on twitter? [11:50:01.0000] <rubys> /me would like a serious answer to that question [11:50:02.0000] <shelleyp> Yes, I did. [11:51:00.0000] <mpilgrim> ok. tip: if you want to directly engage me, try not to lead off with "yanked" next time [11:51:01.0000] <mpilgrim> but anyway [11:51:02.0000] <mpilgrim> my opinion on accessible markup has changed since i wrote "dive into accessibility" [11:51:03.0000] <shelleyp> Why? [11:52:00.0000] <mpilgrim> partly having seen usability studies [11:52:01.0000] <mpilgrim> partly having worked directly with screenreader vendors (while i was at ibm) [11:52:02.0000] <shelleyp> Do you have a link to the studies? [11:52:03.0000] <mpilgrim> partly from having access to real-world data on what the web looks like (from google's cache) [11:52:04.0000] <mpilgrim> they were mentioned on the list a while back [11:53:00.0000] <mpilgrim> lemme check... [11:53:01.0000] <shelleyp> Thanks [11:53:02.0000] <shelleyp> I don't know that's a good idea to base anything on Google cahe [11:53:03.0000] <shelleyp> sorry, cache [11:54:00.0000] <mpilgrim> google's cache gives a view into how markup scales [11:54:01.0000] <mpilgrim> on a global scale [11:54:02.0000] <mpilgrim> which mistakes are common [11:54:03.0000] <mpilgrim> etc. [11:54:04.0000] <shelleyp> But it's a snapshot of a point in time, not necessarily a directive for the future [11:54:05.0000] <rubys> mpilgrim: your opinion has changed, got it; care to state what your opinion now is? [11:54:06.0000] <mpilgrim> you can't get that from anecdotes or accessibility tutorials [11:55:00.0000] <mpilgrim> shelleyp: that is absolutely true [11:55:01.0000] <mpilgrim> it's only a cache of the now [11:55:02.0000] <shelleyp> Sam's impatient -- what is your view now? [11:57:00.0000] <mpilgrim> i can't find the link to the usability studies [11:57:01.0000] <mpilgrim> perhaps someone can jump in? [11:57:02.0000] <mpilgrim> it was mentioned here: http://lists.w3.org/Archives/Public/public-html/2008Feb/0059.html [11:58:00.0000] <mpilgrim> it was just 1 study, IIRC [11:58:01.0000] <mpilgrim> as for my view, i thought i spelled it out pretty well in http://diveintomark.org/archives/2009/03/21/accessibility-is-a-harsh-mistress [11:58:02.0000] <mpilgrim> "The accessibility orthodoxy does not permit people to question the value of features that are rarely useful and rarely used." [11:58:03.0000] <Philip`> Do you mean the http://lists.w3.org/Archives/Public/public-html/2007Sep/0103.html videos? [11:58:04.0000] <mpilgrim> indeed, thanks Philip` [11:59:00.0000] <gsnedders> /me is somewhat scared by how quickly Philip` can find any email on the list [11:59:01.0000] <mpilgrim> i think that there has to be some cost-benefit analysis of proposed accessibility features [11:59:02.0000] <mpilgrim> in fact, of existing accessibility features [11:59:03.0000] <mpilgrim> and there just isn't [11:59:04.0000] <Philip`> gsnedders: It's not exactly hard - I just Bing some words and click the links [11:59:05.0000] <mpilgrim> (that i know of) [12:00:00.0000] <mpilgrim> Philip`++ for using Bing (or at least claiming to) [12:00:01.0000] <mpilgrim> and you can say, "oh, if we can allow just 1 blind person to use the internet, how can you put a price on that?" [12:00:02.0000] <Philip`> /me just wanted to verb its name [12:00:03.0000] <mpilgrim> well that's crap [12:00:04.0000] <mpilgrim> we (as a society) can put a price on anything [12:00:05.0000] <gsnedders> £42. [12:00:06.0000] <mpilgrim> we put a price on a child's life [12:01:00.0000] <mpilgrim> (seatbelt laws) [12:01:01.0000] <rubys> ok, so mpilgrim doesn't care for the "accessibility orthodoxy", got it; still curious as to what mpilgrim's opinion now is... [12:01:02.0000] <mpilgrim> actuaries can put a price on anything [12:01:03.0000] <mpilgrim> but nobody ever has [12:01:04.0000] <mpilgrim> the only effort has been in defining markup (usually, accessibility-specific markup) that purports to solve certain problems [12:02:00.0000] <mpilgrim> but there's no followup to determine if those particular solutions actually DO solve those particular problems [12:02:01.0000] <mpilgrim> long descriptions are a real problem [12:02:02.0000] <rubys> http://esw.w3.org/topic/HTML/SummaryForTABLE/SummarySpecification [12:02:03.0000] <mpilgrim> navigating complex tables is a real problem [12:03:00.0000] <mpilgrim> rubys: i've read all that [12:03:01.0000] <mpilgrim> and i agree with everything except the conclusion [12:03:02.0000] <mpilgrim> because they're missing something [12:03:03.0000] <mpilgrim> the something is "does it actually fucking work?" [12:04:00.0000] <mpilgrim> nobody checks [12:05:00.0000] <mpilgrim> every time somebody does check, the answer is "no" [12:05:01.0000] <mpilgrim> nobody likes that answer, so they ignore it [12:05:02.0000] <mpilgrim> and we go round and round [12:05:03.0000] <rubys> help me prepare the right question, and I'll make sure that it is not ignored. [12:05:04.0000] <mpilgrim> it's been asked [12:05:05.0000] <mpilgrim> it's been explained [12:06:00.0000] <rubys> url? [12:06:01.0000] <shelleyp> philip: All of the links in that one site are dead - I guess hypertext links don't work, either [12:06:02.0000] <mpilgrim> um, most recently, http://lists.w3.org/Archives/Public/public-html/2009Jun/0173.html [12:06:03.0000] <mpilgrim> yeah, shelleyp: that's exactly the kind of rhetorical attitude that other people make fun of you for [12:07:00.0000] <mpilgrim> if you want a real discussion, stick around [12:07:01.0000] <mpilgrim> if you just want to take cheap shots, tell me now [12:07:02.0000] <mpilgrim> i can do that, too [12:07:03.0000] <shelleyp> But that wasn't a snarky comment -- if you look at things at specific points in time, you can find that nothing really works [12:08:00.0000] <mpilgrim> i'm not sure what that means [12:08:01.0000] <shelleyp> What I'm trying to say is that maybe the real problem with accessible markup is we haven't given it enough time to work [12:09:00.0000] <shelleyp> Our sampling size, in this case based on time, is too small -- same as judging links by one page [12:10:00.0000] <mpilgrim> html 4 is almost 10 years old now [12:10:01.0000] <mpilgrim> it was published when google had 40 employees [12:10:02.0000] <mpilgrim> how much more time do you suggest we give it? [12:10:03.0000] <shelleyp> If i remember correctly, it took decades for wheelchair ramps to be added to commercial buildings [12:11:00.0000] <rubys> mpilgrim: would you agree that the key question from Ian's note is """ Could you elaborate on why you believe summary="" serves its purpose well?" """, and acceptable answers would be stated in a form like http://diveintoaccessibility.org/day_20_providing_a_summary_for_tables.html does ? [12:11:01.0000] <mpilgrim> there's a difference: every single physical building you've ever entered in your lifetime was built by a professional and inspected/regulated by a government [12:12:00.0000] <mpilgrim> unless you're suggesting we apply that same standard to building a web site, there will always be infinitely more sites built by amateurs than by professionals [12:12:01.0000] <shelleyp> So? We can't control what amateurs do, but the government is stepping up to commercial site development [12:12:02.0000] <mpilgrim> esp. professionals who give a shit (or are forced by law to give a shit) about small niche audiences like accessibility [12:13:00.0000] <shelleyp> People don't have to have ramps for their homes, either [12:13:01.0000] <mpilgrim> i heard a "joke" (in rather poor taste, but whatever) at an accessibility conference i went to a few years ago [12:13:02.0000] <rubys> Isn't """ When JAWS encounters your calendar, Jackie hears "Summary: Monthly calendar with links to each day's posts." """ a reasonable response to "does it actually fucking work?" ? [12:13:03.0000] <mpilgrim> "if you want accessibility to be taken seriously, stab 100 million people in the eyes" [12:14:00.0000] <mpilgrim> no, the "does it actually fucking work" question needs to apply to the entire web [12:14:01.0000] <mpilgrim> that's the point [12:14:02.0000] <mpilgrim> does it work on a web scale? [12:14:03.0000] <rubys> what sort of answer are you looking for? [12:14:04.0000] <mpilgrim> of course you can design a special-purpose system within a general-purpose system that "works" [12:14:05.0000] <shelleyp> mpilgirm: you're setting a "standard" that none of us have agreed to -- has to have web scale? [12:14:06.0000] <mpilgrim> when you author it correctly [12:15:00.0000] <mpilgrim> and pay vendors to support it correctly [12:15:01.0000] <mpilgrim> and educate the user about it before you turn on the camera in the usability study [12:15:02.0000] <mpilgrim> but it doesn't fucking scale [12:15:03.0000] <shelleyp> mpilgrim: I would imagine there is more use of @summary than there is local database storage or web workers [12:15:04.0000] <mpilgrim> another category error [12:15:05.0000] <hober> shelleyp: Of course it has to have web scale--we're talking about the language of the web. [12:16:00.0000] <hober> if html features don't work at web scale, they're not features at all, they're bugs. [12:16:01.0000] <mpilgrim> if people try to use local storage and they're stupid, it won't work [12:16:02.0000] <mpilgrim> nothing is lost; it just doesn't work [12:16:03.0000] <shelleyp> mpilgrim: But who determines what web scape is? Something like SVG isn't used as widely as a div element [12:16:04.0000] <mpilgrim> maybe they give up and publish their site without oflfine capabilities [12:16:05.0000] <shelleyp> Does that mean SVG is failed? [12:17:00.0000] <mpilgrim> SVG is wildly successful [12:17:01.0000] <mpilgrim> i believe both gnome and KDE use it now [12:17:02.0000] <mpilgrim> to provide scalable icons and graphics on linux desktops [12:17:03.0000] <mpilgrim> on the web, not so much usage [12:17:04.0000] <mpilgrim> no [12:17:05.0000] <hober> Hopefully, the outcome of the SVG-in-HTML effort allows SVG to succeed on the Web. It's succeeded elsewhere, as markp just said. [12:17:06.0000] <shelleyp> mpilgrim: but it doesn't have the same scale on the web as div [12:17:07.0000] <mpilgrim> although you may have heard that we (google) pre-announced our make-SVG-work-in-IE javascript library [12:18:00.0000] <rubys> ... built on flash [12:18:01.0000] <mpilgrim> yes [12:18:02.0000] <mpilgrim> as a shim for a browser that will likely never support it natively [12:18:03.0000] <shelleyp> @hober: I don't agree with you. You're saying that unless an arbitrary percentage of sites use @summary, it's "failed" [12:18:04.0000] <mpilgrim> no [12:18:05.0000] <mpilgrim> that is not what we're saying [12:19:00.0000] <mpilgrim> you have to look at the percentage of people who use it [12:19:01.0000] <mpilgrim> the percentage of people who use it properly [12:19:02.0000] <mpilgrim> and the percentage of people who should use it, but aren't [12:19:03.0000] <mpilgrim> that's the problem with @summary and @longdesc [12:19:04.0000] <mpilgrim> it's not useful for most images [12:19:05.0000] <mpilgrim> it's frequently abused (i.e. wikipedia abused it for many years) [12:19:06.0000] <mpilgrim> it's frequently misused [12:19:07.0000] <shelleyp> But people are capable of learning, both from mistakes and from other people's successful efforts [12:20:00.0000] <mpilgrim> yes [12:20:01.0000] <mpilgrim> individuals are capable of learning [12:20:02.0000] <mpilgrim> groups are not [12:20:03.0000] <mpilgrim> (much) [12:21:00.0000] <mpilgrim> and (finishing my thought on @summary) most of the images that should be using it, don't [12:21:01.0000] <shelleyp> Returning to the topic, specifically @summary, where is the "harm" in its use? [12:21:02.0000] <mpilgrim> even if you educated every single blind person (the alleged target audience for the feature), they wouldn't ever run into a proper use of it [12:21:03.0000] <mpilgrim> except in accessibility tutorials [12:22:00.0000] <mpilgrim> the harm has also been explained [12:22:01.0000] <mpilgrim> i'll ignore the cost to implementers and other non-blind people for now [12:22:02.0000] <mpilgrim> focusing strictly on blind web users [12:23:00.0000] <rubys> summarizing:" summary has been around for a number of years, by this point whether it works or not should not be a matter of assertion from first principles, but from actual measurements. Key questions to be answered: who should use summary? are they? If not, why not? what should user agents do with summary? are they? If not, why not?" Is that a fair summary? [12:23:01.0000] <mpilgrim> if the feature is frequently abused or misused, then blind people are just being bombarded with crap [12:23:02.0000] <mpilgrim> accessi-spam, as it were [12:24:00.0000] <shelleyp> Unfortunately, blind people are most likely bombarded with crap on a regular basis -- including tables for layout [12:24:01.0000] <mpilgrim> if there is a real problem (navigating complex tables), and if @summary is put forth as the ONE TRUE WAY to solve that problem, [12:24:02.0000] <mpilgrim> then blind people are being harmed by the missed opportunity of a better solution [12:24:03.0000] <shelleyp> So what's your solution? [12:24:04.0000] <rubys> I am serious about wanting to find the right question, in today's HTML WG conference call, it looks like it is time to create a vote... [12:24:05.0000] <mpilgrim> i'm not saying i have an easy solution [12:24:06.0000] <hober> rubys: basically, yes. after ten years, we can be empiricists. as tantek said, http://twitter.com/t/status/318565892 [12:24:07.0000] <mpilgrim> i'm still trying to get people to understand the problem [12:25:00.0000] <rubys> /me thanks hober [12:25:01.0000] <mpilgrim> read the last sentence of http://blog.whatwg.org/the-longdesc-lottery [12:25:02.0000] <shelleyp> But if you don't have an alternative, can't we at least let those most committed have the tools they need? [12:25:03.0000] <mpilgrim> "So can we please get past the grandstanding and start talking about a better solution?" [12:26:00.0000] <rubys> forgive me, mpilgrim, but different people have different perceptions as to exactly who is doing the grandstanding. I want to get beyond that question. [12:26:01.0000] <mpilgrim> that would actively dissuade people from attempting a better solution [12:26:02.0000] <mpilgrim> my point is, i acknowledge that some images are complex enough to require further explanation [12:26:03.0000] <mpilgrim> beyond a short phrase [12:27:00.0000] <mpilgrim> i acknowledge that some tables are complex enough to warrant a description of their structure [12:27:01.0000] <rubys> the problem is that there is a growing perception that no matter how good the answers given are, they will never be "good enough" as the goalposts are always being raised. [12:27:02.0000] <mpilgrim> but for many years, the only people trying to solve these problems were only interested in solving them for themselves [12:27:03.0000] <mpilgrim> with accessibility-specific markup [12:27:04.0000] <mpilgrim> which doesn't scale [12:27:05.0000] <mpilgrim> across the world wide web of authors [12:28:00.0000] <mpilgrim> (unless you're serious about stabbing 100 million people in the eyes) [12:28:01.0000] <shelleyp> @mpilgrim, forgive me (and no snark) but that is a very circular argument [12:28:02.0000] <mpilgrim> that mindset is the problem [12:28:03.0000] <mpilgrim> it's why we're talking past each other [12:28:04.0000] <mpilgrim> one side says "if the image is so complex, wouldn't everyone benefit from a longer description of it?" [12:29:00.0000] <mpilgrim> and the other side says "YOU STOLE MY @LONGDESC!" [12:29:01.0000] <mpilgrim> "YOU HATE BLIND PEOPLE!" [12:29:02.0000] <shelleyp> If we're talking past each other, isn't there a chance that you might be wrong? [12:29:03.0000] <mpilgrim> of course [12:29:04.0000] <mpilgrim> i have been wrong about many things [12:29:05.0000] <mpilgrim> but you asked how my opinion had changed [12:29:06.0000] <mpilgrim> and that's it [12:29:07.0000] <mpilgrim> when i wrote "dive into accessibility", i believed (as the w3c believed) that accessibility-specific markup could ever be a solution to accessibility problems [12:30:00.0000] <shelleyp> But in the case of HTML5, "being wrong" and removing attributes that have existed in the past could have far reaching consequences [12:30:01.0000] <mpilgrim> and now i don't believe that [12:30:02.0000] <mpilgrim> in fact, i believe that accessibility-specific markup is actively harmful to accessibility [12:30:03.0000] <rubys> ah, now I finally see an answer to the question shelley actually asked [12:30:04.0000] <mpilgrim> and the more specific the use case, the more harmful it gets [12:30:05.0000] <shelleyp> rubys raised a point: there is a growing perception that no matter how many answers are provided, they'll never enough [12:31:00.0000] <mpilgrim> no one has demonstrated why accessibility-specific markup is the right solution [12:31:01.0000] <mpilgrim> they just keep repeating it [12:31:02.0000] <mpilgrim> they take it on faith that it is [12:31:03.0000] <hober> I think something slightly weaker is true: accessibility-specific markup whose contents aren't shown to users not using AT, that's the harmful kind of markup [12:31:04.0000] <inimino> would that include @alt? [12:31:05.0000] <hober> E.g., seems like @aria-describedby would be much better than @summary [12:31:06.0000] <shelleyp> But you're taking it on faith that enough time has passed that we can determine that accessibility markup has "failed" [12:31:07.0000] <mpilgrim> @alt is the most successful of a very weak lot [12:31:08.0000] <hober> (assuming the element pointed at is visible to everyone) [12:32:00.0000] <mpilgrim> measured in absolute terms, it has been a miserable failure [12:32:01.0000] <shelleyp> What absolute terms? [12:32:02.0000] <mpilgrim> despite being easy to grasp, and REQUIRED FOR VALIDATION [12:32:03.0000] <mpilgrim> and short [12:32:04.0000] <hober> It's been a decade! Half of the lifetime of the Web! How has not enough time passed? [12:32:05.0000] <mpilgrim> and, in many cases, easy for tool vendors to build in (at least images within templates) [12:32:06.0000] <shelleyp> No, I just got through saying it took decades just to get ramps for commercial buildings [12:32:07.0000] <shelleyp> It takes time to do things right [12:32:08.0000] <mpilgrim> (these are the cases i targeted in "dive into accessibility") [12:33:00.0000] <shelleyp> But if you change the game, frequently, nothing ever has a chance to succeed [12:33:01.0000] <mpilgrim> that may be true [12:33:02.0000] <mpilgrim> but that's a weak argument for sticking with solutions that aren't working [12:33:03.0000] <inimino> it seems reasonable to expect that the next decade will see more new HTML authors than the previous decade [12:33:04.0000] <shelleyp> A better approach would to leave the markup in, increase education, and then look at alternatives [12:33:05.0000] <mpilgrim> ("stay the course", etc.) [12:34:00.0000] <shelleyp> But now, we'll have nothing. Nothing at all. How can that be the best solution? [12:34:01.0000] <mpilgrim> that's not true [12:34:02.0000] <mpilgrim> there have been proposals put forth for how to describe structurally complex tables for everyone [12:34:03.0000] <mpilgrim> and the response from the accessibility community is always the same [12:35:00.0000] <mpilgrim> "YOU STOLE OUR @SUMMARY!" [12:35:01.0000] <mpilgrim> "YOU HATE BLIND PEOPLE!" [12:35:02.0000] <mpilgrim> "don't take away what little we have" [12:35:03.0000] <shelleyp> But I'm not saying that mpilgrim [12:35:04.0000] <mpilgrim> (even if the little we have isn't working) [12:35:05.0000] <mpilgrim> what they're really saying is "we're not thinkers interested in solving problems, we're just fans" [12:35:06.0000] <mpilgrim> "go team" [12:35:07.0000] <mpilgrim> "yay mets, boo yankies" [12:35:08.0000] <mpilgrim> whatever [12:36:00.0000] <shelleyp> mpilgrim, but what you're response is, is more of the same "grandstanding" you seem to deplore [12:37:00.0000] <hober> shelleyp: what did you think of my proposal to use @aria-describedby instead of @summary? [12:37:01.0000] <rubys> what's the status of aria in html5? [12:37:02.0000] <mpilgrim> i acknowledge that you are not saying that [12:37:03.0000] <mpilgrim> shelleyp [12:37:04.0000] <hober> It avoids the "<caption> isn't the same as @summary" rathole, and it ensures (to the extent reasonable) that the summary is visible to all users... [12:37:05.0000] <mpilgrim> so, did i answer your question? [12:38:00.0000] <Hixie> rubys: here now [12:39:00.0000] <shelleyp> mpilgrim: Yes, I understand where you're coming from, which helps. Thank you. [12:39:01.0000] <mpilgrim> i acknowledge that i tend to grandstand when talking about so-called accessibility experts [12:40:00.0000] <mpilgrim> i also acknowledge that they have put a lot of work into their answers on @summary and related features [12:40:01.0000] <mpilgrim> i still think they're not asking vital questions ("does it fucking work?") [12:41:00.0000] <rubys> hixie: http://krijnhoetmer.nl/irc-logs/html-wg/20090604#l-452 [12:41:01.0000] <mpilgrim> and i still think that all of their work starts from a fundamentally incorrect premise ("let's add accessibility-specific features to solve accessibility-related problems") [12:41:02.0000] <tantek> mpilgrim, but is calling someone out for grandstanding itself grandstanding? Just as pointing out an ad hominem attack is not itself ad hominem, neither is pointing out grandstanding itself grandstanding. Both are a criticisms of the level of dialog taking place, with the (perhaps optimistic) hope that the level of the dialog can be improved. [12:41:03.0000] <rubys> http://krijnhoetmer.nl/irc-logs/html-wg/20090604#l-468 [12:42:00.0000] <mpilgrim> tantek: i don't want to get into a meta-discussion about it [12:42:01.0000] <mpilgrim> everyone has strong opinions [12:42:02.0000] <mpilgrim> and strong personalities [12:42:03.0000] <Hixie> rubys: what would the vote be? "Should we include a technology that has been proven to not work repeatedly and to actually harm accessibility, or should we include a technology that is recommended by the US gov't and might actually solve the problem?" [12:43:00.0000] <shelleyp> mpilgrim: You're putting forth a solid argument. I don't agree, but it's an argument one can work with? [12:43:01.0000] <rubys> I'd like help in formulating the question. I don't think that's it :-) [12:43:02.0000] <shelleyp> Sorry, forget the question mark -- I'm not especially good at IRC [12:43:03.0000] <tantek> mpilgrim, I've found it difficult to escape conversational loops without raising exceptions to suboptimal conversational behaviors. YMMV. [12:44:00.0000] <mpilgrim> tantek: if i did that, i'd never finish a conversation [12:44:01.0000] <Hixie> rubys: why are we even asking a question? this isn't something where there are two valid views. There's one group asking for something that is technically inferior, and there's one group who has suggested something (that's also suggested by the very gov't agencies the former group is saying are important!) that actually addresses key questions like "does it actually improve accessibility" [12:44:02.0000] <Hixie> rubys: isn't the right technical choice obvious? [12:45:00.0000] <Hixie> rubys: i'm happy to have a vote, assuming the vote fairly outlines the actual options [12:45:01.0000] <Hixie> rubys: but i don't think it's possible to fairly outline the possible options without summary="" advocates thinking they are being shafted [12:45:02.0000] <rubys> No. hearing that " When JAWS encounters your calendar, Jackie hears "Summary: Monthly calendar with links to each day's posts"" is something that I consider compelling. [12:46:00.0000] <hober> I think the same would happen were that text in <caption>, rubys [12:46:01.0000] <tantek> rubys, which version of JAWS, which platform, with which example URL on the Web, and do you have an MP3 of the audio recording? [12:47:00.0000] <rubys> tantek: EXCELLENT questions. those are the ones I'm looking for. [12:47:01.0000] <mpilgrim> he's talking about http://diveintoaccessibility.org/20 [12:47:02.0000] <shelleyp> hober: but isn't caption supposed to be visible? In fact, isn't caption more of a table title? [12:47:03.0000] <Hixie> rubys: for every web page that you can find that actually uses summary="" in a way that actually helps surse of screen readers, i can find you a page that actually _harms_ accessibility of the page. [12:47:04.0000] <Hixie> rubys: and i can also find you two more that harm screen-reader users specifically [12:47:05.0000] <Hixie> rubys: support for the attribute is disabled by expert users because it's so badly used [12:48:00.0000] <Hixie> rubys: and the information should just be included for all users most of the time anyway [12:48:01.0000] <tantek> rubys, those questions MUST be answered for us to actually form a scientific hypothesis that can be tested and corroborated with reproducible results by someone other than those that formed the hypothesis. [12:48:02.0000] <shelleyp> Hixie: I believe the question on the floor is, will you support a vote and abide by the vote decision? [12:49:00.0000] <rubys> tantek: agreed. My (unexpert impression) is that Hixie simply is making alternate assertions. [12:49:01.0000] <Hixie> shelleyp: as an htmlwg member i will support any correct use of process, yes, of course [12:49:02.0000] <Hixie> rubys: when this issue first came up (and second, and third, and fourth, and fifth came up) i included detailed analyses including referencing which studies i was talking about and what the results were [12:49:03.0000] <tantek> Hixie, I don't think you can "fairly outline the actual options" without answering questions such as I posed for each of the options. And any such vote on options without the answers to such empirical questions is no better than randomly voting on theoretical options. [12:50:00.0000] <hober> shelleyp: re: caption and visibility, yes, it is visible, which is the whole point. @summary suffers from metacrap partially because it isn't visible. [12:50:01.0000] <Hixie> rubys: for cases 6 through 139 i stopped doing that because i am tired of repeating myself in the face of "experts" who ignore research and just demand particular solutions that have been proven to not work [12:50:02.0000] <rubys> I'm new to this, but a URL to the analysis (or to your link to the analysis) would be helpful. [12:50:03.0000] <Hixie> look through the archives [12:50:04.0000] <Hixie> i really don't have the inclination to go through this again [12:51:00.0000] <rubys> I plan to put forward a cogent question and a process to resolving this. Anybody who has the inclination to participate will be welcome to do so. [12:52:00.0000] <shelleyp> rubys: I'm no longer part of the HTML WG, but I'd like to participate [12:52:01.0000] <rubys> rejoin! [12:52:02.0000] <shelleyp> Maybe I should [12:52:03.0000] <shelleyp> I can't stop myself from getting involved, anyway [12:52:04.0000] <rubys> there is no "has to participate n hours a month" requirement. [12:52:05.0000] <rubys> well, there is, but it is really, really low. [12:53:00.0000] <mpilgrim> some data: http://lists.w3.org/Archives/Public/public-html/2009Apr/0243.html [12:53:01.0000] <shelleyp> It's not the requirement, it's the fact that I just can't seem to pull away [12:53:02.0000] <mpilgrim> some questions: http://lists.w3.org/Archives/Public/public-html/2009Jun/0173.html [12:54:00.0000] <rubys> data is anecdotal; questions aren't to tantek's level of precision [12:54:01.0000] <mpilgrim> http://lists.w3.org/Archives/Public/public-html/2009Feb/0610.html [12:54:02.0000] <mpilgrim> http://lists.w3.org/Archives/Public/public-html/2009Feb/0738.html [12:54:03.0000] <mpilgrim> http://lists.w3.org/Archives/Public/public-html/2009Feb/0735.html [12:54:04.0000] <shelleyp> rubys has a good point, the data is anecdotal [12:54:05.0000] <mpilgrim> http://lists.w3.org/Archives/Public/public-html/2009Feb/0732.html [12:54:06.0000] <mpilgrim> http://lists.w3.org/Archives/Public/public-html/2009Feb/0727.html [12:54:07.0000] <mpilgrim> http://lists.w3.org/Archives/Public/public-html/2009Feb/0726.html [12:54:08.0000] <mpilgrim> http://lists.w3.org/Archives/Public/public-html/2009Feb/0723.html [12:54:09.0000] <mpilgrim> http://lists.w3.org/Archives/Public/public-html/2009Feb/0722.html [12:54:10.0000] <mpilgrim> http://lists.w3.org/Archives/Public/public-html/2009Feb/0721.html [12:54:11.0000] <mpilgrim> http://lists.w3.org/Archives/Public/public-html/2009Feb/0720.html [12:54:12.0000] <mpilgrim> http://lists.w3.org/Archives/Public/public-html/2009Feb/0695.html [12:55:00.0000] <mpilgrim> http://lists.w3.org/Archives/Public/public-html/2009Feb/0693.html [12:55:01.0000] <Hixie> rubys: there is actually no minimal requirement for participation at all [12:55:02.0000] <mpilgrim> http://lists.w3.org/Archives/Public/public-html/2009Feb/0690.html [12:55:03.0000] <mpilgrim> http://lists.w3.org/Archives/Public/public-html/2009Feb/0680.html [12:55:04.0000] <mpilgrim> important one: http://lists.w3.org/Archives/Public/public-html/2009Feb/0601.html [12:55:05.0000] <mpilgrim> (that last url contains comments on various examples of "good summary usage" put forth by others) [12:55:06.0000] <shelleyp> mpilgrim: thanks for the links. Does anyone know if these are recorded in a wiki somewhere? Luckily logs are kept of the page [12:56:00.0000] <rubys> ok, I like the three questions in the last link mpilgrim posted [12:56:01.0000] <rubys> still not scientifically testable, but relevant questions nevertheless [12:57:00.0000] <tantek> This issue appears to provide an overview with many more links: http://www.w3.org/html/wg/tracker/issues/32 [12:57:01.0000] <tantek> "table-summary / how to provide a summary of a table, e.g. for unsighted navigation?" [12:58:00.0000] <mpilgrim> http://lists.w3.org/Archives/Public/public-html/2008Aug/0505.html [12:58:01.0000] <rubys> tantek: the goal is to get http://esw.w3.org/topic/HTML/SummaryForTABLE/SummarySpecification to the point where it contains the right question. [13:00:00.0000] <shelleyp> rubys, if I've read the wiki entries correctly, the action going forward is to frame the options into a question and then put it to a vote, correct? [13:00:01.0000] <shelleyp> Sorry, IRC entries [13:01:00.0000] <tantek> rubys, IMHO Hixie's analysis in http://lists.w3.org/Archives/Public/public-html/2009Feb/0601.html is the most thorough/scientific so far, and asks many of the tough questions that rigor demands we ask (his 1,2,3 under "several things to consider when evaluating possible solutions") [13:01:01.0000] <rubys> I try really, really, really hard to avoid votes, but unfortunately, it looks like it can't be avoided this time. [13:02:00.0000] <shelleyp> rubys: not unless you want the issue to arise, over and over again [13:02:01.0000] <mpilgrim> here's hixie giving his opinion after watching that usability study: http://lists.w3.org/Archives/Public/public-html/2008Feb/0058.html [13:02:02.0000] <mpilgrim> i don't think there are any earlier links that give hixie's opinion [13:02:03.0000] <mpilgrim> hth [13:03:00.0000] <tantek> rubys, IMHO voting on this issue won't get you a better scientific (ergo technical) result, but merely a representation of who's dogma has better propagated or who can rally the most sports fans to their side. [13:03:01.0000] <Hixie> indeed [13:03:02.0000] <shelleyp> mpilgrim: does anyone know where working links to the videos can be found? [13:03:03.0000] <mpilgrim> i do not know [13:04:00.0000] <tantek> s/who's/whose - ahem. [13:04:01.0000] <mpilgrim> i don't believe i still have them [13:04:02.0000] <rubys> The two sides I see are "I'm not listening" and "I'm not listening". [13:04:03.0000] <shelleyp> mpilgrim: I'll see if I can find them somehow [13:04:04.0000] <Hixie> woah [13:04:05.0000] <Hixie> woah [13:04:06.0000] <Hixie> woah [13:04:07.0000] <Hixie> i'm listening [13:04:08.0000] <Hixie> i've done nothing but listen [13:04:09.0000] <Hixie> did you see the e-mail i sent yesterday? [13:05:00.0000] <Hixie> i even read there dozens-of-pages-long advocacy material [13:05:01.0000] <Philip`> shelleyp: Probably http://www.cfit.ie/html5_video/ [13:05:02.0000] <Hixie> their [13:05:03.0000] <rubys> I want to see a concrete question that, if answered, would get you to change your mind. Not an abstract or untestable theory, but something tangible. [13:05:04.0000] <Hixie> i've looked at their videos [13:05:05.0000] <Hixie> rubys: anything showing that summary="" actually improves accessibility more than it hurts it [13:05:06.0000] <rubys> untestable. try again. [13:05:07.0000] <tantek> rubys, "two sides" is a false dichotomy IMHO that does not actually help the discussion move forward, and it is inaccurate to characterize requests for answers to scientific questions as "I'm not listening" (unclear if you intended that or not). [13:06:00.0000] <Hixie> rubys: no, it's not untestable [13:06:01.0000] <rubys> propose a test? [13:06:02.0000] <Hixie> rubys: it's exactly what my opinion is based on and it is the ONLY thing that really matters! [13:06:03.0000] <Hixie> rubys: did you read the e-mail mark just cited? [13:07:00.0000] <rubys> this one? http://lists.w3.org/Archives/Public/public-html/2009Feb/0601.html [13:07:01.0000] <shelleyp> Hixie: Are you saying then that it doesn't matter what's said, asked, or a vote-you won't change your mind, or include @summary [13:07:02.0000] <Hixie> rubys: we looked at how summary="" is used on the large samples of data, how it's used on small samples representing pages by people who are compelled by law to use it correctly, we looked at videos of blind users using screen readers, we looked at gov't papers on the topic, what more do you want? [13:07:03.0000] <Philip`> I guess you could test how it directly affects users (ignoring costs to authors and implementors) by finding a group of people who use the web with a screen reader and who know how to enable/disable summary support in it, and then see how many choose to enable it and how many disable it [13:07:04.0000] <Hixie> shelleyp: no, i'm saying the exact opposite [13:07:05.0000] <rubys> hixie: which email? [13:07:06.0000] <Hixie> all the ones mark just pasted into irc [13:08:00.0000] <Hixie> rubys: what would it take to change the summary="" advocates' minds? [13:08:01.0000] <rubys> at various times, yes. [13:08:02.0000] <shelleyp> Hixie, are you responding to me publicly or privately? [13:08:03.0000] <tantek> rubys, IMHO Hixie's request for "anything showing that summary="" actually improves accessibility more than it hurts it" provides you with a far wider array of options than your request for a concrete question from him. [13:08:04.0000] <mpilgrim> here's an early discussion about accessibility-specific markup vs. universal markup solutions: http://lists.w3.org/Archives/Public/public-html/2007Jun/0962.html [13:08:05.0000] <Hixie> rubys: have the summary="" advocates indicated whether they would stop asking for summary="" if the vote supported <caption> rather than summary=""? [13:09:00.0000] <mpilgrim> (read the whole thing, or at least search for "This argues for having accessibility features be implicit") [13:09:01.0000] <Hixie> rubys: have the summary="" advocates said what it would take to change _their_ minds? [13:09:02.0000] <shelleyp> Wait a sec -- folks are setting conditions on the vote, that could undermine the integrity of the vote [13:09:03.0000] <shelleyp> A vote is a vote [13:09:04.0000] <Hixie> i'm merely asking of the "other side" as rubys put it the same questions as have just been asked of me [13:09:05.0000] <rubys> tantek: the peception I want to squash is "go Mets". Where any answer that doesn't match the orthodoxy is discounted. From my perception, that applies to "summary adocates" as well as "summary detractors". [13:09:06.0000] <rubys> hixie: that's a given [13:09:07.0000] <rubys> I will follow through on that. [13:09:08.0000] <Hixie> then why are you asking me the same question? [13:10:00.0000] <mpilgrim> and an early thread about solving problems vs. rooting for your favorite solution: http://lists.w3.org/Archives/Public/public-html/2007Jun/1052.html [13:10:01.0000] <tantek> rubys, if you are viewing http://lists.w3.org/Archives/Public/public-html/2009Feb/0601.html as a "summary detractor", then I have yet to see anything even close to that level of scientific questioning/rigor from any other position (whether "summary advocate" or some other position) [13:11:00.0000] <rubys> I want everybody to agree on some sort of scientific test by which we can determine the answer. Yes, Hixie: I'd like your input on that. [13:11:01.0000] <Hixie> /me waves to the people on twitter enjoying the discussion :-) [13:11:02.0000] <Hixie> rubys: i have already given you my input on that [13:11:03.0000] <Hixie> repeatedly [13:11:04.0000] <tantek> rubys, perhaps you should put a call out for scientific tests first, maybe on the mailing list. [13:12:00.0000] <shelleyp> rubys now I have to disagree: I don't think "scientific test" is necessarily the appropriate approach [13:12:01.0000] <tantek> perhaps providing http://lists.w3.org/Archives/Public/public-html/2009Feb/0601.html as a starting point - in other words, any scientific tests proposed should be at least as good as the analysis in that email to be worthy of consideration. [13:12:02.0000] <rubys> I do plan to put out a question on the mailing list, I just wanted to gather some input on IRC first. [13:12:03.0000] <shelleyp> After all, we can't put the web in a petri dish, and turn on a microscope [13:13:00.0000] <rubys> tantek: agreed. [13:16:00.0000] <shelleyp> Then I think it's only fair to say that every last bit of the HTML5 spec be subjected to the same form of scientific tests, to be fair [13:16:01.0000] <Hixie> it has been [13:16:02.0000] <Hixie> that's been one of the cornerstones of html5's development [13:16:03.0000] <Philip`> /me doesn't remember having seen suggested tests for summary that are actually scientific (like with numbers) rather than being subjective judgements on the relative importance of various benefits and costs [13:16:04.0000] <shelleyp> Hixie: Most have been subjected to your interpretation, and a couple of test apps -- that's not a 'scientific test' [13:17:00.0000] <Philip`> (...like the cost of bad summary text vs the value of good summary text) [13:17:01.0000] <Hixie> Philip`: yup, just like the rest of the spec :-) [13:17:02.0000] <Hixie> Philip`: (it's been subject to the same form of scientific tests) [13:17:03.0000] <shelleyp> Philip: Thank you for clarification [13:18:00.0000] <Philip`> Hixie: You mean the unscientific form of scientific tests? [13:18:01.0000] <Hixie> shelleyp: it's language design. interpretation is the name of the game. [13:18:02.0000] <Hixie> Philip`: yup [13:18:03.0000] <shelleyp> In other words, not based on empirical measurements, but the reason and logic of the arguments [13:18:04.0000] <Hixie> based on both [13:18:05.0000] <Hixie> empirical measurements isn't science [13:18:06.0000] <Hixie> but it's still better than nothing [13:18:07.0000] <shelleyp> Hixie you're not designing a language, you're improving a language [13:19:00.0000] <Hixie> why thank you [13:19:01.0000] <tantek> Hixie, empirical measurements are used in some science. And yes, better than nothing. [13:19:02.0000] <Hixie> tantek: empirical measurements is necessary but insufficent to have science [13:19:03.0000] <Hixie> er, s/is/are/ [13:20:00.0000] <shelleyp> Philip: can you suggest a specific scientific question/test on a specific component of the HTML spec I can look at in order to ensure my interpretation of "scientific test"? [13:21:00.0000] <tantek> /me provides "Science is a process" http://evolution.berkeley.edu/evolibrary/article/0_0_0/nature_05 as hopefully a good overview. [13:22:00.0000] <Hixie> (just to clarify, i don't think i ever said that we needed scientific tests. i said we need to base our judgements on research rather than on unqualified opinions.) [13:22:01.0000] <shelleyp> tantek: thank you, but I meant specifically as has been applied in the past to HTML5 [13:22:02.0000] <shelleyp> I work best with given examples of past success [13:22:03.0000] <Hixie> (research shows that summary="" isn't actually solving the problem that needs solving here; therefore we need something better. various people, including some from the US gov't's accessibility teams, have concluded that <caption> is an overall better solution.) [13:23:00.0000] <tantek> Hixie, basing judgments on research rather than unqualified opinions is a scientific method. [13:24:00.0000] <Hixie> tantek: to some extent, yes; but some people (e.g. rubys earlier) believe that it means you can come up with a single test the result of which can directly drive language design without a subjective component [13:24:01.0000] <Hixie> which is clearly not the case [13:24:02.0000] <Hixie> you have to interpret the results of any study [13:24:03.0000] <tantek> shelleyp, Hixie certainly applied "scientific thinking ... observations, questions, multiple hypotheses, and more observations." to the table summary issue, per the cited 601.html email. [13:24:04.0000] <tantek> (quote excerpted from "Science is a progress" link above) [13:25:00.0000] <shelleyp> tantek, thank you, I was looking for an example of a specific test where the result was to add to the HTML5 spec. For instance, what scientific test was used with the microdata section? [13:25:01.0000] <tantek> Hixie, I'm not sure of the "single test" hypothesis either - that's why I simply asked for alternative scientific analyses, at least as rigorous as your 601.html [13:25:02.0000] <Hixie> tantek: right [13:26:00.0000] <Hixie> shelleyp: no one test could be used for something that big, the microdata section involved hundreds of individual decisions each of which had to be based on research, implementation experience, discussions with authors, etc [13:27:00.0000] <Hixie> shelleyp: for example, to test whether the microdata features as described actually solved the problems they set out to solve, i explicitly used them to do so [13:27:01.0000] <Hixie> shelleyp: (most of that isn't documented in the list, i just did it in scratch files) [13:27:02.0000] <Hixie> shelleyp: (some of it became the examples in the spec) [13:27:03.0000] <tantek> shelleyp, it may be that the evaluation process for adding new features (e.g. microdata) is different than the evaluation process for dropping existing features, since existing features can be better evaluated based on observations in the wild. [13:28:00.0000] <shelleyp> Hixie: So an instance of implementation is a successful test of a given component of the HTML5? [13:28:01.0000] <Hixie> tantek: certainly the evaluation process for features for which we have a decade of experience can be far more rigorous [13:28:02.0000] <Hixie> shelleyp: it's a data point, of course. certainly not the only one. it is necessary but not sufficient. [13:29:00.0000] <shelleyp> OK, good, that's something tangible I can work with [13:29:01.0000] <Hixie> /me thought it was pretty obvious that you had to evaluate whether proposals actually solved the problems they set out to solve [13:29:02.0000] <Hixie> shelleyp: for example one use case was to take the events on livebrum.co.uk and make them into paragraphs that can be copied and pasted into blogs while still being drag-and-droppable to calendars [13:29:03.0000] <Hixie> shelleyp: so i worked with the guy who runs that site to come up with something that would satisfy his need [13:29:04.0000] <shelleyp> Yes, but interpretation of whether a problem is "solved" are frequently based on biased, subjective viewpoints [13:30:00.0000] <Hixie> you can see this example in the spec, it's one of those that made it in [13:30:01.0000] <Hixie> shelleyp: if so, then your problem description is far too vague [13:30:02.0000] <Hixie> shelleyp: problem descriptions should be objectively testable [13:32:00.0000] <shelleyp> Hixie: Of course, there's another issue: that you seem to be the one person making a determination of what is, or is not, a viable test [13:32:01.0000] <Hixie> the vast majority of the problem descriptions for microdata were in fact written by other people [13:33:00.0000] <rubys> /me agrees with Philip` [13:33:01.0000] <shelleyp> Hixie: yeeesss, but as I pointed out in a later document, you were incorrect in your assessment of the some of the problem descriptions [13:33:02.0000] <shelleyp> Your interpretation was in error [13:34:00.0000] <Hixie> shelleyp: i am not infallible, hopefully you raised such errors to my attention on the list [13:34:01.0000] <shelleyp> Hixie: Yes, I did. Unfortunately, you've been quite busy since. I still am reviewing your work, for a longer writing [13:35:00.0000] <rubys> hixie: what's the status/plans for aria? [13:36:00.0000] <Hixie> rubys: unchanged since my last "input on the agenda" e-mail [14:16:00.0000] <Hixie> gsnedders: yt? [14:16:01.0000] <gsnedders> /me wonders what now… [14:16:02.0000] <Hixie> /me looks for the sample anolis doc [14:16:03.0000] <Hixie> do you have the link to that again? [14:16:04.0000] <Hixie> we should put it on pimpmyspec.net [14:17:00.0000] <gsnedders> <http://hg.gsnedders.com/anolis/raw-file/tip/example.src.html> [14:17:01.0000] <gsnedders> <http://hg.gsnedders.com/anolis/raw-file/tip/example.html> [14:17:02.0000] <gsnedders> Message-Id: <699948DD-6166-4F55-8171-44D81108A898⊙gc> [14:20:00.0000] <Hixie> awesome thanks [14:32:00.0000] <gsnedders> /me just searched his sent box for to:hixie [14:32:01.0000] <gsnedders> :P [15:11:00.0000] <jgraham> Hixie: Added links to those docs to pms.net [15:33:00.0000] <Hixie> jgraham: sweet thanks [15:34:00.0000] <gsnedders> Weee… [15:35:00.0000] <gsnedders> just got a SPARQL query down from using 1GB RAM in rdflib to around 12MB RAM. [15:35:01.0000] <gsnedders> (It is really, really, really inefficient on some queries) [15:36:00.0000] <Philip`> I think you should rewrite the query to just use grep [15:36:01.0000] <gsnedders> That would be very hard. [15:55:00.0000] <Philip`> gsnedders: Nonsense! Everything is easy with grep [15:59:00.0000] <Philip`> Am I missing some place where <foo xmlns:bar=""/> is clearly a namespace-well-formedness error? [15:59:01.0000] <Philip`> All I see is "the namespace name may not be empty", which isn't a very convincingly accurate conformance requirement [16:04:00.0000] <Philip`> Oh, it's clarified in an errate [16:04:01.0000] <Philip`> s/e/a/ [16:06:00.0000] <Hixie> the average rate of new feedback has increased [16:06:01.0000] <Hixie> interesting [16:53:00.0000] <Hixie> hsivonen: yt? 2009-06-05 [18:18:00.0000] <Hixie> who was i discussing the websocket protocol with recently? [18:18:01.0000] <Hixie> something about how servers should handle errors [18:18:02.0000] <Hixie> was that jwalden? [18:33:00.0000] <franksalim> Hixie, you were recently discussing the use of CONNECT with jwalden [18:34:00.0000] <Hixie> ok [18:34:01.0000] <Hixie> thanks [18:34:02.0000] <Hixie> i'll have to catch him when he gets back [18:54:00.0000] <Hixie> i wish i understood the IETF better [18:54:01.0000] <Hixie> e.g. why does http://tools.ietf.org/html/draft-hixie-thewebsocketprotocol have versions 12, 13, and 15 but not 14? [18:58:00.0000] <franksalim> I don't know, and now I cannot easily see the diff for the new version [18:59:00.0000] <Hixie> the diff is at http://html5.org/tools/web-apps-tracker [18:59:01.0000] <Hixie> http://html5.org/tools/web-apps-tracker?from=3191&to=3192 specifically is today's update [19:01:00.0000] <franksalim> thanks. it is unfortunate that the missing versions broke the diff links on the ietf site, though [21:52:00.0000] <jwalden> Hixie: it'd be nice if you made it easier to download acid3 for local testing and modifications somehow, maybe a zip of the directory and the Apache config files inside it (I'm guessing at your current setup) [21:53:00.0000] <Hixie> i can provide a tarball if you want [21:53:01.0000] <Hixie> btw i made the websocket protocol spec clearer about server-side error handling rules [21:53:02.0000] <jwalden> that'd have been useful a few hours ago, but I hacked one together [21:53:03.0000] <jwalden> spat of wgets and the occasional -S to see what I was missing [21:53:04.0000] <jwalden> /me looks at ws [21:55:00.0000] <Hixie> well let me know if you ever want one, i do make the on demand [21:55:01.0000] <Hixie> i don't really want to encourage people to download it though [21:58:00.0000] <jwalden> why not? [21:58:01.0000] <jwalden> fears that people are likely to screw up the configuration process? [21:59:00.0000] <Hixie> more or less [21:59:01.0000] <jwalden> hm [21:59:02.0000] <Hixie> i don't want to have to field support questions from non-browser developers [21:59:03.0000] <Hixie> non-browser-developers i should say [21:59:04.0000] <jwalden> could put a URL in the acid3 source, might be sufficiently obscure for people not to use it unless they're interested [21:59:05.0000] <Hixie> shrug browser people can just ask me :-P [22:06:00.0000] <ezyang> Do we have updated tests for th latest dash-dash-bang-angle comment terminator? [23:52:00.0000] <hsivonen> Hixie: you asked if I was here [23:52:01.0000] <hsivonen> I just read the log from yesterday [23:52:02.0000] <Hixie> i wanted to ask you your input on the comment stuff, iirc [23:52:03.0000] <hsivonen> I think what Philip` said http://krijnhoetmer.nl/irc-logs/whatwg/20090604#l-1058 is the best way to discover if @summary is useful [23:53:00.0000] <hsivonen> Hixie: context? [23:54:00.0000] <hsivonen> Hixie: I want to avoid reparsing if we can avoid it. I don't have proper data, so tentatively I trust zcorpan's judgment on --!> [23:54:01.0000] <Hixie> http://html5.org/tools/web-apps-tracker?from=3190&to=3191 [23:56:00.0000] <hsivonen> Hixie: seems reasonable [23:56:01.0000] <Hixie> k [23:56:02.0000] <hsivonen> I'm already over the 8000 byte limit anyway :-( [23:58:00.0000] <hsivonen> (aside on the 8000-byte limit: not unsurprisingly, on Stack Overflow, I got more people doubting the premise on my question than contributing towards answering the actual question) [23:59:00.0000] <Hixie> heh [23:59:01.0000] <hsivonen> here http://stackoverflow.com/questions/944492/is-there-a-java-bytecode-optimizer-that-removes-useless-gotos [00:03:00.0000] <Hixie> hsivonen: have you considered just writing one manually? it seems you know what optimisation you want. [00:04:00.0000] <Hixie> (possibly by extending an existing optimiser) [00:14:00.0000] <hsivonen> Hixie: I've even considered writing a custom compiler for the method, but I'd prefer not to spend the time [00:14:01.0000] <Hixie> heh [00:14:02.0000] <Hixie> you've already written a Java to C++ convertor [00:14:03.0000] <Hixie> how much hard could a Java to bytecode convertor be :-) [00:15:00.0000] <hsivonen> I've also considered a per hack for splitting the doctype states out just ahead of compilation without spoiling the mail source [00:15:01.0000] <hsivonen> s/pel/perl/ [00:39:00.0000] <zcorpan> "I would like to kill Hixie" - http://hg.gsnedders.com/anolis/raw-file/tip/example.src.html [00:41:00.0000] <pesla> Morning [00:42:00.0000] <hsivonen> zcorpan: fails at nice and happy :-( [00:55:00.0000] <jgraham> Hixie: It you have something like  [00:57:00.0000] <jgraham> Hixie: I see efg --> outside the comment [00:57:01.0000] <jgraham> (WebKit is different) [01:00:00.0000] <Hixie> http://software.hixie.ch/utilities/js/live-dom-viewer/saved/133 [01:00:01.0000] <Hixie> what do you see? [01:00:02.0000] <Hixie> i'm using Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.5; en-US; rv:1.9.2a1pre) Gecko/20090604 Minefield/3.6a1pre [01:07:00.0000] <zcorpan> Hixie: why is "anything else" a parse error in the bang state? [01:09:00.0000] <zcorpan> Hixie:  has one parse error but  has two [01:09:01.0000] <Hixie> /me denies everything while checking in a fix [01:10:00.0000] <jgraham> Hixie: What I see depends on standards mode v quirks mode [01:10:01.0000] <jgraham> In standards mode dash-dash-space terminates a comment [01:11:00.0000] <Hixie> i've only been checking quirks mode [01:12:00.0000] <zcorpan> in gecko [01:12:01.0000] <Hixie> cos that's what the e-mail i responded to earlier today said to check :-) [01:13:00.0000] <zcorpan> from the pages that were analyzed, it was reasonably clear that pages would render better if -- > closed comments [01:14:00.0000] <zcorpan> certainly there was no page in the set that would render worse [01:15:00.0000] <jgraham> Hixie: It was opera quirks mode that does reparsing. I didn't realise that gecko behaviour depended on mode [01:15:01.0000] <Hixie> what exactly is the proposal? Same as for bang, but if you hit a space, stay in the state? How do you know how many spaces to emit when you hit a bogus character afterwards? [01:15:02.0000] <Hixie> seems like this would involve tracking more state than we'd like [01:16:00.0000] <jgraham> Hixie: Yes and I doubt it matters [01:16:01.0000] <jgraham> Oh maybe it does matter [01:17:00.0000] <Hixie> collapsing all the spaces everytime you hit two --s in a row in a comment would be mighty weird [01:17:01.0000] <Hixie> especially if one of the spaces is a newline [01:17:02.0000] <Hixie> and the comment is something like: [01:17:03.0000] <Hixie>  [01:18:00.0000] <zcorpan> i guess you could just append the two dashes to the comments data upon seeing whitespace so that " on the last non-whitespace line of scripts? [07:12:00.0000] <zcorpan> ie and opera have interesting behavior where the last line is stripped if it ends with --> [07:13:00.0000] <zcorpan> in opera the line doesn't get stripped if it also contains // [07:13:01.0000] <Lachy> which versions of IE have you tested? [07:13:02.0000] <zcorpan> Lachy: ie8 [07:13:03.0000] <jgraham> Lachy: Consider http://arxiv.org/pdf/astro-ph/0603449v2 Table 2 has a rather long caption. Many other tables do not. Insisting that they should be different elements doesn't make much sense [07:13:04.0000] <zcorpan> in ie it doesn't get stripped if the line contains // or  isn't commented out with // [07:14:00.0000] <Philip`> zcorpan: Do you mean something like "^\S+-->\s</script>"? [07:15:00.0000] <zcorpan> Philip`: there could be whitespace on the line, just not only whitespace followed --> [07:15:01.0000] <zcorpan> foo = 'bar'; --> [07:15:02.0000] <zcorpan> </script> [07:16:00.0000] <Lachy> jgraham, the captions in that contain a lot of content that I would have thought more appropriate for the main content of the paper where referring to the relevant figures [07:16:01.0000] <Philip`> zcorpan: Oops, yes [07:17:00.0000] <jgraham> Lachy: That is a typical style for those papers. They are rahter optimised for people using figures and tables as references without also having to reread the whole paper [07:17:01.0000] <Philip`> zcorpan: Do you want to ignore lines that use //--> (or similar)? [07:17:02.0000] <zcorpan> Philip`: yes, ignore lines that have // [07:17:03.0000] <Lachy> well, I'd consider that bad form, but ok [07:19:00.0000] <Philip`> zcorpan: What if they start with </script>) [07:22:00.0000] <Lachy> Philip`, I've seen that done in order to prevent some broken backend systems from mistakenly outputting empty scripts as <script src="..." /> [07:22:01.0000] <Lachy> so they put in a comment to make it appear non-empty [07:22:02.0000] <zcorpan> Philip`: if they are overwhelming then yeah it would be good to remove those that have  but no \s</script> [07:25:01.0000] <Philip`> (but only searching individual lines, so it won't find anything if the </script> is on the next line after the content) [07:26:00.0000] <Lachy> if you did the search using just (?i)^\s(?!//|\s* would you get too many false positives? [07:27:00.0000] <jgraham> Philip`: If possible it would be nice to get cases where the </script> is on the next line too [07:27:01.0000] <jgraham> Although I guess that would be slower [07:27:02.0000] <Philip`> Lachy: I expect that would find a bazillion HTML comments that have nothing to do with scripts [07:28:00.0000] <Lachy> oh, right, so you need a way to limit the search to within script elements [07:35:00.0000] <Philip`> /me tries running it across multiple lines, but it'll take forever (since the single-line one hasn't even finished yet) [07:36:00.0000] <Philip`> I really ought to parallelise my grep :-/ [07:38:00.0000] <zcorpan> Philip`: i'm not in a hurry. :) sorry for using your cpu [07:39:00.0000] <jgraham> "Aplologies for increasing the entropy of the universe" [07:39:01.0000] <jgraham> Apologies [07:40:00.0000] <Philip`> Ooh, the first one finished [07:40:01.0000] <Philip`> zcorpan: http://philip.html5.org/data/comment-close-at-end-of-script.txt [07:41:00.0000] <zcorpan> Philip`: thanks [07:51:00.0000] <zcorpan> Philip`: the first few pages of those use //--> but your .txt says /--> [07:53:00.0000] <zcorpan> though www.mp3music.co.il/forums/topic/?id=6&guid=MNRQFBZGBU12252008235139 does use a single slash [07:54:00.0000] <zcorpan> so maybe www.oclc.org/services/preservation/ has changed [07:54:01.0000] <Philip`> zcorpan: The pages like http://www.oclc.org/services/preservation/http:/www.oclc.org/programsandresearch/kilgouraward/http:/www.oclc.org/research/http:/www.oclc.org/support/systemalerts/default.asp say "/--></script>" twice when I look at them [07:54:02.0000] <Philip`> s/twice/thrice/ [07:54:03.0000] <Philip`> at the bottom [07:55:00.0000] <zcorpan> ah [07:55:01.0000] <zcorpan> i thought your script emitted several urls [07:55:02.0000] <Philip`> Only one per line [07:56:00.0000] <Philip`> That is a pretty weird URL, though [07:56:01.0000] <zcorpan> i guess it's not too bad if a 404 page doesn't work correctly :) [07:58:00.0000] <Philip`> Hmm, my other grep has reached chunk 6 out of f [07:58:01.0000] <Philip`> so it might take a long time [07:58:02.0000] <zcorpan> http://www.onpointevents.com/PPF/Page/1/resources.asp has a script that sets the today variable but it's not used anywhere, so it wouldn't break [07:59:00.0000] <Philip`> zcorpan: http://philip.html5.org/data/comment-close-at-end-of-script-across-multiple-lines.txt is some of the multi-line grep's output [08:00:00.0000] <Philip`> (Apologies for the barely-readable output format) [08:00:01.0000] <zcorpan> Philip`: thanks. it's readable enough :) [08:01:00.0000] <zcorpan> the script at mp3music.co.il/lyrics2/text.aspx?id=26601 would stop working [08:02:00.0000] <zcorpan> but it doesn't seem like a vital script [08:10:00.0000] <zcorpan> it breaks a script at http://www.marika.com/productinfo_item_43-dc_614-md_614-pd_564-cd_on-sn_5.htm but i don't know what the script does - the page seems to work in firefox [08:12:00.0000] <zcorpan> if(lt){d.write('<'+'!-- ');}//--></script><noscript> [08:12:01.0000] <zcorpan> <img src="http://adlik.akavita.com/bin/lik?id=28254" [08:12:02.0000] <zcorpan> border=0 height=31 width=88 alt="Akavita: каталог, рейтинг, счетчик для сайтов Беларуси"> [08:12:03.0000] <zcorpan> </noscript><script language="JavaScript"></script> [08:12:05.0000] <zcorpan> http://mogbiz.by/modules.php?name=pages&pa=list_pag&cid=32 [08:13:00.0000] <zcorpan> won't break since the second <script> tag will be part of an html comment [08:13:01.0000] <Philip`> Ouch [08:17:00.0000] <zcorpan> yay i've gone through http://philip.html5.org/data/comment-close-at-end-of-script.txt - none of them actually breaks in functionality afaict, a few of them would cause a script error [08:17:01.0000] <zcorpan> on to multiline [08:18:00.0000] <zcorpan> <script> [08:18:01.0000] <zcorpan>  [08:18:05.0000] <zcorpan> </script> [08:18:06.0000] <zcorpan> sigh [08:19:00.0000] <zcorpan> http://www.ctax.org.cn/was40/search?channelid=39843 does break [08:22:00.0000] <Philip`> Oops, X crashed, which killed my SSH session, which killed my grep [08:22:01.0000] <Philip`> I hope that partial list is adequate for now [08:22:02.0000] <zcorpan> yeah it's good enough [08:22:03.0000] <zcorpan> thanks again [08:23:00.0000] <zcorpan> http://www.modelenium.co.uk/models/TACA-A319--GJTAI149.aspx just tries to disable right click so not too serious if the script doesn't run :) [08:24:00.0000] <zcorpan> http://www.stadtguide24.de/index.php?myGeoId=14556&site=edit&lang=de has an ie-specific script with document.all check so doesn't matter [08:27:00.0000] <gsnedders> http://s.ytimg.com/yt/js/base_all_with_bidi-vfl101077.js is being really slow taking YouTube down here :\ [08:27:01.0000] <gsnedders> s.ytimg.com is slow in general :\ [08:28:00.0000] <zcorpan> http://www.meganime.org/foro/viewtopic.php?p=316494 PreloadImages() will break, but only breaks the preload so not too serious [08:30:00.0000] <zcorpan> http://www.fise.org/associazioni/assioma/index.php?we_objectID=9643&we_objectTID=425 - breaks the image swap in the menu. not too serious, either [08:31:00.0000] <zcorpan> http://www.wholesalemegamall.com/index.php?c=office&n=1084128&i=B000LYAX1G&a=buy&m=All&p=1&x=Panasonic_Dect_60_Series_3_Handset_Cordless_Phone_System_with_Answering_System_KX_TG1033S has ie-specific script [08:33:00.0000] <zcorpan> http://www.diadora.de/planetnewsplus/misc/tell_a_friend/tell.php?id=20070301120209 also breaks preload / image swap (but i don't know which images) [08:35:00.0000] <zcorpan> http://www.foodireland.com/Merchant2/merchant.mvc?Screen=PROD&Product_Code=w580420&Category_Code=caseflour&Product_Count=4 - breaks the script that prints out today's date [08:38:00.0000] <zcorpan> http://www.emumax.com/emumax_tvnews_detail.asp?id=188 - seems to break a tracker script or something [08:42:00.0000] <zcorpan> seems to possibly break the comment form in http://www.langya.cn/travel/lydx/200801/t137487.htm [08:45:00.0000] <zcorpan> the gay.sexe-photos site uses <script language="vbscript"> [08:45:01.0000] <zcorpan> /me realises that he just watched porn at work [08:46:00.0000] <gsnedders> /me facepalms [08:48:00.0000] <zcorpan> community.bravo.de/boards/viewtopic.php?t=134678: / ]]> /--> [08:49:00.0000] <zcorpan> that works in spidermonkey but breaks in nitro and v8 [08:52:00.0000] <zcorpan> http://www.rapido.nl/care-plus-click-away-bite-relieve-travel-accessoire-artikelnummer-91654.html probably breaks some progress bar script [08:55:00.0000] <zcorpan> well the rest just seem to be more of the same stuff [08:56:00.0000] <virtuelv> well, Opera summer party coming up in 9 seconds. [08:58:00.0000] <zcorpan> crap, http://www.bbsone.com/html.php?file=bbs/business/gb/messages/908.htm would break a link [09:00:00.0000] <zcorpan> http://www.hjsflower.com/flower/show.asp?id=A3004&bookid=961&anid=3&nid=34 might break the form [09:01:00.0000] <zcorpan> http://www.yeongnam.com/yeongnam/html/edu/university/article.shtml?id=20090112.010330807390001 breaks form validation (not too serious) [09:04:00.0000] <zcorpan> i'm through the list! [09:04:01.0000] <zcorpan> so let's see [09:06:00.0000] <Philip`> Let me know if you want a longer list ;-) [09:07:00.0000] <zcorpan> 2 with missing content, 1 with broken link, 2 something with forms, 2 with a progress bar something, a few with preload/imageswap and the rest didn't really break anything [09:08:00.0000] <zcorpan> so that's about 5 pages breaking in functionality [09:12:00.0000] <zcorpan> (since i didn't see any progress bar and the imageswap isn't really functionality) [09:13:00.0000] <zcorpan> wonder if the other vendors want to implement this [09:19:00.0000] <zcorpan> /me updates http://wiki.whatwg.org/wiki/Web_ECMAScript#HTML_comments [12:09:00.0000] <gsnedders> /me grumbles at the W3C's own inability to follow its own schemas [12:11:00.0000] <jgraham> gsnedders: are you surprised? [12:24:00.0000] <gsnedders> jgraham: Knowing how many bugs I've filed on this already, no :P [12:24:01.0000] <gsnedders> Actually, in some ways, I'm surprised I haven't found all the issues yet. [12:24:02.0000] <gsnedders> /me wonders if he's doing a Philip` and reporting issues on his own suggestions [13:59:00.0000] <hober> http://radar.oreilly.com/2009/06/fbml-yml-osml-oh-my.html [15:09:00.0000] <gsnedders> ezyang: We need to deal with stuff at a Unicode level in the TreeBuilder so we can coerce what we have into an XML infoset, which means we will either need to decode the data _again_, or we just use an array of codepoints, which eats memory… [15:09:01.0000] <gsnedders> /me sighs [15:11:00.0000] <ezyang> Oh, we want XML infoset support? [15:11:01.0000] <gsnedders> ezyang: We need to, if we are to use DOM… [15:11:02.0000] <gsnedders> ezyang: It's the cause of all 10 current exceptions thrown [15:11:03.0000] <ezyang> Well, I have no intentions of staying on DOM [15:12:00.0000] <gsnedders> ezyang: How are you going to write anything in userland with anywhere near decent perf? :) [15:12:01.0000] <ezyang> Write it in C and PHP [15:12:02.0000] <gsnedders> ezyang: Not re-distributable. [15:12:03.0000] <gsnedders> (Well, sure, it is, but if you want to ship it in normal software…) [15:12:04.0000] <ezyang> Performance is "not re-distributable" [15:12:05.0000] <ezyang> We'll obviously have a PHP drop-in replacement [15:13:00.0000] <ezyang> A userland tree implementation can have decent performance, given some concessions on the tree-builder and API side [15:13:01.0000] <ezyang> Decent enough, at least, for anyone who wants to casually use the library [15:13:02.0000] <gsnedders> What would be nice would be to get DOM to do no checks when strictErrorHandling = false [15:14:00.0000] <ezyang> What's more likely to happen is that userspace SimpleTree will blow up memory usage [15:14:01.0000] <ezyang> (not that DOM doesn't do that already, but you can't see that if you profile) [15:14:02.0000] <ezyang> gsnedders: We filed a bug on that a long time. Didn't work. [15:14:03.0000] <ezyang> a long time ago [15:15:00.0000] <gsnedders> Didn't work in what way? [15:15:01.0000] <bdodson> Hi, can anyone tell me how websockets will handle/support binary data? Will you be able to download the bytes of a jpeg through it and load it in the DOM for example? [15:15:02.0000] <Hixie> eventually, yes [15:15:03.0000] <Hixie> not in the first version though [15:16:00.0000] <ezyang> As in, they were like, WONTFIX, and proceeded to add the check for the place they missed [15:16:01.0000] <ezyang> Let me dig up the bug [15:16:02.0000] <ezyang> (we had this conversation before) [15:16:03.0000] <gsnedders> http://bugs.php.net/bug.php?id=44648&edit=1? [15:17:00.0000] <gsnedders> That's the only thing I find looking for http://bugs.php.net/bug.php?id=44648&edit=1 [15:17:01.0000] <ezyang> Ya, that's it [15:17:02.0000] <gsnedders> strictErrorHandling [15:17:03.0000] <gsnedders> Clipboard fail. [15:17:04.0000] <gsnedders> I got told that it could be done… if I wrote the patch. [15:17:05.0000] <gsnedders> Which is all fine and well… if it isn't for the fact I don't know C. :P [15:17:06.0000] <ezyang> It's not that difficult a patch [15:17:07.0000] <ezyang> PHP insulates you sufficiently from actually C-ness [15:17:08.0000] <ezyang> The hard problem is figuring out how to compile PHP [15:17:09.0000] <gsnedders> I couldn't work out how to get the strictErrorHandling property [15:18:00.0000] <ezyang> Hm. [15:18:01.0000] <gsnedders> If I could work that out, I could probably write a patch :P [15:18:02.0000] <ezyang> That's slightly non-trivial [15:18:03.0000] <ezyang> Anyway, I still argue, that's not tenable for the sake of portability [15:18:04.0000] <gsnedders> How so? Older shipped versions? [15:19:00.0000] <ezyang> Yep [15:19:01.0000] <gsnedders> Agreed, but in the longer-term it would be nice to be able to use it. [15:19:02.0000] <ezyang> Sure. [15:19:03.0000] <ezyang> I wonder if SimpleXML does checking... [15:19:04.0000] <gsnedders> In places. [15:20:00.0000] <bdodson> does anyone have basic websocket support now? [15:20:01.0000] <ezyang> :-( [15:20:02.0000] <gsnedders> /me wishes they would fix XMLWriter so you could actually rely on it outputting XML. [15:20:03.0000] <Philip`> An XML serialiser that outputs XML? I think you're asking for too much [15:21:00.0000] <gsnedders> http://bugs.php.net/bug.php?id=41800 [15:21:01.0000] <gsnedders> See, it's not a bug! [15:21:02.0000] <ezyang> gsnedders: If you want to take a stab at it again (I probably won't have time for it this weekend), the first thing you'll want to figure out is whether or not it's something stored in the PHP layer, or the libxml layer [15:22:00.0000] <gsnedders> ezyang: I don't care enough. I have too many bad experiences to think I have any chance of it ever being committed. [15:23:00.0000] <gsnedders> http://bugs.php.net/bug.php?id=44619 [15:24:00.0000] <ezyang> They're going to fix it The Right Way! No Really! [15:24:01.0000] <ezyang> More seriously, [15:25:00.0000] <gsnedders> /me adds XMLWriter grips to his list of problems with PHP [15:25:01.0000] <gsnedders> gripe [15:25:02.0000] <ezyang> I think the PHP dev community is not too receptive about patches unless you bug them a lot about it [15:25:03.0000] <gsnedders> gripes [15:25:04.0000] <ezyang> You will almost certainly not get a patch that big into PHP5.2. Maybe PHP5.3 [15:25:05.0000] <Philip`> gsnedders: Are you foolishly expecting the XML serialiser to deal with the issues of encoding data so it can be serialised as XML? Clearly that should be the user's responsibility [15:26:00.0000] <gsnedders> Philip`: I am expecting an XML serializer to always output something that meets the document production in the XML spec. [15:26:01.0000] <Philip`> It's a bug that they've even got any methods other than $xml->write('<test>blabla</test>') [15:26:02.0000] <Hixie> bdodson: webkit and mozilla are working on it, i believe [15:27:00.0000] <gsnedders> ezyang: I bet you 5.3 will be in commit freeze by the time I get around to anything like writing a patch… [15:27:01.0000] <Philip`> (It's up to you to make sure the data is encoded properly) [15:28:00.0000] <ezyang> So, if PHP didn't have binary strings by default, this would make sense [15:28:01.0000] <bdodson> Hixie: great, I guess I'll wait patiently for a bit :) [15:28:02.0000] <gsnedders> ezyang: s/by default/always/ [15:30:00.0000] <Hixie> bdodson: you and me both! [15:30:01.0000] <ezyang> Well, if most of the strings whizzing around are UTF-16, and then I want to put a binary string into the XML writer [15:30:02.0000] <ezyang> I'd say it has two options: refuse to do anything with it, or base64 encode it [15:30:03.0000] <bdodson> Hixie, huh, here's a video claiming minefield has WebSocket support ... [15:30:04.0000] <bdodson> http://www.youtube.com/watch?v=CrV9jJQn_MA [15:31:00.0000] <ezyang> "Fixing it" so that it's well-formed UTF-whatever is totally not the correct behavior. [15:31:01.0000] <gsnedders> ezyang: I'd expect it to die. [15:31:02.0000] <ezyang> And then, you can't even do that the smart way in PHP, because it doesn't have any notion of types [15:31:03.0000] <ezyang> So this won't get fixed until PHP 6, when we get Unicode strings [15:32:00.0000] <Hixie> bdodson: that's a demo of the work they're doing, it's not in yet i don't think [15:32:01.0000] <ezyang> PHP is currently operating an abstraction level too low [15:32:02.0000] <ezyang> And this is a leak. [15:38:00.0000] <gsnedders> Up to 12 items on my list of problems with PHP… [15:39:00.0000] <ezyang> i,i Only twelve? [15:39:01.0000] <gsnedders> So far :) [15:39:02.0000] <gsnedders> These are mainly fairly high-level issues. [16:43:00.0000] <Hixie> wow, shelley just sent what i believe is the most honest answer as to why people want to keep summary="" despite it having demonstrably failed to improve accessibility [16:48:00.0000] <annevk2> 'They've spent, what seems to me, a lot of time promoting this attribute, and accessibility in general. And now, a few people just want to yank it because "it's not being used right".' -> http://en.wikipedia.org/wiki/Sunk_costs [16:52:00.0000] <Hixie> yup [16:53:00.0000] <annevk2> nn [16:53:01.0000] <Hixie> nn [16:54:00.0000] <Dashiva> It could be that some people believe that the bad use just doesn't count. It's just a tiny waste of time each occasion. 2009-06-06 [17:01:00.0000] <Philip`> "wow. fascinating. depressing." - that's an appropriate response to pretty much any attempt to look at a random collection of real-world markup [17:02:00.0000] <Dashiva> Or humanity in general :) [17:02:01.0000] <Philip`> I hope nobody is going to notice that there's some weird euc-jp decoding issues on my table-summary-values-dotbot page where it inexplicably misses the '"' at the end of an attribute value [17:03:00.0000] <Philip`> (I assume it's an euc-jp decoding issue because all the offending values look kind of Japanese or something foreign like that) [17:03:01.0000] <Philip`> (I can't reproduce the problem with parsetree.validator.nu but I don't care enough to debug it further) [17:04:00.0000] <Dashiva> Shouldn't the URL show whether it's Japanese or not? [17:04:01.0000] <Philip`> I guess so [17:04:02.0000] <Philip`> but I didn't look very closely [17:05:00.0000] <Philip`> and I'm not going to look now because it's a 5MB page and I have to pay the equivalent of half a chocolate digestive biscuit per megabyte of traffic [17:06:00.0000] <Dashiva> Well, I just clicked the expand link in your summary table :) [17:07:00.0000] <Philip`> You could look for some of the values lower down that have strings like "<tr>" in them [17:08:00.0000] <Philip`> and see if those are .jp pages [17:08:01.0000] <Dashiva> Yup [17:08:02.0000] <Hixie> Philip` needs a better network service provider [17:08:03.0000] <Hixie> one that charges a flat rate [17:08:04.0000] <Lachy> I'm surprised by how well Shelly wrote those emails about summary. Is she the same Shelly who's been complaining about RDFa stuff? [17:08:05.0000] <Hixie> and not per bit [17:09:00.0000] <Dashiva> Maybe they charge per byte, so it still costs the same if he only downloads 7 bits [17:09:01.0000] <Philip`> Hixie: I have one of those, but the wireless router appears to be dead, so I'm using a mobile broadband thingamy instead [17:10:00.0000] <Hixie> Philip`: fix the router :-P [17:10:01.0000] <Dashiva> Philip`: I checked a few different ones, they all contain Japanese sites. I can't say every site listed is Japanese, but it seems likely. [17:10:02.0000] <Philip`> Hixie: I can't fix it myself - all I can do is complain to the owners, and they're asleep now :-p [17:11:00.0000] <Lachy> although, I really hate it when people claim that "... the folks who really want this attribute gone, really have no alternative solution in mind." because that's just complete nonsense. [17:11:01.0000] <Hixie> Philip`: ah [17:12:00.0000] <Philip`> /me found that most "flat rate" mobile broadband thingamy plans had a limit of 3GB per month, and cost £15/month, which isn't so good [17:13:00.0000] <Philip`> (so I got one that's £15/GB and doesn't expire for 6 months, which is better value when I don't use it much) [17:13:01.0000] <takkaria> Philip`: hm, is that 3G USB stick-type stuff? who did you get yours from? [17:14:00.0000] <Philip`> takkaria: It is [17:14:01.0000] <Dashiva> Bah. Why is it so hard to get that stuff like autoplay are author hints, the UA is perfectly capable of overriding. [17:14:02.0000] <Philip`> takkaria: It's a Vodafone one [17:15:00.0000] <takkaria> Philip`: ah cool, sounds like what I want really [17:16:00.0000] <Philip`> takkaria: (Also it works on Linux, with a couple of configuration files found via Google) [17:30:00.0000] <jjlee> what's the best way to get character indices of tags out of html5lib? [17:31:00.0000] <jjlee> I had a quick look a few weeks back, and suspected that copy and pasting chunks of html5lib into my code was likely my best option [17:31:01.0000] <jjlee> any other ideas? [18:20:00.0000] <ezyang> So, is the dash-dash-bang-angle comment terminator change solid now? [18:20:01.0000] <ezyang> (If it is, we should add tests) [18:22:00.0000] <Hixie> i think so [18:23:00.0000] <ezyang> Great! [21:33:00.0000] <Hixie> uh oh [21:33:01.0000] <Hixie> i've reached microdata e-mails [21:34:00.0000] <Hixie> time to go home i think! [21:35:00.0000] <Dashiva> But it's not even midnight yet! [22:21:00.0000] <Hixie> http://digg.com/comedy/W3C_s_Response_to_the_sarcasm_Tag_in_HTML_5_PIC is sad [22:22:00.0000] <Hixie> first of all, whoever is behind the w3c twitter account clearly doesn't know html5 well [22:23:00.0000] <Hixie> secondly, clearly the people commenting there don't either :-P [22:27:00.0000] <ezyang> Yaya [22:27:01.0000] <ezyang> But you should take a deep breath [22:28:00.0000] <jruderman> you disagree with the reply from the w3c account? [05:44:00.0000] <zcorpan> Hixie: maybe <hgroup> without <h1>-<h6> descendant should be ignored altogether instead of being equivalent to <h1> [08:54:00.0000] <gsnedders> /me wants resizable text fields in all browsers… [08:54:01.0000] <gsnedders> /me grumbles [10:25:00.0000] <Philip`> <li><a href...>Menu item </a>  <ul> Sub-menu items... </ul>  </li> [10:25:01.0000] <Philip`> Conditional comments let people do scary things [10:52:00.0000] <krijnh> Philip`: almost every example on http://www.cssplay.co.uk/ uses that [10:54:00.0000] <krijnh> I think it started with http://www.cssplay.co.uk/menus/dd_valid_2.html [11:02:00.0000] <gsnedders> Philip`: WTF? [11:26:00.0000] <jgraham> /me wonders if he should clean up the table summary long terms solution page so that it isn't so biased [11:27:00.0000] <jgraham> Or failing that at least so that it makes sense [11:28:00.0000] <jgraham> And actually has the proposal I made rather than a worse variation on it [13:47:00.0000] <gsnedders> Is there anyway to get a a Document object from an HTML string? [13:47:01.0000] <gsnedders> a string containing HTML [13:48:00.0000] <Philip`> document.write into an iframe? [13:49:00.0000] <gsnedders> how do I do it in an iframe? [13:50:00.0000] <gsnedders> I mean, the iframe will already have html/head/body… [13:52:00.0000] <Philip`> Do what the Live DOM Viewer does, I guess [13:52:01.0000] <gsnedders> What does that do? [13:52:02.0000] <gsnedders> :P [13:55:00.0000] <Philip`> document.writes into an iframe, I think [13:56:00.0000] <Philip`> (after calling document.open) [13:57:00.0000] <gsnedders> /me always gets confused when he looks at the Live DOM Viewer's source :P [14:16:00.0000] <webben> (OT): http://www.w3.org/TR/CSS2/syndata.html#charset Anybody know what CSS2.1 means by "byte" here? [14:17:00.0000] <webben> I'm trying to implement CSS2.1 character encoding sniffing, and I'm just looking at the character encodings that involve transcoding. [14:17:01.0000] <webben> Starting with GSM 03.38. [14:17:02.0000] <gsnedders> /me notes this would be easier if he hadn't just crashed his web browser [14:19:00.0000] <webben> What's confusing me is that whereas you can grab octets for (say) UTF-8 bytes (i.e. octets) and match against their regexes, GSM 03.38 takes packs 7-bit codepoints into octets. [14:19:01.0000] <webben> so if you just naively grab the octets, they won't match that regex [14:19:02.0000] <gsnedders> Well, if it is codepoints then doesn't that give you the Unicode string? [14:19:03.0000] <webben> you actually have to decompose them to septets. [14:20:00.0000] <webben> gsnedders: I'm using codepoints loosely. These aren't unicode codepoints. [14:20:01.0000] <gsnedders> /me thinks this would be easier to answer if he actually knew what GSM did :) [14:21:00.0000] <webben> There the abstract hex values in the GSM 03.38 character encoding, which can be represented into 7 bits, which are then packed into octets such that one octet contains bits from multiple codepoints. [14:21:01.0000] <webben> http://pda.etsi.org/exchangefolder/ts_100900v070200p.pdf is the GSM 03.38 spec. [14:22:00.0000] <webben> and http://www.dreamfabric.com/sms/ is a English-language description of what it does. [14:22:01.0000] <webben> I've been wrestingly with this in increasing confusion all day; figured there's quite a lot of spec-compliance and character encoding knowledge in this channel ;) [14:23:00.0000] <gsnedders> /me would ignore that table in CSS21 :) [14:23:01.0000] <gsnedders> Other things make more sense for GSM 03.38, IMO. [14:23:02.0000] <gsnedders> Or maybe I'm wrong [14:23:03.0000] <gsnedders> /me shrugs [14:23:04.0000] <gsnedders> Don't ask me! [14:23:05.0000] <webben> I've also been wondering: who delivers CSS in GSM 03.38? I can't find any reference to this anywhere else. [14:24:00.0000] <webben> /me is also amused that i18n's test cases for CSS and charset don't actually cover all these random encodings. [14:24:01.0000] <webben> (http://www.w3.org/International/tests/list-html-css#charset) [14:30:00.0000] <Philip`> /me discovers that if you don't have setuptools installed, Trac will silently choose not to attempt to load any plugins [14:48:00.0000] <virtuelv> GSM 03.38 brings back bad memories [16:17:00.0000] <ezyang> Hixie, you update the tree-construction and tokenization parts of the spec too much :-( [16:22:00.0000] <ezyang> In other news, it would be totally awesome if the diffs would actually tell you what section you were, instead of some bogus function. [16:23:00.0000] <gsnedders> Uh, yeah. We discussed this at TPAC. If someone writes the code to do this :P [16:23:01.0000] <ezyang> We could cheeze it by adding  [16:53:00.0000] <gsnedders> ezyang: We're above that :P 2009-06-07 [23:32:00.0000] <ezyang> Huh. The new comment end space stuff has the ability to produce a lot of errors. [23:33:00.0000] <ezyang> Also, switching the state is strictly unnecessary. [23:59:00.0000] <ezyang> Sweet, PHP spec up to space [03:36:00.0000] <zcorpan_> ezyang: did you see the next checkin? http://html5.org/tools/web-apps-tracker?from=3195&to=3196 [03:48:00.0000] <gsnedders> /me has just been using git blame and going through stuff more recent than when it was last updated [05:30:00.0000] <maikmerten> hsivonen, you remember our ActiveX control chat a while ago? Perhaps something can be slammed together using http://wiki.videolan.org/ActiveX and a JS wrapper [05:40:00.0000] <maikmerten> I'm currently trying to evaluate how well this works with a clean WinXP image [05:41:00.0000] <maikmerten> if things happen to take twenty mouse-clicks, three registry hacks and five reboots things may not be user-friendly enough [07:33:00.0000] <maikmerten> hsivonen, okay, it seems IE only allows the installation of signed controls by now (wise move) - that's something VLC can't provide due to associated costs [07:34:00.0000] <maikmerten> hsivonen, but even when lowering security settings in IE (shoot me now, thanks) http://people.xiph.org/~maikmerten/demos/bigbuckbunny-activex.html won't work for me (VLC will install, but I just see solid black where the player should be). This could be some problem with VirtualBox. [07:59:00.0000] <gsnedders> w00t [07:59:01.0000] <gsnedders> /me has US iTunes account, so can finally download Truth in 24 [10:06:00.0000] <ezyang> zcorpan_: Aha [13:31:00.0000] <annevk2> WWDC 2003 Keynote Coverage -- Live [13:31:01.0000] <annevk2> Macworld - ‎1 hour ago‎ [13:31:02.0000] <annevk2> FAIL [15:36:00.0000] <jgraham> /me wonders why <legend> and <caption> don't allow block-level children [15:36:01.0000] <jgraham> Or rather I think it is a bug [15:37:00.0000] <jgraham> s/block-level/flow content/ if you like [15:38:00.0000] <annevk2> Technology Quarterly, yay! [15:39:00.0000] <jgraham> ? [15:52:00.0000] <jgraham> http://james.html5.org/table_descriptions/ [15:53:00.0000] <jgraham> Work in progress, please don't email public-html about it yet [15:53:01.0000] <jgraham> Feedback here or by email welcome (but I won't deal with it until after I sleep) [15:54:00.0000] <jgraham> testcases for different suggestions for marking up table descriptions [15:55:00.0000] <jgraham> ideas taken from the wiki page although I ignored <description> or whatever it's called since that has unrealistic requirements [15:55:01.0000] <jgraham> (several paragraphs of detailed descriptions) [15:56:00.0000] <jgraham> and I ignored <table><details> because that is just silly [15:56:01.0000] <jgraham> Although I can add that if people want to see it [15:56:02.0000] <jgraham> Also I haven't validated the pages or anything so they are probably full of basic errors [15:58:00.0000] <jgraham> Have I already mentioned that <legend> is a rather irritating element to use with <details> and <figure>? It really makes all the examples much worse [15:59:00.0000] <jgraham> Right that is enough soloquiy for now. Otherwise I will have gsnedders writing an essay on me [15:59:01.0000] <jgraham> Which would just be embarassing, frankly [15:59:02.0000] <gsnedders> … [16:02:00.0000] <gsnedders> /me wonders what to say… [16:02:01.0000] <gsnedders> jgraham: When did I ever write an essay on anyone!? [16:05:00.0000] <jgraham> gsnedders: e.g. for your english ah [16:05:01.0000] <gsnedders> That wasn't on anyone. [16:05:02.0000] <jgraham> .me really is going now [16:05:03.0000] <gsnedders> It was on several works of fictio! [16:05:04.0000] <gsnedders> fiction [16:06:00.0000] <jgraham> All written by the same someone [16:06:01.0000] <jgraham> gn [16:06:02.0000] <gsnedders> nn 2009-06-08 [17:30:00.0000] <heycam> anyone know if there's a flattened dtd for mathml somewhere? [17:32:00.0000] <Dashiva> Aren't there programs for creating those? [17:33:00.0000] <heycam> probably. i don't use dtds enough to know though. [17:33:01.0000] <heycam> maybe i can just substitute in for the <!ENTITY> bits and hope it works [22:47:00.0000] <Hixie> Lachy_: where are you these days, geographically? [02:07:00.0000] <hsivonen> It seems to me that when parsing as UTF-16 tentatively, one might as well do it confidently always, because there's no way to get a reasonable reason to revise the guess. [02:07:01.0000] <Hixie> yeah i considered changing the spec to say that instead [02:07:02.0000] <Hixie> i believe it's equivalent to what the spec says now, actually [02:08:00.0000] <Hixie> but i couldn't find a good way to say it -- it turns out i'd have to poke things in way more places than the way the spec does now [02:08:01.0000] <hsivonen> UTF-16 is such a mistake [02:08:02.0000] <hsivonen> is the series of learning experiences that affect a lot of people [02:09:00.0000] <jgraham> hsivonen: Nice explaination of web-scale btw [02:09:01.0000] <hsivonen> jgraham: thanks [02:10:00.0000] <Hixie> not as bad a mistake as UTF-32 and UTF-7! (though they, like UTF-16, both have their place in practice) [02:10:01.0000] <Hixie> if othermaciej was here i'd say he should take the text of http://www.w3.org/mid/4B6CB076-271A-4E9E-AAF8-891C8F4B4A72⊙if and put it in the design principles [02:33:00.0000] <hsivonen> I wish there were some kind of submarine patent craziness 101 that people were required to read before entering a codec debate [02:34:00.0000] <Hixie> and i wish there were no patents, but... [02:34:01.0000] <Hixie> Philip`: yt? [02:35:00.0000] <hsivonen> Hixie: that, too, preferably [02:37:00.0000] <franksalim> how did the patent issues with img play out? [02:37:01.0000] <franksalim> i know about gif [02:37:02.0000] <franksalim> aren't there a lot of image formats that we largely avoid due to patent concerns? [02:37:03.0000] <Hixie> png and jpeg didn't have patent issues that i know of [02:37:04.0000] <hsivonen> franksalim: IIRC, Netscape's lawyers figured that the patents didn't cover decode and Netscape continued shipping decode [02:38:00.0000] <hsivonen> franksalim: IIRC, about boxes in Microsoft software suggest that MS took a license but they also shipped encode in some of their products. [02:38:01.0000] <hsivonen> franksalim: all IIRC and IANAL [02:39:00.0000] <franksalim> hsivonen, pretty much nobody is a lawyer (PMNIAL) [02:39:01.0000] <hsivonen> Hixie: there was the Forgent thing about JPEG, but Forgent went after digital camera manufacturers and not against all kinds of software [02:40:00.0000] <Hixie> dannyb who was posting on whatwg recently is a lawyer, fwiw [02:40:01.0000] <Hixie> expert on software licenses [02:40:02.0000] <hsivonen> franksalim: some software (IIRC, Ghostscript and GIMP) shipped an encoder that didn't compress but produced a bitstream that was readable with an LZW decoder [02:40:03.0000] <jgraham> On an entirely different topic, the requirement that the <caption> be the first child of <table> is non-obvious to the point of uselessness, I guess [02:43:00.0000] <jgraham> /me is sad that Google think that H.264 decoding is more valuable than patent licenses not being required to read web content [02:44:00.0000] <jgraham> At the very least it makes new entrants to the browser market not backed up by some big company less likely [02:44:01.0000] <Hixie> google supports theora [02:44:02.0000] <Hixie> or rather, chrome does [02:44:03.0000] <jgraham> Hixie: Sure [02:45:00.0000] <jgraham> It's not clear if that's enough [02:45:01.0000] <hsivonen> jgraham: I'm eager to see how Opera plays this [02:45:02.0000] <jgraham> hsivonen: me too :) [02:46:00.0000] <jgraham> Actually I'm interested to see what happens with youtube. Because if they start doing HTML5+H.264 then it puts significant pressure on everyone else to but patent licenses [02:47:00.0000] <jgraham> buy [02:47:01.0000] <Hixie> youtube is already doing 264 for flash 10 and iphone and apple tv, iirc [02:47:02.0000] <hsivonen> indeed, YouTube weilds a lot of power so let's hope they follow "Don't be Evil" [02:48:00.0000] <Hixie> /me tries to understand http://www.w3.org/Submission/2008/SUBM-ccREL-20080501/#SECTION00071000000000000000 and fails [02:48:01.0000] <hsivonen> /me tried doing some back-of-the-envelope calculations on how much H.264 is going to cost YouTube when the grace period runs out, but I probably misunderstood things [02:53:00.0000] <annevk2> http://torrentfreak.com/pirate-party-wins-and-enters-the-european-parliament-090607/ is cool [02:54:00.0000] <Hixie> indeed [02:57:00.0000] <hsivonen> /me expects the Ogg controversy Wikipedia page to grow in whatwg list references in due course :-/ [03:00:00.0000] <archtech> Sorry for the OT, but anyone using LaTeX here? Is the syntax for ` and ' turning into quotes getting in the way? [03:00:01.0000] <annevk2> "I'm not so sure. YouTube is very popular despite the fact that its video clips resemble the transmission from the moon landing in 1969." :) [03:01:00.0000] <hsivonen> /me disagrees with howcome's characterization of JPEG2000 [03:02:00.0000] <hsivonen> JPEG2000 is pretty sad as a whole [03:03:00.0000] <hsivonen> but it's nice that JPEG2000 shows how good the old JPEG is after all these years [03:06:00.0000] <hsivonen> JBIG, for all its patent flaws, at least improves more upon CCITT Group 4 than JPEG2000 improves over JPEG [03:08:00.0000] <jgraham> http://james.html5.org/table_descriptions/ has been updated a little. If someone could look over that and tell me if I have done anything silly I will email public-html about it [03:12:00.0000] <hsivonen> http://twitter.com/summary must be enjoying HTML5 tweets [03:31:00.0000] <hsivonen> maikmerten: yeah, the unsigned ActiveX control gets blocked for me, too, in IE8 on XP SP3 on Parallels [03:32:00.0000] <maikmerten> maikmerten, according to some chatting in the videolan irc channel there's a chance VLC will get a signed control soon [03:32:01.0000] <hsivonen> maikmerten: cool [03:32:02.0000] <hsivonen> /me has no idea how much the signing process costs [03:33:00.0000] <Hixie> how on earth do i express <div about="x"><img src="x">...</div> in RDFa without repeating "x" [03:33:01.0000] <Hixie> where ... is a bunch of property/value pairs that hang on the about="x" bit [03:33:02.0000] <hsivonen> maikmerten: Are there releases (deployable binaries with corresponding source drops) for the wikimedia version of cortado? [03:34:00.0000] <hsivonen> I've only found the svn repo and noticed wikimedia's own binary deployment [03:34:01.0000] <maikmerten> hsivonen, vastly outdated ones only. Xiph.org currently is in the process of taking over development http://www.theora.org/cortado/ [03:34:02.0000] <hsivonen> maikmerten: ok [03:34:03.0000] <maikmerten> hsivonen, we hope to have releases available "soon" [03:35:00.0000] <hsivonen> maikmerten: great! [03:35:01.0000] <maikmerten> hsivonen, Xiph.org also provides a signed version of Cortado at http://theora.org/cortado.jar. You can directly embed this applet without having a local copy. [03:35:02.0000] <hsivonen> cool [03:35:03.0000] <maikmerten> that thingie is signed with a Xiph.org certificate, though [03:35:04.0000] <maikmerten> not by a "real" CA [03:36:00.0000] <hsivonen> maikmerten: is it a cost thing or a principle thing? [03:36:01.0000] <maikmerten> hsivonen, definately not a principle thing [03:36:02.0000] <maikmerten> hsivonen, it may be the cost or the effort to get through a proper CA [03:38:00.0000] <hsivonen> maikmerten: donated. (btw, xiph would have benefited from making the amount of donation editable :-) [03:39:00.0000] <maikmerten> (on a side note: I did see the latest wave of codec/patent/FUD/not-FUD discussion on the list but currently think I should only join into the fun if I have something constructive to say.) [03:39:01.0000] <maikmerten> hsivonen, thanks! [03:43:00.0000] <maikmerten> (I still remember embarassing myself by doing way-over-the-top stupid posts discussing with Apple representatives. Happens in a heated discussion, I guess, but doesn't help the cause) [03:44:00.0000] <maikmerten> I forwared your donation feedback to xiphmont and also asked for more status on Xiph.org's CA status [03:49:00.0000] <maikmerten> hsivonen, can you elaborate on your donation-amount problem? xiphmont appears surprised - there should be a way to adjust the amount. [03:52:00.0000] <maikmerten> hsivonen, ah, apparently a "theora donation" is a fixed item at $10.00 - now, that's suboptimal [03:52:01.0000] <maikmerten> (when clicking the donta-button on theora.org) [03:52:02.0000] <maikmerten> donate [03:53:00.0000] <maikmerten> when clicking the button on xiph.org/donate/ I get to homepage-thingie on paypal... hmmmm.... broken? [04:03:00.0000] <annevk2> yeah, I got the same on xiph.org [04:03:01.0000] <annevk2> it asked me to login but I don't have a PayPal account and normally you can pay without having an account [04:19:00.0000] <Hixie> ok bed time [04:19:01.0000] <Hixie> nn [04:20:00.0000] <annevk2> g'night! [04:24:00.0000] <hsivonen> maik|lunch: yeah, the issue was the when clicking the donation button on the cortado page, the amount was fixed to $10 [04:29:00.0000] <zcorpan> Hixie: NETWORK_NO_SOURCE is inconsistently marked up (sometimes with <code> sometimes not) [04:46:00.0000] <annevk2> the ways in which some people use "IMHO" make it a completely useless initialism in practice :/ [04:54:00.0000] <annevk2> wow, Xbox Project Natal looks pretty cool [04:56:00.0000] <Rik|work> annevk2: but pretty vaporware for the moment [05:05:00.0000] <annevk2> yeah [05:06:00.0000] <jgraham> hsivonen: It is worth noting that many users don't change default settings even if they could improve their user experience by doing so [05:06:01.0000] <jgraham> Even expert users [05:06:02.0000] <hsivonen> jgraham: which way do the defaults on summary reading go? [05:07:00.0000] <jgraham> hsivonen: I think "on" but I'm not sure [05:09:00.0000] <hsivonen> it would be helpful if screen reader users and vendors (if they've done usability studies) participated in these threads [05:34:00.0000] <maikmerten> hsivonen, donations should be fixed once the webpage mirrors catch up [05:34:01.0000] <maikmerten> hsivonen, thanks for the catch [05:34:02.0000] <maikmerten> (and for the donation, of course) [05:35:00.0000] <hsivonen> maikmerten: OK. thanks. I'll make another attempt later [05:47:00.0000] <jgraham> /me wonders if he should explicitly point out that the <details> proposal has more or less no new conformance requirements relative to <details> [05:47:01.0000] <annevk2> it seems a bit complicated to me [05:48:00.0000] <jgraham> annevk2: It is the best I can manage that takes account of the positions of both "sides" [05:49:00.0000] <jgraham> But if yoou can think of something less complex... [05:49:01.0000] <annevk2> http://www.zeldman.com/2009/06/08/not-safe-for-work-tag-in-html-5/ lol [05:50:00.0000] <annevk2> it's all over twitter too [05:50:01.0000] <Dashiva> I can't wait for the "We must allow users to encode which kind of work it's not safe for" talk to start [05:51:00.0000] <jgraham> (well "just use <caption>" is less complex but doesn't account for the argument that @summary is needed because authors don't want to burden all users with information that only a small constituency require) [05:52:00.0000] <zcorpan> <details><legend>Not safe for work</legend><img src=badkitten.jpg></details> [05:52:01.0000] <Dashiva> That tiny minority that supply summary info, and that tiny minority of those who actually supply non-universal summary info, wouldn't they be competent enough to use a hidden block of text linked with ARIA? [05:53:00.0000] <jgraham> Dashiva: They're not competent enough to use @summary [05:54:00.0000] <Dashiva> Then what's the problem? [05:54:01.0000] <jgraham> I never understand Zeldman when he starts talking in the first person plural. Does he have multiple personality disorder or does he think he's the Queen? [05:54:02.0000] <Dashiva> I hear kings do it too [05:57:00.0000] <jgraham> (also if <caption> allowed flow content children, being able to use <details> would just fall out naturally so there would be no problem except documenting the best-practice) [06:12:00.0000] <jgraham> Things I hate about javascript number 0156: The difference between undefined meaning "no property" and undefined meaning "a property with the value undefined" [06:14:00.0000] <Dashiva> Meet the in operator [06:15:00.0000] <annevk2> I want to post a short follow-up post on the video thing [06:15:01.0000] <annevk2> Chrome can distribute FFmpeg because its patent license that it may or may not have covers the Chrome product and not the library, right? [06:15:02.0000] <jgraham> Dashiva: I know /how/ to tell. But it is still stupid and unintuitive (this is based on debugging code where someone else got this wrong) [06:17:00.0000] <Dashiva> jgraham: Wouldn't you say that code treating the two differently is a bug, though? [06:18:00.0000] <jgraham> Dashiva: In that case the ECMAScript spec is really buggy [06:18:01.0000] <jgraham> (see e.g. Array.prototype.sort) [06:21:00.0000] <Dashiva> The part about undefined values coming before missing indexes? [06:21:01.0000] <jgraham> Yes [06:21:02.0000] <Dashiva> That's invisible to code that doesn't separate the two, though [06:21:03.0000] <jgraham> In general array methods behave rather differently with undefined and missing values [06:22:00.0000] <jgraham> (this is not the only area of the spec that is differnt of course) [07:14:00.0000] <hsivonen> oh. the chrome/ffmpeg/lgpl thing hit slashdot yesterday [08:22:00.0000] <hsivonen> hmm. If nearly all non-calendar tables on Philip's list are layout tables and JAWS detects them as layout tables, I could understand how users hear none of the whole pile Philip had and instead hear summaries only on .gov sites [08:23:00.0000] <hsivonen> that seems like a huge waste ratio :-/ [08:24:00.0000] <hsivonen> Hixie: maybe you should put a JAWS-compatible summary suppression algorithm in the spec [08:26:00.0000] <jgraham> hsivonen: AFAIK we haven't got as far as reverse engineering how JAWS (or others) detect layout tables [08:26:01.0000] <jgraham> Nor is there any evidence that they have reached a maximum of efficiency to the point at which the algorithm should be standardised [08:27:00.0000] <jgraham> well, that didn't quite make sense but you see what I mean, maybe [08:27:01.0000] <hsivonen> ok, but today's back-and-forth may actually have teased out a reasonable explanation that reconciles the data with the claims that users aren't burdened with accessispam [08:28:00.0000] <jgraham> That JAWS and others have layout-table-detecting algorithms that cause @summary to be supressed? I thought we already knew that [08:29:00.0000] <jgraham> We don't know what the efficiency of those algorithms is [08:29:01.0000] <hsivonen> what I didn't realize before is how predominantly the summaries are used on layout tables outside .gov [08:30:00.0000] <hsivonen> what does it say about the success of an accessibility feature that the leading screen reader has code to suppress it? [08:31:00.0000] <jgraham> (there was a machine-learning based paper that claimed something like 95% although I can't remember quite what that referred to) [08:32:00.0000] <jgraham> (clearly it can't be "right in 95% of all cases because you cn do rather better than that just by saying all tables are layout tables) [08:33:00.0000] <jgraham> s/s/s"/ [08:33:01.0000] <hsivonen> anyway, I can buy the argument that summary should be 'in', because .gov sites may manage to use it right and that abuse by others is suppressed by algorithms [08:34:00.0000] <hsivonen> but then why does the .gov guide recommend <caption> [08:34:01.0000] <hsivonen> and can we have @summary for .gov without having all these other people write summaries only to be suppressed? [08:36:00.0000] <hsivonen> I wonder what kind of effect would a validator that whined about non-empty summary on tables that JAWS would consider layout tables would have [08:36:01.0000] <hsivonen> minus one 'would' [08:38:00.0000] <jgraham> hsivonen: It is worth noting that the data was not so conclusive about .gov managing to use @summary right [08:38:01.0000] <hsivonen> jgraham: that, too [08:39:00.0000] <jgraham> I don't think we should be keeping heavily abused features just because some small subset of people might get them right sometimes and the bad effects of people getting them wrong can be mitigated by significant client side effort [08:40:00.0000] <jgraham> Not least because it makes developing clients significantly harder; to compete with JAWS you have to reverse engineer its table heuristics [08:40:01.0000] <jgraham> or build better ones [08:40:02.0000] <jgraham> or offer a worse user experience [08:40:03.0000] <hsivonen> jgraham: unless HTML5 saves prospective developers the cost by having a canned explanation of what JAWS does [08:41:00.0000] <jgraham> hsivonen: Since FS have no interest in giving that to us we would have to reverse engineer it. But that sort of thing should not be codified in the spec [08:41:01.0000] <jgraham> It would be an intersting and useful excersie though [08:42:00.0000] <jgraham> although if they have used e.g. machine learning the algorithm could be rather complex to pin down [08:42:01.0000] <Dashiva> "Structure is nothing if it is all you got. Skeletons spook people if they try to walk around on their own. I really wonder why XML does not." [08:54:00.0000] <webben> jgraham: "should not be codified in the spec" ... why not? [08:54:01.0000] <webben> Seems like an important area of interoperable consumption of the web corpus. [09:02:00.0000] <jgraham> webben: Because it seems like a UI issue [09:02:01.0000] <jgraham> (sort of) [09:05:00.0000] <jgraham> and because making a dejure method of categorising tables into layout and non-layout unnecessarily freezes the state of the art [09:05:01.0000] <hsivonen> jgraham: I disagree. I think it's a semantic issue for answering "what's the summary if any of this table" [09:07:00.0000] <webben> jgraham: I don't think it's a UI issue. If one wants to allow conforming UAs to implement even better algorithms, it could always be a suggested algorithm. [09:08:00.0000] <webben> that lowers the bar to creating a UA, while still allowing evolution of better heuristics. [09:15:00.0000] <jgraham> I am not opposed to an informative algorithm, perhaps in a seperate document. I would be very worried about defining semantic issues based on things that have a certian known failure rate and no manual override [10:39:00.0000] <annevk42> gsnedders, "all other specs"? mwaha [10:39:01.0000] <gsnedders> :P [10:39:02.0000] <gsnedders> all the other specs, I meant [10:39:03.0000] <gsnedders> And I said that! [10:40:00.0000] <gsnedders> There is a difference between "all other" and "all the other" :P [10:41:00.0000] <annevk42> still, CORS, Selectors API, various CSS specs, ... [10:41:01.0000] <gsnedders> the is the definite article, hence it doesn't apply to absolutely all. [10:42:00.0000] <annevk42> maybe I'm being dense, but it's not clear to me how "the" includes XMLHttpRequest, but not those [10:42:01.0000] <annevk42> but it also does not really matter and I have to go :) [10:42:02.0000] <jgraham> /me onders what is being discussed [10:42:03.0000] <jgraham> wonders [10:42:04.0000] <gsnedders> Nuances of English :) [10:43:00.0000] <jgraham> damn double-u [10:43:01.0000] <jgraham> if it was spelt uuonders I could get it right [10:45:00.0000] <gsnedders> jgraham: See -archive, though [11:24:00.0000] <gsnedders> Does anyway have a test suite for XML serializers, apart from giving Philip` access to it? :P [12:18:00.0000] <hsivonen> scary. google's page optimization docs mention netscape 4.04 [12:36:00.0000] <gsnedders> Uh, xml-names doesn't define what a valid value for an xmlns actually is, apart from saying it must be a URI reference (whatever that is) or an empty string. [12:37:00.0000] <ezyang> I'm pretty sure anything for xmlns is fair game. [12:37:01.0000] <ezyang> I've used urns for xmlns before. [12:37:02.0000] <gsnedders> Yes, but what's a URI? [12:38:00.0000] <ezyang> Uniform Resource Identifier [12:38:01.0000] <gsnedders> What RFC3986 says? What RFC 2396 says? [12:38:02.0000] <ezyang> See RFC 3986 [12:38:03.0000] <ezyang> 3986 supercedes 2396 [12:39:00.0000] <gsnedders> /me finally finds a reference in the spec saying what a URI is. [12:39:01.0000] <gsnedders> (I mean what the XML namespaces spec means by "URI", I don't care about what RFC obsoletes what :P) [12:39:02.0000] <ezyang> jaja [12:40:00.0000] <gsnedders> (I mean, if I wanted to throw fatal errors on invalid namespace names, then it'd be better that it's behaviour was constant and didn't depend on what RFC I chose to implement, as there should hopefully be a static reference in the spec.) [12:41:00.0000] <gsnedders> Oh, more fun: xml-names first ed. references RFC 2396 and second edition RFC 3986 [12:41:01.0000] <gsnedders> /me wonders which to be compatible with [12:41:02.0000] <ezyang> That makes sense [12:41:03.0000] <ezyang> 3986 is mostly a strict superset of 2396, so you should use that [12:41:04.0000] <gsnedders> The problem is it means a document using namespaces may throw parse errors if it implements the first edition. [12:42:00.0000] <ezyang> Unlikely. [12:42:01.0000] <gsnedders> But possible. Bear in mind the big issue of XML 1.0 Fifth edition. [12:43:00.0000] <gsnedders> (i.e., a document valid under that can be not even well-formed according to another XML 1.0 processor, because it implements an earlier edition) [12:43:01.0000] <gsnedders> /me wonders how to implement URI validation [12:43:02.0000] <ezyang> I'm pretty sure they only sanity check URIs, and don't do an in-depth check. [12:44:00.0000] <gsnedders> Processors normally do not checking :) [12:44:01.0000] <gsnedders> For a serializer, it ought to be impossible to output anything that isn't conforming, though [12:44:02.0000] <ezyang> btw, you should read "D.2. Modifications" [12:45:00.0000] <gsnedders> /me thought from memory there were actual changes [12:47:00.0000] <gsnedders> Is it true that any string that matches unreserved / reserved _and_ where any "%" are followed by 2HEXDIG will match URI-reference? [12:47:01.0000] <jgraham> gsnedders: I feel you are being over-analytic somewhere [12:48:00.0000] <gsnedders> That's untrue. [12:48:01.0000] <gsnedders> /me wonders if he can do this without implementing a full URI parser [12:49:00.0000] <jgraham> gsnedders: Really, worrying that a document may implement(?) an absolete RFC is over-analysing something [12:49:01.0000] <gsnedders> :P [12:49:02.0000] <gsnedders> OK, but now I've moved on from that! [12:54:00.0000] <hsivonen> gsnedders: serializers should target 4th ed for compat [12:54:01.0000] <gsnedders> hsivonen: Indeed. [12:54:02.0000] <gsnedders> hsivonen: And what edition of xml-names? [12:54:03.0000] <gsnedders> (Or does that not matter?) [12:55:00.0000] <hsivonen> gsnedders: 1.0 [12:55:01.0000] <hsivonen> edition, dunno [12:55:02.0000] <gsnedders> hsivonen: first or second edition, though? [12:56:00.0000] <gsnedders> /me sighs [12:57:00.0000] <gsnedders> I'm subclassing PHP's built-in XMLWriter to fix the fact it can output non well-formed XML (this is a "bogus" bug). This is gonna be a big subclass. :( [12:58:00.0000] <ezyang> Oh, if this is PHP, HTML Purifier has RFC-compliant code for parsing URIs [12:58:01.0000] <gsnedders> LGPL though [12:58:02.0000] <ezyang> HTML Purifier is LGPL? [12:59:00.0000] <gsnedders> Is it not? [12:59:01.0000] <ezyang> Yep, it is. What do you need to license it as? [13:00:00.0000] <gsnedders> Ideally MIT [13:01:00.0000] <gsnedders> But it needs to at least be includable within an Apache product. [13:02:00.0000] <gsnedders> Also: URI parsing isn't the hard part. Validation is. [13:02:01.0000] <ezyang> Sure. [13:02:02.0000] <ezyang> But I thought you just wanted to freak out when the URI wasn't compliant [13:04:00.0000] <gsnedders> Yeah. [13:04:01.0000] <gsnedders> But that means validating it :P [13:04:02.0000] <gsnedders> (Or using a needlessly complex parser) [13:10:00.0000] <myakura> /me is installing Safari 4 [13:39:00.0000] <jwalden> man, this chrome/ffmpeg thread is like the gift that keeps on giving [13:50:00.0000] <jgraham> jwalden: You get odd gifts [13:57:00.0000] <jwalden> what can I say, my needs are few :-) [14:22:00.0000] <gsnedders> The XMLWriter extension really is useless :( [14:25:00.0000] <sayrer> gsnedders, is that php? [14:25:01.0000] <gsnedders> sayrer: yeah [14:26:00.0000] <gsnedders> (And yes, I know PHP is all useless :D) [14:26:01.0000] <sayrer> oh, it is what it is :) [14:26:02.0000] <sayrer> gsnedders, is there a genx wrapper? that's always been pretty good [14:26:03.0000] <gsnedders> /me wonders how much he would lose by making BetterXMLWriter not a subclass XMLWriter [14:27:00.0000] <sayrer> after all, any bozo can generate well-formed xml [14:27:01.0000] <gsnedders> I mean, I re-implement so much… [14:27:02.0000] <gsnedders> sayrer: That would have to be a non-standard extension, which makes it virtually impossible to rely upon for shippable code [14:27:03.0000] <sayrer> oh, I see. I didn't realize there were "standard extensions" [14:28:00.0000] <gsnedders> Oh, sure. More or less anything useful is in an extension. And, naturally, it can be disabled. [14:28:01.0000] <sayrer> what are you writing, btw? [14:29:00.0000] <gsnedders> A subclass of XMLWriter that can be relied upon to always output XML. [14:29:01.0000] <sayrer> XMLForSeriousWriter? [14:29:02.0000] <gsnedders> BetterXMLWriter. [14:29:03.0000] <sayrer> ActuallyXMLWriter? [14:30:00.0000] <sayrer> I have been using Venus for some stuff lately [14:30:01.0000] <gsnedders> /me finds startPI actually works! [14:30:02.0000] <gsnedders> Like, actually does enough sanitization to be usable. [14:30:03.0000] <sayrer> I have to say... it's the first time I've actually derived value from an XML publishing pipeline [14:33:00.0000] <gsnedders> Awesome. U+FFFF is allowed as an element name. [14:33:01.0000] <gsnedders> And when you use U+FFFF as a namespace URI you get ï¿¿ [14:33:02.0000] <gsnedders> (EF BF BF is U+FFFF in UTF-8) [14:34:00.0000] <sayrer> There’s just no nice way to say this: Anyone who can’t make a syndication feed that’s well-formed XML is an incompetent fool [14:34:01.0000] <sayrer> quote unquote [14:34:02.0000] <sayrer> :) [14:34:03.0000] <gsnedders> /me gets the quote :) [14:35:00.0000] <gsnedders> I mean, it's hardly as if WP can be made to produce a feed that isn't XML well-formed… [14:36:00.0000] <sayrer> I that is my second favorite XML quote ever [14:36:01.0000] <gsnedders> (This is, of course, a bug that has been open for 16 months, but has existed since before it was called WP :P) [14:36:02.0000] <gsnedders> sayrer: What's your favourite? [14:36:03.0000] <jgraham> I asume we are expected to ask what your favourite XML quote is [14:36:04.0000] <jgraham> Oh gsnedders already said that [14:36:05.0000] <gsnedders> jgraham: You're slow. [14:37:00.0000] <jgraham> Well that's no surprise to anyone [14:37:01.0000] <sayrer> No doubt I can write a routine to parse this, but look at how deep they went to re-invent, XML itself wasn't good enough for them, for some reason (I'd love to hear the reason). Who did this travesty? Let's find a tree and string them up. Now. [14:39:00.0000] <gsnedders> Ah, Dave Winer. That makes the quote seem quite mild and calm. [14:39:01.0000] <gsnedders> /me wonders what the hell XMLWriter is doing here [14:39:02.0000] <gsnedders> XMLWriter::writePI('a', "\xEF\xBF\xBF") [14:39:03.0000] <gsnedders> <?a Ôø\BF?> [14:40:00.0000] <gsnedders> Ô is U+00D4 and ø is U+00F8 [14:41:00.0000] <jgraham> gsnedders: You do realise that using a language without proper unicode support in 2009 is just silly, right? [14:41:01.0000] <gsnedders> /me points to the first point of his rant on PHP's problems [14:41:02.0000] <jgraham> (or at least the ability to invent unicode support in a reasonably performant manner) [14:42:00.0000] <gsnedders> /me watches jgraham get confused by gsnedders pointing at an unpublished draft :P [14:42:01.0000] <jgraham> gsnedders: Don't rant, boycott :) [14:43:00.0000] <gsnedders> jgraham: If you want the state of XML to get better on the web you need to get it working in PHP. Sad reality. [14:44:00.0000] <jgraham> gsnedders: I'm not that concerned with the state of XML on the web [14:44:01.0000] <gsnedders> /me grumbles, finding more ways to break XMLWriter [14:44:02.0000] <jgraham> It hasn't worked very well so far, it will continue to not work very well in the future [14:44:03.0000] <sayrer> it might be more effective to get genx in the php distribution [14:45:00.0000] <gsnedders> sayrer: Should I point out WP supports versions of PHP shipped over six years ago? [14:46:00.0000] <sayrer> first you fix the ones actually hosted on wordpress.com [14:46:01.0000] <sayrer> then you get the ones that actually update for bug fixes and perf [14:46:02.0000] <sayrer> and the others die out [14:46:03.0000] <sayrer> it will take a while, I agree [14:49:00.0000] <gsnedders> Even getting WP to use a broken XML serializer would probably help [14:58:00.0000] <jgraham> gsnedders: Did you implement the comment-end-bang stuff in php-html5lib? [14:59:00.0000] <ezyang> I did. [14:59:01.0000] <gsnedders> jgraham: No. [15:01:00.0000] <jgraham> ezyang: Yeh I just noticed on hg log :) So, where do the two parse errors for  come from? [15:02:00.0000] <ezyang> first parse error is from !, second parse error is from a [15:02:01.0000] <jgraham> It seems like the tests assume a in comment end bang state is a parse error but that's not what the spec says [15:02:02.0000] <ezyang> Unless Hixie, like, changed it. [15:02:03.0000] <ezyang> Oh? [15:03:00.0000] <ezyang> "Parse error. Switch to the comment end bang state." [15:03:01.0000] <jgraham> Right [15:03:02.0000] <ezyang> (the parse error results when you transition into the state, not in the actual state) [15:03:03.0000] <jgraham> then why does the a give a second parse error? [15:04:00.0000] <jgraham> s/a/"a"/ [15:04:01.0000] <ezyang> oh ho, that's not right [15:05:00.0000] <ezyang> I can fix that, or you can? [15:05:01.0000] <ezyang> (I have the html5lib-php fix, anyway) [15:05:02.0000] <jgraham> You can if you like [15:07:00.0000] <sayrer> olliej: congrats on Safari 4 (I hate the codecs, but I know you guys don't have much to do with that choice) [15:08:00.0000] <olliej> sayrer: we had those in S3.1 :D [15:08:01.0000] <annevk42> http://twitter.com/daringfireball/status/2078539734 [15:08:02.0000] <sayrer> it's much nicer than the beta [15:08:03.0000] <olliej> sayrer: but omg i wih there was some sane codec solution [15:08:04.0000] <olliej> sayrer: or failing that [15:08:05.0000] <olliej> sayrer: people would stop arguing about chrome specifics in the whatwg list [15:08:06.0000] <ezyang> Pushed. [15:08:07.0000] <sayrer> well, I think the argument is healthy [15:08:08.0000] <sayrer> easy to skip [15:10:00.0000] <annevk42> just filter on chrome + ffmpeg if you don't care about it [15:10:01.0000] <jgraham> ezyang: Thanks [15:11:00.0000] <jgraham> olliej: We could get people to diss apple too, if you like :) [15:12:00.0000] <olliej> jgraham: they already do [15:12:01.0000] <ezyang> jgraham: I think it would be generally useful for people do adopt the SPEC convention [15:12:02.0000] <olliej> jgraham: and then they claim that chrome has actually implemented stuff [15:13:00.0000] <sayrer> olliej? [15:14:00.0000] <jgraham> /me wonders what a SPEC convention is [15:14:01.0000] <olliej> sayrer: oh google acting like (and convincing people that) they are responsible for significant amounts of the browser feature set [15:14:02.0000] <sayrer> oh I see [15:14:03.0000] <olliej> sayrer: which i also take to be bitching about apple :D [15:14:04.0000] <ezyang> jgraham: Oh yeah... [15:15:00.0000] <sayrer> don't you think that's how open source goes? [15:15:01.0000] <olliej> sayrer: what, bitching about apple? [15:15:02.0000] <sayrer> no, claiming you have a feature [15:15:03.0000] <sayrer> since you just import an open source library [15:15:04.0000] <olliej> sayrer: ah right [15:15:05.0000] <sayrer> sort of like we claim we have ogg [15:15:06.0000] <ezyang> jgraham: It's putting a SPEC file in the root of your project directory, and setting it's contents to be 1234, where that's the last revision of HTML5 that you've checked the library against. [15:15:07.0000] <sayrer> but we didn't write the ogg libraries [15:16:00.0000] <jgraham> ezyang: Tha would be hard to do since I will often miss things that aren't covered by tests [15:16:01.0000] <olliej> sayrer: but yes, there's a difference between claiming support for ogg, and claiming to have added support for video/audio/appcache/etc when all you did was pickup existing or enabled from a standard trunk build [15:16:02.0000] <sayrer> what is the difference? [15:17:00.0000] <jgraham> Although it would be helpful to know what the last revision I looked at was [15:17:01.0000] <sayrer> like, what could they say that would be ok? [15:17:02.0000] <olliej> sayrer: eg. compare "Firefox now supports ogg" to "Flock has now added support for html5" [15:17:03.0000] <sayrer> hmm. I can understand that they want people to know <audio> will work in Chrome [15:17:04.0000] <olliej> (Flock is that geco/firefox based social browser?" [15:17:05.0000] <sayrer> how do they say that the right way? [15:18:00.0000] <jgraham> ezyang: <a a=a<> should have a parse error, right? (< in attribute value, unquoted, state) [15:18:01.0000] <olliej> sayrer: enabled would be preferable -- i would prefer "we have now updated to a new version of webkit that provides ..." [15:18:02.0000] <ezyang> jgraham: That's fine; but you really don't want to be re-scanning the spec every time something changes [15:18:03.0000] <olliej> sayrer: but anyway, this isn't relevant to the channel [15:19:00.0000] <jgraham> olliej: See /topic [15:19:01.0000] <olliej> sayrer: this started with me just wishing people would continue arguing about whether google is legally allowed to do what they're doing [15:19:02.0000] <sayrer> this channel is pretty wide open :) [15:19:03.0000] <ezyang> jgraham: Umm, that's a new test, I think it needs a parse error [15:19:04.0000] <sayrer> or so I hear [15:19:05.0000] <olliej> errr [15:19:06.0000] <olliej> s/continue/stop [15:19:07.0000] <ezyang> Whatever it says there is what I think it should be, since I wrot ethe test [15:19:08.0000] <olliej> subtle difference there :D [15:19:09.0000] <ezyang> wrote the [15:19:10.0000] <sayrer> haha [15:20:00.0000] <olliej> sayrer: i have JSON.stringify implemented now [15:20:01.0000] <olliej> sayrer: not yet landed (Waiting for review) [15:20:02.0000] <sayrer> cool... mostly the same as mine? [15:20:03.0000] <jgraham> ezyang: I think there should be a parse error but the test disagrees [15:20:04.0000] <sayrer> I know I have a few deviations [15:20:05.0000] <sayrer> stupid spec changes every meeting [15:20:06.0000] <jgraham> I guess ECMA don't believe in testsuites? [15:21:00.0000] <sayrer> actually [15:21:01.0000] <sayrer> Microsoft donated the beginnings of one [15:21:02.0000] <sayrer> at the last meeting [15:21:03.0000] <jgraham> Awesome [15:21:04.0000] <sayrer> host on their codeplex site [15:21:05.0000] <sayrer> (like google code or sourceforge) [15:21:06.0000] <sayrer> it's dual licensed BSD / MSPL [15:21:07.0000] <jgraham> But there is no interoperable implementations requirement like with W3C? [15:22:00.0000] <sayrer> there is an agreement in the group [15:22:01.0000] <jgraham> sayrer: Do you have a pointer? [15:22:02.0000] <olliej> sayrer: mostly the spec/json2.js behaviour [15:22:03.0000] <ezyang> Ok, spec agrees with what you say. Checking code. [15:22:04.0000] <sayrer> olliej: I'm sure you'll find json2.js changes over time with no changelog [15:22:05.0000] <sayrer> I certainly have :) [15:23:00.0000] <olliej> sayrer: hehe [15:23:01.0000] <olliej> sayrer: i've tried to go for "sane" [15:23:02.0000] <Philip`> Hixie: I'm not here (and will be similarly not here for the next week, so if it can't wait then now is as good a time as any) [15:23:03.0000] <olliej> sayrer: in most cases the issues i pointed out should never matter [15:23:04.0000] <sayrer> yeah, I was pretty happy about hat [15:23:05.0000] <sayrer> that [15:23:06.0000] <sayrer> clearly edge cases [15:23:07.0000] <olliej> sayrer: yeah [15:23:08.0000] <olliej> sayrer: i was happy with that :D [15:24:00.0000] <ezyang> I think the test is right. [15:24:01.0000] <sayrer> jgraham: I think there is a message on the es-discuss list [15:24:02.0000] <olliej> jgraham: both mozilla and webkit have fairly substantial js test suites as well [15:24:03.0000] <ezyang> Hu, no. [15:24:04.0000] <sayrer> olliej: I think we're going to try and upstream to microsoft's for es5 [15:25:00.0000] <sayrer> and have hg pull it in [15:25:01.0000] <olliej> sayrer: righto [15:25:02.0000] <sayrer> I'm going to donate my JSON tests at least [15:25:03.0000] <jgraham> olliej: I am aware of the mozilla testsuite. It is nice but not always a test of the spec :) [15:26:00.0000] <olliej> sayrer: cool [15:26:01.0000] <sayrer> there are a few that test mozilla-specific extensions like scripted iterators that I will have to remove [15:26:02.0000] <ezyang> Oh hey, the test fails for my impl too [15:26:03.0000] <ezyang> Yep, that should have a parse error [15:26:04.0000] <ezyang> Wow, how'd I miss that? [15:26:05.0000] <ezyang> Can you fix that? [15:27:00.0000] <annevk42> hmm, nothing in the keynote was not predicted by daringfireball [15:27:01.0000] <annevk42> no fun [15:27:02.0000] <jgraham> ezyang: Sure [15:43:00.0000] <Philip`> sayrer: I guess you mean genx as in http://lists.w3.org/Archives/Public/www-archive/2009Mar/0060.html ? :-) [15:44:00.0000] <sayrer> Philip`: class [15:44:01.0000] <sayrer> who did this XML travesty? It's not even JSON! 2009-06-09 [18:18:00.0000] <jwalden> hm, is aside the correct element to use to mark up parentheticals, even if those parentheticals should be |display: inline|? [18:22:00.0000] <Hixie> jwalden: no, use Unicode characters U+0028 and U+0029 to mark up parentheticals. [18:46:00.0000] <Hixie> anyone got any opinions on bruce's e-mail regarding whether to use <figure> or <aside> for pullquotes? [18:46:01.0000] <Hixie> i guess <aside> is better [19:18:00.0000] <jwalden> Hixie: thanks; might be worth adding an informative sentence to <aside> mentioning that anti-use case [19:19:00.0000] <Hixie> added a comment, dunno if it'll confuse more than it helps, though [19:21:00.0000] <jwalden> /me looks [19:23:00.0000] <jwalden> /me waits for http://html5.org/tools/web-apps-tracker (presumably) to pick up the change [00:05:00.0000] <jwalden> /me reads the parts of the ffmpeg/chrome thread he hadn't read yet, merely seen as having happened, and is bemused at how quickly things went south when some people chose not to shut up [00:06:00.0000] <jwalden> frankly, I'm amazed it took so long to catch flame :-D [00:28:00.0000] <hsivonen> so predictable: http://twitter.com/jdowdell/status/2085821360 [01:08:00.0000] <jgraham> HTML5 fully speciying something and being silent on whether it is conforming or not seems difficult [01:09:00.0000] <jgraham> Unless HTML5 made no claims about document conformance at all, I suppose [01:09:01.0000] <jgraham> But that seems... unhelpful [02:02:00.0000] <annevk42> "To call out just one change, note that Snow Leopard now defaults to the same display gamma as Windows — 2.2 instead of 1.8." does this mean PNG now works the same everywhere? [02:43:00.0000] <annevk42> <di> ftw [02:46:00.0000] <jgraham> annevk42: I don't think so [02:46:01.0000] <jgraham> (the png thing) [02:48:00.0000] <jgraham> (at least I think that change alone doesn't solve the problem that hsivonen always brings up which is that png colours may not match CSS colours in browsers that have support for PNG colour profiles) [02:48:01.0000] <annevk42> http://hsivonen.iki.fi/png-gamma/ suggests it is about gamma [02:50:00.0000] <annevk42> it seems it was not much of a problem anyway anymore [02:51:00.0000] <jgraham> annevk42: I don't understand this much but there is also the colour space issue [02:51:01.0000] <jgraham> Oh wait, I am thinking about JPEGs although the same applies to PNG afaik [02:54:00.0000] <hsivonen> when browsers do ICC stuff, you can now enjoy the old png-gamma problem with JPEGs, too! [02:55:00.0000] <hsivonen> annevk42: the main old problem was that some old browsers didn't make PNG color "management" opt-in [02:55:01.0000] <jgraham> hsivonen: For JPEGs it makes much more sense since fewer people are using JPEGs as part of their site design [02:56:00.0000] <hsivonen> w00t! finally gamma sanity at Apple with the 2.2 default. [02:56:01.0000] <hsivonen> time to party. at last! [02:56:02.0000] <hsivonen> /me reads logs backwards [02:57:00.0000] <hsivonen> existing content and degrading gracefully FTW! [02:57:01.0000] <hsivonen> Mac-created content degrading gracefully on Windows, that is [02:57:02.0000] <hsivonen> and existing Windows-created content working on Mac [03:16:00.0000] <Lachy> hsivonen, PNG Gamma article of yours seems a little out of date. It's mentions browsers like Safari 2.0, IE5 Mac and Opera 7 as if they were still relevant [03:16:01.0000] <Lachy> if you have time one day, it would be nice if you could update it to discuss the current situation [03:17:00.0000] <hsivonen> Lachy: yeah, I should probably update it to say that the article discusses a problem that has now pretty much gone away as a browser interop problem but is now relevant as a photoshop/browser problem [03:17:01.0000] <hsivonen> /me has a lot of outstanding feedback on the article :-( [03:18:00.0000] <Lachy> yeah, I always had trouble getting photoshop to export PNGs in ways that played nicely with browsers [03:19:00.0000] <hsivonen> the great thing about OS X is that old releases fade into irrelevance much faster than old Windows releases [03:19:01.0000] <Lachy> I had much more luck with Fireworks, which somehow seems to get it right [03:19:02.0000] <hsivonen> so pre-Tiger Safari doesn't matter anymore [03:20:00.0000] <Lachy> so, is the best way to get PNGs with consistent colour to omit the colour profile information, and rely on current browsers matching the CSS colours? [03:21:00.0000] <hsivonen> Lachy: yeah. with unlabeled color for everything, CSS, bitmaps and plug-ins match [03:21:01.0000] <Lachy> good [03:22:00.0000] <Lachy> that would have been useful to know when I was doing web development a couple of years ago. [03:22:01.0000] <hsivonen> Lachy: you should have googled for my article! [03:23:00.0000] <Lachy> I did read it a long time ago, but didn't really understand it too well [03:23:01.0000] <Lachy> I just knew if I exported PNGs with Fireworks, they were right, and if I used Photoshop, they'd be wrong in some browser/platform combinations [03:24:00.0000] <beowulf> you can use software to remove the gamma black from pngs, that's what i do anyway [03:25:00.0000] <Lachy> I still don't really understand the concept of colour profiles too well. They seem like a hack to make poor quality hardware give good colour, instead of getting hardware vendors to produce better quality colour [03:25:01.0000] <hsivonen> most of the feedback I haven't addressed yet revolves around endorsing various removal tools [03:26:00.0000] <Lachy> the pngcrush command the article mentions seems good enough [03:26:01.0000] <Lachy> although, I haven't used it [03:26:02.0000] <jgraham> Lachy: define "good colour" [03:27:00.0000] <beowulf> i use pngthing, but i can't imagine it's any better worse than something else that does the same thing, considering [03:27:01.0000] <Lachy> jgraham, good colour = colour that looks like what it's supposed to look like [03:29:00.0000] <jgraham> Lachy: Define "what it's supposed to look like" [03:31:00.0000] <sprasanna> im trying to use "string".is_utf8? in my code but it gives false all the time... i dunno why it happens [03:35:00.0000] <jgraham> sprasanna: You probably would need to give more context. But at a guess you want a more ruby-related channel [04:47:00.0000] <MikeSmith> hsivonen: question about v.nu backend behavior: As far as I can see now, there's not currently a way to overwrite contents of the local-entities subdir with freshly downloaded copies, right? [04:49:00.0000] <hsivonen> MikeSmith: do you mean 3rd party ones? [04:50:00.0000] <MikeSmith> hsivonen: no, I mean for the case where the contents of the upstream resource might have changed [04:51:00.0000] <hsivonen> MikeSmith: but resources that aren't in the whattf repo? [04:51:01.0000] <hsivonen> MikeSmith: I'm also curious which upstream resources have changed [04:52:00.0000] <MikeSmith> hsivonen: stepping back, I'm thinking about it in the context of trying figure out the best way to handle running v.nu in "local only" .. so the upstream resources are the language-subtag-registry file, MicrosyntaxDescriptions file, Validator.nu_alt_advice file [04:53:00.0000] <MikeSmith> what I want to add is some way to have the build download copies of those and store them locally [04:54:00.0000] <hsivonen> ah. there's no code for that download action right now, but build.py should have parameters that can take file URL once downloaded [04:54:01.0000] <MikeSmith> hsivonen: so what I was thinking is, it'd be easiest to just use your existing entity-map.txt and downloadLocalEntities() mechanism [04:55:00.0000] <MikeSmith> but - [04:56:00.0000] <MikeSmith> I would want to make it possible for a user to make changes to the remote sources of those, and then re-run v.nu and incorporate those remote changes [04:57:00.0000] <MikeSmith> but I think if I use the existing entity-map.txt and downloadLocalEntities() mechanism as-is, then I can't do that [04:57:01.0000] <hsivonen> ok. perhaps there should be a build.py action for redownloading [04:57:02.0000] <MikeSmith> OK, I can add that [04:57:03.0000] <hsivonen> since downloadLocalEntities() never redownloads unless you delete the local files [04:58:00.0000] <MikeSmith> hsivonen: Yep. I'm just wondering if it'd be preferable to you for me to add that redownload mechanism, or just to add separate, new handling for the language-subtag-registry, MicrosyntaxDescriptions, Validator.nu_alt_advice stuff [04:59:00.0000] <MikeSmith> I mean, the upstream DTDs don't change, so I suppose it's a bit of a waste to re-download them [05:00:00.0000] <hsivonen> MikeSmith: it might be best to commit microsyntax and alt advice into svn and handle languag tags like 3rd-party dtds [05:32:00.0000] <jgraham> public-html totally needs the /topic from here to be appeneded to the end of every message [05:33:00.0000] <jgraham> It ould make it so much easier to follow [05:47:00.0000] <hsivonen> I don't buy it that accessibility were a topic that weren't tractable to empirical research methods [05:50:00.0000] <jgraham> hsivonen: Shell'y argument appeared to me to be "we should keep features that ere designed for accessiblity independent of their actual usefulness because their presence makes disabled people feel better even if they provide no tangible benefit" [05:51:00.0000] <jgraham> Which seems like a pretty poor way to design a language (I would prefer to provide features that have actual benefits) not to mention rather patronising [05:55:00.0000] <Lachy> hey, in WebStorage, it says "Multiple separate objects implementing the Storage interface can all be associated with the same list of key/value pairs simultaneously" -- http://dev.w3.org/html5/webstorage/#the-storage-interface [05:56:00.0000] <Lachy> Is the only way that can happen when there are two pages open from the same site simultaneously, either in separate tabs or frames? [07:30:00.0000] <jgraham> /me fails to find the microsoft ES5 testsuite [08:55:00.0000] <zcorpan_> Hixie: http://www.w3.org/Bugs/Public/show_bug.cgi?id=6880 [08:57:00.0000] <jgraham> /me has never got accessed denied on W3C bugzilla before [08:59:00.0000] <annevk42> /me now wonders what that bug is about [09:02:00.0000] <zcorpan_> oh i didn't know it had limited access [09:02:01.0000] <Lachy> hmm, damn, I had no luck finding a bugmail archive that would have a copy of the report. I guess not all bugs go to a list like public-html-bugzilla [09:03:00.0000] <zcorpan_> it's an aria bug that's pending input from the html wg [09:03:01.0000] <zcorpan_> wonder how the html wg can provide input if html wg participants can't access the bug! [09:03:02.0000] <jgraham> Oh well good luck on getting input if no one in the WG can see the bug :) [09:03:03.0000] <jgraham> Oh you just said that :( [09:05:00.0000] <annevk42> WAI should really fix their transparency issues :/ [09:06:00.0000] <Lachy> oh, where did you find the bug mentioned? [09:08:00.0000] <jgraham> Isn't zcorpan in the wg or something? [09:08:01.0000] <zcorpan_> i'm subscribed to aria-ua-impl [09:14:00.0000] <remysharp> Hi - I'm looking for how to markup a slider in HTML5 (type="range") - but going by the documentation, I can't see anywhere where it says to use "range" as the type: http://www.whatwg.org/specs/web-apps/current-work/multipage/forms.html#range-state [09:14:01.0000] <remysharp> I was just wondering if there were docs/specs I had missed and you might be able to point me in the right direction [09:17:00.0000] <annevk42> in http://www.whatwg.org/specs/web-apps/current-work/#attr-input-type it defines how type attribute values map to states [09:17:01.0000] <remysharp> cheers. [09:18:00.0000] <remysharp> from a developers point of view - don't you think it might be good to have either a link to that page or a practical example implementation? [09:18:01.0000] <remysharp> I would have found that link sooner or later, [09:18:02.0000] <remysharp> but I'm thinking newbies to html. [09:18:03.0000] <annevk42> that link is present from your section [09:18:04.0000] <annevk42> type is a link [09:18:05.0000] <remysharp> ah - from the 'type' [09:19:00.0000] <annevk42> I think at some point Hixie will add more examples and fluff to make it more understandable [09:19:01.0000] <remysharp> okay, not sure that's obvious, but at least it's there - [09:19:02.0000] <remysharp> and good to hear that more examples will be added - definitely worth it for learning from. [09:19:03.0000] <remysharp> cheers. [09:20:00.0000] <annevk42> you can play with range in Opera/Safari fwiw [09:20:01.0000] <remysharp> I know ;-) [09:36:00.0000] <a-ja> anyone else noticing changes in <br /> in Opera 10 beta ? [09:38:00.0000] <a-ja> as in not starting a new line [09:39:00.0000] <jgraham> a-ja: Testcase? [09:39:01.0000] <a-ja> checking....think it's within an inline-block [09:41:00.0000] <a-ja> NSFW: http://stlrope.info/grue2/ [09:44:00.0000] <jgraham> a-ja: Is that a testcase? [09:45:00.0000] <a-ja> not a reduced one....it's an example [09:46:00.0000] <jgraham> OK [09:46:01.0000] <annevk42> /me finds http://www.openscreenproject.org/ [09:46:02.0000] <a-ja> jgraham: witihin the vevent div's [10:07:00.0000] <gsnedders> /me needs to get an SSD for /… IO slows me down way too much. [10:09:00.0000] <jgraham> gsnedders: More than IRC? [10:09:01.0000] <gsnedders> jgraham: Yes, I run more than just IRC :P [10:10:00.0000] <gsnedders> (This is when starting everything up after booting up, which is a side-effect of having so much open) [10:10:01.0000] <jgraham> gsnedders: That's not what I meant :p [10:11:00.0000] <gsnedders> jgraham: Hey, I don't look at the sort of websites zcorpan does at work! [10:17:00.0000] <Philip`> gsnedders: Why care about disk IO when you can just get lots of RAM and everything should be magically cached? [10:17:01.0000] <gsnedders> Philip`: It still has to get from IO to RAM in the first place, and that is my problem :P [10:18:00.0000] <Philip`> gsnedders: That should happen about twice a year when you reboot your machine [10:18:01.0000] <gsnedders> Philip`: The problem is I do it closer to 52 times a year. [10:18:02.0000] <Philip`> gsnedders: Then stop doing that :-p [10:29:00.0000] <jgraham> gsnedders: Or stop worring about optimizing something that ony happens once a week [10:29:01.0000] <jgraham> worrying [10:40:00.0000] <gsnedders> /me doesn't think Opera is meant to do that (i.e., crash) [10:47:00.0000] <jgraham> http://abcdefu.wordpress.com/2009/06/09/introducing-yip-an-unified-notification-system-for-the-web/ [10:47:01.0000] <jgraham> Seems prople want a standardised notification API [10:47:02.0000] <shelleyp> jgraham, it would be helpful that if you have a comment about something I write at the HTML WG mailing list, you respond to it there, instead of here [10:47:03.0000] <shelleyp> That way we have a discussion, rather than asides that pass each other in the internet night [10:49:00.0000] <jgraham> shelleyp: Sure I was planning to, but not till this evening [10:49:01.0000] <shelleyp> jgraham: cool, thanks [10:49:02.0000] <jgraham> I have other things to do fortunately / sadly (delete as required) [10:49:03.0000] <jgraham> :) [10:50:00.0000] <shelleyp> I understand, we're all overbooked [10:51:00.0000] <jgraham> shelleyp: FWIW the comment I was planning on making is basically "I don't think that solutions that make people feel good but don't actually work should be an acceptable design goal" [10:51:01.0000] <jgraham> (irrespective of whether @summary actually works) [10:51:02.0000] <shelleyp> jgraham: I don't want to get too far into here, I think the discussion should happen in the mailing list but... [10:52:00.0000] <shelleyp> My point was that measurements of success aren't always based on empirical observations [10:52:01.0000] <jgraham> /me doesn't really understand how you can have a non-empirical measurement [10:53:00.0000] <jgraham> Purely on semantic grounds [10:54:00.0000] <annevk2> that YIP is basically what showNotification() in HTML5 was [10:54:01.0000] <jgraham> (I can see logically ho you could say something like "success shouldn't be defined on empirical grounds" meaning "we won't measure anything before declaring a success or not") [10:55:00.0000] <gsnedders> /me grumbles about Adobe's ordering. They're saying they will cancel the order if they don't receive proof of academic identification tonight. It's was submitted over a week ago. [10:55:01.0000] <jgraham> annevk2: Whatever happened to showNotification? [10:56:00.0000] <jgraham> gsnedders: Didn't you just email / web form it? [10:56:01.0000] <gsnedders> jgraham: I did. [10:56:02.0000] <gsnedders> jgraham: They claim to have not got it! [10:56:03.0000] <jgraham> gsnedders: Oh. WFM [10:56:04.0000] <gsnedders> WFM in the past too. [10:56:05.0000] <jgraham> gsnedders: Resubmit it? [10:56:06.0000] <gsnedders> I am. Oh well. They get two open cases for the same order. Their problem. [10:57:00.0000] <shelleyp> jgraham: Interesting viewpoint, looking forward to reading more in your email response later tonight [10:57:01.0000] <annevk2> jgraham, only Google was interested and wanted full HTML support [10:57:02.0000] <jgraham> Oh [10:57:03.0000] <jgraham> Well obviously that's not quite true [10:58:00.0000] <a-ja> jgraham: opera 10 beta doesn't seem to be handling br:before{content:"\A";} [10:59:00.0000] <gsnedders> "You can expect a response within 1 business day." — bullshit. :\ [11:00:00.0000] <jgraham> a-ja: Oh interesting. I will check if there is a bug already reported (but not right now because I'm going home, so if someone else wants to do it, feel free :) ) [11:01:00.0000] <a-ja> pretty sure it worked in prior weekly.....wil back out to that and verify, in a bit [11:07:00.0000] <annevk2> a-ja, maybe we finally fixed the bug that requires you to add white-space:pre-line for that to work? [11:11:00.0000] <a-ja> annevk2: will give that a try......wasn't aware that pre-line 's required [11:12:00.0000] <annevk2> or pre [11:12:01.0000] <annevk2> \A is just a newline character [11:15:00.0000] <a-ja> there's some white-space:nowrap stuff in same <p>. will mess with that a bit and see what happens [11:34:00.0000] <gsnedders> /me grumbles at PHP some more [13:01:00.0000] <takkaria> gsnedders: are you at Opera too this summer? [13:03:00.0000] <jgraham> takkaria: Yeah you are both interning :) [13:03:01.0000] <jgraham> takkaria: When do you start? [13:04:00.0000] <takkaria> I'm there on the 22nd [13:04:01.0000] <takkaria> until the 18th September [13:04:02.0000] <jgraham> OK, nice [13:04:03.0000] <jgraham> gsnedders starts the week after iirc [13:05:00.0000] <takkaria> do you know if there are any plans on what we'll be working on? [13:05:01.0000] <jgraham> takkaria: Nothing specific. You are QAing, right? [13:06:00.0000] <takkaria> I'm not sure, I specified that I'd be happy to code or QA [13:06:01.0000] <jgraham> (I mean I don't know anything speific, not that you won't be woking on anything specific) [13:06:02.0000] <takkaria> yeah, not working on anything specific seems like a pretty easy job :) [13:07:00.0000] <jgraham> OK. Well if you are QAing I can think of at least three or four things that you could work on and that's just things that I happen to be interested in :) [13:07:01.0000] <takkaria> fair enough [13:07:02.0000] <takkaria> what do you do, out of interest? [13:09:00.0000] <jgraham> QA. I've been working with the JS engine a bit and on general document-related bugs (as opposed to layout-related bugs) [13:10:00.0000] <takkaria> ah, cool [13:12:00.0000] <gsnedders> takkaria: 28th [13:12:01.0000] <gsnedders> takkaria: I'm doing QA [13:12:02.0000] <gsnedders> takkaria: What office you at? [13:15:00.0000] <Hixie> i am SO glad i gave xhr to anne :-) [13:16:00.0000] <jgraham> gsnedders: I am pretty sure he is in Linkoping [13:18:00.0000] <annevk42> Hixie, I recall I just took it; I'm also not glad with it :p [13:18:01.0000] <annevk42> sync/async is a bitch to define properly [13:19:00.0000] <annevk42> and now I'm looking at XHR2 it gets even worse because then it's sync same-origin/sync cross-origin/... [13:20:00.0000] <jgraham> annevk2: Just define that sync cross-origin requests always fail :) [13:20:01.0000] <gsnedders> jgraham: I'm pretty sure he is too [13:20:02.0000] <gsnedders> takkaria: So I arrive after you and leave before you :P [13:20:03.0000] <annevk42> jgraham, compelling, but it's "needed" for Workers [13:21:00.0000] <jgraham> annevk42: Oh. [13:23:00.0000] <annevk42> seamless sandboxed iframes are neat [13:23:01.0000] <annevk42> especially the origin feature [13:46:00.0000] <slightlyoff> hey Hixie [13:46:01.0000] <slightlyoff> is there an SVN repo w/ acid3 in it somewheres? [13:49:00.0000] <slightlyoff> n/m, dug it out of webkit sources... [13:55:00.0000] <annevk42> Hixie, are we getting <iframe document> at some point before LC? [13:58:00.0000] <rubys> hsivonen: ping? [13:59:00.0000] <rubys> java.lang.ClassCastException: org.xml.sax.InputSource cannot be cast to nu.validator.xml.TypedInputSource [14:05:00.0000] <Hixie> annevk42: depends on implementors. I don't want to add it until we have experience with sandbox. [15:36:00.0000] <jgraham> Hmm I thought people had grasped the difference between UA requiements and author requirements, but I guess not [15:36:01.0000] <sayrer> jgraham: I think people grasp them [15:37:00.0000] <sayrer> they just don't think the concept is useful [15:38:00.0000] <sayrer> it all depends on how effective you think prescriptive author requirements will be [15:39:00.0000] <sayrer> if there were any advantage to complying with that sort of thing, websites would do it [15:39:01.0000] <jgraham> sayrer: I think you may be projecting your view onto other people :) [15:40:00.0000] <sayrer> well maybe, but I also want to know why they are useful [15:40:01.0000] <sayrer> smart people seem to think they are [15:40:02.0000] <sayrer> so I am willing to entertain the notion that I am wrong [15:41:00.0000] <jgraham> sayrer: Generally they are useful as QA tools. If you stay within conformance you generally won't hit sharp edges that are likely to get you burnt [15:41:01.0000] <sayrer> there is a lint aspect to them [15:41:02.0000] <sayrer> I see that [15:41:03.0000] <sayrer> but that is different than a conformance requirement [15:41:04.0000] <jgraham> There is also the question of simplifying the language [15:41:05.0000] <sayrer> hmm? [15:42:00.0000] <sayrer> <font> is much simpler than css [15:42:01.0000] <sayrer> for authors [15:42:02.0000] <jgraham> By restricting the set of all working ways of achieving an effect we can focus resources on those that have the best properties for the whole ecosystem [15:42:03.0000] <sayrer> ah, but we can't [15:43:00.0000] <sayrer> since the spec gives quite precise requirements for everything, whether they are conformant for authors or not [15:43:01.0000] <sayrer> that is something I fully support [15:44:00.0000] <jgraham> sayrer: <font> doesn't scale well as a skill and tends to lead to pages that are hard to repurpose beyond the original author's intentions (e.g. to a user using a different type of UA or a screen scraper or...) [15:44:01.0000] <sayrer> I'm not sure that's true [15:44:02.0000] <sayrer> given the example at hand [15:44:03.0000] <sayrer> <font color="blue"> [15:45:00.0000] <jgraham> sayrer: There is far from enough context in that example o deduce anything [15:45:01.0000] <sayrer> well, I agree that it's possible you could do something more reusable in some situations [15:45:02.0000] <jgraham> If you could show me that on real pages people only use <font> when they "should" have used <span>+CSS I would be more convinced [15:46:00.0000] <jgraham> and very surprised, frankly [15:46:01.0000] <sayrer> but a class name + a css file might also buy you an extra file and CSS [15:46:02.0000] <sayrer> when all the author wants is a blue word [15:47:00.0000] <sayrer> jgraham: anything syndicated generally goes for font [15:47:01.0000] <gsnedders> Oh man. That did not just happen. In a race on GTA 4, my car broke down while in the lead coming out of the final corner. [15:48:00.0000] <jgraham> sayrer: I guess the times that "all the auhor wants is a blue word" with no context for the blueness are rather few [15:48:01.0000] <jgraham> Compared to the cases where <font> is used [15:49:00.0000] <sayrer> that seems quite difficult to quantify [15:49:01.0000] <jgraham> sayrer: Sadly lots of useful things are hard to quantify in practice [15:50:00.0000] <sayrer> sure, we don't go making MUST requirements about them :) [15:50:01.0000] <jgraham> It is at least clear that "leading" web designers voluntarily forgo the use of <font>, suggesting that the advntages to not using it are non-theoretical [15:51:00.0000] <jgraham> (and, I should add, they can back up their decision with cognent reasoning about mantainability and so on) [15:56:00.0000] <jgraham> /me is going to try sleeping now [15:56:01.0000] <sayrer> jgraham: they also write "xhtml" [15:56:02.0000] <sayrer> so I think your point is questionable [15:56:03.0000] <sayrer> night [15:57:00.0000] <Dashiva> And use floats to make table layouts [16:03:00.0000] <gsnedders> And use floats for tabular data, because tables are bad. [16:13:00.0000] <sayrer> snicker [16:13:01.0000] <sayrer> google.com [16:13:02.0000] <sayrer> perhaps the most closely measured page in the history of the world [16:13:03.0000] <sayrer> contains a font element [16:14:00.0000] <ezyang> jaja [16:14:01.0000] <sayrer> somebody better tell them [16:14:02.0000] <sayrer> about their suboptimal markup [16:15:00.0000] <ezyang> It's probably really portable! [16:34:00.0000] <hober> I don't think I understand Shelley's latest email. [16:35:00.0000] <sayrer> hober, why not? [16:35:01.0000] <sayrer> it seems to contain a good point [16:35:02.0000] <hober> which was? [16:36:00.0000] <sayrer> you don't want to edit a font element out of a document in order to produce "conformant" markup [16:37:00.0000] <hober> Who is 'you' in this case? (I think I would want to do that...) [16:37:01.0000] <sayrer> hober, if you agree that a <span style=""> is no better [16:38:00.0000] <sayrer> then you would want to move the CSS to a style element [16:38:01.0000] <sayrer> that is madness [16:38:02.0000] <sayrer> since you would be changing the "lexical scope" that a given piece of markup depends on [16:38:03.0000] <hober> Well, I'd like to know why the text is blue, so I could write <strong>...</strong>, or <span class="some-semantic-missing-from-the-default-set-of-elements">...</span> [16:39:00.0000] <sayrer> hober, well, I don't agree that the semantics of the blue text are always worth writing down, but where do you put the styling information if it's clear that the element should be <strong> [16:40:00.0000] <hober> In an external stylesheet, linked to with <link> [16:41:00.0000] <hober> (ideally the selector for applying the blue style is simply "strong", too) [16:41:01.0000] <sayrer> so you introduce a dependency like that via a wysiwyg editor? [16:42:00.0000] <hober> I use Emacs to write markup. :) [16:43:00.0000] <sayrer> hober: I take it you forfeit? [16:43:01.0000] <sayrer> ;) [16:44:00.0000] <sayrer> hober, but even with Emacs [16:44:01.0000] <sayrer> are you aware of any editor that would perform that transformation? [16:44:02.0000] <hober> I think the question of how to build a WYSIWYG web page editor is a separate one from what we were talking about [16:44:03.0000] <sayrer> hober, not according to the spec [16:44:04.0000] <sayrer> but I am willing to engage on Emacs editing modes [16:44:05.0000] <sayrer> are there Emacs editing modes that perform the transformation you described? [16:45:00.0000] <hober> What do you mean by transformation? [16:45:01.0000] <hober> the author types in < s t r o n g > ... [16:45:02.0000] <sayrer> <font color="blue"> --> <strong> + css file [16:45:03.0000] <sayrer> the editor doesn't always create the pages it edits [16:46:00.0000] <hober> The author doesn't always create the page he/she is editing, sure. [16:46:01.0000] <sayrer> so, if I copy and thenpaste a section with <font> [16:46:02.0000] <sayrer> should the editor paste in a font? [16:46:03.0000] <sayrer> a font element, I mean [16:46:04.0000] <hober> I mean, this is generally true of any document editing. How does an author change foo to bar in a document of type baz? Well, he/she finds all the foos and changes them to bar.. [16:47:00.0000] <sayrer> authoring tools are subject to the same requirements that authors are [16:47:01.0000] <hober> If you copy source from somewhere and paste it in somewhere, I would be really annoyed if it wasn't the same source... [16:47:02.0000] <sayrer> yes, I agree [16:48:00.0000] <sayrer> so it will be quite impossible for authoring tools to meet the current conformance requirements [16:50:00.0000] <hober> so you've got markup in an intermediate state, of unknown conformance. I don't see why it would be "quite impossible" to have a "export to conformant html5" button [16:51:00.0000] <hober> with an annoying "some formatting may be lost" confirmation dialog [16:51:01.0000] <sayrer> so it would require data loss? [16:52:00.0000] <sayrer> lame [16:52:01.0000] <hober> user writes unkown attribute or element (which are non-conformant), what does an editor replace them with? [16:52:02.0000] <sayrer> nothing, obviously [16:53:00.0000] <hober> <unknown> could become <div class="com.example.unknown"> [16:53:01.0000] <sayrer> but it might suggest something [16:53:02.0000] <hober> unknown attributes though, get lost I think [16:53:03.0000] <sayrer> a far more likely example might be <dvi></div> [16:53:04.0000] <hober> data-com-example-unknown? [16:53:05.0000] <Hixie> hober: you could use data-com.example.attr="foo" [16:53:06.0000] <Hixie> yeah [16:53:07.0000] <Hixie> what you said [16:54:00.0000] <sayrer> but if we veer into wysiwyg, this is irrelevant [16:54:01.0000] <sayrer> it looks how it looks [16:54:02.0000] <hober> ... and when you click that button, and OK that dialog, the result is conformant HTML5, right? [16:55:00.0000] <hober> I don't see why it wouldn't be, or rather why it can't be. [16:55:01.0000] <sayrer> two things [16:55:02.0000] <sayrer> I am not aware of any editor that would do that [16:55:03.0000] <sayrer> it might be undesirable [16:56:00.0000] <sayrer> the HTML5 standard is an astonishingly bad place to speculate on Editors Of The Future [16:58:00.0000] <sayrer> it is sort of antithetical to the "don't mess with headers you don't understand" ethos of email [16:59:00.0000] <sayrer> and more complicated given the external dependencies that a "proper" file would have 2009-06-10 [17:05:00.0000] <hober> well, think of it in terms of postel's law. a UA is at the 'liberal in what you accept' end, matching your "don't mess" behavior. [17:05:01.0000] <hober> an editor / authoring tool is at the 'conservative in what you send' end [17:06:00.0000] <hober> hence authoring requirements being stricter than UA requirements [17:16:00.0000] <sayrer> hober, I am saying that those requirements are incompatibl [17:16:01.0000] <sayrer> e [17:17:00.0000] <hober> how are they incompatible? [17:20:00.0000] <roc> they're incompatible if you require round-tripping [17:20:01.0000] <roc> which is what most people want from an editor [17:23:00.0000] <hober> hmm. conformant documents round-trip with conformant tools. non-conformant documents round-trip if your tools are also non-conformant. [17:24:00.0000] <takkaria> gsnedders: yup, Linkoping [17:46:00.0000] <ojan> Hixie: ping [18:11:00.0000] <Hixie> ojan: here [18:21:00.0000] <ojan> Hixie: i'm looking at http://www.whatwg.org/specs/web-apps/current-work/#attr-fe-disabled [18:21:01.0000] <ojan> it's not clear to me from that description whether text selection in a disabled input should work [18:22:00.0000] <Hixie> looking... [18:23:00.0000] <Hixie> it's up to the UA [18:23:01.0000] <ojan> FWIW, IE allows text selection and Gecko/WebKit don't from my quick test [18:23:02.0000] <ojan> ok [18:23:03.0000] <ojan> thx [18:24:00.0000] <Hixie> do you want a note in there to that effect? [18:25:00.0000] <ojan> meh. i guess text selection stuff is mostly up to the UA, right? [18:25:01.0000] <ojan> so a note just in that one instance doesn't seem necessary [18:25:02.0000] <Hixie> ok [19:46:00.0000] <adu> hi [00:39:00.0000] <hsivonen> Hixie: yes, I pointed out the absolute URL thing to DanC [00:58:00.0000] <annevk42> seems that Web ECMAScript needs to define (new Date(NaN).toString()) as well [01:02:00.0000] <Mrmil> /me received a xhtml list of features coded in <br />'s from a top-notch application developer. sigh [01:05:00.0000] <Hixie> hsivonen: yt? [01:06:00.0000] <hsivonen> Hixie: yes [01:07:00.0000] <Hixie> hsivonen: http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2009-May/019843.html and http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2009-May/019852.html [01:07:01.0000] <Hixie> hsivonen: how would you like me to proceed? [01:07:02.0000] <Hixie> hsivonen: do you think i should define a mapping to infoset for microdata as you describe? [01:08:00.0000] <hsivonen> If it's proven by implementation, yes. However, at this point it's not proven by implementation. [01:08:01.0000] <hsivonen> I guess you could put it in and cut it at CR if it sucks. [01:10:00.0000] <Hixie> i'm leaning towards waiting and seeing if there's demand [01:10:01.0000] <Hixie> would that be ok? [01:11:00.0000] <Hixie> and re your latest e-mail, i now understand what you meant by "literal non-characters", my bad [01:14:00.0000] <hsivonen> Hixie: yeah, it would be OK to wait with the XML mapping [01:14:01.0000] <annevk42> oh yay, Gecko puts elements in the XHTML namespace too now [01:19:00.0000] <Hixie> ok, thanks [01:20:00.0000] <hsivonen> annevk42: I really hope the change proves to be a non-event as far as Web compat goes [01:20:01.0000] <annevk42> from prior ruby feedback it seems better to just obsolete the fallback mechanism [01:21:00.0000] <annevk42> (and require everyone to support ruby properly) [01:21:01.0000] <hsivonen> annevk42: what about degrading gracefully? [01:24:00.0000] <hsivonen> https://twitter.com/plhw3org/statuses/2090392921 [01:26:00.0000] <annevk42> hsivonen, e.g. myakura said the fallback was not good [01:26:01.0000] <annevk42> hsivonen, heh, I see you replied as well, me too: https://twitter.com/annevk/status/2090405819 [01:30:00.0000] <annevk42> cheap shot I suppose [01:31:00.0000] <remysharp> hi chaps (and chapesses?), I wanted to sign up to the mailing lists, but which one is best to give feedback on implementing the markup? [01:31:01.0000] <remysharp> is it the web designers and html authors - or [01:31:02.0000] <remysharp> implementations (though I think this might be the browser implementors) [01:31:03.0000] <remysharp> implementers [01:34:00.0000] <annevk42> remysharp, questions go to help, feedback goes to whatwg, and questions on implementing features in user agents (e.g. html5lib, browsers, etc.) go to implementors [01:35:00.0000] <remysharp> okay, so feedback on how elements work or are used from an authors point of view, I want whatwg⊙wo then. [01:35:01.0000] <remysharp> cheers. [01:35:02.0000] <Hixie> macs really need to get better at handling hard disk failures [01:36:00.0000] <Hixie> hanging the entire OS when the disk fails to respond is not acceptable [01:36:01.0000] <Hixie> (external disk) [02:08:00.0000] <annevk42> maybe we should introduce the concept "legacy fallback" [02:08:01.0000] <annevk42> to make it clear how various elements differ in fallback model [02:10:00.0000] <Hixie> hsivonen: doesn't seem to be much point pushing <video> for freeness if one then uses java as the fallback [02:10:01.0000] <hsivonen> making Hixie allow <applet> and/or <object classid> in <video> content would be a start... [02:10:02.0000] <hsivonen> Hixie: Java is royalty-free, no? [02:11:00.0000] <hsivonen> Hixie: also, it's about making video play in IE and Opera which are both already proprietary [02:11:01.0000] <Hixie> if all you care about is money, it might be ok, i dunno [02:12:00.0000] <hsivonen> Hixie: I care about Freedom [02:12:01.0000] <Hixie> (money isn't the part of freedom that i particularly care about) [02:12:02.0000] <Hixie> java is basically a single-vendor technology for all intents and purposes [02:12:03.0000] <hsivonen> Hixie: to have Free as in Freedom implementations of the platform, it's useful to enable the same videos to play in IE [02:13:00.0000] <hsivonen> and that particular mechanism may be non-Free but royalty-free [02:13:01.0000] <Hixie> didn't we establish that java could use <object> or <embed> [02:14:00.0000] <hsivonen> Hixie: have we established that Java could use <object> in IE without classid? [02:14:01.0000] <Hixie> no idea [02:14:02.0000] <Hixie> i'm finding it difficult to care about java fallback in IE [02:15:00.0000] <hsivonen> Hixie: is your story that Cortado fallback should focus on IE only? (If you don't allow <applet>, it's pretty obvious that authors who put in a fallback won't bother tweaking it for Opera) [02:15:01.0000] <annevk42> are you sure <object classid> does not work in Opera? [02:15:02.0000] <hsivonen> Hixie: I think I have an extraordinarily legitimate use case for Java applets here [02:15:03.0000] <hsivonen> annevk42: I am not sure [02:15:04.0000] <Hixie> frankly personally i don't really see why anyone is using video at all ye, since we don't have a codec yet [02:18:00.0000] <hsivonen> hmm. if it turns out that Opera pays attention to classid, classid would no longer be single-vendor technology [02:19:00.0000] <Hixie> classid is already defined in html5 [02:19:01.0000] <hsivonen> whoa. I just looked [02:19:02.0000] <Hixie> it's just not conforming, because it's values aren't necessary to obtain the effect it does [02:19:03.0000] <hsivonen> /me looks again [02:19:04.0000] <hsivonen> well, then [02:19:05.0000] <Hixie> (type="" does everything classid does) [02:20:00.0000] <Hixie> s/it's/its/ [02:20:01.0000] <hsivonen> Hixie: do you have a demo for bootstrapping Java in IE using object without classid? [02:20:02.0000] <Hixie> no, i don't really care enough about java to have a demo [02:21:00.0000] <hsivonen> Hixie: how do you know that "(type="" does everything classid does)" [02:21:01.0000] <hsivonen> in the present tense [02:21:02.0000] <hsivonen> if you haven't demos [02:21:03.0000] <Hixie> classid="" is a dispatch mechanism. type="" is a dispatch mechanism. we only need one dispatch mechanism. [02:21:04.0000] <hsivonen> but the present tense has all the relevance to fallback* [02:22:00.0000] <Hixie> there is no present tense with <video> [02:22:01.0000] <Hixie> <video> has no defined codec yet. [02:22:02.0000] <Hixie> so nobody should be using it. [02:22:03.0000] <Hixie> and if they do, they are on their own [02:22:04.0000] <hsivonen> Hixie: there is a present tense to having a fallback in IE8 [02:25:00.0000] <Hixie> there is no need for fallback when there is nothing to fallback _from_ [02:26:00.0000] <hsivonen> Hixie: if we get a codec in the future, falling back in IE8 will be relevant [02:26:01.0000] <hsivonen> Hixie: by your logic, we could make <video> itself non-conforming while pondering the codec [02:27:00.0000] <Hixie> who knows what the fallback needs will be then [02:27:01.0000] <Hixie> <video> _is_ non-conforming today [02:27:02.0000] <Hixie> html5 is a draft [02:27:03.0000] <annevk42> you're talking past each other [02:27:04.0000] <annevk42> right [02:27:05.0000] <Hixie> hasn't even reached last call [02:27:06.0000] <hsivonen> Hixie: yet, in practice, the browsers that have shipped <video> or are about to have exactly one common codec set [02:27:07.0000] <annevk42> Hixie is talking about the future and hsivonen is talking about now [02:27:08.0000] <Hixie> hsivonen: they are? [02:28:00.0000] <hsivonen> Hixie: with XiphQT for Safari, yeah [02:28:01.0000] <Hixie> hsivonen: one browser doesn't even have <video> [02:28:02.0000] <Hixie> hsivonen: and if we can ask people to download a new codec, why not ask the other people to download a custom plugin? [02:29:00.0000] <hsivonen> Hixie: you still get out-of-the-box goodness in products from two other vendors and you bypass the plug-in prison on Snow Leopard (AFAICT) [02:30:00.0000] <Hixie> (for some definition of "goodness") [02:43:00.0000] <Hixie> ok i can't get java to work even with <applet>, so the odds of me getting it to work with anything else are minimal [02:44:00.0000] <Hixie> it crashed firefox twice, IE refuses to render anything but a red cross, and safari claims java can't be enabled. [02:44:01.0000] <Hixie> which from what i hear is a good thing since i don't think apple has fixed the java exploit yet [02:50:00.0000] <Hixie> ah, i don't seem to even have java installed. [02:51:00.0000] <Hixie> oh jesus java tried to install the MSN toolbar [02:51:01.0000] <Hixie> hsivonen: i'm less and less convinced that this is a useful route to follow for video fallback [02:52:00.0000] <hsivonen> Hixie: note that this is for casual windows users who have already installed the JRE for other reasons [02:53:00.0000] <hsivonen> Hixie: not for non-windows users who take a closer look at what they install [02:54:00.0000] <Hixie> the installer seems to have hung [02:54:01.0000] <Hixie> so the odds of casual windows users having it installed seem low to me :-) [02:55:00.0000] <hsivonen> Hixie: I'd agree with you based on my own installation experience, but random XP boxen seem to have the Java thingy in tray surprisingly often [02:55:01.0000] <hsivonen> Hixie: I don't know how it ends up installed, but it seems it does [03:02:00.0000] <Hixie> hsivonen: http://damowmow.com/playground/demos/java/001.html [03:03:00.0000] <hsivonen> Hixie: the second </applet> should be </object>, right? [03:03:01.0000] <Hixie> oops [03:03:02.0000] <Hixie> fixed [03:04:00.0000] <Hixie> man windows 7 sounds more like the wii than a serious OS [03:04:01.0000] <annevk42> of course, "application/x-java-applet" is not really conforming either, but I guess the validator does not go that far [03:05:00.0000] <Hixie> i'll let sun worry about _that_ [03:06:00.0000] <Hixie> notwithstanding the fact that the video doesn't actually play for me, i got it working for <applet>, <object>, and <embed> in IE8 without violating HTML5 for <object> or <embed>. [03:06:01.0000] <hsivonen> Hixie: OK. cool.. [03:06:02.0000] <hsivonen> I guess I shouldn't trust Sun's documentation [03:06:03.0000] <hsivonen> at all [03:07:00.0000] <annevk42> none of it works for me btw; guess I don't have Java [03:07:01.0000] <Hixie> i get as far as cortado loading and saying "buffering... 25%" (for some arbitary %) [03:07:02.0000] <Hixie> sometimes cortado crashes with a NullPointerException [03:07:03.0000] <Hixie> and it never plays the video back [03:08:00.0000] <Hixie> but i figure those problems are besides the point here [03:11:00.0000] <hsivonen> http://hsivonen.iki.fi/test/moz/video-fallback-validation/object-type.html WFM in IE8 [03:11:01.0000] <hsivonen> and validates [03:13:00.0000] <nessy> for me, safari on the mac buffers 1% of each of the three videos; firefox 3.5 doesn't display any of them [03:13:01.0000] <hsivonen> http://hsivonen.iki.fi/test/moz/video-fallback-validation/object-type.html WFM in Opera on Windows, too. [03:14:00.0000] <hsivonen> I'll close the bug. sorry about the false alarm [03:14:01.0000] <annevk42> is already closed [03:14:02.0000] <hsivonen> oh. Hixie closed it already. [03:14:03.0000] <Hixie> :-) [03:15:00.0000] <Hixie> no idea why it works for you but not me [03:15:01.0000] <Hixie> that's weird [03:15:02.0000] <Hixie> oh wel [03:15:03.0000] <Hixie> l [03:15:04.0000] <Hixie> (for your page but not mine, i should say) [03:16:00.0000] <hsivonen> Hixie: on surface, I notice that I used .ogg instead of .ogv and I have carefully configured Apache's types in advance [03:16:01.0000] <hsivonen> Hixie: Also, I obtained cortado fresh from svn [03:16:02.0000] <hsivonen> I'm positively surprised that it compiled with no classpath hell whatsoever [03:19:00.0000] <Hixie> i used the cortado build on their ftp site from just now, and use the same mime type as you [03:19:01.0000] <Hixie> do you think they examine the filename? [03:19:02.0000] <Hixie> i could change that i guess... [03:20:00.0000] <Hixie> same result [03:20:01.0000] <Hixie> <applet> crashed, <object> says "Buffering...", and <embed> is blank. [03:20:02.0000] <Hixie> /me continues to experience hell whenever dealing with java [03:30:00.0000] <jgraham> "[@summary] may just be a one trick pony. It does that trick well". Well I hope it is now obvious why I think lots of accessibility people are uninterested in enganing in a feedback loop about its usefulness [03:30:01.0000] <jgraham> engaging [03:32:00.0000] <Hixie> they seem very interested in engaging in a feedback loop if by "feedback loop" you mean a loop where they give the same feedback repeatedly ignoring all responses [03:32:01.0000] <jgraham> Heh. That wasn't quite what I had in mind... [03:33:00.0000] <Hixie> (and by "they" i mean specifically a dozen or so people in public-html, not the whole accessibility community, many of whom have privately told me that they agree that summary="" has failed) [03:34:00.0000] <annevk42> Hixie, unfortunately private communication cannot be verified by anyone so is hardly useful for common understanding of things (imo) [03:34:01.0000] <Hixie> indeed [03:35:00.0000] <Hixie> hence why i have been focusing on the research and not making claims to authority in my arguments [03:36:00.0000] <annevk42> you did make a few with respect to the US gov, to be fair [03:36:01.0000] <Hixie> hm? [03:36:02.0000] <annevk42> but they were not that relevant to the overall point [03:36:03.0000] <Hixie> not sure which case you're talking about [03:39:00.0000] <annevk42> I recall something like "even the US gov discourages summary=" [03:40:00.0000] <annevk42> but alas, I cannot find it [03:42:00.0000] <Hixie> there was a recent post where i said that, but that was citing the documentation that the fpwg gave [03:50:00.0000] <Hixie> ok bed time [04:26:00.0000] <Dashiva> Is there a term for giving examples that are nothing at all like the relevant subject at hand? [04:28:00.0000] <jgraham> Dashiva: For example? [04:32:00.0000] <Dashiva> Like talking about handicap parking in the @summary discussion [04:33:00.0000] <jgraham> Oh I'm not sure what that's called [04:33:01.0000] <jgraham> Although I think Shelly was just wrong [04:34:00.0000] <jgraham> If you measured the impact of disabled spaces then I'm pretty sure that they would come out well [04:34:01.0000] <jgraham> compared to the alternatives [04:35:00.0000] <Dashiva> What is the parking equivalent of summary on a layout table? [04:36:00.0000] <jgraham> I guess it is a disabled person using the space. But to be fair I think she was trying to make a general (and as I said, IMHO, wrong) point so there need not be a 1:1 correlation [04:38:00.0000] <Dashiva> Well, yeah. Any example to something regulated by business and law is somewhat irrelevant by default when it comes to a volunteer and amateur web [04:39:00.0000] <jgraham> My assumption is that she believes that an "empiracist" would regard disabled parking spots as a failure if it could be shown that they were more often used by non-disabled people than by disabled people [04:39:01.0000] <jgraham> It might be true that they are used in that way. But that doesn't make them a failure [04:40:00.0000] <jgraham> Because they might nevertheless increase the accessibility of shops to disabled people by increasing the chance that they will get a spot near the shops [04:41:00.0000] <jgraham> Thus allowing them to go shoppiing when they would otherwise not be able to [04:42:00.0000] <Dashiva> I've been avoid calling it an analogy because it's a very bad one :) [04:42:01.0000] <jgraham> On the other hand they would be a failure if they were less cost effective at achieving this goal than some other solution that could be used instead. [04:42:02.0000] <jgraham> And were preventing that alternative solution being deployed [04:43:00.0000] <Dashiva> You also have to look at the overall result. Non-disabled use of disabled parking still produces utility. [04:44:00.0000] <Dashiva> Whereas an ignored attribute is pure author waste [04:44:01.0000] <jgraham> True [04:46:00.0000] <jgraham> (as an example of an alternative solution, we can imagine requiring businesses to ferry disabled users directly to the door from any parking space. This would have greater benefits (door to door service) but also much greater costs) [05:19:00.0000] <annevk42> I have the feeling some of the ECMAScript folks have been working on "obsoleting" the origin security infrastructure whereas we are working on strengthening it [05:20:00.0000] <annevk42> E.g. they are working on safe subsets of ECMAScript that can run within the same origin somehow for ads etc. whereas we have the sandboxed <iframe> feature to give those ads a distinct origin [06:00:00.0000] <annevk42> (My main worry though is that I am bringing this up and nobody else.) [06:02:00.0000] <jgraham> annevk42: I agree ith you but I don't kno if it is a problem [06:04:00.0000] <annevk42> jgraham, maybe it's just a problem for me then [06:04:01.0000] <jgraham> annevk42: Why is it a problem to have two possible approaches to the same problem, in this case? [06:04:02.0000] <annevk42> jgraham, some of the ECMAScript folks are asking for specific XMLHttpRequest changes that do not seem to take into account the feature set that sandboxed <iframe>s offer [06:05:00.0000] <jgraham> Oh [06:05:01.0000] <jgraham> Well I guess it depends on the merits of those changes [06:05:02.0000] <annevk42> jgraham, one security researcher makes claims about CORS based on the assumption that the Web will go in the direction of Caja/etc. [06:06:00.0000] <annevk42> jgraham, the merits heavily depend on where we end up, security wise [06:06:01.0000] <jgraham> I think it is bad to assume that either technique will dominate [06:07:00.0000] <annevk42> you also think it is bad to have several ways to do the same thing [06:08:00.0000] <annevk42> and given the API bloat that XMLHttpRequest already has I'm not inclined to sprinkle around some more [06:09:00.0000] <jgraham> In general I agree but with something like security there is more of an argument for more than one way to do it [06:09:01.0000] <jgraham> Assuming the ways are complementary and don't crowd each other out of the market [06:10:00.0000] <jgraham> (with different browsers implementing different solutions) [06:11:00.0000] <jgraham> Or cause undue confusion about the right approach to use [06:12:00.0000] <annevk42> I think you need one approach as it affects how ad platforms have to work, how you have to do communication securely, whether origins can be trusted, etc. [06:13:00.0000] <jgraham> I don't think we will (in general) want to assume different origins can be trusted any time soon [06:13:01.0000] <annevk42> it's about whether you can trust your own [06:13:02.0000] <jgraham> ad platforms should just pick whichever solution meets their needs best [06:15:00.0000] <jgraham> annevk42: Well I don't really understand the details... I assume it has something to so with embedding external script in your own origin but using caja to limit what it can do? [06:16:00.0000] <gsnedders> /me grumbles at Adobe [06:16:01.0000] <annevk42> yes [06:16:02.0000] <annevk42> versus embedding it inside a sandboxed iframe [06:16:03.0000] <gsnedders> /me has sent a form of verification they say is valid and they're saying it isn't [06:16:04.0000] <jgraham> (it is possible that my opinions on this topic are nonsensical since I don't really understand it all that well and it is rather complex) [06:17:00.0000] <gsnedders> "Official, current report card indicating name of school and student" — uh, yes, what I sent is that. [06:17:01.0000] <jgraham> annevk42: So what, in detail, is the problem with allowing ad networks / authors to choose whichever approach they feel is best [06:18:00.0000] <annevk42> it complicates XHR and potentially other APIs [06:18:01.0000] <jgraham> How much? [06:18:02.0000] <gsnedders> Oh, duh. [06:18:03.0000] <jgraham> Does it break back compatibility for example? [06:18:04.0000] <annevk42> needs a new constructor that changes behavior [06:19:00.0000] <jgraham> That seems like a pretty big change. What is the new behaviour [06:19:01.0000] <jgraham> ? [06:19:02.0000] <annevk42> acting as if the sending origin was null and not include any credentials in the request [06:20:00.0000] <annevk42> a sandboxed iframe gives you exactly that [06:20:01.0000] <jgraham> Yeah [06:21:00.0000] <jgraham> Well maybe it is worth trying to reduce the overlap between the two solutions if thay are going to cause significant duplication of functionality but in different places [06:24:00.0000] <annevk42> my current tactic is "look, sandboxed iframes" [08:39:00.0000] <tantek> Hixie, classid and type do have different semantics though (what piece of code is preferred to run this thing, what is the data type of this thing), that have been shown to have utility in various systems, even before the web. [08:39:01.0000] <tantek> E.g. MacOS files have the notion of Type and Creator. If the Creator is found for opening a file then it is used, otherwise, the system looks for what other applications support the Type of document, and give the user the option of opening the document in one of those apps. [08:40:00.0000] <tantek> And you could certainly base support for classid on existing web content that uses it. [08:41:00.0000] <tantek> /me is not a huge supporter of classid, but there are the facts behind it that should be considered from a scientific perspective before rejecting/removing it. [08:47:00.0000] <gsnedders> In title case, should à be capitalized (en français)? [10:52:00.0000] <annevk42> someone joked about grids for the Web and it taking ten years [10:52:01.0000] <annevk42> first time it was discussed on a public list was early '95 apparently: http://lists.w3.org/Archives/Public/www-style/1995Jun/0000.html [10:53:00.0000] <annevk42> where "early" means "halfway" [11:03:00.0000] <beowulf> don't we already have grids in tables... [11:04:00.0000] <beowulf> /me gets his coat [11:44:00.0000] <jgraham> annevk42: The lack of grid layouts is far and away the biggest failing of CSS [11:44:01.0000] <jgraham> Probably one of the biggest failings in the web standards world [11:45:00.0000] <annevk42> The group responsible for CSS mostly appears interested in print and documents [11:46:00.0000] <annevk42> Though to be fair grid layout is more seriously being worked on now [11:47:00.0000] <jgraham> annevk42: It has been "being worked on" for a long time. There have been drafts since at least 2000 right? [11:47:01.0000] <annevk42> I don't think so [11:48:00.0000] <annevk42> I think flex is really cool, but there's nobody to really drive it :/ [11:48:01.0000] <krijnh> Wasn't Hixie planning to pull a WHATWG on the CSSWG? [11:49:00.0000] <jgraham> It is pretty sad that Mozilla havn't managed to drive it to completion [11:49:01.0000] <annevk42> Though we do have a draft of it now: http://dev.w3.org/csswg/css3-flexbox/ [11:49:02.0000] <krijnh> Somewhere around 2023 [11:49:03.0000] <annevk42> But that is the old draft [11:49:04.0000] <annevk42> krijnh, we need an editor first :) [11:50:00.0000] <krijnh> By that time we can clone people [11:50:01.0000] <jgraham> annevk42: Is dbaron not working on it anymore? [11:51:00.0000] <annevk42> jgraham, I believe he is looking into it now, yes [11:51:01.0000] <jgraham> awesome [11:51:02.0000] <jgraham> I hope he is able to finish it soon [11:53:00.0000] <annevk2> problem with flex is that it's 1D; it's not grids [11:54:00.0000] <annevk2> there was some 2D flex proposal floating around based on tables, but only in email and very vague [11:54:01.0000] <jgraham> Does XUL have more than that? [11:54:02.0000] <jgraham> If webapps can have good enough layout primitives to do everything that XUL can do it is a huge win [11:54:03.0000] <jgraham> even if better ones would be nice [11:55:00.0000] <annevk2> XUL has some kind of weird grid model besides flex [11:55:01.0000] <annevk2> from what I heard we don't wanna copy that [12:00:00.0000] <krijnh> What's wrong with position: inline-absolute? http://www.shauninman.com/archive/2006/05/22/clearance_position_inline_absolute [12:09:00.0000] <_trace> question re:html5... [12:09:01.0000] <_trace> any good examples on 'article' and 'section' element usage? [12:10:00.0000] <_trace> should I be using a single 'article' element with multiple 'section's for each article on the same page, or multiple 'article's? [12:11:00.0000] <jgraham> _trace: multiple <article> elements [12:11:01.0000] <jgraham> One per article [12:11:02.0000] <hober> _trace: "for each article" suggests multiple <article>s [12:11:03.0000] <jgraham> <section> for subsecions of the article, if any [12:11:04.0000] <_trace> jgraham: thanks, that's what I thought, but I was seeing conflicting info in various articles [12:12:00.0000] <_trace> hober: thanks [12:12:01.0000] <jgraham> _trace: Got any examples of articles you found confusing? [12:15:00.0000] <_trace> well, http://www.alistapart.com/articles/previewofhtml5/ didn't specifically mention multiple articles, http://www.w3schools.com/tags/html5_article.asp mentioned 'article' was for content from an 'external' provider... [12:15:01.0000] <jgraham> Oh wow. w3schools ftw [12:16:00.0000] <jgraham> (it is totally misleading) [12:16:01.0000] <_trace> but the current draft of the HTML5 spec was vague, but I read it as meaning a single 'article' per article... [12:16:02.0000] <_trace> yeah, I ignored that since I didn't read the HTML5 draft that way at all [12:16:03.0000] <jgraham> If you have suggestions for clarifying the spec, you should talk to Hixie [12:17:00.0000] <_trace> jgraham: excellent [12:23:00.0000] <annevk2> krijnh, looks complex [12:27:00.0000] <hober> I only use Google's SearchWiki feature to remove w3schools results... [12:28:00.0000] <annevk2> the fact is though that nobody has come up with a good competitor to w3schools [12:28:01.0000] <annevk2> you'd think that a wiki/Web 2.0 like approach would be an instant hit [12:30:00.0000] <annevk2> maybe the competition gets all up in terminology all the time though and competes using "element" rather than "tag" [12:31:00.0000] <krijnh> annevk2: complex? [12:31:01.0000] <krijnh> :S [12:34:00.0000] <annevk2> krijnh, besides it doesn't do a whole lot of things that grids/flex would allow for [12:34:01.0000] <sayrer> grids/flex god [12:34:02.0000] <krijnh> True [12:34:03.0000] <sayrer> please [12:35:00.0000] <sayrer> I can't stand to read another article on lame CSS to simulate that [13:12:00.0000] <rubys> there is a required audio format, at the moment there is no required video formats, is (are?) there any required image format(s)? [13:15:00.0000] <annevk2> they're not listed [13:15:01.0000] <annevk2> but I believe the plan is to list them at some point [13:17:00.0000] <annevk2> PNG is required for <canvas> [13:17:01.0000] <annevk2> at this point [13:18:00.0000] <rubys> png as an output* format [13:18:01.0000] <annevk2> good point [13:19:00.0000] <jgraham> We should probably list gif/jpeg/png since they are required to support the web [13:19:01.0000] <annevk2> (though everyone expects roundtripping to work of course) [13:19:02.0000] <rubys> svg's status seems unclear [13:20:00.0000] <annevk2> unclear how? [13:21:00.0000] <rubys> can a browser correctly claim to support HTML5, but support only gif/jpeg/png as image formats? [13:21:01.0000] <jgraham> An image format in <img>? [13:22:00.0000] <jgraham> I don't see why not at the moment [13:22:01.0000] <annevk2> SVG and MathML are part of HTML5 in a way [13:22:02.0000] <sayrer> I think it needs to support ico [13:22:03.0000] <rubys> at the moment, that appears to also be true for png; but I'm not asking about "at the moment", but the intent [13:22:04.0000] <jgraham> I don't know if we have defined what should happen for svg-in-<img> yet [13:23:00.0000] <jgraham> sayrer: .ico in <img> or for favicons? [13:23:01.0000] <sayrer> jgraham: hmm, not sure about img, but browsers can also view favicons in the content window [13:23:02.0000] <jgraham> (svg-in-<img> has differnt behaviour compared to <svg>-in-html with respect to e.g. scripting) [13:24:00.0000] <rubys> jgraham: if it has any behavior at all :-) [13:24:01.0000] <annevk2> heh [13:24:02.0000] <jgraham> sayrer: How? Is it needed for webcompat? You could make a browser today that didn't have favicons [13:25:00.0000] <jgraham> and the UI would probably suck but otherwise it would be OK, no? [13:25:01.0000] <annevk2> I'm not sure what the complete list is rubys, but I imagine APNG to be there [13:25:02.0000] <annevk2> and BMP [13:27:00.0000] <annevk2> heh, this <nfsw> thing has a whole share of followers suddenly [13:27:01.0000] <annevk2> e.g. http://twitter.com/corewarrior/statuses/2107076157 [13:28:00.0000] <annevk2> it's also being reported as part of HTML5: http://twitter.com/silner/statuses/2106913972 [13:28:01.0000] <rubys> s/nfsw/nsfw/ [13:29:00.0000] <jgraham> We totally need a National Freedom of Speech Week tag [13:31:00.0000] <annevk2> and also nfslw then [13:32:00.0000] <annevk2> to please our anti-fanboy [13:32:01.0000] <jgraham> /me can't work out all the letters [13:33:00.0000] <annevk2> Last Week! [13:33:01.0000] <jgraham> It was the nfs I was having trouble with... [13:34:00.0000] <annevk2> on the one hand it seems kind of bad for HTML5 to require specific formats, but on the other hand not doing it for things like e.g. <canvas> would be very impractical [13:34:01.0000] <annevk2> oh heh, nsflw [13:34:02.0000] <annevk2> jgraham, I just based it on yours [13:35:00.0000] <jgraham> Oh OK. I was hoping for "Not Safe For Mr Last Week" [13:36:00.0000] <jgraham> <nsfmlw>Hixie is my God and the W3C sucks</nsfmlw> [13:36:01.0000] <jgraham> That sort of thing [13:36:02.0000] <jgraham> Then when he read the logs it would appear as "Mr last week rules!" [13:39:00.0000] <annevk2> it's amazing there's still 73 people in this channel with all the nonsense that's going around :p [13:41:00.0000] <jgraham> I assume there are 71 people ignoring this channel [13:42:00.0000] <annevk2> prolly true, although I'm here twice [13:57:00.0000] <Hixie> did the pfwg just announe their intention to violate w3c process? [13:57:01.0000] <gsnedders> /me points out that we did that ages ago [13:58:00.0000] <Hixie> we did? [13:59:00.0000] <jgraham> Hixie: Pointer? [13:59:01.0000] <gsnedders> Is not basing decisions on their logical merits against the process as it isn't consensus? [13:59:02.0000] <Hixie> http://www.w3.org/mid/20090610153947.GE3371⊙srn [14:00:00.0000] <jgraham> gsnedders: We haven't published anything at the W3C without consensus [14:00:01.0000] <Hixie> gsnedders: but that's not the process the wg is following, that's only the process i'm following to make my chartered first proposals [14:00:02.0000] <gsnedders> That is true. [14:02:00.0000] <jgraham> is insure an en-US spelling of ensure? [14:04:00.0000] <Hixie> it appears to be an often used alternative spelling in US-english [14:04:01.0000] <Hixie> (insure also means to obtain insurance, but that's separate) [14:05:00.0000] <Hixie> 37 new e-mails overnight. Man. Every day 100 steps forward, 50 steps back. [14:05:01.0000] <jgraham> It's one of those things that grates ever so slightly when I see it even though I know that is an illogical reaction [14:05:02.0000] <jgraham> (insure that is, not new email) [14:05:03.0000] <gsnedders> /me wonders whether the latest 50 emails on public-html are worth reading [14:06:00.0000] <Hixie> i don't know, i just put them on my pile and will read them next month or so [14:07:00.0000] <gavin_> "insure" as "ensure" bothers me too [14:11:00.0000] <annevk2> please tell me if I make a mistake like that [15:14:00.0000] <Lachy> people who use insure in place of ensure simply don't know how to spell. Anyone who claims it's a valid alternative spelling is wrong [15:14:01.0000] <roc> many dictionaries do [15:16:00.0000] <Hixie> i was in fact quoting the dictionary macs ship with [15:17:00.0000] <gavin_> see usage note at http://www.answers.com/topic/assure [15:22:00.0000] <Lachy> Wow, that's even crazier than people who spell "percent" as "per cent". My Australian dictionary only has the insurace related meanings for insure. [15:25:00.0000] <gsnedders> Lachy: You realize in en-gb per cent is normal? [15:25:01.0000] <Lachy> gsnedders, yes, it's the one case where en-gb is stupid [15:26:00.0000] <jgraham> Lachy: You realise that once upon a time percent was probbly spelt per cent, right? People get things "wrong", language changes, grammer nazis find something new to complain about [15:26:01.0000] <gsnedders> Lachy: No, it's just we haven't got around to putting it together yet, right? [15:26:02.0000] <jgraham> probably [15:26:03.0000] <gsnedders> Lachy: It was once everywhere two words [15:26:04.0000] <Hixie> unless en-au has a normative spec, there's not much point arguing over what is right [15:27:00.0000] <Hixie> that's why i have a normative spec for en-gb-x-hixie :-) [15:27:01.0000] <Lachy> it makes no sense as two words because percentage is one word [15:27:02.0000] <gsnedders> /me points out there is o normative spec for en-gb either :P [15:27:03.0000] <Lachy> actually, en-au is wrong on that case too [15:27:04.0000] <gsnedders> en-gb-x-sneddy really ought to have a spec [15:27:05.0000] <jgraham> Woah, back up there. You seem to be trying to apply logic to spelling rules [15:27:06.0000] <Philip`> Lachy: "percentage" makes no sense as one word because "per cent" is two [15:28:00.0000] <jgraham> It isn't going to work [15:29:00.0000] <Lachy> it also makes no sense because the word cent on its own has its own meaning as a unit of currency, and when you write it as per cent it makes it appear as though it's referring to the currency, not % [15:29:01.0000] <jgraham> Fun fact: There is research into double negatives that shows that languages go in cycles on whether double negatives are used for negation or not [15:30:00.0000] <Philip`> Lachy: You think words should only ever have one meaning? [15:30:01.0000] <Lachy> no, I didn't say that [15:30:02.0000] <jgraham> So a language will start with single negation and people will start using double negaion for emphasis. Then people will again start to favour single negation for brevity and the cycle will repeat [15:30:03.0000] <jgraham> Or something [15:31:00.0000] <gsnedders> Lachy: cent also means one hundred :P [15:31:01.0000] <gsnedders> (At least in Latin, and Latin is a proper language) [15:31:02.0000] <Lachy> Philip`, consider the phrase "I got 5 per cent". Depending on context, that could mean "I got 5 [items] per cent (currency)" or "I got 5%" [15:32:00.0000] <Lachy> writing it without the space removes the ambiguity and makes it easier to read [15:33:00.0000] <Philip`> ("per cent" is apparently an abbreviation of "per centum", so the two-word form seems the more original) [15:34:00.0000] <jgraham> gsnedders: The phrase "proper language" used without a smilie is sure to raise ire [15:34:01.0000] <Philip`> Lachy: Trying to make all sentences unambiguous independent of context seems like a battle that was lost millennia ago [15:35:00.0000] <Philip`> Context matters, and this isn't an example that seems ambiguous when you do have context [15:36:00.0000] <jgraham> I would wager that thre has never been a language that could be understood in an entirely context free way [15:36:01.0000] <jgraham> Because the difficulty of arriving at such a thing would be huge yet its value would be rather small [15:38:00.0000] <Philip`> Not even Lojban [15:43:00.0000] <jgraham> I guess languages with no (native) speakers don't count. But even so I wonder if it is truly context free 2009-06-11 [17:35:00.0000] <Hixie> location.reload() always does a GET? [20:42:00.0000] <bosky101> please leave in your comments for http://news.ycombinator.com/item?id=652202 - a library launched by my startup today [20:42:01.0000] <bosky101> and would anyone knwo the channel for talk on the semantic web [20:42:02.0000] <jcranmer> don't we hate that here? [20:43:00.0000] <bosky101> in anycase i'm not sure in which cateogry start.hover.in lies in ...its basically about links using the URI to render behaviour within a webpage [20:44:00.0000] <bosky101> <a href="http://onhover.hover.in/hoverlet/hover.in/relatedyoutube/election"> election videos</a> will therefore render youtube videos when you hover over the link [20:44:01.0000] <bosky101> available events are onhover and onclick [20:45:00.0000] <bosky101> and there are around 20 kinds of return types available now, looking to build a community of it [03:53:00.0000] <RdeWilde> Q: Anyone can tell me where's the right place to ask this question, or has an answer to it? I'd like to know if this header: http://pastebin.ca/1456435 is somehow legal http 1.1 (as in specs). Tried to figure out, but not sure yet.. [04:16:00.0000] <archtech> which part of a tag is considered the "tag body" actually: <foo>bar</baz> [04:16:01.0000] <archtech> it is <foo> or bar [04:16:02.0000] <jgraham> archtech: WWhat are you quoting? [04:16:03.0000] <jgraham> I don't recall hearing the term "tag body" [04:21:00.0000] <archtech> jgraham, kinda myself. I was trying to cauge if the term is intuitive [04:21:01.0000] <archtech> jgraham, it's confusing to talk about "content type of the content, and content type of the attributes" [04:21:02.0000] <archtech> content type of the body and the attributes, now that is kinda more natural [04:25:00.0000] <archtech> gauge I mean [04:26:00.0000] <jgraham> /me has never found himself talking about the "content type" of the content or attributes [04:28:00.0000] <archtech> jgraham, well, you know, inline, block, literal... [04:28:01.0000] <archtech> say <script> is a literal [04:28:02.0000] <archtech> an implicit CDATA if you will [04:36:00.0000] <jgraham> Oh content model. [04:36:01.0000] <jgraham> /me should have inferred that [05:13:00.0000] <annevk2> /me reads backlog; spelling, startups, and a question about HTTP... [05:16:00.0000] <jgraham> annevk2: See /topic [05:17:00.0000] <annevk2> I lost that long ago [05:17:01.0000] <annevk2> In other news, http://norman.walsh.name/2009/06/10/ajax is interesting considering the Oracle proposals in WebApps [05:25:00.0000] <MikeSmith> annevk2: yeah [05:27:00.0000] <jgraham> /me wonders what the Oracle proposals are [05:28:00.0000] <gsnedders> Yay. Adobe have now cancelled my order for not providing proof of eligibility, despite it waiting around on their website the entire time. [05:28:01.0000] <gsnedders> grumblegrumblegrumble… [05:35:00.0000] <annevk2> hmm, I now get Swedish spam [06:12:00.0000] <gsnedders> /me would expect them to only cancel an order after the time limit if it wasn't waiting for _them_ to respond [06:21:00.0000] <gsnedders> /me is meant to provide non-sensitive photo ID [06:21:01.0000] <gsnedders> The only photo ID I have is a passport which is sesitive… [06:21:02.0000] <gsnedders> /me grumbles… [06:22:00.0000] <beowulf> /me wonders what a non-sensitive photo id would be? [06:23:00.0000] <Lachy> gsnedders, how do you define what is and isn't a sensitive photo ID? [06:23:01.0000] <gsnedders> Lachy: http://kb2.adobe.com/cps/407/kb407013.html [06:23:02.0000] <beowulf> maybe it's a photo he's not sensitive about? you know, nice haircut, cheeky grin [06:27:00.0000] <gsnedders> /me considers taking the fuck-it option and submitting a copy of passport… [06:29:00.0000] <hsivonen> gsnedders: yay for trying to be a paying customer [06:29:01.0000] <gsnedders> hsivonen: And then I'll end up pirating because they refuse to take my order! [08:16:00.0000] <hsivonen> /me wonders what kind of UI OOo will have for the ODF 1.2 fork of RDFa [08:16:01.0000] <annevk42> office formats are such a mess [08:17:00.0000] <annevk42> people say HTML is bad, but seriously [08:18:00.0000] <jgraham> /me bets on a UI that eposes a feature set that could easilly be replicated by a much simpler underlying technology [08:19:00.0000] <hsivonen> Word has a very CSS-like layout model, but the UI for tweaking the properties is horribly badly exposing the underlying model [08:20:00.0000] <annevk42> From CSS WG meetings I get the impression MS Office uses the same layout engine as MSIE [08:20:01.0000] <Rik|work> I'd like a word processor that would help to do structure first then style, like what good web devs do [08:21:00.0000] <annevk42> You can do that with MS Office [08:21:01.0000] <jgraham> Right but almost no-one actually does [08:21:02.0000] <annevk42> Just takes a bit of learning. Personally I write my documents in plain text or HTML :) [08:22:00.0000] <Rik|work> annevk2: you can, but the UI doesn't encourage you to do so [08:22:01.0000] <Lachy> jgraham, I do [08:22:02.0000] <Lachy> that is, when I use Word, which I don't do so frequently anymore [08:24:00.0000] <Lachy> the problem with Word is, from what I've seen of Word 2007, is that they took a backwards step with regards to encouraging structure-based editing instead of presentational editing [08:27:00.0000] <jgraham> Lachy: You don't really count as the inverse of "almost no-one" [08:28:00.0000] <jgraham> Plus, even hen people do use it, the temptation to use the bold and italics buttons is still rather strong e.g. the ES5 fraft has a bunch of things that could usefully be some sort of inline style but afaict aren't [08:28:01.0000] <jgraham> (it's quite hard to tell becase the document is read-only and I don't know how to change that) [08:29:00.0000] <jgraham> (and it would be useful if they were some sort of inline class and if that got preserved in the HTML export because then it would be easier to hypertextulise the document in a post-hoc fashion) [08:30:00.0000] <Lachy> I never had any temptation to use them for normal editing. I only used them when customising the style of the available structures [08:33:00.0000] <jgraham> /me realises that the file system permissions on the document were set to read only [08:37:00.0000] <Lachy> It's interesting that these two articles about the same scientific calculations chose to focus on two different aspects. 1) Mars being ejected from the solar system and 2) Mars colliding with Earth. - and both articles mention the other issue as a minor note. [08:37:01.0000] <Lachy> http://arstechnica.com/science/news/2009/06/kicking-a-planet-out-of-the-solar-system-physically.ars [08:37:02.0000] <Lachy> http://news.bbc.co.uk/2/hi/science/nature/8093005.stm [08:38:00.0000] <Lachy> what's even more interesting is that both conclusions being so radically different from each other indicate that the scientists really have no clue what's going to happen. [08:39:00.0000] <hsivonen> whee! Windows CE C++ compiler claims "UNICODE" to itself as a reserved word [08:39:01.0000] <hsivonen> or built-in constant [08:40:00.0000] <annevk42> Lachy, they are about the same article and seem to report different aspects of it [08:41:00.0000] <Lachy> annevk42, yeah, that's basically what I said [08:41:01.0000] <annevk42> Lachy, the BBC focuses on the anomaly and Ars on what is more likely to happen [08:41:02.0000] <annevk42> Lachy, no, you said it indicates scientists have no clue what is going to happen which seems like the wrong conclusion [08:41:03.0000] <Lachy> no, read the first sentence of what I wrote [08:44:00.0000] <jgraham> Actually the "no clue" thing is probably more accurate. But not in a negative way [08:45:00.0000] <jgraham> You can run these models and get an answer but they are obviously rather densitive to the initial conditions which you can't know precisely, and the exact dynamics, which you can't know precisely [08:45:01.0000] <jgraham> So the best that you can do is get a grasp on some of the possibilities [08:46:00.0000] <hsivonen> If I want to get proper hosting for a hg repo, what do I need to know about Google Code and BitBucket? [08:46:01.0000] <jgraham> and understand that things are far from being as simple as "all the planets go around in nice perfer ellipses from now until the end of time" [08:46:02.0000] <hsivonen> should I consider something else, too? [08:47:00.0000] <jgraham> hsivonen: AFAIK html5lib has not had any serious problems using Google code. Although the issue with the @ sign in email addresses is annoying. [08:48:00.0000] <jgraham> You get 1GB of space by default (I think) [08:48:01.0000] <jgraham> The main restriction is that you have to choose one of their approved licenses [08:48:02.0000] <hsivonen> oh, right. that [08:48:03.0000] <hsivonen> V.nu is a bit of a mix of licenses [08:49:00.0000] <annevk42> put it in the public domain! [08:49:01.0000] <Lachy> what's the issue with the @ sign in email addresses? [08:49:02.0000] <hsivonen> annevk42: It's not because of me that it is a mix of licenses [08:50:00.0000] <hsivonen> Lachy: the supposedly copy-pasteable snippet is bogus and you need to double-escape @ in your hgrc [08:50:01.0000] <jgraham> hsivonen: bitbucket is 150Mb free and you have to pay for more. Although there might be some open source plans if you contact them [08:51:00.0000] <hsivonen> jgraham: have you used bitbucket? [08:51:01.0000] <hsivonen> does the Google Code license policy extend to dependency jars? [08:51:02.0000] <jgraham> hsivonen: No. I was thinking of moving html5lib just as google announced hg support [08:52:00.0000] <hsivonen> are other projects violating or following rules by checking in differently-licensed 3rd party jars? [08:53:00.0000] <jgraham> /me is tempted to clone the html5lib repository on bitbucket anyway [08:55:00.0000] <ezyang> hsivonen: You could cheese it and have your build script download the jars [08:56:00.0000] <hsivonen> ezyang: I still want to check in MIT and BSD code to one repo [08:56:01.0000] <ezyang> Oh, ok, so it's not just JAR problems [08:56:02.0000] <hsivonen> ezyang: also, the C++ translator uses the Mozilla Tri-license [08:57:00.0000] <annevk42> the ToS doesn't say much about having to adhere to license policies [09:00:00.0000] <gsnedders> /me needs to get email working on VPS [09:11:00.0000] <hsivonen> is HTML+TIME used on public sites at all? [09:11:01.0000] <hsivonen> is it used on intranets? [09:13:00.0000] <Lachy> is it supported by any browser? [09:14:00.0000] <Lachy> or at least a plugin? [09:28:00.0000] <annevk42> Lachy, IE [09:30:00.0000] <gsnedders> /me laughs at Hixie's reply to him last night :) [10:30:00.0000] <hober> http://upload.wikimedia.org/wikipedia/commons/e/e4/Web_development_timeline.png [10:32:00.0000] <gsnedders> That says quite a bit about how client side open web stagnated [10:33:00.0000] <hober> It's weird that the XHTML2 WG is on there, but not the HTML WG (nor WHATWG). Then again, it's odd the XHTML2 WG is on there at all. [10:34:00.0000] <gsnedders> Where's XHTML2 on that? [10:34:01.0000] <gsnedders> Oh, there, at the top [11:59:00.0000] <annevk42> weird that Opera is placed so late [12:46:00.0000] <gsnedders> /me ponders whether to get open or closed headphones [12:47:00.0000] <gsnedders> /me isn't sure whether he really wants the bulk of closed ones, but could probably do with no sound leakage [12:48:00.0000] <jgraham> gsnedders: For what purpose [12:49:00.0000] <gsnedders> jgraham: For listening to music when I can't practically carry speakers around with me :P [12:50:00.0000] <jgraham> gsnedders: Right. But are you going for "I have to work in an office and I want to cut out background noise" headphones or "I'm a wanker^H audiophile headphones and I want the best possible sound quality" [12:50:01.0000] <jgraham> ? [12:50:02.0000] <gsnedders> jgraham: the latter. [12:50:03.0000] <gsnedders> jgraham: But the former is useful too :P [12:50:04.0000] <jgraham> Oh well I can't help you then. [12:50:05.0000] <gsnedders> Hence my issue. [12:50:06.0000] <gsnedders> jgraham: Are you allowed headphones there? [12:50:07.0000] <jgraham> Yes [12:51:00.0000] <gsnedders> Then I guess I probably should get closed ones :P [12:51:01.0000] <gsnedders> (More for the sound leakage POV than incoming sound) [12:54:00.0000] <jacobolus> gsnedders: get open headphones [12:54:01.0000] <jacobolus> gsnedders: in fact, just go pick out the set of grados that are in your budget :) [12:54:02.0000] <gsnedders> jacobolus: :) [12:54:03.0000] <jacobolus> gsnedders: these are great, I think http://www.gradolabs.com/product_pages/sr60.htm [12:55:00.0000] <jacobolus> definitely for the price [12:58:00.0000] <gsnedders> jacobolus: I had been looking at the Sennheiser HD 595 or 555 [13:00:00.0000] <archtech> Can someone comment on what is going on with XHTML2? [13:00:01.0000] <annevk42> http://twitter.com/jdowdell/statuses/2120530436 -- isn't this exactly what we're doing? [13:01:00.0000] <annevk42> archtech, this is not the right place for that question, we're not working on it [13:01:01.0000] <archtech> annevk2, all right. I was just wondering if it was abandoned completely in favor of HTML5.. [13:02:00.0000] <Hixie> archtech: no, there is still a working group working on xhtml2 [13:02:01.0000] <Hixie> archtech: i believe they are even planning on releasing an updated draft soon [13:03:00.0000] <archtech> Hixie, I see. [13:03:01.0000] <gsnedders> /me thought they were planning on releasing an updated draft before TPAC last year [13:03:02.0000] <archtech> Is HTML5 intending to move towards paragraphs not just for inline content? [13:04:00.0000] <archtech> Like XHTML2 is, apparently. I can't understand that concept. [13:04:01.0000] <Hixie> archtech: you mean like <p> containing <ol>? [13:05:00.0000] <archtech> Hixie, yes, that is one of the examples in the XHTML2 draft it seems [13:05:01.0000] <archtech> However I have yet to see a documentation system that has this kind of concept. [13:05:02.0000] <Hixie> archtech: we actually tried that a few years ago, but we couldn't get it to work in a way compatible with browsers today [13:06:00.0000] <archtech> Hixie, if compatibility wasn't a problem, I wonder if it would work conceptually? [13:06:01.0000] <Hixie> archtech: yeah, seems like a pretty reasonable idea (though it makes WYSIWYG editing much harder to do) [13:09:00.0000] <archtech> Does LaTeX define something like that? [13:09:01.0000] <Hixie> no idea [13:09:02.0000] <archtech> Thanks. [13:09:03.0000] <gsnedders> jacobolus: Peh. Looking at a Grado compared with the Sennheiser, the Senheiser does better at the bass end, but then the Grado has too much treble and the Sennheiser too little :( [13:10:00.0000] <gsnedders> archtech: LaTeX has no concept of a list being part of a paragraph, from memory. [13:10:01.0000] <jacobolus> gsnedders: what do comparison reviews say? [13:11:00.0000] <gsnedders> jacobolus: That there's little to choose between them :P [13:11:01.0000] <gsnedders> (in terms of quality) [13:11:02.0000] <gsnedders> /me is tempted by the Sennheiser mainly because it is more comfy [13:11:03.0000] <jacobolus> gsnedders: if you want good bass, you should go for the SR80s, and buy an amp to plug in between your laptop/ipod/whatever and your headphones [13:12:00.0000] <gsnedders> jacobolus: The Sennheiser HD 595 seems to do better with bass with/without amp [13:12:01.0000] <gsnedders> jacobolus: But I don't care that much about bass alone [13:14:00.0000] <gsnedders> Generally I'm still tempted by the Sennheiser [13:14:01.0000] <jacobolus> go for it!' [13:14:02.0000] <jacobolus> I'm sure you'd be plenty happy with them [13:15:00.0000] <jacobolus> those are pretty pricey, no? [13:15:01.0000] <gsnedders> No [13:15:02.0000] <archtech> gsnedders, it seems yes, LaTeX defines lists as their own block elements. [13:15:03.0000] <archtech> I don't know. [13:15:04.0000] <gsnedders> Oh, wait, from having been looking at both the HD 595 and HD 555 I've got mixed up between the two :P [13:15:05.0000] <archtech> Maybe it's a better model to keep it as is, anyway. [13:16:00.0000] <jacobolus> (well, by my super-cheap standards; amazon has $260 marked down to 178) [13:16:01.0000] <gsnedders> archtech: LaTeX is really aimed to get the semantics needed to print correctly, it isn't concerned with actual document semantics, so it makes no sense for LaTeX to support both. [13:16:02.0000] <jacobolus> archtech: LaTeX doesn't care about semantic purity [13:16:03.0000] <archtech> It doesn't, but semantics bear some overlap with presentation. [13:17:00.0000] <archtech> I can't conceptually wrap my mind around a list in a paragraph. [13:17:01.0000] <jacobolus> but inline lists don't need special structure in latex [13:17:02.0000] <archtech> At least the way literary paragraphs are defined. [13:17:03.0000] <gsnedders> The HD 555 I can get quickly (which is good seeming I leave soon). [13:17:04.0000] <gsnedders> And there isn't overly much difference between them and the HD 595 (which cost almost twice as much) [13:17:05.0000] <gsnedders> (£68.41 v. £112.95, which is a lot less than twice) [13:18:00.0000] <gsnedders> /me was rounding both numbers down in his head to nearest multiple of ten :P [13:25:00.0000] <gsnedders> jgraham: There again, I guess people at work would prefer people to have closed headphones :P [13:29:00.0000] <shelleyp> I've posted an email about the WAI CG Consensus Resolutions on Text alternatives in HTML 5 to the HTML WG email list. Hopefully we can pull together the threads from both IRC and twitter into one place [13:32:00.0000] <archtech> Is HTML5 going to tweak the definition of inline/block or leave it as is? [13:32:01.0000] <archtech> I see some voices that say this should be purely a styling concept. [13:32:02.0000] <archtech> And not an element concept. [13:34:00.0000] <rubys> hsivonen: ping? [13:38:00.0000] <annevk42> archtech, HTML5 does not mention inline/block-level [13:38:01.0000] <archtech> annevk2, as in? It doesn't define them or it does away with them? [13:39:00.0000] <annevk42> archtech, it groups elements using different terms [13:39:01.0000] <annevk42> archtech, so yeah, it does away with them, effectively [13:39:02.0000] <archtech> annevk2, which section should I check for more info on this? [13:39:03.0000] <annevk42> seems section 3.4 is about this [13:40:00.0000] <archtech> Thanks :) [13:49:00.0000] <beowulf> so, nice weather out... [13:50:00.0000] <beowulf> the heat generated from accessibility discussions makes me try very hard to not use the img element [13:51:00.0000] <jgraham> gsnedders: Given where you are likely to be sitting I can sey with some certianty that there will be a non zero number of people who will not want to hear your music [13:52:00.0000] <jgraham> Especially if you start listening to christian rock [13:52:01.0000] <gsnedders> /me was expecting that :P [13:54:00.0000] <archtech> annevk2, I can't find how is the <br> tag classified in this section? [13:54:01.0000] <jgraham> Hixie: I don't see how we get to tell people waht is intuitive [13:55:00.0000] <jgraham> We have to guess what is intuitive to make the most usable language possible [13:55:01.0000] <Hixie> language design is all about making choices based on what we think is more intuitive [13:55:02.0000] <Hixie> it's like UI design [13:55:03.0000] <jgraham> Sure. That's not the same as telling people what is intuitive [13:55:04.0000] <annevk42> archtech, br is both flow and phrasing [13:55:05.0000] <Hixie> to some level you can test it, but at the end of the day it's a judgement call, even if you base it on some level of testing [13:55:06.0000] <Hixie> how is it not? [13:55:07.0000] <annevk42> archtech, the element definition lists the categories it is in [13:56:00.0000] <annevk42> archtech, the other way around might be added at some point, but that is currently not in the draft [13:56:01.0000] <archtech> isn't it confusing to have each element in multiple categories though? [13:56:02.0000] <jgraham> Hixie: Because "telling people" implies some level of control over the definition of "intuitive". We have no control, we can only infer what is intuitive (e.g. by studying behaviour) and specify from that [13:56:03.0000] <annevk42> HTML4 had a similar concept, called flow [13:57:00.0000] <archtech> I mean, it's useful for classification in a manual, but seems to not be best suited for implementing HTML5 parsers. [13:57:01.0000] <archtech> annevk2, sure, but it was simpler [13:57:02.0000] <jgraham> (the flow of information is in the opposite direction) [13:57:03.0000] <annevk42> seems natural for things to fit in multiple categories [13:57:04.0000] <annevk42> you don't implement HTML5 parsers based on content models :) [13:58:00.0000] <annevk42> you use the section that specifies parser requirements for that [13:58:01.0000] <annevk42> 9.2 iirc [13:58:02.0000] <Hixie> jgraham: my interpretation was that the guy meant that we don't get to decide on what we think is intuitive [13:59:00.0000] <jgraham> Hixie: That wasn't really my interpretation. At least I think your reply is more confusing than helpful [13:59:01.0000] <Hixie> jgraham: ah well. [13:59:02.0000] <Hixie> jgraham: what did you think he was saying then? [14:00:00.0000] <jgraham> Hixie: I thought he was trying to say "as language designers you can't mandate that something ought to be intuitive even if user behavior shows otherwise" [14:01:00.0000] <Hixie> yay, shelley spilt the beans on where the html4all cabal moved to [14:01:01.0000] <Hixie> jgraham: that seems like a truism, how can anyone mandate what is intuitive? [14:03:00.0000] <jgraham> Hixie: They can't obviously. But they could think that they can. Lots of bad UI designers do, for example :) [14:03:01.0000] <jgraham> s/do/think that/ [14:03:02.0000] <Hixie> i don't think they do, i think they just do exactly what we do, which is make assumptions about what is intuitive [14:03:03.0000] <jgraham> s/think that/seem to think that/ [14:03:04.0000] <Hixie> i don't think they they they do either [14:04:00.0000] <Hixie> er [14:04:01.0000] <Hixie> i don't think they think they do either [14:04:02.0000] <jgraham> Well it's not really important. The point is that it is possible to think that you can do the impossible [14:04:03.0000] <jgraham> (you might back it up with arguments like "we can use education to make people understand our model") [14:05:00.0000] <Hixie> education can't affect what's intuitive, by definition :-) [14:06:00.0000] <jgraham> Right but you often hear people say "x is intuitive once you understand principle y" which can be true or can be false. So people think that by teaching y they can make x intuitive [14:07:00.0000] <Hixie> i guess [14:07:01.0000] <jgraham> (but this is a bit of the point I was originally trying to make) [14:07:02.0000] <jgraham> (which is that I think the comment was meant in a different way to the one in which you interpreted it) [14:08:00.0000] <Hixie> that's certainly possible [14:09:00.0000] <archtech> Is this valid according to HTML5: <b>foo<br>bar</b>? [14:09:01.0000] <archtech> This is a bold tag that crosses theboundary of a line, although that line is an implicit tag [14:09:02.0000] <archtech> defined by a break [14:10:00.0000] <Hixie> aww, their archives are member-only http://wilbur.bytowninternet.com/mailman/listinfo/talk_html4all.org [14:10:01.0000] <Hixie> i guess the cabal doesn't believe in the level of transparency we believe in! [14:10:02.0000] <Hixie> what could they possibly have to hide [14:11:00.0000] <jgraham> I guess irony is lost on some people [14:15:00.0000] <Hixie> right, bbiab, lunch [14:17:00.0000] <jacobolus> archtech: why wouldn't it be? [14:19:00.0000] <archtech> jacobolus, the line break defines two implicit line elements, say I have this: Hello, this is <b>some bold text,<br>and we continue on a new line.</b> Done. [14:19:01.0000] <archtech> Now those are two lines [14:19:02.0000] <archtech> <line>Hello, this is <b>some bold text,</line><line>and we continue on a new line.</b> Done.</line> [14:19:03.0000] <archtech> There's no valid way to nest this. [14:20:00.0000] <jacobolus> the same could be said about the '.' character creating two implicit 'sentence' elements [14:20:01.0000] <jacobolus> etc. [14:20:02.0000] <archtech> My point is that a break is a metacharacter. [14:21:00.0000] <archtech> The actual objects are lines, are they not? [14:21:01.0000] <jacobolus> I don't think so, no [14:21:02.0000] <archtech> I see. [14:21:03.0000] <jacobolus> but I don't really know the browser internals that well [14:22:00.0000] <jacobolus> in any case, no, html5 doesn't to my knowledge mandate that "implicit line elements" not included anywhere in the dom be treated in any special fashion [14:34:00.0000] <archtech> thanks [14:38:00.0000] <shelleyp> Hixie, html4all cabal? Isn't that a pretty silly thing to say? [14:39:00.0000] <shelleyp> You know, you might try communicating with people, rather than make snarky asides in the IRC [14:48:00.0000] <annevk42> jacobolus, per CSS you do get multiple line boxes, but you can get that with a single element as well [14:48:01.0000] <annevk42> (e.g. by just having line wrapping; <br> is nothing more than a forced wrap) [14:49:00.0000] <annevk42> (that's a lie to be honest, <br> has a few other tricks up its sleave) [14:49:01.0000] <jacobolus> annevk42: so you can style a particular line (besides the first)? [14:49:02.0000] <jacobolus> I mean, of course at some level lines, glyphs, etc. are objects [14:50:00.0000] <annevk42> there are proposals [14:50:01.0000] <jacobolus> but are the opposite sides of a <br> actually treated as special, "implicit objects"? [14:50:02.0000] <Philip`> hsivonen: I've used BitBucket to a trivial extent, and it worked trivially and I haven't seen any problems [14:51:00.0000] <annevk42> jacobolus, no, <br> in that sense is just a forced linebreak [14:56:00.0000] <Hixie> shelleyp: i was just using the term that html4all participants themselves used for themselves: http://html4all.org/pipermail/list_html4all.org/2007-August/000004.html [14:57:00.0000] <Hixie> shelleyp: and i have been talking directly to everyone involved; i would involve myself in more discussions but since e.g. html4all is a closed group, it is not possible for me to do so [14:59:00.0000] <shelleyp> Hixie, the context of your use of the term belies your later justification [15:00:00.0000] <annevk42> html4all has two forums I think shelleyp [15:00:01.0000] <annevk42> one public and one private [15:00:02.0000] <annevk42> the former is no longer in use and is the one Hixie is pointing to [15:00:03.0000] <Hixie> i really was just using the term because that's how the group of people involved self-identifies [15:00:04.0000] <shelleyp> OK, and since the html4all group does not have controlling interest over any aspect of any W3C document, what's the point annevk42? [15:00:05.0000] <Philip`> hsivonen: (I suppose there's the question of what you mean by "proper hosting", seeing as Hg isn't reliant on a central repository like SVN and so the requirements can be looser) [15:00:06.0000] <Hixie> if you have a problem with the way the people you are associating with self-identify, i urge you to take it up with them [15:01:00.0000] <annevk42> shelleyp, point? [15:01:01.0000] <annevk42> /me was just trying to explain something [15:01:02.0000] <shelleyp> annevk42, Ah, OK, thanks for the explanation. [15:02:00.0000] <Philip`> hsivonen: (e.g. maybe the goal is that people can write "hg clone some-url-which-won't-change-ever", in which case BitBucket is nice because it can apparently do CNAMEs and you could move the same domain to a different host later) [15:03:00.0000] <shelleyp> Hixie, why you're concerned about this one group, I have no idea. There's plenty of people to communicate with on the HTML WG. [15:04:00.0000] <Hixie> i'm not in the least bit concerned about the html4all group [15:04:01.0000] <shelleyp> But communication has to be more than telling people that no, their idea won't work, and here's what you're going to do instead [15:04:02.0000] <Hixie> i agree entirely [15:05:00.0000] <shelleyp> Actually, I don't want to come across as pugnacious, but I'm not sure that you do agree with me [15:05:01.0000] <Hixie> i agree entirely that ignoring feedback and simply asserting positions is not a good way to participate in any group [15:06:00.0000] <shelleyp> Hixie, one could say that what you just described is quite similar to many of your recent communications with folks [15:06:01.0000] <Hixie> indeed many do, i'm sure [15:06:02.0000] <Hixie> mr last week for instance [15:07:00.0000] <Hixie> though it's interesting that he hides behind a screen of anonymity to do so [15:07:01.0000] <shelleyp> I'm not talking about an anonymous individual. [15:08:00.0000] <Hixie> he is merely one example among many, indeed [15:08:01.0000] <Hixie> maybe even dozens :-) [15:08:02.0000] <shelleyp> No, he's not. The folks I'm talking about are those who participate in the groups, and identify themselves by name [15:09:00.0000] <Hixie> i am talking about all the people who claim that i ignore them despite my writing e-mails literally hundreds of lines long explaining my positions based on detailed research and reasoning [15:10:00.0000] <Hixie> mostly, if not exclusively, people who have found themselves not agreeing with my reasoning, sadly [15:10:01.0000] <shelleyp> Hixie, take a break, have a beer, come back and read the line you just wrote. You just about summed up the entire problem. [15:10:02.0000] <Hixie> i don't drink and i'm on a break right now, but thanks for the suggestion :-) [15:10:03.0000] <jacobolus> shelleyp: but quite often what happens, that I've observed, is that people re-write the same emails, completely ignoring Hixie's responses, and then pretend that he's ignoring them [15:11:00.0000] <jacobolus> it's frustrating even to just watch from the sidelines [15:11:01.0000] <Hixie> the sad truth of the matter is that on many topics, there are people with a variety of opinions, and i could never make the spec agree with them all [15:11:02.0000] <Hixie> so merely disagreeing with any one group doesn't mean i'm ignoring them, it merely means that they were not the group that was most convincing [15:12:00.0000] <shelleyp> Hixie, no one says you should. But when you're the only author of the spec, your being in a position of power means you have to go the extra mile [15:12:01.0000] <Hixie> it's not even that i'm just doing what i want -- there are a variety of things in the spec where the group that "won" is actually the group i disagree with [15:12:02.0000] <Hixie> (many things, in fact) [15:12:03.0000] <shelleyp> jacoblus, then that's incorrect -- people should specifically address Ian's comments in response [15:12:04.0000] <hober> He's already going the extra distance-unit-much-longer-than-a-mile [15:14:00.0000] <hober> I mean, seriously, exchanges of the form 1. initial feedback, 2. hixie replies with what's changed, rationale, etc., 3. "new" feedback which doesn't take anything in (2) into account and is in fact indistinguishable from (1) happen all too frequently [15:15:00.0000] <hober> Most people would stop replying with (2) at all, because all the work going into it doesn't seem to have any effect. Clearly Hixie isn't normal, to continue to engage like he does. [15:15:01.0000] <shelleyp> hober, Hixie has worked his butt off, and I appreciate his effort. But Ian also has a habit of discounting other's expertise, which is disrespectful [15:16:00.0000] <shelleyp> I only checked in because I was not happy to see the term "cabal", especially related to a hot topic like accessibility [15:16:01.0000] <shelleyp> I'm sure this group would rather get back to tech stuff [15:16:02.0000] <shelleyp> But I will remember what was said about people not responding to Ian's comments, and make sure that I do [15:18:00.0000] <sicking_> shelleyp, what i feel happens a lot is that people join and say "Do X. You should listen to me because i'm an expert". What would be much more useful is people saying "here is the argument for why you should do X" [15:21:00.0000] <annevk42> I did not know: http://www.olsenhome.com/gif/ [15:22:00.0000] <annevk42> I'm not sure what to do with this knowledge though, pronouncing it as "gift" seems more understandable... [15:22:01.0000] <shelleyp> sicking: I've not see that behavior. Can you point out where someone said that, specifically? [15:23:00.0000] <Philip`> annevk42: That's just prescriptivist pronunciation - you should ignore it and say it the sensible way with a hard 'g' [15:24:00.0000] <sicking> shelleyp, every time someone tries to argue from authority. If you have expertise you shouldn't have to point it out i'd think. It should be obvious from your well layed out arguments [15:29:00.0000] <jruderman> the simple solution is to switch from GIF to PNG ... except that PNG can also be pronounced in two ways [15:30:00.0000] <annevk42> yeah, they should've called it PING [15:30:01.0000] <hober> PNG gets pronounced like "ping" (and GIF like "jiff") [15:31:00.0000] <hober> What's the other way to say PNG? "pong"? [15:31:01.0000] <annevk42> s/gets/is supposed to be/ [15:31:02.0000] <gavin_> pee en gee [15:31:03.0000] <annevk42> that's how I call it usually [15:32:00.0000] <Philip`> You could pronounce it like "snog" but with a p [15:34:00.0000] <gsnedders> Philip`: Instead of the s? I can't get it out as one syllable then :P [15:34:01.0000] <Hixie> "spog"? [15:34:02.0000] <Hixie> "snpg"? [15:36:00.0000] <Philip`> gsnedders: Yes [15:37:00.0000] <Philip`> gsnedders: It's not any harder than plenty of syllables in other languages :-p [15:50:00.0000] <shelleyp> sicking: your argument makes no sense. I, for one, am more likely to listen to someone with proven experience and expertise in the field, then someone who doesn't [15:51:00.0000] <sicking> shelleyp, right, but to most people on the list, someone saying that they have expertise is not a "proven expert" [15:52:00.0000] <sicking> shelleyp, since they haven't spent time in that field [15:52:01.0000] <shelleyp> sicking: again, I've not seen anyone imply an expertise they don't have -- not in any of the debates I've followed [15:53:00.0000] <sicking> shelleyp, the word "expertiese" have been brought up several times. Though it could be that most of the time in the form of "people on this list isn't listening to experts" [15:53:01.0000] <sicking> shelleyp, i don't want to get into a meta discussion here though, that's just a time waster [15:54:00.0000] <shelleyp> sicking: I agree. If you find specific examples, twitter or email me with them. [15:54:01.0000] <sicking> shelleyp, next time i see it i'll email it to you [15:55:00.0000] <shelleyp> sicking: thanks. [15:58:00.0000] <sicking> shelleyp, i'll note though, that the page you linked to today, from WAI CG, contains very little rationale [15:58:01.0000] <sicking> shelleyp, i happen to agree with most of it, but that's beside the point [16:00:00.0000] <sicking> shelleyp, either they think that the rationale is obvious, which it IMHO it isn't, or they are arguing that we should listen to them because they have expertise. Though I'm hoping it's because of some other reason that I can't think of [16:02:00.0000] <arun_> I have a rough editor's draft of the File API spec, which I'd welcome discussion on. There are a few issues with it still (Callbacks and* ErrorCallbacks have to be handled; editorial nits; sample code nits) but I'd love to get feedback from the channel: http://dev.w3.org/2006/webapi/FileUpload/publish/FileAPI.xhtml [16:03:00.0000] <annevk42> .xhtml? o_O :p [16:03:01.0000] <arun_> annevk42, it was just convenient to edit. [16:03:02.0000] <arun_> I'll change that [16:03:03.0000] <annevk42> I was joking, no worries [16:04:00.0000] <arun_> In particular, I'd like to hear about FileDialog. I think it's a good idea, but I've had nits to the effect that honing <input type="file" ... > behavior is better. [16:05:00.0000] <annevk42> "This interface describes a single file in a FileList," I think something else for this would work better since we might not always want to have them in lists [16:05:01.0000] <sicking> yeah, i had some comments like that too [16:05:02.0000] <Hixie> to repeat the comments i made privately: you'll want to make better use of rfc2119, and i recommend splitting File into two, FileData with everything but the name, and File with the name, where File inherits from FileData [16:05:03.0000] <sicking> it also assumes that lists come from the FileDialog [16:06:00.0000] <annevk42> would FileData be like Blob? [16:06:01.0000] <sicking> that was my question too [16:07:00.0000] <sicking> i'm not sure if Blob will have a mimetype though? [16:08:00.0000] <Hixie> File doesn't have a mime type either right now does it? [16:08:01.0000] <arun_> Hixie, agree on rfc2119. I'm amenable to splitting File into two. You'd want a slice() on FileData at some point, I take it, for "blobs"? [16:08:02.0000] <annevk42> I'd rather tie it to HTMLInputElement and DataTransfer than to FileDialog to be honest [16:08:03.0000] <Hixie> i agree that whatever File inherits from shouldn't have a type [16:08:04.0000] <sicking> Hixie, hmm.. good point, i think they should [16:09:00.0000] <sicking> Hixie, or rather, in the firefox implementation they do [16:09:01.0000] <Hixie> arun_: yeah that would be one of the ways this would get extended [16:09:02.0000] <Hixie> arun_: similarly binary data coming from a WebSocket would be a FileData (or Binary or Blob or whatever you prefer), not a File [16:09:03.0000] <sicking> Hixie, oh, no, they don't [16:09:04.0000] <annevk42> spawning file picker dialogs to users seems open to abuse and we already have somewhat sandboxed way for picking files we can reuse [16:09:05.0000] <arun_> OK; you also suggested that obliging developers to handle both the data callback and the error callback was a stylistic choice. No strong opinions either way? [16:10:00.0000] <sicking> Hixie, ah, now i remember, we figured it was mostly useful for piecing together a data: uri, but since we expose a data: uri we didn't feel that mimetype was that important [16:11:00.0000] <Hixie> arun: that's just an api design decision, no right answer really. personally i design callback-style apis such that they call either the success or the failure callbacks, never both, but it depends on the way you'd expect them to be used. [16:11:01.0000] <annevk42> while we brainstorm, it would be sort of nice to just like with XMLHttpRequest you can get a Document out of it and maybe even HTMLImageElement / ImageData / JSON [16:12:00.0000] <Hixie> getting an HTMLImageElement would be nice, but long term what would be even nicer is being able to get a short (not data:) URL out of it [16:12:01.0000] <arun_> annevk42, I agree it is open to abuse. I had an ednote about conditions where implementations would want to define when to raise a FileError with a SECURITY_ERR, and FileDialog abuse seemed like a prime choice. [16:12:02.0000] <sicking> annevk42, since it's an async API, the UA can show a unintrusive warning if it suspects misuse. A'la GeoLocation [16:12:03.0000] <annevk42> sicking, but why have two ways to do things? [16:13:00.0000] <Hixie> e.g. .getAsUrl() and it returns "magic-file-or-whatever://origin.example.com/0" where that URL works only for browsing contexts in the same origin [16:13:01.0000] <arun_> Basically, major web applications like Flickr and GMail use the Flash uploader. Sure, we could address some of this by changing how user agents handle input type="file", but a programmatic FileDialog seemed useful. [16:13:02.0000] <sicking> annevk42, because a lot of sites are working around the uglyness of the current <input type=file> [16:13:03.0000] <Hixie> that would let us automatically take a FileData object and plug it into anything like <video> or <img> or whatever [16:13:04.0000] <sicking> annevk42, gmail being an obvious example [16:13:05.0000] <annevk42> I'd argue we should solve that with CSS, not ECMAScript [16:14:00.0000] <annevk42> the reason Gmail/Flickr use Flash for uploading is mostly multiple files (<input type=file multiple>) and I suppose async upload (using XMLHttpRequest) [16:14:01.0000] <arun_> annevk42, it's not just ugliness, it's also functionality, but HTML5 does allow input type="file" to have multiple selects, etc. I'm in two minds about this. [16:14:02.0000] <sicking> annevk42, no-one has been able to crack that nut for many many years now [16:14:03.0000] <annevk42> did we even try? [16:15:00.0000] <annevk42> for many years progress on CSS has been so-so [16:15:01.0000] <sicking> styling UI elements have been attempted many times i think [16:15:02.0000] <sicking> webkit is currently attempting scrollbars [16:15:03.0000] <sicking> there has been suggestions of various pseudo-elements [16:15:04.0000] <arun_> Hixie, wouldn't that mean defining yet another URI scheme? [16:16:00.0000] <sicking> Hixie, an alternative would be to have streams [16:16:01.0000] <Hixie> arun_: yes [16:16:02.0000] <Hixie> sicking: well i imagine we'd need streams to use the same feature [16:17:00.0000] <arun_> Hixie, that doesn't seem desirable. I'd like that use case better articulated. [16:17:01.0000] <sicking> Hixie, i agree it's something we should look into, not sure if it needs to be part of this spec though. At least not initial version [16:17:02.0000] <Hixie> sicking: unless we want to add an API to everything that accepts a url :-) (e.g. background-image in CSS, list-style-image... SVG...) [16:17:03.0000] <Hixie> sicking: absolutely, not for this version [16:17:04.0000] <annevk42> Hixie, how can the URL be shorter than with data URLs? [16:17:05.0000] <Hixie> annevk42: it wouldn't include the data [16:18:00.0000] <Hixie> annevk42: just a locally minted pointer to the data [16:18:01.0000] <sicking> Hixie, agreed, we might need both [16:18:02.0000] <Hixie> annevk42: this would not be the kind of URL that ever goes over the wire [16:18:03.0000] <annevk42> oh I see [16:18:04.0000] <arun_> Hixie, hmmm... now I get that better [16:18:05.0000] <sicking> annevk42, magical elves [16:18:06.0000] <annevk42> fancy [16:18:07.0000] <sicking> annevk42, with pixie dust [16:18:08.0000] <sicking> it's pretty awesome [16:18:09.0000] <annevk42> hixie dust [16:18:10.0000] <Hixie> anyway i wouldn't expect this for this version [16:19:00.0000] <Hixie> but i think it's a better solution long term than getting HTMLImageElements out of FileData objects [16:19:01.0000] <sicking> yeah [16:19:02.0000] <sicking> i think we can stick to data-uris for now [16:19:03.0000] <arun_> OK, this is good stuff. I'll leave FileDialog in for now, but clean up File a bit more, maybe break it into two (File, FileData or something) [16:19:04.0000] <sicking> otherwise we'll have to define things like what File.magicalHTMLImageProducer().src returns and such [16:20:00.0000] <annevk42> i'm still opposed to FileDialog fwiw [16:20:01.0000] <annevk42> I'd rather wait and see if <input type=file multiple> + XHR can do the trick [16:20:02.0000] <arun_> annevk42, so's Hixie. But I'm unmoved for now ;-) [16:20:03.0000] <sicking> annevk42, XHR?? [16:20:04.0000] <Hixie> i'm not a fan of FileDialog either, i think leaving the security to the last minute is a bad way of doing things [16:20:05.0000] <sicking> annevk42, you mean CSS? [16:20:06.0000] <Hixie> i don't mind something new, but security needs to be first and foremost in mind [16:21:00.0000] <sicking> Hixie, how is it different from geolocation? [16:21:01.0000] <Hixie> such that the entire design precludes abuse [16:21:02.0000] <annevk42> sicking, people articulated the desire for async upload / fragment upload [16:21:03.0000] <arun_> Hixie, I'm not sure we're quite leaving it to the last minute; we're discussing it out of the starter gate :) [16:21:04.0000] <Hixie> sicking: geolocation doesn't pop up a modal dialog [16:21:05.0000] <sicking> Hixie, neither would this need to [16:21:06.0000] <arun_> Hixie, even here, we're NOT suggesting a modal dialog [16:21:07.0000] <arun_> Hixie, in fact I mention a non-modal one [16:21:08.0000] <Hixie> arun_: what i mean is that you have the API designed before the security solution [16:21:09.0000] <sicking> Hixie, no, we talked about this a lot [16:22:00.0000] <sicking> Hixie, not sure if all possible implementation strategies ended up in the draft, but we have several in mind [16:22:01.0000] <Hixie> arun_: what flash does is a modal dialog. If it's supposed to be something else, then that's better, but the spec needs to make that crystal clear, not just leave it up to hte UA. [16:22:02.0000] <Hixie> arun_: and if it's not what flash does, then it's not clear it addresses the same use case :-) [16:23:00.0000] <arun_> Hixie, ahh, there are two things here. 1. The actual File Picker and 2. The user interaction for permission. But yeah, that must be made clearer. [16:23:01.0000] <sicking> Hixie, what I want to do is to leave it up to the UA to decide what to do, but give a few non-normative possible strategies that the implementation can use [16:24:00.0000] <annevk42> modal or non-modal it's a new API for authors, new UI for users (sounds more complex already) and more complexity for implementors [16:24:01.0000] <annevk42> for something that's already possible [16:25:00.0000] <arun_> annevk42, good points -- I think really explaining the use case will get us on the way to resolution. [16:25:01.0000] <annevk42> the main use case I'm hearing is styling [16:25:02.0000] <annevk42> i'm not at all convinced that's a good reason [16:26:00.0000] <annevk42> there's been no attempt whatsoever by any browser at letting authors style the file input control [16:26:01.0000] <arun_> annevk42, we could say that the existing input type ="file" mechanism in browsers today is ugly and causes developers of big web apps to use Flash. FileDialog is one solution, and doesn't need to be bound to using an element in your application. [16:26:02.0000] <annevk42> people use Flash because they can allow users to pick multiple files [16:26:03.0000] <annevk42> that's far more a reason than styling [16:26:04.0000] <sicking> annevk42, It's a complaint I hear a lot, that people think that <input type=file> is ugly and wanting to use something else. [16:26:05.0000] <arun_> But multiple selections (as spec'd in HTML5) is a big reason. [16:27:00.0000] <sicking> annevk42, i don't think it's the whole truth that people use flash just for multiple [16:27:01.0000] <arun_> annevk42, yes -- multiple selection is a big one, and if that's the prime use case, I agree that it can be solved within input type = "file" ... but I'm not sure it is [16:28:00.0000] <annevk42> sicking, yes, but address the styling problem where it should be addressed [16:28:01.0000] <sicking> annevk42, first of all I think that many people aren't able to select multiple files in a filepicker anyway (requires keyboard shortcuts on windows) [16:28:02.0000] <annevk42> arun_, my suggestion is to first fix multiple selections and async upload for files and then see if it's good enough [16:28:03.0000] <annevk42> sicking, no, you can just drag around with your mouse iirc [16:29:00.0000] <sicking> annevk42, ah, yeah, if files have neighboring names [16:29:01.0000] <annevk42> for photo sites that seems likely [16:29:02.0000] <sicking> annevk42, true [16:30:00.0000] <sicking> annevk42, for gmail, not so sure [16:30:01.0000] <annevk42> depends on what people attach primarily [16:30:02.0000] <annevk42> hixie would be in a better place to answer that one I guess :) [16:30:03.0000] <sicking> annevk42, and again, i'm not convinced people generally know about that [16:30:04.0000] <sicking> annevk42, we can of course ask google :) [16:31:00.0000] <annevk42> but you're talking about users, i was suggesting why web developers want it [16:31:01.0000] <sicking> i mean we can ask google why they use flash [16:31:02.0000] <sicking> i.e. what made them not consider HTML good enough [16:31:03.0000] <annevk42> when i log in to gmail it doesn't use flash [16:34:00.0000] <Lachy> wow, so html4all moved to a whole new secret list now? That explains why I've had no mails from the old list for a few months [16:34:01.0000] <annevk42> but yeah, there should definitely be substantive evidence that we really need to introduce another API for opening a file dialog [16:34:02.0000] <annevk42> imo [16:35:00.0000] <sicking> arun_, it occurred to me that we should specify the .files property on HTMLInputElement as well though [16:35:01.0000] <sicking> arun_, if you aren't already [16:36:00.0000] <Hixie> i'll do that in html5 [16:36:01.0000] <Hixie> i've been waiting for this spec [16:36:02.0000] <Hixie> i have a whole folder of issues to deal with that are blocked on not having this spec [16:38:00.0000] <sicking> Hixie, ah, cool [16:40:00.0000] <ojan> annevk42: I see a flash uploader in gmail. [16:41:00.0000] <annevk42> I don't :) [16:42:00.0000] <sicking> annevk42, do you have flash installed and enabled? They detect if you don't and then just use a normal <input type=file> [16:42:01.0000] <annevk42> yes, when I go to youtube.com I can see videos of weird people [16:43:00.0000] <sicking> weird people = flash. Check! [16:43:01.0000] <sicking> annevk42, note that you have to click on 'attach' for any of this to happen [16:43:02.0000] <annevk42> I know [16:43:03.0000] <annevk42> I checked the source [16:43:04.0000] <annevk42> <input id=":jz" name="f_fvu266070" class="LIODqc" size="50" tabindex="2" type="file"> [16:43:05.0000] <sicking> annevk42, and that there won't be any weird people indicating flash on gmail [16:43:06.0000] <sicking> annevk42, dunno, appears they're failing to detect flash for you then [16:44:00.0000] <sicking> annevk42, WFM [16:44:01.0000] <arun_> annevk42, for me too. It seems to invoke Flash pretty consistently. [16:44:02.0000] <arun_> annevk42, and the experience when i turn off Flash is dramatically different. [16:45:00.0000] <arun_> annevk42, though I should say, I'm a weird person. [16:45:01.0000] <annevk42> to test this I'm using Firefox 3.0 for Ubuntu which is the only browser where I have Flash [16:45:02.0000] <arun_> annevk42, OK, I'll bet there's a "Flash version doesn't do foo on Ubuntu" issue which causes you to always default to the HTML input element for file selection. [16:47:00.0000] <annevk42> anyway, whether Gmail does or does not use Flash does not really sway my opinion one way or another :) [16:49:00.0000] <arun_> annevk42, it shouldn't. i just want a good sense for why. Sure, there's multiple selection and style, but I'm interested in other features. [16:50:00.0000] <arun_> So this has been great. I'm going to tweak it and report back :) 2009-06-12 [17:02:00.0000] <jwalden_> hsivonen: #undef UNICODE /* Stupid Windows CE C++ compiler. / ? [17:10:00.0000] <annevk42> jwalden_, it's a reserved word or something [17:12:00.0000] <jwalden_> annevk42: you sure? my experience is that most collisions are from idiotic macro names in the Windows SDK [17:13:00.0000] <jwalden_> and an #undef takes care of them, as long as you don't need to use the name in a header that might shared a large number of places [17:14:00.0000] <annevk42> that's what https://bugzilla.mozilla.org/show_bug.cgi?id=497580 suggests [17:14:01.0000] <annevk42> /me follows bugs with "html5" [17:21:00.0000] <jwalden_> it really is quite appropriate that the abbreviation for that platform is WINCE [17:21:01.0000] <annevk42> is WINCE Windows Mobile? [17:32:00.0000] <jcranmer> Windows Compact Edition [17:32:01.0000] <jcranmer> so yes [18:47:00.0000] <doublec> Windows Mobile is based on WINCE but the version numbers are different. Windows Mobile 6 is not based on WINCE 6 for example. [18:47:01.0000] <doublec> it's based on WINCE 5 [01:36:00.0000] <jgraham> Finally an <audio> demo: http://evilbrainjono.net/piano.html Bets on when the first song made like this reaches the charts? [02:10:00.0000] <jwalden> /me wishes so many keyboard-using demos didn't have issues when the browser in use has a default action associated with the key in question [02:11:00.0000] <jwalden> find as you type horks things like that pretty hard [02:15:00.0000] <hsivonen> jwalden: yeah, I can probably just #undef UNICODE. silly me. (shows my newbieness with approaching Windows compilation issues) [02:16:00.0000] <jwalden> not being corrupted in that manner is not necessarily bad :-) [02:19:00.0000] <Philip`> Re last night (UTC): Surely a URL that has totally different meanings in different contexts would be violating some fundamental architecture of HTTP? (like the whole "Uniform" bit of "URL") [02:21:00.0000] <hsivonen> Philip`: if it's not an HTTP URL, you wouldn't be violationg HTTP but maybe AWWW [02:22:00.0000] <annevk42> context-dependent identifiers are acceptable [02:22:01.0000] <annevk42> e.g. localhost [02:30:00.0000] <jwalden> heck, all of DNS [04:03:00.0000] <zcorpan> http://forums.whatwg.org/viewtopic.php?t=49#5022 - in a couple of days, another newly registered user will reply with a URL, and i will delete both posts as spam [04:05:00.0000] <zcorpan> http://www.google.com/search?q=%22What+is+the+best+GPS+Navigation+system+I+can+buy+for+my%22 [04:05:01.0000] <zcorpan> /me deletes it now instead [04:43:00.0000] <annevk42> whoever invented sync XHR be doomed [04:45:00.0000] <jgraham> annevk42: We're all* doomed (for some sutiable definition of doomed) [04:55:00.0000] <annevk42> jgraham, that's not helpful [05:02:00.0000] <jgraham> annevk42: I think introducing "imminent doom" as the punishment for inventing bad apis would rapidly deplete the already worryingly empty pool of wwilling specification editors [05:02:01.0000] <jgraham> Although maybe it is a fairer punishment for introducing bad non-standard apis that everyone else than has to adopt [05:02:02.0000] <jgraham> Since it would encourage people to seek peer review [05:09:00.0000] <annevk42> again, not helping! [05:09:01.0000] <annevk42> :) [05:24:00.0000] <hsivonen> /me finds reftests failing due to HTML5 now discarding incomplete tag tokens [05:24:01.0000] <hsivonen> apparently not trying to test incomplete tokens, though [05:26:00.0000] <jgraham> hsivonen: How come the reftests have all these weird dependencies on parsing? [05:26:01.0000] <jgraham> (apart from ones that test parsing) [05:26:02.0000] <hsivonen> jgraham: they are written by people [05:27:00.0000] <jgraham> hsivonen: OK, so it is just bugs in the tests rather than something fundamental [05:27:01.0000] <jgraham> That's basically what I was wondering [05:28:00.0000] <annevk42> from some earlier conversation about this it was mostly autogenerated tests from Philip` that had this issue [05:37:00.0000] <hsivonen> anyway, lacking ">" in the string set as the value of innerHTML is something to watch out for in case it matters for Web compat [05:40:00.0000] <rubys> hsivonen: http://intertwingly.net/blog/2009/06/12/Validator-Nu-on-GCJ [05:41:00.0000] <hsivonen> rubys: cool! [05:43:00.0000] <jgraham> rubys: Did you run the testsuite? [05:44:00.0000] <rubys> against HTML2XML? That's all I have built so far. [05:44:01.0000] <jgraham> Against the htmlparser [05:44:02.0000] <jgraham> /me guesses no then [05:45:00.0000] <jgraham> (Also I make not be making much sense since I don't really know how hsivonen's code is organised) [05:45:01.0000] <rubys> /me mutters people are so impatient. :-) [05:46:00.0000] <hsivonen> /me deals with new reftests that assume <col> becomes a child of <table> [07:07:00.0000] <hsivonen> /me finds a test case failing because <div> breaks out of <p><span> in HTML5 [07:52:00.0000] <gsnedders> /me buys headphones [07:52:01.0000] <gsnedders> /me can't believe he really just spent that much on headphones :P [07:56:00.0000] <Lachy> gsnedders, what type of headphones did you buy and how much? [07:56:01.0000] <gsnedders> Lachy: Audio-Technica ATH-M50 [07:57:00.0000] <gsnedders> just under £120 [07:59:00.0000] <Lachy> why did you need such expensive headphones? [08:00:00.0000] <gsnedders> Lachy: audiophilia? :P [08:00:01.0000] <gsnedders> Lachy: Also needing something with very little sound leakage, and comfy [08:01:00.0000] <Lachy> Do you have exceptional hearing ability, and are you able to hear the difference in quality? [08:01:01.0000] <gsnedders> Lachy: yes [08:02:00.0000] <gsnedders> Lachy: Also, a lot of cheaper headphones (esp. those under £60, the real reason for the high price is getting one which meets all three requirements) have either weak bass of treble which almost anyone will notice [08:03:00.0000] <Lachy> can you also tell the difference between a lossless recording and a high quality, though compressed, 256kbps AAC or 320kbps MP3? [08:03:01.0000] <gsnedders> No, my hearing isn't that good. [08:03:02.0000] <annevk42> whoa, I thought my EUR 30 headphones were expensive [08:04:00.0000] <Lachy> annevk42, that's about the price range of my headphones, I think. They're good enough for me [08:05:00.0000] <gsnedders> /me would get annoyed at the sound quality of more or less all 30 EUR headphones :P [08:11:00.0000] <Lachy> I want to get a good quality headset to replace my USB headset, which instead works with the 3.5mm optical output on my Mac. But I haven't found one yet. [08:23:00.0000] <jgraham> /me has lousy hearing but quite nice 80EURish in ear headphones [08:23:01.0000] <jgraham> Which I bought for the noise reduction rather than the playback quality [08:24:00.0000] <jgraham> Although the build quality is insufficent [08:25:00.0000] <Lachy> anyone know of a good quality, stereo, wireless headset/mic, suitable for general use, like skype calls or youtube videos? [08:39:00.0000] <gsnedders> jgraham: in ear headphones I tend to find uncomfortable, and also normally a lot of sound leakage [08:39:01.0000] <gsnedders> (And also tend to be too small to get any really good quality) [08:43:00.0000] <jgraham> gsnedders: WFM [08:43:01.0000] <ezyang> I have PX 200 headphones: they're small, but surprisingly good quality for their size. [08:44:00.0000] <jgraham> I mean they're not super comfortable for many hour periods but for being on a noisy office or, especially, on a plane they make much more sense than anything else [08:44:01.0000] <gsnedders> ezyang: But they're Sennheiser headphones, so that's not overly surprising :P [08:44:02.0000] <ezyang> I <3 my headphones [08:45:00.0000] <jgraham> /me doesn't understand why gsnedders thinks they wwould have a lot of sound leakage [08:46:00.0000] <gsnedders> jgraham: They tend not to have much to stop the sound from getting out [08:46:01.0000] <ezyang> If it sounds good, does it really matter? [08:46:02.0000] <jgraham> gsnedders: The ones I'm talking about are the ones that are basically an earplug with a headphone in the middle [08:47:00.0000] <gsnedders> jgraham: I know [08:47:01.0000] <gsnedders> ezyang: If you're in an office, you don't want sound leakage. [08:47:02.0000] <ezyang> I have an officemate who plays music on speakers, so YMMV [08:47:03.0000] <jgraham> gsnedders: They have a ear plug thing to stop sound getting out [08:48:00.0000] <jgraham> /me hasn't actually measured this so could be wrong [08:48:01.0000] <jgraham> Also I tend to be able to play music at lower volumes because it is quieter [08:48:02.0000] <gsnedders> /me tends to listen to music fairly quietly too [08:48:03.0000] <ezyang> /me cranks it up [08:53:00.0000] <Lachy> I can't listen to music on headphones and work at the same time. It's too distracting [08:53:01.0000] <Lachy> though, I have less difficulty if the music is on relatively low in the background [09:05:00.0000] <Lachy> Apparently Apple are changing the default gamma in Snow Leopard to match Windows. http://www.tuaw.com/2009/06/12/make-your-displays-gamma-in-leopard-match-snow-leopard/ [09:05:01.0000] <Lachy> so at least in theory, colours should start to look the same on both platforms [09:07:00.0000] <jgraham> Lachy: We had this conversation a few days ago. For images it still depends if the image has a built-in colour profile and if your viewing software supports it or not [09:09:00.0000] <Lachy> jgraham, I know [09:31:00.0000] <jgraham> A hint to people writing testsuites: the more complex you make your harness, the more complex it is to figure out what is actually going on [09:51:00.0000] <anne-m> mpt, if I want Ubuntu to get the wireless driver issues to be sorted out; where would the best place be to donate money? [09:52:00.0000] <mpt> /me blinks [09:52:01.0000] <anne-m> mpt, due to some upgrade my mother's computer suddenly switched to the non-proprietary drivers that made DNS resolving suck bigtime [09:52:02.0000] <mpt> from the Random Questions Department :-) [09:53:00.0000] <mpt> hm, one moment, I'll ask around [09:53:01.0000] <anne-m> mpt, took me a long time (20min or so) to figure out it was a driver issue and how I could switch back to the proprietary driver [09:53:02.0000] <anne-m> mpt, you're basically the only person I know that is closely involved :) [09:55:00.0000] <mpt> anne-m, do you happen to know which driver it is? [09:56:00.0000] <mpt> (Different people work on different drivers) [09:56:01.0000] <anne-m> the proprietary one is madwifi and the non-proprietary one (that I had to disable) is ath5k [09:58:00.0000] <mpt> madwifi and ath5k are both open-source drivers produced by the MadWifi project <http://madwifi-project.org/> [09:58:01.0000] <mpt> "ath5k is a relatively new and emerging driver and ... is intended to replace MadWifi in the long run" [09:59:00.0000] <gauthierm> What events should be fired when dragging the scrubber on a media element? FIrefox alphas fire play/pause events as well as seeking and seeked. Safari fires only seeking/seeked. [09:59:01.0000] <mpt> anne-m, so the answer to your question is <http://madwifi-project.org/wiki/Donations> [10:03:00.0000] <anne-m> thanks mpt (though weird that "Hardware Drivers" says madwifi is proprietary) [10:03:01.0000] <virtuelv> /me applauds the "where can I donate" attitude instead of "xxx sucks" [11:49:00.0000] <thin> how many lines of code to draw a triangle, stroke it black and fill it grey? [11:49:01.0000] <thin> with canvas [11:50:00.0000] <thin> approximately [11:50:01.0000] <annevk42> 1 [11:51:00.0000] <thin> will it someday be possible to generate pixel data with canvas/javascript, drop that into the dom as one of those data nodes, then use that image data as a background image in my css? [11:52:00.0000] <gauthierm> probably already possible, but you'd have to write an encoder to encode the pixel data. [11:55:00.0000] <thin> for certain small images do you think doing what I described might be a faster option than using an actual image file? [11:59:00.0000] <thin> Are there any open source, web-friendly image formats trying to compete with jpg/gif/png like ogg is with video/audio? [11:59:01.0000] <annevk42> I think WebKit has an extension to CSS for that [11:59:02.0000] <annevk42> thin, JPEG/PNG/GIF are not open enough for you? [12:00:00.0000] <thin> wasn't there some hub-bub about gif files a few years back? [12:00:01.0000] <annevk42> But you can already do toDataURL() on your <canvas> and then plug that into your CSS through the CSSOM [12:00:02.0000] <annevk42> thin, I think that patent expired by now [12:00:03.0000] <annevk42> thin, or was invalidated [12:00:04.0000] <annevk42> /me forgot [12:01:00.0000] <thin> are you anne van kesteren? [12:03:00.0000] <annevk42> thin, yes [12:03:01.0000] <thin> cool, I like your blog [12:03:02.0000] <annevk42> :) [12:03:03.0000] <gauthierm> thin: png was created as a response to the gif patent. [12:36:00.0000] <annevk42> donated [12:42:00.0000] <jgraham> /me wonders how donations to small software development efforts works [12:42:01.0000] <jgraham> Like how does the money affect the development [12:43:00.0000] <jgraham> I can understand "here is $x to do y" [12:44:00.0000] <jgraham> But "Here is $x to make it better" doesn't have any obvious impact unless sum($x_i) is enough for the developers to give up (part of) their outside job [12:45:00.0000] <gauthierm> Is anyone here working on the HTMLMediaElement spec? [12:45:01.0000] <annevk42> jgraham, I'd imagine part of the money is used for buying the hardware [12:46:00.0000] <annevk42> jgraham, but I trust them to do the right thing [12:46:01.0000] <jgraham> Oh I guess for a hardware driver project that does make sense [12:46:02.0000] <annevk42> gauthierm, as in editing it? I guess Hixie [12:47:00.0000] <Hixie> gauthierm: i'm editing that part of html5 right now [12:47:01.0000] <jgraham> annevk42: I'm not suggesting you did anything silly. I'm just curious what the dynamics are when one donates to a (pure software) project [12:47:02.0000] <gauthierm> Hixie: Cool. I'm having some trouble with media events when the scrubber is clicked and dragged. [12:48:00.0000] <gauthierm> Some browsers fire a 'pause' when you click and hold the scrubber and others do not. [12:48:01.0000] <annevk42> jgraham, I hope it actually works; I've had issues using my MasterCard online [12:48:02.0000] <gauthierm> Hixie: Is one implementation correct, or is that not covered by the spec? [12:50:00.0000] <Hixie> gauthierm: that's basically a user interface issue, so it depends on whether the browser implementor wants the video to continue playing when you let go of the scrubber or not [12:50:01.0000] <Hixie> gauthierm: or whether they want it to play as you're dragging [12:51:00.0000] <annevk42> Hixie, so the way this "Geopriv" works is that the user sets a bunch of variables and the author then either honors them or does his own thing? [12:51:01.0000] <gauthierm> ok. The extra events are causing a problem for me in Firefox in particular. Perhaps I should bug them to fire extra events when the interface is interacted with. [12:51:02.0000] <gauthierm> afaik, both Firefox and Safari's interface is the same, but the events fired are different. [12:52:00.0000] <Hixie> annevk42: as far as i can tell, except that the settings get set in the browser's ui instead of the page's ui, so the user thinks it's the browser who's going to control it (and who is to blame, of course, when it inevitably gets ignored) [12:52:01.0000] <Hixie> gauthierm: which events? [12:52:02.0000] <annevk42> Hixie, wow :/ [12:52:03.0000] <annevk42> /me sends email [12:52:04.0000] <gauthierm> Hixie: Firefox fires play/pause when you drag the scrubber. Safari does not. [12:53:00.0000] <Hixie> annevk42: note that the proposal isn't really geopriv, it's something else. Geopriv itself is an XML format for transmitting geo data with privacy controls, which i think makes a lot of sense (and has nothing to do with UI) [12:53:01.0000] <Hixie> gauthierm: does safari keep playing the video if you just hold the scrubber without moving it? [12:53:02.0000] <gauthierm> Hixie: checking now [12:53:03.0000] <annevk42> Hixie, mkay, somewhat confusing title then [12:53:04.0000] <Hixie> annevk42: yeah [12:54:00.0000] <gauthierm> Hixie: Safari does pause playback when you drag teh scrubber. [12:56:00.0000] <Hixie> gauthierm: then that's a safari bug, it should be pausing the video in the api too [12:58:00.0000] <gauthierm> ok. It would be nice from a JavaScript developer's perspective if there was a way to know whether the video was paused through clicking the pause button or paused through dragging the scrubber. [12:59:00.0000] <gauthierm> The way it's implemented in Mozilla at least, the two are impossible to distinguish. [13:03:00.0000] <gauthierm> Hixie: If I'm filing a bug with Webkit, is that behaviour specified in section 4.8.10.11? [13:06:00.0000] <Hixie> gauthierm: yeah "Where possible (specifically, for starting, stopping, pausing, and unpausing playback, for muting or changing the volume of the audio, and for seeking), user interface features exposed by the user agent must be implemented in terms of the DOM API described above, so that, e.g., all the same events fire." [13:14:00.0000] <Philip`> /me got "here is $x to do y" for his font subsetting tool, which was nice and unexpected [13:15:00.0000] <Philip`> hsivonen / annevk42: Re URLs: Oh, good point(s) [13:19:00.0000] <annevk42> URLs? My memory is failing me... :/ [13:20:00.0000] <Philip`> annevk42: http://krijnhoetmer.nl/irc-logs/whatwg/20090612#l-305 [13:21:00.0000] <annevk42> aah [13:35:00.0000] <annevk42> Hixie, the second of "similar note in audio codecs section" should have s/audio/video [13:40:00.0000] <Hixie> d'oh [13:41:00.0000] <Hixie> btw [13:41:01.0000] <Hixie> i hate software patents [13:41:02.0000] <Hixie> 16bit wave? really? that's the best we can do? [13:41:03.0000] <Hixie> and nothing at all for video? [13:41:04.0000] <Hixie> what bs [13:42:00.0000] <sicking> flac is patented? [13:43:00.0000] <sicking> /me is surprised that microsoft doesn't have patents on wave [13:43:01.0000] <Hixie> apple won't implement codecs that have a submarine patent risk that they haven't already taken on [13:43:02.0000] <sicking> that is bs [13:44:00.0000] <sicking> and a lie [13:44:01.0000] <sicking> everything has submarine patent risk [13:44:02.0000] <annevk42> lots of big companies have taken the risk with vorbis, including Microsoft IIRC [13:44:03.0000] <Hixie> mozilla won't implement codecs that require patent licenses unless they can cover their redistributors, which they can't [13:44:04.0000] <Hixie> opera won't implement codecs that require paying money [13:44:05.0000] <Hixie> sicking: "that they haven't already taken on" is a key part of that sentence [13:45:00.0000] <Hixie> and i'm not arguing about who's right or wrong [13:45:01.0000] <Hixie> because frankly it doesn't matter [13:45:02.0000] <sicking> Hixie, no, i mean implementing cross-site XHR has submarine patent risk [13:45:03.0000] <Hixie> they could not want to implement codecs that didn't start with an "A" and we'd still be in the same boat [13:45:04.0000] <sicking> Hixie, implementing @font-face has submarine patent risk [13:45:05.0000] <Hixie> sicking: argue with them, not me [13:46:00.0000] <sicking> Hixie, implementing anything new has submarine patent risk [13:46:01.0000] <sicking> Hixie, i know [13:46:02.0000] <Hixie> you'll note my employer is implementing everything :-) [13:46:03.0000] <Hixie> ogg theora, vorbis, 264, you name it [13:46:04.0000] <sicking> well, I can't say that I'm happy your employer is implementing things with known patents [13:46:05.0000] <sicking> but that's a separate conversaion [13:46:06.0000] <sicking> well, actually it's related [13:47:00.0000] <Hixie> yeah, well [13:47:01.0000] <Hixie> if the unpatented formats didn't suck... [13:47:02.0000] <Hixie> but anyway [13:47:03.0000] <Hixie> i don't want to argue about it [13:47:04.0000] <Rik|work> Hixie: i really think that's no big deal for the spec, we'll see something similar to what happened with jpg, png and gif [13:47:05.0000] <Hixie> i just want to whine that this makes my life hard from the html5 perspective [13:48:00.0000] <sicking> the thing is, the "we won't implement anything with submarine patent risk" argument is BS. What it really is is "the risk/reward ratio isn't good enough right now" [13:48:01.0000] <Hixie> Rik|work: maybe, though the codec space is _much_, _much_ more patent-encumbered. [13:48:02.0000] <Hixie> sicking: yeah i was simplifying their position for effect, you're right that it's really about risk/reward. [13:48:03.0000] <sicking> we can't do anything to lower the risk part, so we'll have to up the reward part [13:48:04.0000] <Rik|work> gif was patent-encumbered and yet everybody used it [13:48:05.0000] <sicking> which means that we have to get people to use ogg, and only ogg [13:48:06.0000] <annevk42> the GIF patent only covered encoding [13:48:07.0000] <annevk42> not decoding [13:49:00.0000] <Hixie> Rik|work: (a) it wasn't thought to be originally, (b) it was only encoding that was covered, so browsers were safe, and (c) yeah, that's why people are paranoid now [13:49:01.0000] <Hixie> anyway [13:49:02.0000] <Hixie> i should go get lunch [13:49:03.0000] <Hixie> bbl [13:50:00.0000] <sicking> Rik|work, what hixie said. GIF isn't a proof doing patented stuff is fine. It's a lesson costly learned [13:55:00.0000] <gauthierm> Mozilla's doing some good work to make theora suck less. [14:03:00.0000] <sayrer> sicking: the key part is "not arguing about right and wrong" [14:05:00.0000] <sayrer> oh wouldn't it be nice to take an amoral stance regarding <video> [14:06:00.0000] <sayrer> Mozilla could drop Windows Media in there [14:06:01.0000] <sayrer> and we'd have a lot of the Web taken care of [14:09:00.0000] <annevk42> sayrer, isn't that sort of what enabling GStreamer does? [14:10:00.0000] <sayrer> yes, but if we did (have? not sure if we went for it) I would and do openly disagree that it is the right thing [14:11:00.0000] <sayrer> I wouldn't make up some absurd rationalization [14:12:00.0000] <annevk42> not yet apparently [14:12:01.0000] <sayrer> yeah, I don't like it, but I also haven't spent much time working on the mobile product or the video element [14:13:00.0000] <annevk42> (I actually thought it was in, but https://bugzilla.mozilla.org/show_bug.cgi?id=422540 suggests it isn't) [14:13:01.0000] <sayrer> mozilla also doesn't publish mass quantities of h264 video [14:13:02.0000] <sayrer> so there is that [14:14:00.0000] <sayrer> anyway, the whole "some people have paid MPEG-LA, some people haven't, so it's a tossup" argument is just such bullshit [14:15:00.0000] <sayrer> try applying that logic to any number of historical dilemmas [15:08:00.0000] <darbour> I'm having a problem with htmllib5 sanitizer if anyone can help [15:09:00.0000] <darbour> when I give "<p class="some-p">LALALLALA</p>" the sanitizer returns "<p class="some-p">blarg\n" [15:09:01.0000] <darbour> * replace LA* with blarg [15:11:00.0000] <jgraham> darbour: In theory I can help but a) I'm about to go to bed and b) that seems rather odd [15:12:00.0000] <jgraham> Can you give an actually running example? [15:13:00.0000] <sayrer> I think he means that it eliminated his closing tag [15:13:01.0000] <darbour> exactly [15:14:00.0000] <darbour> sorry I was a little vague there [15:14:01.0000] <sayrer> it did that because it is not strictly necessary [15:14:02.0000] <sayrer> also, the sanitizer is pedantic [15:14:03.0000] <sayrer> there you go [15:14:04.0000] <sayrer> :) [15:14:05.0000] <jgraham> Oh. It sounds like it's the serializer setup you have or something [15:15:00.0000] <jgraham> Running code would still help :) [15:15:01.0000] <sayrer> I wonder if the html5lib sanitizer replaces font elements [15:15:02.0000] <darbour> for <p class="first-p">1<br/>2</p> it returns <p class="first-p">1<br>2 [15:15:03.0000] <darbour> where would i begin to look to stop this behavior? [15:16:00.0000] <jgraham> darbour: What do you mean "returns" usually the sanitizer returns an in-memory tree, not a string. How are you creating the string? [15:17:00.0000] <annevk2> hmm, dropping the closing tag seems problematic [15:17:01.0000] <jgraham> annevk2: I think that somewhere somethin is being serialized with omit-optional-tags=True [15:18:00.0000] <jgraham> Er but not quite called that [15:18:01.0000] <darbour> should I look at the html5lb serializer code? [15:20:00.0000] <jgraham> darbour: In the HTMLSerializer there is an attribute omit_optional_tags which is True by default. I think if you set this to False on your serializer instance your problems will go away [15:21:00.0000] <darbour> awesome, I'll try that. Thanks for the help! [15:47:00.0000] <annevk2> I tend to think http://intertwingly.net/blog/2007/05/08/Dont-Break-The-Web#c1178698369 is still a pretty good answer to part of the conformance question [16:26:00.0000] <annevk2> sicking, could be that I missed something, but I believe it's just Mozilla that pays hsivonen for the validator work [16:26:01.0000] <sicking> annevk2, i thought w3c was planning on using some of his work too. Not sure if that has been decided for the HTML5 parser yet [16:27:00.0000] <Hixie> validator.w3.org uses some of the validator.nu code [16:27:01.0000] <Hixie> though in my experience hsivonen's instance is more reliable [16:31:00.0000] <annevk2> http://groups.google.com/group/native-client-announce/browse_thread/thread/e69fe64e8decbe16 -- interesting, albeit scary [16:33:00.0000] <annevk2> Hixie, a DOMString can contain a zero byte too, that'd be problematic as well for Web Sockets [16:33:01.0000] <Hixie> why? [16:33:02.0000] <annevk2> isn't that the end marker? [16:34:00.0000] <Hixie> no, oxFF is [16:34:01.0000] <Hixie> 0xFF even [16:34:02.0000] <olliej> annevk2: Hixie: is there not a length property instead [16:34:03.0000] <Hixie> for binary data yes [16:34:04.0000] <olliej> all data should be considered binary [16:34:05.0000] <Hixie> not for text though [16:34:06.0000] <olliej> 0xff can occur in valid unicode [16:34:07.0000] <Hixie> not in utf-8 [16:35:00.0000] <olliej> what about utf16 [16:35:01.0000] <sayrer> annevk2: none of hsivonen's points seem to apply to @summary or <font> very well [16:35:02.0000] <Hixie> websocket uses utf-8 [16:35:03.0000] <olliej> urgh [16:35:04.0000] <olliej> so we have to do a utf16-utf8 conversino :-/ [16:35:05.0000] <annevk2> sayrer, yup [16:36:00.0000] <Hixie> olliej: i'd rather you have to do it than have every single web author out there have to do it [16:36:01.0000] <annevk2> sayrer, Hixie made a few additional points that seem to cover those [16:36:02.0000] <sayrer> annevk2: not really [16:36:03.0000] <sayrer> but that's ok, I don't have to argue it [16:37:00.0000] <annevk2> me neither :) [16:37:01.0000] <sayrer> I'm just going to fork it :) [16:37:02.0000] <sayrer> now that I have some time [16:38:00.0000] <annevk2> good idea [16:38:01.0000] <annevk2> hopefully the extra competition will turn into something useful [16:39:00.0000] <sayrer> maybe [16:39:01.0000] <annevk2> Hixie, ah ok [16:39:02.0000] <sayrer> I'm just going to cut, and insert text that lets anyone extend it. hopefully the extra competition turns into something useful. [16:39:03.0000] <Hixie> annevk2: hm? [16:40:00.0000] <annevk2> Hixie, my misreading of Web Sockets [16:40:01.0000] <Hixie> ah [16:42:00.0000] <annevk2> IETF drafts are distinctly unreadable :/ [16:42:01.0000] <sayrer> yeah, they need decorations about how they don't work [16:43:00.0000] <annevk2> they could use some hypertext, for starters :) [16:45:00.0000] <sayrer> and some newspeak [16:46:00.0000] <sayrer> they already of the amoral stance towards patents [16:46:01.0000] <sayrer> half way there [16:46:02.0000] <sayrer> of the, have the [16:46:03.0000] <sayrer> you get it [16:46:04.0000] <annevk2> not really [16:47:00.0000] <sayrer> ok [16:51:00.0000] <Hixie> hm, annotations in rfcs describing which parts aren't actually implemented would be really useful 2009-06-13 [17:02:00.0000] <sayrer> I love that no one has explained why font is bad [17:03:00.0000] <sayrer> "you might be able to create inaccessible websites" [17:03:01.0000] <sayrer> still true with html5 [17:03:02.0000] <sayrer> don't let that stop you [20:57:00.0000] <ezyang> Gah, more changes :-) [02:51:00.0000] <jgraham> /me notes he did try to explain some problems with <font> [07:05:00.0000] <gsnedders> archtech: The actual parsing of entities we can't really change (because of need for compat. with existing content), what we can change is what is conforming, and making & legal on its own gets fairly complex. [07:05:01.0000] <archtech> gsnedders, the format of entities is that above today [07:06:00.0000] <annevk42> e.g. ©= does turn into ©= [07:06:01.0000] <gsnedders> archtech: It isn't. Things like © do parse like that, and need to for compatibility, even though they are non-conforming. [07:07:00.0000] <archtech> gsnedders, not every non-conforming change should be supported [07:07:01.0000] <gsnedders> archtech: Browser vendors will not make changes to parsing that breaks existing content. [07:08:00.0000] <gsnedders> (And that is absolutely non-negotiable: if the spec changes to something incompatible, we'll just break the spec) [07:08:01.0000] <archtech> gsnedders, does that mean <b> foo <i> bar </i> baz </b> is part of HTML5 [07:08:02.0000] <archtech> ops [07:08:03.0000] <archtech> I mean b i b i [07:08:04.0000] <gsnedders> archtech: It has a defined parse tree. It is non-conforming. [07:08:05.0000] <gsnedders> archtech: Any character stream has a defined parse tree. [07:09:00.0000] <archtech> So is that valid HTML5? [07:09:01.0000] <gsnedders> archtech: No, it is non-conforming. [07:09:02.0000] <archtech> gsnedders, but ©= is? [07:09:03.0000] <gsnedders> archtech: No, that is non-conforming too. [07:09:04.0000] <archtech> I see. [07:10:00.0000] <archtech> So HTML5 defines the exact behavior of non-conforming code. [07:10:01.0000] <gsnedders> Yes. [07:10:02.0000] <archtech> And & as a literal is supported, but non-conforming. [07:10:03.0000] <archtech> For example: M & M [07:11:00.0000] <gsnedders> archtech: That is legal because a whitespace character follows it [07:12:00.0000] <archtech> I see. [07:12:01.0000] <archtech> Thanks for the info, guys [07:13:00.0000] <gsnedders> archtech: The only place you're allowed a & followed by anything within script and style [07:17:00.0000] <archtech> gsnedders, sure [07:17:01.0000] <archtech> those are defined as literal blocks terminated by strictly </script> and </style> I suppose [07:17:02.0000] <archtech> this is their de-facto behavior. [07:17:03.0000] <gsnedders> Yeah [07:17:04.0000] <archtech> nice [07:17:05.0000] <gsnedders> (case insensitively) [07:17:06.0000] <archtech> oh yea [07:20:00.0000] <archtech> I gotta check how html5 deals with textarea [07:20:01.0000] <archtech> ;) [07:20:02.0000] <archtech> entities are supported.. but nothing else :) [07:21:00.0000] <archtech> curious one [07:21:01.0000] <gsnedders> /me is kinda running around, but can probably answer most of these questions :) [08:49:00.0000] <Philip`> archtech: Actually they're terminated by "</script" and "</style" (case-insensitively) followed by whitespace or "/" or ">" or EOF [08:53:00.0000] <gsnedders> Philip`: Not EOF, I thought, after the latest change. [08:54:00.0000] <gsnedders> Philip`: Well, I guess the contents to remain, we just drop the end tag [08:58:00.0000] <Philip`> gsnedders: I'm just going by what http://www.whatwg.org/specs/web-apps/current-work/multipage/syntax.html#close-tag-open-state says, for look-ahead recognition of the end tag [08:58:01.0000] <gsnedders> Philip`: You're right. [08:58:02.0000] <gsnedders> Philip`: Ignore me, I'm ignorant. [09:00:00.0000] <Philip`> annevk42: Having not thought about it very much, I dislike the idea of NaCl because it provides a significant drawback to non-x86 CPUs [09:00:01.0000] <gsnedders> NaCI? [09:01:00.0000] <Philip`> (and currently you can use non-x86 CPUs reasonably happy if you're running open source software and web stuff and not Flash, so if lots of people used NaCl then it'd have a significant effect on that) [09:01:01.0000] <Philip`> s/happy/happily/ [09:02:00.0000] <Philip`> gsnedders: http://code.google.com/p/nativeclient/ [09:02:01.0000] <Philip`> /me suggests gsnedders gets a font where l and I look different :-p [09:02:02.0000] <gsnedders> ah [09:03:00.0000] <gsnedders> They do, they have slightly different heights :P [09:06:00.0000] <Philip`> (You could emulate x86 on other ISAs but it's going to be pretty rubbish at performance, particularly if the code uses SSE, and it's marketed for computation-intensive processes) [09:06:01.0000] <Philip`> s/it's/NaCl is/ [09:10:00.0000] <annevk42> can someone run http://xmlhttprequest.testsuite.org/dump/2009/status/001.htm in IE8? [09:11:00.0000] <annevk42> same for http://xmlhttprequest.testsuite.org/dump/2009/status/statustext.htm [09:12:00.0000] <Philip`> annevk42: "1xundefined,1xundefined,2x200,3x200,4x200" on 001.htm [09:12:01.0000] <itpastorn> FFox: 1xundefined,1xundefined,2xOK,3xOK,4xOK [09:12:02.0000] <itpastorn> MSIE 8: 1xundefined,1xundefined,2xundefined,3x200,4x200 [09:13:00.0000] <Philip`> annevk42: "1xundefined,1xundefined,2xOK,3xOK,4xOK" on statustext.htm [09:14:00.0000] <itpastorn> My MSIE result was from 001.htm [09:15:00.0000] <annevk42> your MSIE results are different :/ [09:15:01.0000] <itpastorn> Let me try again [09:16:00.0000] <Philip`> I definitely get "2x200" on 001.htm [09:16:01.0000] <annevk42> I like your result better :) [09:16:02.0000] <Philip`> not in compatibility view [09:16:03.0000] <Philip`> (I think) [09:17:00.0000] <annevk42> oh wait [09:18:00.0000] <itpastorn> 1xundefined,1xundefined,2x200,3x200,4x200 in non compat view. 1xundefined,1xundefined,2x200,3x200,4x200 in compat view [09:18:01.0000] <annevk42> cheers, can you run them again? I made a small change to see where that undefined comes from [09:19:00.0000] <itpastorn> 1x[object Error],1x[object Error],2x200,3x200,4x200 Compat view [09:19:01.0000] <annevk42> cool [09:20:00.0000] <itpastorn> 1xError: Odefinierat fel.,1xError: Odefinierat fel.,2x200,3x200,4x200 Non-compat view (Swedish for "undefined error") [09:20:01.0000] <annevk42> seems everyone is pretty much the same then apart from Opera :/ [09:21:00.0000] <annevk42> and only WebKit throws the correct exception [10:48:00.0000] <archtech> Philip`, so this would work? </script/ [10:48:01.0000] <archtech> Interesting. [10:49:00.0000] <archtech> I only now realise how painful it is to reverse engineer a standard from a bunch of shoddy implementations ;) [10:49:01.0000] <Philip`> archtech: I don't know whether there's something to throw away incomplete tags and throw away unclosed script elements, in which case you might need </script/> instead [10:50:00.0000] <Philip`> (I remember some discussion about those issues a while ago but don't know what the spec says now) [10:50:01.0000] <archtech> Philip`, I have a slight complaint about the content models. [10:50:02.0000] <archtech> HTML4 seems to define those more clearly, and they match better what display property in CSS exposes [10:51:00.0000] <archtech> Also textarea has content model "text" which is not covered anywhere in the content model section it seems. [10:51:01.0000] <Philip`> archtech: You should probably complain to Hixie :-) [10:51:02.0000] <archtech> Hehe :) [10:51:03.0000] <Philip`> archtech: Also, are you looking at the "HTML 5: The Markup Language" document, rather than the main spec? [10:51:04.0000] <Philip`> in which case "text" is probably a RELAX NG type or something [10:52:00.0000] <archtech> Philip`, I think I'm looking at the last published spec [10:53:00.0000] <Philip`> /me actually looks at the spec [10:53:01.0000] <Philip`> Oh, yes, it says "Content model: Text." [10:53:02.0000] <archtech> Actually I'm looking at "W3C Working Draft 23 April 2009" [10:53:03.0000] <archtech> Here: http://www.w3.org/TR/html5/ [10:53:04.0000] <archtech> Check section 3.4 Content models [10:53:05.0000] <archtech> There's no "text" in there :P [10:53:06.0000] <archtech> I'm not sure also why phrasing/heading content are differentiated? [10:54:00.0000] <archtech> Headings seem to behave mostly like paragraphs. [10:54:01.0000] <Philip`> (It's probably better to look at http://whatwg.org/html5 or http://www.w3.org/html/wg/html5/ to get the latest version) [10:54:02.0000] <archtech> Ah ok./ [10:55:00.0000] <Philip`> /me doesn't really know anything about content models [10:55:01.0000] <annevk42> is "#PCDATA" really that much clearer than "Text"? [10:56:00.0000] <archtech> annevk2, hehe. There's link and PCDATA is defined ;) [10:57:00.0000] <annevk42> there's no link for me under http://www.w3.org/TR/html4/interact/forms.html#h-17.7 [10:57:01.0000] <annevk42> I agree though that "Text" needs a definition [10:58:00.0000] <archtech> :) [10:59:00.0000] <gsnedders> archtech: CDATA can end with the </ + tagname followed by tab, lf, ff, space, >, /, or EOF [11:01:00.0000] <annevk42> that's not an exhaustive definition [11:01:01.0000] <annevk42> e.g. it does not cover </script foobar> [11:02:00.0000] <Philip`> Why doesn't it cover that case? That's the tag name followed by space, and it ends CDATA [11:02:01.0000] <Philip`> (unless I'm hopelessly mistaken) [11:03:00.0000] <annevk42> archtech, oh, Text is defined in http://www.whatwg.org/specs/web-apps/current-work/#text-0 [11:03:01.0000] <gsnedders> annevk2: it's followed by space, so it does cover that [11:03:02.0000] <gsnedders> annevk2: (The fact that the space is then followed by other stuff is beside the point :P) [11:03:03.0000] <annevk42> I see.. [11:04:00.0000] <annevk42> archtech, but I'd still email if I were you because a) it isn't linked and b) it's defined in a section specific to HTML [11:57:00.0000] <Philip`> If I wrote something like "The smallest positive numbers are 1, 2, 3, <abbr title='4, 5, 6, ...'>...</abbr>", so you can (partially) expand the "..." text via the tooltip (or equivalent in your independent medium), would I be shot for violations against semantics or is it fine? [12:31:00.0000] <jjlee> is it a known issue that html5lib trunk's html5lib.serializer.htmlserializer is broken? [12:32:00.0000] <jjlee> seems somebody changed string constants into numeric ones [12:34:00.0000] <jjlee> I should say, branch "default", I think (I'm new to hg) [12:35:00.0000] <jjlee> hmm, runtests.sh fails with SyntaxError. Presumably that is a known issue. [12:36:00.0000] <jjlee> in which case, I guess there are two reasons why http://code.google.com/p/html5lib/ shouldn't say "The SVN trunk is currently rather stable in anticipation of 0.12" [12:36:01.0000] <jjlee> first, it's not stable [12:36:02.0000] <jjlee> second, it's not in SVN any more (I assume) [12:40:00.0000] <annevk2> hmm, jgraham would know more [14:55:00.0000] <Lachy_> gsnedders, re twitter, what exactly are you expecting Audi to win or lose? is it a race or something? http://twitter.com/gsnedders/status/2157634810 [14:55:01.0000] <gsnedders> Lachy_: 24 Heures du Mans [14:56:00.0000] <gsnedders> Lachy_: They have won every year from 2000 to 2008, except for 2003 which they did not enter. [14:57:00.0000] <Lachy_> is that the french name for 24 Hours of Le Mans [14:57:01.0000] <gsnedders> Lachy_: Oui [14:58:00.0000] <Lachy_> I thought Le Mans was already French. I wonder why they say "du Mans" instead [14:58:01.0000] <gsnedders> "le" is the masculine form of the definite article [14:59:00.0000] <Lachy_> ok [14:59:01.0000] <gsnedders> For masculine "of", "du" is used, which means either "of" or "of the" [14:59:02.0000] <gsnedders> So the "le" vanishes [15:00:00.0000] <Lachy_> anyway, I've not seen that race before. What kind of track do they race on? [15:01:00.0000] <Lachy_> found it http://en.wikipedia.org/wiki/24_Hours_of_Le_Mans#The_circuit [15:02:00.0000] <gsnedders> Lachy_: A high speed one. [15:02:01.0000] <gsnedders> Lachy_: 24 Heures du Mans [15:02:02.0000] <gsnedders> Lachy_: Oh, that on my clip board? [15:02:03.0000] <gsnedders> Lachy_: http://en.wikipedia.org/wiki/Circuit_de_la_Sarthe [15:02:04.0000] <gsnedders> Lachy_: There we go, that's what I meant to paste :P [15:05:00.0000] <Lachy> ok, so it seems similar to the Bathurst 24 Hour, except it [15:05:01.0000] <Lachy> it's been running for a lot longer [15:05:02.0000] <gsnedders> A lot quicker cars :) [15:06:00.0000] <Lachy> Are you saying we drive slow race cars in Aus? [15:06:01.0000] <gsnedders> I think it's true that the quickest cars at Bathurst are the slowest at Le Mans [15:07:00.0000] <Lachy> do you know about the Bathurst race? [15:07:01.0000] <gsnedders> a bit [15:07:02.0000] <Lachy> wow, you're the first non-Australian I know who's heard of it [15:08:00.0000] <gsnedders> hehe [15:14:00.0000] <annevk2> per wikipedia Bathurst 24 Hour only ran for years [15:14:01.0000] <annevk2> two years* [15:15:00.0000] <gsnedders> Lachy: (We're talking that the top cars have over 700 hp/tonne) [15:15:01.0000] <gsnedders> annevk2: That's untrue [15:15:02.0000] <gsnedders> No, I'm being dumb. It's 12 hour and 1000km races it's known for. [15:17:00.0000] <gsnedders> (The race that became the 1000km race goes back to 1963) [15:20:00.0000] <jgraham> /me is disturbed that there seems to be a useful productive discussion about <video> going on [15:21:00.0000] <jgraham> That violates one of the unshakable pillars of my existance [15:22:00.0000] <jgraham> jjlee: It is known that the html5lib trunk is somewhat broken [15:22:01.0000] <annevk2> that sounds a bit dramatic jgraham [15:23:00.0000] <jgraham> I merged in support for namespaced content in the main treebuilder and it has broken all the bits that I didn't fix on the branch e.g. all the stuff that relies on treewalkers [15:23:01.0000] <jgraham> annevk2: Really? [15:25:00.0000] <annevk2> your existance is in part based and in fact relies on unproductive discussions about <video>? [15:25:01.0000] <jgraham> jjlee: So I should fix it soon I guess. Sorry for the inconvenience [15:27:00.0000] <jgraham> annevk2: It's more that it invalidated my theory that it was no longer possible to have a sensible conversation about any remotely controversial HTML-related topic without one or more of the participants acting like a complete arsehat [15:28:00.0000] <jgraham> So I don't know if this says good things about Mike Shaver and Chris DiBona or bad things about everyone else [15:29:00.0000] <annevk2> lets go for somewhere inbetween [15:30:00.0000] <jgraham> I was going for "both" which was a bit depressing becuase I'm in the set of everyone else [15:53:00.0000] <jjlee> jgraham: np, but if development works that way, maybe it's easier to change that sentence on the web page than to change your style of development :-) [15:59:00.0000] <jgraham> jjlee: Will do; when the sentnce was written it was true. The idea of merging the branch was to fix all the issues rather than just let it sit but that hasn't really happened yet [16:02:00.0000] <jjlee> sure. I live in a glass house, shouldn't throw stones [16:16:00.0000] <jgraham> jjlee: Thanks for complaining, I am much more likely to work on something if it seems like someone actually needs it o work [16:17:00.0000] <jjlee> jgraham: any chance of a patch to expose character indices of tokens being accepted? 2009-06-14 [00:37:00.0000] <tiglionabbit> hi [00:39:00.0000] <tiglionabbit> html5 specifies scripting features too, right? [01:55:00.0000] <Hixie> wow, not only does w3c want to charge us $250 each to help improve the web at tpac (in addition to the annual $60k for being able to put their brand on specs), but they require that we pay them through paypal, which i can't use [03:01:00.0000] <Lachy> Hixie, why can't you use PayPal? [03:01:01.0000] <tiglionabbit> what exactly are you paying for? Endorsement? [03:02:00.0000] <annevk42> TPAC attendance I guess [03:03:00.0000] <annevk42> PayPal doesn't work for me either [03:05:00.0000] <Lachy> annevk42, if Opera does send us, then I assume they would deal with the payment issues anyway [03:07:00.0000] <tiglionabbit> do you guys specify javascript changes too? [03:07:01.0000] <annevk42> we keep a page about where browsers have to be incompatible with the ECMAScript spec, but that's about it [03:08:00.0000] <tiglionabbit> hm [03:09:00.0000] <tiglionabbit> what would you guys think about specifying a standard bytecode as an alternative to javascript? Something that is easier to compile into, so people don't have to do silly things like GWT [03:17:00.0000] <annevk42> I wouldn't be a huge fan given the increased complexity, attack surface, etc. [04:22:00.0000] <annevk42> /me finds http://www.w3.org/TR/SVGParam/#DOMInterfaces [04:54:00.0000] <annevk42> http://people.xiph.org/~greg/video/ytcompare/comparison.html is interesting [04:55:00.0000] <annevk42> was that pointed out on the list already? [04:59:00.0000] <Philip`> It was [04:59:01.0000] <Philip`> http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2009-June/020405.html [05:02:00.0000] <annevk42> cheers [05:03:00.0000] <Philip`> It always seems kind of amusing when people talk about the video quality of YouTube, because high video quality isn't something YouTube is historically known for [05:04:00.0000] <annevk42> well, the argument was bandwidth + quality [05:05:00.0000] <Philip`> Stage6 did well on video quality, but not so well on bandwidth and on not-losing-tens-of-millions-of-dollars-per-year [05:06:00.0000] <annevk42> "The website never went beyond beta status, and was shut down by DivX Inc. on February 29, 2008 because of apparent inability to support Stage6 financially, or other officially unspecified reasons." [05:10:00.0000] <annevk42> "According to the web-news blog TechCrunch, the prime reason for the shutdown were not the Stage6-generated bandwidth costs (approx. 1 Million USD per month), but an internal conflict on the DivX, Inc. board." [05:10:01.0000] <Philip`> annevk42: I don't intend to disagree with the real argument; it just seems easy to caricature it as a (slightly) amusing situation of people engaging in deep technical discussions of the relative merits of Theora vs H263 vs H264 for encoding music slideshows and mobile phone videos of cats [05:10:02.0000] <annevk42> but TechCrunch is usually a lot of bs so whether that's true is unclear [05:11:00.0000] <Philip`> I can imagine that spending a million dollars a month with no revenue would be a significant cause of internal conflict :-) [05:14:00.0000] <annevk42> it also says they made enough cash [05:15:00.0000] <annevk42> but yeah, youtube is great for things like http://en.wikipedia.org/wiki/Keyboard_Cat [05:17:00.0000] <Philip`> Hmm, I'm not sure how much Stage6 would have driven downloads of the DivX Web Player [05:17:01.0000] <Philip`> (I just used the mplayer plugin, but I don't know what Windows users would do) [05:25:00.0000] <Lachy> why is it that all studies I've ever seen comparing video and audio codecs always rely on subjective analysis by a person to rate the quality? Surely, there are algorithms available that could be implemented in a program to objectively compare the original lossless video with the compressed versions [05:26:00.0000] <Lachy> and then whichever compressed version is closest to the original is superior [05:30:00.0000] <Philip`> It's easy to make quantitative comparisons (like PSNR), but very hard to make ones that closely correspond to what a human would subjectively perceive as quality [05:31:00.0000] <Philip`> e.g. a human would care much more about blurriness in the compression of human faces than of background details [05:32:00.0000] <Philip`> and a human might prefer an image that looks sharp than one that's more faithful to the original [05:49:00.0000] <jgraham> /me is amazed that people are suggesting that removing conformance requirements would reduce rathole discussions [05:50:00.0000] <jgraham> Seriously if we start telling people that HTML5 makes presentational markup conforming (whether or not we can justify such a thing logically) there will be the other of all flamewars [05:51:00.0000] <jgraham> mother [05:53:00.0000] <Philip`> It wouldn't make presentational markup conforming - it would just mean the answer to "is this page conforming?" would always be "mu", as far as I can tell [05:55:00.0000] <Philip`> (You could still ask a question like "does this page follow the Best Practices described in this other document that isn't the HTML5 specification?") [05:56:00.0000] <Philip`> (and there could be competition between different groups trying to define the most useful sets of best practices for authors, rather than codifying one set in the spec itself) [05:57:00.0000] <Philip`> Anyway, it certainly would reduce rathole discussions, because they would all be consumed by a single giant Sarlacc-hole [06:55:00.0000] <annevk42> http://twitter.com/sarahebourne/statuses/2158980327 o_O so much for using <strong>! [09:29:00.0000] <gsnedders> /me gets another "bug report" of an XML parse error [10:59:00.0000] <annevk42> can someone run http://xmlhttprequest.testsuite.org/dump/xhr-doc/001.htm and http://xmlhttprequest.testsuite.org/dump/xhr-doc/002.htm in IE8? [11:08:00.0000] <annevk42> man, this stuff is complicated :/ [11:41:00.0000] <heycam> is there a spec/rfc that defines a mapping of utf-16 code units to a sequence of unicode characters (in some way other than aborting for things like unmatched surrogates)? [11:43:00.0000] <gsnedders> AFAIK no, it is defined in Unicode to be undefined (when invalid) [11:43:01.0000] <gsnedders> (It isn't defined to abort, it just plain is undefined, explicitly so.) [11:44:00.0000] <heycam> k [11:44:01.0000] <gsnedders> Same for UTF-8 and UTF-32 [11:44:02.0000] <heycam> /me should do some testing, he supposes! [11:46:00.0000] <heycam> rfc 2781 says "the sequence is in error" and to "terminate" for bad input [11:46:01.0000] <gsnedders> I'm more or less certain Unicode 5.1 does what I say [11:47:00.0000] <annevk42> for XHR I need DOMString to UTF-8 [11:47:01.0000] <annevk42> actually, potentially a plethora of encodings with the proposal of making it dependent on inputEncoding [11:48:00.0000] <heycam> annevk2, would an algorithm to obtain a sequence of unicode characters be sufficient? [11:48:01.0000] <heycam> unicode characters to utf-8 is already defined [11:48:02.0000] <heycam> /me is working on http://www.w3.org/mid/op.uuzo0yff64w2qv@annevk-t60 btw [11:48:03.0000] <annevk42> that seems like a good solution [11:49:00.0000] <annevk42> if we have valid Unicode characters mapping is an issue for the relevant encoding specs [11:49:01.0000] <heycam> yup [11:49:02.0000] <heycam> i'll do that [11:50:00.0000] <heycam> but i won't make the "converting ES String to IDL DOMString" algorithm invoke this [11:50:01.0000] <heycam> it'd just be a definition that other specs could use as desired [11:50:02.0000] <annevk42> yeah, you don't want to invoke it there I believe [11:51:00.0000] <annevk42> you want to use it when you need to specifically do something with a DOMString [11:51:01.0000] <annevk42> like sending it over the wire [11:51:02.0000] <annevk42> or writing it to disk maybe [11:51:03.0000] <heycam> or choosing what chars to render or something [11:52:00.0000] <annevk42> true [12:13:00.0000] <heycam> annevk2, do you know what xhr requires in this case? [12:14:00.0000] <heycam> inserting as many U+FFFDs as there are bad code points (for some value of "bad")? [12:17:00.0000] <annevk2> no sorry [12:17:01.0000] <annevk2> i have not actually played with it, just found the issue affected xhr too [12:17:02.0000] <annevk2> (in some way) [12:18:00.0000] <heycam> k [13:42:00.0000] <Hixie> anyone got a good reference for GB2312? [13:44:00.0000] <Hixie> i guess rfc1345 will do [14:04:00.0000] <Hixie> how about big5 [14:04:01.0000] <ezyang> zomg big5 [14:04:02.0000] <ezyang> I think I ended up using iconv as a reference implementation. [14:25:00.0000] <jruderman> mmm, a video codec and html5 slashdot flamewar. http://tech.slashdot.org/story/09/06/14/1649237/YouTube-HTML5-and-Comparing-H264-With-Theora?art_pos=1 [14:27:00.0000] <gsnedders> Peh, there are virtually no comments. [14:37:00.0000] <Hixie> i guess i'll reference wikipedia for big5 [14:37:01.0000] <Hixie> since i can't find a useful normative reference [14:49:00.0000] <annevk2> Hixie, http://tools.ietf.org/html/rfc1922 seems to define it [14:49:01.0000] <annevk2> Hixie, it also has a further reference: http://tools.ietf.org/html/rfc1922#ref-BIG5 2009-06-15 [18:51:00.0000] <heycam> Hixie, how many objects need to do this? [18:51:01.0000] <heycam> (wondering whether it's worth having a special xattr for it) [18:51:02.0000] <heycam> i think the sentence should be fine [18:52:00.0000] <Hixie> right now i only know of one, and i wouldn't say it needs it per se -- just that i would have used such a feature if it was available [18:52:01.0000] <heycam> "In the ECMAScript binding, host objects that implement interface Blah have a [[Call]] method that does ...." [18:52:02.0000] <heycam> k [18:56:00.0000] <Hixie> what would "...." be? [18:56:01.0000] <Hixie> "invokes the [[Get]] operator with the same argument"? [18:56:02.0000] <Hixie> (just putting this in a comment for now in the spec in case i find we do need it) [18:56:03.0000] <heycam> well not quite, you'll have to deal with different numbers of arguments i suppose [18:56:04.0000] <heycam> what happens if you pass 0 or more than 1 arg? [18:57:00.0000] <Hixie> dunno, haven't tested [19:03:00.0000] <heycam> so i'd do "... [[Call]] method that behaves as follows: 1. If called with 0 or more than 1 argument, throw a TypeError. 2. Let arg be the single argument passed. 3. Invoke [[Get]] on the object with ToString(argument) as the property name. 4. Return Result(3)." [19:03:01.0000] <heycam> or so [19:03:02.0000] <Hixie> k, thanks [19:05:00.0000] <heycam> c [19:05:01.0000] <heycam> ww sorry [00:09:00.0000] <Hixie> annevk42: yt? wondering what the timeline you're expecting for cssom is, and whether you think that'd be a good spec to have define when the 'resize' and 'scroll' events fire [00:10:00.0000] <hsivonen> argh. lots of email about validation over the weekend [00:24:00.0000] <sinclair> hello all [00:25:00.0000] <sinclair> so, anyone in here using O3D? [00:28:00.0000] <sinclair> im trying to find a online chat community who is working on O3D [00:28:01.0000] <sinclair> hello [01:47:00.0000] <annevk42> Hixie, I basically haven't found time to work on CSSOM for the past year or so [01:48:00.0000] <annevk42> Hixie, though resize and scroll could I suppose be part of the subset that is CSSOM Vie [01:48:01.0000] <annevk42> w [01:53:00.0000] <jgraham> What happens if I have some HTML like <body bgcolor="black"><article><font color="white">... and the <article> contents ends up in a feed? ill it be unreadable? Taking some but not all style information seems like it could be dangerous [01:53:01.0000] <jgraham> s/ill/will/ [01:55:00.0000] <Dashiva> jgraham: Well, that's "wrong" to begin with [01:55:01.0000] <jgraham> Dashiva: In what way? [01:55:02.0000] <jgraham> Under the assumption that anything that works in a UA is conforming [01:55:03.0000] <Dashiva> You should set text color on the body element as well [01:56:00.0000] <Dashiva> Always specify foreground and background together [01:56:01.0000] <jgraham> Dashiva: According to whom? Not according to the supposed no-author-requirements HTML spec [01:57:00.0000] <Dashiva> According to the best practices that are supposed to replace them [01:57:01.0000] <Dashiva> No requirements includes the freedom to shoot yourself in the foot [01:58:00.0000] <jgraham> Yeah but I'm trying to work out why that's a good thing [01:58:01.0000] <Dashiva> Has anyone claimed that? [01:59:00.0000] <jgraham> That having no author requirements is a good thing? Sure [01:59:01.0000] <Dashiva> That's a choice between evils, not goods [01:59:02.0000] <jgraham> Can you explain more? [01:59:03.0000] <Dashiva> Requirements come with ratholes, bikesheds, and aren't enforced [02:00:00.0000] <Dashiva> The argument (as I read it) goes that avoiding all those weighs up for the loss of the guideance authority that could be provided [02:00:01.0000] <jgraham> Lack of requirments will come with a huge rathole, at the bottom of which will be the original ratholes but renamed "best practices" [02:01:00.0000] <jgraham> So I don't see any possibility for gain here [02:01:01.0000] <Dashiva> That is the gain [02:01:02.0000] <Dashiva> That huge rathole will always be there, that's what good authoring is all about [02:01:03.0000] <Dashiva> But without requirements, we outsource the war over where the hole should be [02:02:00.0000] <jgraham> As far as I can tell the main agenda here is that some people don't like the HTML5 draft because it doesn't endorse something that they hold dear to them. So they have decided that the only way out is to try to remove all restrictions and be told "do what you like" [02:02:01.0000] <Dashiva> Well, the basic complaint holds: What use are requirements that no one actually requires? [02:03:00.0000] <jgraham> Sometimes this is wrapped up in theoretical RFC 2119 arguments [02:04:00.0000] <jgraham> Dashiva: The point of authoring requirements in a language with error correcting clients is not to describe the only possible way of doing things [02:04:01.0000] <jgraham> It is to reduce the authoring surface of the language so that it does not grow without bounds and so there are as few sharp edges as possible [02:05:00.0000] <Dashiva> But nobody reads those requirements [02:05:01.0000] <jgraham> Dashiva: Demonstrably untrue [02:05:02.0000] <Dashiva> They read second- or third-hand references that also includes local author intent, best practices, misconceptions, urban legends, etc etc [02:06:00.0000] <jgraham> There has been a huge movement toward following standards. Yes there is a lot of cargo cult nonsense. But there is enough essentially good stuff that we should no thro the whole thing away just to make the lives of spec authors better [02:06:01.0000] <jgraham> We even have a design principle about that [02:06:02.0000] <Dashiva> Nobody's saying we should throw it away [02:06:03.0000] <jgraham> Dashiva: They really are [02:07:00.0000] <Dashiva> They may think so [02:07:01.0000] <Dashiva> But those references people use will still exist [02:07:02.0000] <Dashiva> It's only mark's assholes that will be weakened [02:09:00.0000] <jgraham> Dashiva: I would expect one effect of removing conformance requirements to be that people will start to reject HTML5 for cargo-cult reasons [02:09:01.0000] <Dashiva> Oh, sure. But there wouldn't be anything for them to reject. It would be a user agent spec full and whole. [02:09:02.0000] <jgraham> Dashiva: Not really. [02:10:00.0000] <jgraham> It still will define a language [02:10:01.0000] <jgraham> It just won't tell you which bits are a really bad idea [02:10:02.0000] <Dashiva> Now we're back at html4 ground [02:10:03.0000] <jgraham> ? [02:10:04.0000] <Dashiva> The language in the spec does not matter for authors, authors and users only go via user agents [02:11:00.0000] <jgraham> But we just discussed this. Authors learn from somewhere and that somewhere often has looked at the spec [02:11:01.0000] <jgraham> And authors do use vaidators [02:12:00.0000] <jgraham> and validators have read the spec for sure [02:12:01.0000] <Dashiva> But again the authors don't go directly to the spec [02:12:02.0000] <Dashiva> They see the validator's rules, which would include all kinds of best practices [02:13:00.0000] <jgraham> Dashiva: What kind of best practices are encoded by validator.w3.org or validator.nu, beyond what is in the spec? [02:13:01.0000] <Dashiva> All the author requirements [02:14:00.0000] <jgraham> I mean today [02:14:01.0000] <Dashiva> Today there isn't much need, henri gets it put into the spec instead [02:14:02.0000] <Dashiva> In particular the downplayed errors, he seems to be the main driving force there [02:15:00.0000] <jgraham> Dashiva: The concept of downplayed errors is silly [02:15:01.0000] <jgraham> That is a UI issue [02:15:02.0000] <Dashiva> If so, that strenghtens the point [02:16:00.0000] <Dashiva> These are things he doesn't want the validator to mark as errors. Now he's working inside the spec, if the spec dropped author reqs he would just go ahead and do it directly [02:16:01.0000] <Philip`> What if someone else writing a different validator has different ideas of what should be considered an error or a warning? [02:17:00.0000] <jgraham> Dashiva: So in the future what do you envison? validator.w3.org uses best practices document A. commercial.validator.org uses weaker best practices document B to iucrease marketshare. Validator validate.net is written by areholes and has a whole load of additional requirements. An author tries to validate her document. What does she conclude? [02:17:01.0000] <Dashiva> Then you get competition in who validates the best [02:17:02.0000] <Philip`> Putting those requirements in the spec prevents the benefits of competition, because there will only be one official blessed notion of validity [02:17:03.0000] <Philip`> which seems similar to the argument against putting machine-readable schemas in the spec [02:17:04.0000] <jgraham> Dashiva: Re: dowwnplayed errors, Henri can already choose to downplay any error he likes. It is a pure UI issue [02:18:00.0000] <Dashiva> jgraham: Again, sure. It's not material to my point. The point is that he puts effort into convincing Hixie to mark them as downplayed. That means he cares. [02:19:00.0000] <jgraham> Philip`: The problem with putting schemas in the spec is that it is not possible to express all conformance requirements in schema and there is the possibility of divergence between the actual requirements and the schema (due either to the limitations or to bugs) [02:20:00.0000] <Dashiva> This situation exists even now, of course [02:20:01.0000] <Dashiva> There are requirements outside the HTML5 spec [02:20:02.0000] <jgraham> Dashiva: And I still maintain it is silly to have the requirements in the spec. So I don't see that I'm taking an inconsistent position [02:20:03.0000] <Dashiva> This whole thing would just be moving one set of requirements into some other (possible more than one) location [02:21:00.0000] <jgraham> s/requirements/downplayed errors/ [02:21:01.0000] <Dashiva> It's not your position. His effort suggests that even if the spec did not have author requirements, he personally cares enough about these issues that his validator would have best practices for them [02:22:00.0000] <Dashiva> So no, validators would not just be windows onto the UA spec [02:22:01.0000] <jgraham> Dashiva: I don't understand your argument at all [02:22:02.0000] <jgraham> specifically the sentence "validators would not just be windows onto the UA spec [02:22:03.0000] <jgraham> " [02:23:00.0000] <Dashiva> [10:17:05] <jgraham> Dashiva: What kind of best practices are encoded by validator.w3.org or validator.nu, beyond what is in the spec? [02:23:01.0000] <jgraham> "The spec" in this case is not just a "UA spec" though [02:24:00.0000] <Dashiva> Now we're just repeating ourselves [02:24:01.0000] <Dashiva> The validator doesn't have its own requirements because they're in the spec. If they're not in the spec, they would be in the validator instead (so the validator itself wouldn't change) [02:26:00.0000] <jgraham> There is no disagreement about what the requirements are. There is only the issue of whether some errors should be displayed with greater weight. That, to me, is a UI issue and I think that it shouldn't be in the spec [02:27:00.0000] <Dashiva> Leave that issue. It's just an example to show that henri personally cares about some of the requirements, to support the claim that the validator would still have requirements if they weren't in the spec [02:27:01.0000] <jgraham> So I disagree with Henri here. But I don't feel bad about it too much because I disagree with Henri less than almost anyone else involved with HTML5 :) [02:28:00.0000] <jgraham> Dashiva: Your argument seems totally unjustified. Unless you can point to things that are not in the spec but Henri has included as requirements anyway I don't think that you can conclude that he would do such a thing [02:29:00.0000] <jgraham> I also don't think it's very relevant to whether, overall, there is any benefit to be gained by removing all the author requirements from HTML5 [02:29:01.0000] <jgraham> Anyway, I guess I should do something more useful... [02:30:00.0000] <Dashiva> We were there earlier [02:31:00.0000] <Dashiva> It's related because it claims that validators wouldn't change much. And validators and non-primary sources are where authors go [02:31:01.0000] <Dashiva> => Things would not change for authors and users [02:32:00.0000] <jgraham> I still disagree but there se go [02:32:01.0000] <jgraham> we [02:32:02.0000] <Dashiva> Sure, you might get small civil wars over what best practices are best, but that would be hidden to most people. And it already happens, just look at WCAG Samurai [02:35:00.0000] <jgraham> Actually the accessibility space is a great example of wwhy this kind of thing can be a disaster. There are so many documents relating to WCAG that it is almost impossible to tell where you should look for any given thing. And so you ask an "expert" and get an answer. And ask another "expert" and get a different answer. And so on... [02:36:00.0000] <Philip`> Does anyone take any notice of the WCAG Samurai's suggestions, or is it evidence that everyone just follows the most official sounding guidelines (from the W3C) and ignores everything else regardless of merits? [02:36:01.0000] <Dashiva> I think the evidence is that people ignore everything, period [04:23:00.0000] <hsivonen> Dashiva: actually, I don't get to put stuff in the spec. If I did, td@align and object@classid would be valid by now. [04:24:00.0000] <hsivonen> v.nu has non-spec-backed warnings especially around the area of authors shooting themselves in the foot by not using UTF-8 [04:59:00.0000] <annevk42> hsivonen, sync XHR is superevil [04:59:01.0000] <annevk42> also to spec [05:00:00.0000] <annevk42> though I think I have it nailed down now [05:06:00.0000] <hsivonen> I wonder if the script execution code in any browser is clear to someone who just reads it [05:06:01.0000] <hsivonen> anyway, it seems I managed to make http://mxr.mozilla.org/mozilla-central/source/content/base/test/test_bug461555.html not crash [05:06:02.0000] <hsivonen> yay [05:09:00.0000] <hsivonen> /me observes that it would be appropriate to call Lamborghinis dump trucks if the large majority of Lamborghini drivers used them as dump trucks [05:37:00.0000] <gsnedders> /me guesses that is some reference to a post on the list [05:53:00.0000] <jgraham> gsnedders: No actually it is true that most owners of italian sportscars use them mainly for taking stuff to the dump. Researchers say "the have such small boots that it can require hundreds of trips to dispose of the cuttings every time you trim your garden hedge" [05:53:01.0000] <gsnedders> :P [05:57:00.0000] <beowulf> i looked at a disabled car parking space the other day, it'd be almost impossible to implement something like that incorrectly [05:58:00.0000] <beowulf> and easy to do as well, just paint some lines [06:14:00.0000] <annevk42> I will toally include <supercalifragilisticexpialidocious> in all my documents from now on [06:14:01.0000] <jgraham> hsivonen: I am interested to know what you think of the argument that making <font> non conforming has positive implications for the learnability of the language (due to making it smaller) [06:14:02.0000] <jgraham> And by helping authors avoid doing things that are belived to be bad [06:15:00.0000] <annevk42> (why: http://www.w3.org/mid/5D1706DF-0909-418F-8592-24535A5C19A1⊙if ) [06:15:01.0000] <jgraham> Specifically the difference beteen <font> and <span style> is that authors using <span> must be aware of CSS and have taken a concious decision to use inline style in place of seperated style [06:15:02.0000] <jgraham> That seems like a positive ting [06:16:00.0000] <jgraham> If a tool emits <span style> in place of <font> it is neutral [06:16:01.0000] <jgraham> So therefore removing <font> seems to have an overall positive balance [06:17:00.0000] <hsivonen> somehow I'm more annoyed when OpenOffice.org or Nvu emit style='' [06:17:01.0000] <jgraham> hsivonen: Is that just an aethetic preference though? [06:17:02.0000] <jgraham> aesthetic [06:18:00.0000] <hsivonen> jgraham: I tend to think that HTML already has enough weird corners that learnability needs to be achieved by prioritizing what to reveal and what order to a learner [06:18:01.0000] <hsivonen> jgraham: no, if you use TextWrangler's grep find&replace to zap style='', you then need to separately zap spans, too [06:18:02.0000] <jgraham> hsivonen: What weird corners are part of the conforming language? Optional tags I guess [06:18:03.0000] <Philip`> It seems a pain if you use <font> for most styling until you reach a point where you want e.g. small-caps and you have to switch to <span style> and now have to understand how to use both and how they interact, rather than simply using the more powerful mechanism from the start [06:19:00.0000] <hsivonen> jgraham: optional tags is the #1 thing, yes [06:19:01.0000] <jgraham> Philip`: Indeed [06:19:02.0000] <hsivonen> #1 weird corner, not #1 thing to learn [06:20:00.0000] <hsivonen> Philip`: I agree. <font color> makes some sense for GUI tool-generated visual empahasis as long as the GUI tool is simple enough not to support all of CSS [06:20:01.0000] <hsivonen> also, with <font face> and <font size>, you lose [06:20:02.0000] <jgraham> I note that we tried the wwhole <font>-for-GUI-tools only thing and people didn't go for it [06:21:00.0000] <hsivonen> jgraham: we tried a slightly different thing [06:21:01.0000] <hsivonen> jgraham: we tried <font style> for GUI tools [06:21:02.0000] <hsivonen> jgraham: also, I see sayre's point about survival in syndication [06:22:00.0000] <jgraham> hsivonen: I may misremembert but I'm not sure if the distincion makes a big difference [06:22:01.0000] <jgraham> hsivonen: It seems relatively dangerous to keep some but not all style during syndication since you could e.g. render some text invisible [06:23:00.0000] <hsivonen> jgraham: yeah, it only works for <font color> red [06:23:01.0000] <hsivonen> if a user hits select all, picks a color and the editor generates a zillion <font>s, you lose [06:24:00.0000] <hsivonen> if the user selects two words and picks red, not generating <font color> is quite silly [06:24:01.0000] <Philip`> jgraham: Why is rendering some text invisible dangerous? [06:25:00.0000] <Philip`> jgraham: You could achieve the same effect by replacing all the text in your feed with whitespace [06:25:01.0000] <jgraham> hsivonen: Like I said I think the latter case is a wash compard to <span style>. At least, as Philip` said, if you select two words and select small caps you end up with the same markup [06:25:02.0000] <Philip`> jgraham: ...unless you mean "dangerous" as in "likely to go wrong", rather than "permitting malicious behaviour" [06:25:03.0000] <jgraham> Philip`: Not dangerous in the security sense, but in the UI sense [06:25:04.0000] <hsivonen> jgraham: I think small caps is on a more advanced level of typography than red [06:26:00.0000] <hsivonen> jgraham: different users. different classes of editor UIs [06:26:01.0000] <jgraham> hsivonen: I'm not sure there is any correlation between people who would use small caps and people who would use an advanced editor [06:27:00.0000] <hsivonen> jgraham: note how in my email, I didn't take the time to argue Sam's <font color> bait per se [06:27:01.0000] <jgraham> And if the use case is people who don't read the markup, I don't really see how <font> is an advantage [06:28:00.0000] <Philip`> hsivonen: Maybe a better example would be highlighting (as in <span style=background:yellow>...</span>), which in (at least) OO.o has almost exactly the same UI exposure as font colour [06:28:01.0000] <hsivonen> <font color> survives syndication and legacy tools better [06:28:02.0000] <hsivonen> Philip`: indeed. much better example. [06:30:00.0000] <jgraham> /me is being slow and doesn't understand why that is a bette example [06:31:00.0000] <Philip`> jgraham: Because it's conceptually very similar to font colour (as evidenced by word processor UIs), and can be implemented with very similar markup if you use <span style>, but not if you use <font> [06:31:01.0000] <Philip`> so using <font> introduces an unnecessary distinction between those similar concepts [06:32:00.0000] <hsivonen> /me notes that <font color> is one special thing compared to style="" like @summary is one special legacy thing compared to aria-describedby [06:32:01.0000] <Philip`> which makes everything harder to understand, and is more fragile when e.g. syndicators strip style attributes but not font elements [06:32:02.0000] <jgraham> Oh. OK that's what I though. But I wasn't sure if it was supposed to support hsivonen's point [06:32:03.0000] <hsivonen> or maybe aria-label considering that it's human-readable text hidden in an attribute [06:32:04.0000] <Philip`> I haven't been concentrating enough to read what hsivonen's point was so I don't know [06:33:00.0000] <hsivonen> my point is that if you want to make a principled markup about stuff that was in HTML 4, you'd better make sure your principle applies to <font color>, too [06:33:01.0000] <hsivonen> otherwise, you are just picking favorites while pretending to have a principle [06:34:00.0000] <hsivonen> s/principled markup/pricipled point/ [06:34:01.0000] <hsivonen> or something [07:00:00.0000] <annevk42> hsivonen, you could group all the font messages into one [07:01:00.0000] <annevk42> hsivonen, your page uses a legacy element called "font" please don't [07:27:00.0000] <annevk42> licensing⊙fo: will we now get an influx of laywers on whatwg⊙wo? [07:27:01.0000] <annevk42> oh, I guess blocking from third parties is still blocked [07:37:00.0000] <gsnedders> annevk2: Licensing⊙fo is mainly volunteers, not all lawyers [07:40:00.0000] <annevk42> s/blocking/email/ [08:34:00.0000] <gsnedders> /me squints [08:34:01.0000] <gsnedders> There's a mysterious gap of trains between 15:42 and 17:42 from Nyköping to Linköping, right in the rush hour :\ [08:35:00.0000] <jgraham> gsnedders: You really really want to take the bus [08:35:01.0000] <gsnedders> Ah, OK [08:35:02.0000] <gsnedders> How long does that take? [08:36:00.0000] <jgraham> 1.5 hours [08:36:01.0000] <jgraham> http://www.flygbussarna.se/ [08:36:02.0000] <gsnedders> Yes, I've found taht already :P [08:36:03.0000] <gsnedders> that [08:36:04.0000] <jgraham> http://www.flygbussarna.se/Default.aspx?lang=EN is even better [08:37:00.0000] <gsnedders> Yes, I found that actually :P [08:37:01.0000] <gsnedders> (I didn't find the Swedish.) [08:37:02.0000] <jgraham> Oh well wwhy are you asking how long it takes then? [08:38:00.0000] <gsnedders> Because I hadn't clicked through the three pages to the timetable and hadn't downloaded it yet, and I thought you'd know [08:38:01.0000] <gsnedders> /me guesses jgraham has probably flown from STN to NYO like he will be doing next week [08:39:00.0000] <jgraham> gsnedders: Just a little [08:40:00.0000] <gsnedders> /me was just emailing Bibbi because she has the key to his flat, and he won't get to Linköping until a bit before 19:00 [08:40:01.0000] <jgraham> gsnedders: What day of the wweek? [08:40:02.0000] <gsnedders> jgraham: Friday. [08:40:03.0000] <gsnedders> /me is really inconvenient [08:41:00.0000] <jgraham> /me wonders if he can fit a new keyboard in his hand luggage [08:41:01.0000] <jgraham> gsnedders: Oh everyone will still be here for Friday beer [08:41:02.0000] <gsnedders> Ah [08:41:03.0000] <jgraham> If you mean Fri. 26th [08:41:04.0000] <gsnedders> Isn't a keyboard a bit big [08:41:05.0000] <gsnedders> yeah [08:41:06.0000] <jgraham> gsnedders: Probably :( [08:42:00.0000] <gsnedders> hmm, fits in my laptop bag (which holds my 17" MBP) [08:42:01.0000] <jgraham> gsnedders: It is highly likely that bibbi will still be at the office then. You should of course check [08:42:02.0000] <gsnedders> /me just emailed her [08:43:00.0000] <jgraham> gsnedders: Do you know your address? [08:43:01.0000] <jgraham> /me is curious [08:43:02.0000] <gsnedders> Off the top of my head? No [08:56:00.0000] <gsnedders> jgraham: But yeah, then I have the weekend to come to terms with where I am in the world :P [09:11:00.0000] <jgraham> gsnedders: I think I am busy that weekend but we should organise something social for the first weekend in July [10:16:00.0000] <takkaria> Opera does seem to gobble up whatwg people [10:21:00.0000] <Philip`> There must be sinister motives [10:35:00.0000] <gsnedders> /me remembers talking with MikeSmith over lunch at TPAC about the percentage of Opera employees involved in W3C WGs [10:54:00.0000] <dglazkov> Philip`: do you have a bug in Chromium tracker for the canvas failures? [11:00:00.0000] <Philip`> dglazkov: No - I've never properly looked at Chromium's behaviour on tests at all [11:06:00.0000] <jgraham> takkaria: When are you coming over? [11:20:00.0000] <takkaria> jgraham: Sunday morning [11:21:00.0000] <annevk42> are you all still there in August? [11:21:01.0000] <takkaria> I'm there til mid-September [11:22:00.0000] <annevk42> at one point I made this remark I'd visit Linkoping on another time then during the Christmas party, but I haven't made the time yet [11:22:01.0000] <takkaria> btw, do you pronounce Linköping as "linchurping"? I can't figure it out from my phraseboo [11:23:00.0000] <Philip`> How exciting, an IE compat list update [11:24:00.0000] <Philip`> Sites added: 7m.cn beijing.com.cn benq.com.cn btbroadbandinformation.com btopenworld.com chinanews.com.cn chinatelecom.com.cn donews.com foundertech.com gmail.com kooxoo.com mydown.com narrowad.com tobacco.gov.cn [11:24:01.0000] <annevk42> "linkshohping" or so [11:24:02.0000] <annevk42> http://en.wikipedia.org/wiki/Linköping might have better tips [11:24:03.0000] <Philip`> Sites removed: about.com desjardins.com guardian.co.uk hyves.net hyves.nl intuit.com me.com mozilla.com mozilla.org [11:24:04.0000] <krijnh> hyves.nl :S [11:24:05.0000] <jgraham> takkaria: Ah OK. I just thought I should check in cae you were on the same flight as me on Saturday [11:25:00.0000] <jgraham> Although I guess you are not flying from Stanstead [11:25:01.0000] <dglazkov> Philip`: ok [11:25:02.0000] <dglazkov> Philip`: I'll look over. Thanks for the test suite :) [11:25:03.0000] <Philip`> The IE compat list has very little activity, given that it's got three thousand entries and only a few dozen changes per month [11:26:00.0000] <takkaria> jgraham: no, I'm coming from manchester [11:26:01.0000] <takkaria> jgraham: are you doing Stanstead->Schiphol->Linkoping? [11:26:02.0000] <jgraham> takkaria: It has a soft k pronounciation so you ronounce it like lin-sher-ping [11:26:03.0000] <takkaria> right, gotcha [11:26:04.0000] <jgraham> or so [11:26:05.0000] <Philip`> Soft k? Crazy language :-( [11:26:06.0000] <jgraham> (imagine you are slurring the word shopping...) [11:27:00.0000] <annevk42> just write it on a piece of paper or laptop and show that to people [11:27:01.0000] <jgraham> takkaria: No STN-NYO w/ryanair [11:27:02.0000] <annevk42> they'll know what you mean :) [11:27:03.0000] <jgraham> Philip`: You're telling me. I still can't pronounce anything [11:28:00.0000] <jgraham> the sj and tj sound is... interesting [11:28:01.0000] <gsnedders> annevk42: I'm there until the 31st [11:29:00.0000] <gsnedders> jgraham: This Sat? [11:29:01.0000] <jgraham> gsnedders: yes [11:29:02.0000] <gsnedders> jgraham: You getting the 6:30am flight? :P [11:29:03.0000] <jgraham> gsnedders: No the on in the evening [11:30:00.0000] <gsnedders> jgraham: Hehe, I was expecting not :) [11:30:01.0000] <gsnedders> /me wonders why he can't print [11:30:02.0000] <takkaria> I'm getting the 0600 out of manchester on sunday morning [11:30:03.0000] <takkaria> hellish really [11:30:04.0000] <gsnedders> Ow. [11:31:00.0000] <gsnedders> /me is flying at 13:30 from STN -> NYO [11:31:01.0000] <takkaria> the alternative is getting to Linkoping airport at 11 [11:31:02.0000] <gsnedders> (Next Friday) [11:31:03.0000] <takkaria> (11pm, that is) [11:31:04.0000] <gsnedders> Oh well, I get to cross another person off the list of cabal members to meet :P [11:36:00.0000] <annevk42> takkaria, just party until 3:30 and then take an hour and a half to get to the airport :) [11:56:00.0000] <annevk42> I wonder if http://www.cssquirrel.com/comic/?comic=22 influences my chances of getting invited to future BBQ parties [12:01:00.0000] <jgraham> /me wonders if it is churlish to point out that alt=comic doesn't really help blind users much [12:01:01.0000] <gsnedders> hah [12:02:00.0000] <annevk42> surely it has longdesc? [12:02:01.0000] <jgraham> Curiously not [12:02:02.0000] <annevk42> I'm so glad I live in the Netherlands [12:03:00.0000] <jgraham> annevk42: Why? [12:03:01.0000] <annevk42> I can already feel the wrath of John Foliot for these highly inappropriate jokes [12:03:02.0000] <Hixie> writing alt text for comics is actually surprisingly easy [12:03:03.0000] <annevk42> alt=":-)" [12:05:00.0000] <annevk42> Hixie, Google didn't do it though, for their huge Chrome comic book [12:05:01.0000] <Hixie> (photographs are a pain) [12:05:02.0000] <Hixie> annevk42: google put the chrome comic book online? [12:05:03.0000] <Hixie> sweet [12:05:04.0000] <Hixie> uri? [12:05:05.0000] <annevk42> like ages ago [12:05:06.0000] <Hixie> /me has a book version of that [12:06:00.0000] <annevk42> seems to be here among other places: http://www.google.com/googlebooks/chrome/ [12:06:01.0000] <Philip`> Hixie: http://www.google.com/search?q=chrome+comic+book [12:06:02.0000] <annevk42> fancy [12:06:03.0000] <Hixie> hm yeah, no alt text [12:06:04.0000] <Hixie> that is sad [12:07:00.0000] <Hixie> on another note, is it possible to remove duplicates from a list of strings in O() less than sorting a string? [12:07:01.0000] <Hixie> er [12:07:02.0000] <Hixie> sorting the list of strings [12:08:00.0000] <Philip`> Yes, if you don't care about memory usage [12:09:00.0000] <Philip`> (e.g. have a hash-set containing all the strings you've seen so far, and for each new string you do an O(1) test to see if it's new) [12:09:01.0000] <Philip`> (so it ends up as O(n) if you've got n strings) [12:09:02.0000] <Philip`> (vs O(n log n) for sorting) [12:10:00.0000] <Philip`> (which doesn't seem a hugely important difference in practice, since n will usually be 1 or sometimes 2) [12:26:00.0000] <Philip`> /me decides to adopt a policy of not reading analogies [12:29:00.0000] <annevk42> /me wonders if he should point out that the new iPhone compass allows you to switch between true North and magnetic north [12:36:00.0000] <jgraham> /me wonders wtf the anaology was supposed to mean [12:36:01.0000] <jgraham> and also wtf he can't spell [12:37:00.0000] <Hixie> Philip`: hmm, thanks [12:38:00.0000] <Hixie> wait was my compass analogy bad? or is this someone else's compass analogy [12:38:01.0000] <annevk42> you got a reply [12:38:02.0000] <Hixie> ah [12:40:00.0000] <jgraham> Hixie: FWIW I think "modifying user behaviour" is something that conformance requirements are supposed to do in the sense that the state "author behaviour given conformance requirements" is expected to be different from "author behaviour given no conformance requirements" [12:41:00.0000] <annevk42> people often get the compass interface wrong; don't fix the interface because experts tell you so; instead educate the masses [12:41:01.0000] <annevk42> so as anology that'd mean [12:42:00.0000] <annevk42> people often get summary wrong; don't fix summary because experts tell you so; instead educate the masses [12:42:01.0000] <Hixie> jgraham: i agree that, to the same extent as a compass is intended to modify user behaviour, a validator (and conformance requirements) are intended to modify user behaviour. [12:43:00.0000] <jgraham> Right [12:43:01.0000] <jgraham> /me is way too tired for the compass analogy [12:45:00.0000] <annevk42> it seems to me that true noth + gps obsoletes the need for a compass (maybe even gps by itself if you can walk around a bit) and any kind of trouble the masses might have in understanding it, but maybe that takes the anology too far [12:46:00.0000] <Philip`> I would expect the masses would have somewhat more trouble understanding a GPS device than a compass... [12:47:00.0000] <annevk42> I'm thinking of a map + arrow [12:49:00.0000] <annevk42> I'm not sure what's wrong with getting a better compass though. According to Wikipedia ships widely use a Gyrocompass, which points to true north [12:50:00.0000] <jgraham> annevk42: http://en.wikipedia.org/wiki/Gyrocompass should explain the problem [12:50:01.0000] <Philip`> http://en.wikipedia.org/wiki/File:Kreiselkompass_Schnitt_Anschütz.jpg - looks like a great device to take on an orienteering trip [12:50:02.0000] <annevk42> there's a lot of issues with the magnetic north: http://en.wikipedia.org/wiki/North_Magnetic_Pole#Magnetic_North_and_Magnetic_Declination [12:51:00.0000] <annevk42> jgraham, as I said the new iPhone does true north too [12:51:01.0000] <Philip`> I prefer the approach of not really caring much about where I'm going, and just looking at the sun to work out which direction is more likely to be north than south [12:51:02.0000] <jgraham> annevk42: It is kind of a pain if your phone battery dies when your are days from anywhere :) [12:52:00.0000] <jgraham> Philip`: To be fair it took you months to find Sainsburys so I don't really look to you as a source of wisdom on navigation [12:52:01.0000] <Hixie> lol [12:53:00.0000] <Philip`> jgraham: True :-( [12:54:00.0000] <gsnedders> /me wonders how hard it is to find Sainsbury's [12:54:01.0000] <gsnedders> (I presume the Cambridge one, this is) [12:54:02.0000] <gsnedders> I mean, it's hardly as if it's that far from King's… [12:55:00.0000] <Philip`> It's easy to find - you just walk to Sidney Sussex, then turn around, and that's it [12:56:00.0000] <Philip`> It's right behind the Big Issue seller [12:57:00.0000] <gsnedders> I know where it is [12:57:01.0000] <gsnedders> I just wonder how it can be hard! [12:57:02.0000] <Philip`> It isn't [12:57:03.0000] <gsnedders> /me must confess he constantly forgets where Sidney Sussex is though [12:57:04.0000] <Philip`> It's right next to Sainsbury's [12:58:00.0000] <gsnedders> (And then I get told, "it's opposite Sainsbury's", and I headdesk at forgetting that again.) [12:58:01.0000] <Philip`> /me only knows where Sidney Sussex is because he had some supervisions there; he has no idea where any other colleges are [12:59:00.0000] <gsnedders> Philip`: Do you know where Clare is? Caius? [12:59:01.0000] <gsnedders> Catz? Queens'? [13:01:00.0000] <gsnedders> /me is still mildly amused by the fact that he seems to know Cambridge better than Philip` [13:02:00.0000] <Philip`> I could probably find them by walking around a bit and looking at the signs outside them, but I'd forget within a few hours [13:16:00.0000] <aroben> any idea where the definition of the Storage interface went? [13:17:00.0000] <Hixie> Web Storage [13:17:01.0000] <Hixie> http://dev.w3.org/html5/webstorage/ [13:18:00.0000] <aroben> thanks! [13:43:00.0000] <gsnedders> Philip`: Hint: They are all adjacent to King's [14:14:00.0000] <annevk2> Hixie, "contains-p" does not work as variable in ECMAscript; I suggest containsP ;) [14:28:00.0000] <Hixie> annevk2: that's not better :-) [14:28:01.0000] <annevk2> at least it works [14:30:00.0000] <Hixie> well, contains-p does too (nobody said that was js, after all) [14:30:01.0000] <Hixie> bbiab, lunch [14:30:02.0000] <virtuelv> contains_p [14:31:00.0000] <annevk2> yeah, seems very likely authors will try it somewhere else o_O [14:33:00.0000] <Philip`> Kristof would try it in VBScript, I guess [14:36:00.0000] <annevk2> so does it work there? [14:38:00.0000] <Philip`> No [14:40:00.0000] <annevk2> heh, that was rhetorical question [14:41:00.0000] <Philip`> And that was a rhetorical answer [15:26:00.0000] <Hixie> annevk2: fine fine :-P [15:43:00.0000] <annevk2> oops, forgot to look for DOMStringList usage within HTML5 2009-06-16 [17:05:00.0000] <Hixie> annevk42: is there anywhere i can log issues for cssom/cssom-view ? [18:44:00.0000] <Hixie> /me redefines contact information using the "declarative" style instead of the "imperative" style (along with fixing it at the same time) [18:44:01.0000] <Hixie> easier to understand my foot. [21:20:00.0000] <jwalden> /me wonders how much longer people will incorrectly argue about the complexity of an ordered set implemented as a hash table [00:43:00.0000] <hsivonen> ah the moral high ground of awareness but not solving the problem and alt="Comic" [01:12:00.0000] <Philip`> /me can't quite work out the point of Opera Unite [01:12:01.0000] <Dashiva> Hold that thought [01:13:00.0000] <Madness> I still like face gestures better. [01:13:01.0000] <hsivonen> /me wonders if there's a royalty-free spec [01:14:00.0000] <Philip`> It seems the only real advantage is that it runs the server-side stuff on your own computer, which means you don't have to pay someone for hosting (disk space and CPU) (though Opera still has to pay for all the bandwidth) [01:14:01.0000] <hsivonen> or is it HTTP and not some kind of new p2p protocol? [01:14:02.0000] <Philip`> but that doesn't seem a particularly fundamental difference [01:14:03.0000] <Madness> Philip`, when I first saw it, it reminded me of dropbox. [01:15:00.0000] <Philip`> hsivonen: It's HTTP (with a possibly-custom protocol for proxying) - you write a widget with a web server that handles GETs and POSTs etc [01:15:01.0000] <Philip`> (as far as I can tell) [01:19:00.0000] <hsivonen> the Opera Unite video reminds me of the Du bist Terrorist video [01:19:01.0000] <hsivonen> (the graphical style and narration) [01:25:00.0000] <annevk42> Hixie, the CSS WG does have a bug tracking system [01:59:00.0000] <jgraham> Hmm Mozilla are proposing restricting when proto can be used to mutate an object's prototype. I'll be interested to see how that works out [02:01:00.0000] <fakeolliej> jgraham: their performance argument isn't really valid as it's an issue specific to their implementation, much like there are issues that only occur in implementations like JSC [02:01:01.0000] <hsivonen> MikeSmith: is there a particular reason to change the order of assertions vs. RELAX NG? [02:01:02.0000] <hsivonen> jgraham: is this on an Ecma list somewhere? [02:02:00.0000] <jgraham> fakeolliej: I wondered if that might be the case. But I wonder if they have done a detailed web compat analysis. I understand that proto is really needed if you are not IE [02:02:01.0000] <jgraham> hsivonen: es-discuss [02:02:02.0000] <hsivonen> jgraham: thanks [02:03:00.0000] <fakeolliej> jgraham: the irony of crockford saying that people will only use "safe browsers" that "don;t support proto" [paraphrased] is not lost on me [02:03:01.0000] <jgraham> https://mail.mozilla.org/pipermail/es-discuss/2009-June/009523.html [02:03:02.0000] <hsivonen> uh oh. I have an <isindex> Heisenbug [02:03:03.0000] <fakeolliej> jgraham: because time and time again people have chosen "site working" over security, speed, etc [02:04:00.0000] <jgraham> fakeolliej: Indeed. I wonder if I should ask/say any of this on this list [02:04:01.0000] <fakeolliej> jgraham: we can ask hixie to tell us how many sites use proto [02:04:02.0000] <MikeSmith> hsivonen: no, no current reason I can think of. When I asked you about that before, it was because at that time, I thought maybe your reason for having legacy.rnc was to control the order in which the obsolete-element errors get reported. But I realize now that the reason it be able to report errors in the subtree. [02:05:00.0000] <hsivonen> MikeSmith: ok [02:07:00.0000] <hsivonen> oops. not a Heisenbug [02:08:00.0000] <hsivonen> fortunately [02:14:00.0000] <hsivonen> /me wishes the public-html threads didn't go off on tangents based on analogies [02:15:00.0000] <jgraham> hsivonen: You made ti to the compasses I see ;) [02:15:01.0000] <hsivonen> jgraham: yes [02:19:00.0000] <hsivonen> did anyone ship DOMTokenLish.has() yet? [02:20:00.0000] <MikeSmith> are tokens in HTML5 allowed to contain character references? [02:20:01.0000] <hsivonen> MikeSmith: element and attribute names aren't [02:21:00.0000] <MikeSmith> hsivonen: but tokens in attribute values are, right? [02:21:01.0000] <hsivonen> MikeSmith: the character references are expanded before the token is created, so no in that sense [02:22:00.0000] <Philip`> http://devfiles.myopera.com/articles/608/jsrest-url.htm indicates proto appears on something like 1% of pages, unless I'm misinterpreting the numbers [02:22:01.0000] <MikeSmith> hsivonen: well, I mean serialized attribute values -- e.g., in what the spec defines as "space-separated tokens" [02:23:00.0000] <MikeSmith> or rather "set of space-separated tokens" [02:23:01.0000] <hsivonen> MikeSmith: you can have character references in the serialization, yes [02:23:02.0000] <MikeSmith> k [02:23:03.0000] <hsivonen> /me was talking about different tokens [02:24:00.0000] <Philip`> http://philip.html5.org/data/proto.txt [02:25:00.0000] <annevk42> Philip`, Microsoft uses it in their compat layers for various properties [02:26:00.0000] <Philip`> The Prototype JS library seems to use it for something [02:28:00.0000] <hsivonen> grr. <isindex> is an evil special case for explicit memory management [02:28:01.0000] <hsivonen> aarrgh [02:29:00.0000] <hsivonen> now I'm annoyed at both <isindex> and at C++ not having a garbage collector [02:30:00.0000] <hsivonen> <isindex> is fast appoaching the most-buggy feature of the V.nu parser [02:30:01.0000] <hsivonen> if it isn't already [02:30:02.0000] <hsivonen> approaching [02:32:00.0000] <hsivonen> should I add refcounting for this edge case or add a special hack with 'here be dragons' comments... [02:32:01.0000] <hsivonen> somehow I don't like the idea of refcounting just for this one special corner case [02:32:02.0000] <hsivonen> anyway, yay for unit tests [02:33:00.0000] <Philip`> Is it very unlikely that any other feature will ever trigger the same edge case? [02:33:01.0000] <hsivonen> Philip`: extremely unlikely [02:33:02.0000] <hsivonen> Philip`: <isindex> is the only parser macro [02:34:00.0000] <hsivonen> where attribute names get referenced from more than one attribute holder [02:34:01.0000] <hsivonen> everywhere else custom attribute name objects never leave their original holder [02:35:00.0000] <hsivonen> I guess I'll implement ownership transfer for this case instead of burning cycles refcounting all over [02:37:00.0000] <hsivonen> in case anyone is interested, the problem is that non-well-known attributes are copied from the isindex token to the generated <input> element and then get freed from both [02:38:00.0000] <hsivonen> I'll change things so that they get moved rather than copied [02:38:01.0000] <hsivonen> so that the isindex token no longer has attributes when its attributes are freed [04:15:00.0000] <zcorpan> hmm firefox sets tabindex="0" on video and audio elements [04:37:00.0000] <hsivonen> I have a hard time believing that the masses of Web authors could be able to deal with the codecs parameter [04:40:00.0000] <zcorpan> hsivonen: in canPlayType or in HTTP Content-Type? [04:41:00.0000] <hsivonen> zcorpan: anywhere [04:41:01.0000] <hsivonen> zcorpan: mostly in <source> [04:42:00.0000] <zcorpan> hsivonen: what do you suggest instead? [04:43:00.0000] <zcorpan> look at file extension? [04:43:01.0000] <hsivonen> I don't have a better suggestion [04:43:02.0000] <hsivonen> codecs=theora,vorbis vs. codecss=dirac,vorbis is sane [04:43:03.0000] <hsivonen> the MP4 profile stuff is just too weird [04:44:00.0000] <hsivonen> (IIRC last time I asked, even Apple QuickTime guys weren't sure what the right values for their encoder output would be) [04:44:01.0000] <hsivonen> I guess that's solved by only using it for theora vs. dirac and not using encumbered ISO stuff :-) [04:45:00.0000] <hsivonen> which remins me that we should really be beating the marketing drum about the lack of crazy AVC levels and profiles in Theora [05:15:00.0000] <Philip`> hsivonen: You mean the lack of any way to know in advance that you'll get smooth playback in a wide range of devices when using Theora? :-) [05:17:00.0000] <jgraham> Philip`: That applies to all codecs, for some sutaible definiton of "wide range" [05:19:00.0000] <jgraham> (but yeah Vorbis plays on fewer devices than mpeg-something) [05:19:01.0000] <jgraham> s/vorbis/theora/ [05:20:00.0000] <Philip`> jgraham: When there are levels and profiles, any device can say "I support this set of profiles" and any video can say "I can be played on this set of profiles" and if the intersection is non-empty then you know in advance it'll work, and that seems to be independent of the wideness of the range [05:20:01.0000] <Philip`> and I assume that's the idea of the MPEG stuff, and it's entirely lacking in Theora [05:21:00.0000] <Philip`> (I remember DivX had a similar profile system too) [05:21:01.0000] <jgraham> In practice I expect authors to put in on <source> that wworks on desktop (maybe two) and one that wworks onm the iPhone [05:22:00.0000] <jgraham> The profile system seems like a plausible solution to a technical problem that is highly likely to fail on the web [05:22:01.0000] <jgraham> s/plausible/plausible sounding/ [05:23:00.0000] <Philip`> Oh, I'd agree it seems unlikely to actually work, because it requires authors to understand and care about lots of details that they have no motivation to understand or care about [05:24:00.0000] <jgraham> Right, so not solving the problem doesn't seem like a big deal [05:24:01.0000] <jgraham> Or at least not solving it in that way [05:25:00.0000] <jgraham> In fact it can be a good thing because it leaves the oppertunity open for a solution that does work [05:26:00.0000] <Philip`> It just seems potentially dangerous to use as a marketing claim, because it could be easily twisted into saying Theora doesn't have support for a wide range of mobile devices [05:31:00.0000] <gsnedders> /me wonders what this whole Unite thing is [05:33:00.0000] <Dashiva> Are the profiles really usable that way, though? [05:33:01.0000] <Dashiva> I seem to remember there being about fifteen different ones related to h264 [05:33:02.0000] <annevk42> hsivonen, yeah, I'm not convinced at all that the whole codecs thing and canPlayType will work for anything but the simplest of cases [05:37:00.0000] <hsivonen> Dashiva: as far as I can tell, the h264 profiles are a nightmare for interop [05:38:00.0000] <hsivonen> Dashiva: if you are doing something other than encoding with QuickTime, you can spend a week tweaking parameters so that you interop with anything you can think of [05:45:00.0000] <Dashiva> While still retaining the same quality? [05:45:01.0000] <Philip`> gsnedders: It's something that's top secret and you mustn't talk about it else Opera will fire you [05:45:02.0000] <gsnedders> Philip`: But, but, but I haven't started work at Opera! :P [05:45:03.0000] <gsnedders> /me knows as much about Unite as is publicly available! [05:50:00.0000] <jgraham> gsnedders: That somewhat contradicts your earlier statement [05:52:00.0000] <gsnedders> jgraham: You assume I haven't read anything inbetween those two statemnets [05:52:01.0000] <jgraham> gsnedders: No I assume you haven't read /everything/ in between those two statements [05:53:00.0000] <gsnedders> Well, that's probably true [06:04:00.0000] <Philip`> Most of the suggested uses of Unite seem to be overestimating the uptime of a typical person's home computer [06:06:00.0000] <gsnedders> I't all these engineers assuming everyone is like them [06:07:00.0000] <Philip`> Is "I't" Klingon? [06:08:00.0000] <Philip`> I don't even have a high-uptime home computer myself - I've just got some headless servers that can't run graphical browsers, and some laptops that are switched off or are offline for at least half the day [06:09:00.0000] <Philip`> so they're not exactly great for hosting a web site or sharing music from [06:09:01.0000] <beowulf> i think it's a good idea for showing stuff quickly, when I'm at my computer [06:10:00.0000] <beowulf> saves me the bother of ftp'ing [06:10:01.0000] <beowulf> did i just say that? how lazy have i become? [06:11:00.0000] <Philip`> I suppose it could be useful in situations where someone wants to send me a file by MSN and it doesn't work because I'm using a crazy client, and if they were using Opera then they could share it that way and I could download it trivially [06:12:00.0000] <Philip`> except those people don't use Opera, so that's not going to help much [06:14:00.0000] <Philip`> (I just upload files with the same process as I use for copying local files, i.e. navigating to the right directories in Krusader and pressing F5 to copy between them, and it's no different if some of the directories are over FTP/SFTP/etc) [06:14:01.0000] <annevk42> sshfs ftw [06:15:00.0000] <annevk42> /me uses cd/cp [06:16:00.0000] <Philip`> Things like "Fridge" and "Web Server" and "Photo Sharing" don't make much sense unless you're keeping your computer up all day, else people will get frustrated when they're down, assuming I'm not failing to notice some clever magic in Unite that makes it work when you're offline [06:16:01.0000] <Philip`> annevk42: I tried sshfs for a while but it always seemed a bit unreliable, particularly if I disconnected without unmounting it cleanly first [06:17:00.0000] <Philip`> and I like doing file management in Krusader much more than in bash :-) [06:18:00.0000] <jgraham> I don't think there is any special magic [06:19:00.0000] <beowulf> my example was nonsense [06:19:01.0000] <jgraham> (although I am far from being an expert on unite so I am avoiding saying much lest it be construed as accurate) [06:20:00.0000] <gsnedders> jgraham: But you work for Opera :P [06:20:01.0000] <gsnedders> /me has new headphones [06:20:02.0000] <jgraham> gsnedders: I noticed [06:21:00.0000] <gsnedders> jgraham: What? My new headphones? Where's the camera? [06:21:01.0000] <Philip`> /me is happy to treat Unite in the same way he treats widgets, i.e. by ignoring it entirely because he doesn't see how it would be of any value to him [06:23:00.0000] <jgraham> gsnedders: It is possible you misinterpreted my statement. It is also possible I track your post. [06:23:01.0000] <gsnedders> jgraham: I did work out what you meant, but that's boring in comparison [06:24:00.0000] <beowulf> what's the difference with mobile widgets and offline applications? [06:26:00.0000] <gsnedders> /me wonders how much SEK he should get before leaving [06:27:00.0000] <gsnedders> jgraham: Do you know whether the return tickets for the coach to/from Skavsta have a time limit on them? [06:29:00.0000] <jgraham> gsnedders: I don't recall [06:30:00.0000] <jgraham> gsnedders: There are plenty of cash machines here that will give you SEK [06:30:01.0000] <gsnedders> jgraham: You must be joking :P [06:31:00.0000] <jgraham> gsnedders: Well unless you have a silly bank account that charges you for withdrawls abroad there is little point in getting much cash [06:31:01.0000] <gsnedders> jgraham: As of a few months ago every bank/buidling society here does, AFAIK :( [06:31:02.0000] <jgraham> gsnedders: Even the nationwide? They didn't used to [06:32:00.0000] <gsnedders> jgraham: They do since the 1 June, I think the date was [06:32:01.0000] <gsnedders> 1 May, actually [06:33:00.0000] <gsnedders> "We pass on a Visa charge for processing transactions in certain currencies other than sterling. The fee is currently 0.84% and it will be included in the sterling amount shown on your statement." [06:33:01.0000] <gsnedders> Oh, there's no charge in some countries. [06:33:02.0000] <gsnedders> WTF? [06:33:03.0000] <jgraham> gsnedders: Sweden - No fee charged [06:33:04.0000] <gsnedders> Indeed [06:33:05.0000] <gsnedders> That wasn't the implication of the letter they sent out several months ago [06:34:00.0000] <jgraham> It seems like there is no fee in EU countries and a few others [06:34:01.0000] <gsnedders> /me has to pay for his Visa card though, as that's from Lloyds TSB (as they are, as far as I could tell, the only place to give under 18s a debit card) [06:34:02.0000] <gsnedders> (Whereas I only have a cash card for the Nationwide) [06:35:00.0000] <Philip`> Hasn't anybody written a non-Flash file upload system that has progress bars? (Surely you could get the server to record how much it's received already, and send that back to the client via XHR or something, but I've only ever heard of ones that do it in Flash) [06:37:00.0000] <gsnedders> There are such things [06:38:00.0000] <Philip`> http://www.lonelycode.com/2009/06/09/django-file-upload-progress-bar-without-flash/ - oh, okay [06:38:01.0000] <gsnedders> http://bluga.net/projects/uploadProgressMeter/ too [06:41:00.0000] <Philip`> "It uses the upload progress meter patch and extension from: http://pdoru.from.ro/" [06:41:01.0000] <Philip`> You have to install a patch and recompile PHP? [06:41:02.0000] <Philip`> That really doesn't sound good [06:44:00.0000] <gsnedders> You can't normally do anything until the request has been fully received [06:45:00.0000] <Philip`> Seems like the language is somewhat lacking in flexibility [06:46:00.0000] <gsnedders> Who ever said PHP was a good language? [06:46:01.0000] <Philip`> I presume the people who designed it thought it was [06:47:00.0000] <Philip`> else they would have stuck with Perl [06:47:01.0000] <gsnedders> Wait, PHP was designed? [06:47:02.0000] <Philip`> s/designed/implemented/ [08:58:00.0000] <gsnedders> takkaria: yt? [09:54:00.0000] <takkaria> gsnedders: hi [10:29:00.0000] <gsnedders> takkaria: What do you think of Manchester (university)? [10:38:00.0000] <mpt> Is there a referencable summary/explanation anywhere of the principle that the accuracy of metadata is proportional to its visibility? [10:44:00.0000] <tantek> mpt I'm fairly certain this blog post is (one of?) the first to introduce/describe the principle: http://tantek.com/log/2005/06.html#d03t2359 "Principles of visibility and human friendliness" [10:45:00.0000] <mpt> thanks tantek, that's pretty much what I'm looking for [10:47:00.0000] <tantek> mpt - more microformats principles (in addition to visibility and human friendliness) documented here: http://microformats.org/wiki/principles [11:09:00.0000] <hober> mpt: related is ruby's postulate (accuracy of metadata inversely proportional to the square of the distance between the data and the metadata) [11:09:01.0000] <mpt> ah, that's the one I was thinking of [11:09:02.0000] <mpt> well, it's the form I was thinking of, but a slightly different idea [11:09:03.0000] <hober> from http://www.intertwingly.net/blog/2004/09/23/Copy-and-Paste [11:10:00.0000] <tantek> hober do you have a citation for that? we've been stating that for quite some time in the microformats community re: minimizing impact of DRY violations etc. [11:10:01.0000] <hober> that link is the source [11:13:00.0000] <tantek> thanks hober - crossed messages :) [11:16:00.0000] <mpt> tantek, http://intertwingly.net/slides/2004/devcon/68.html apparently dates from a few months earlier than that post (but is missing any context) [11:17:00.0000] <hober> visible metadata principle also relates to 'people are lazy': http://www.well.com/~doctorow/metacrap.htm#2.2 [11:18:00.0000] <mpt> hm, no, actually, that slide is from a month later [11:19:00.0000] <tantek> note that per the Ruby postulate: "The accuracy of metadata is inversely proportional to the square of the distance between the data and the metadata." [11:19:01.0000] <tantek> if the metadata IS the data itself (c.f. microformats, visible data, DRY), then the distance is 0 [11:19:02.0000] <hober> indeed [11:20:00.0000] <tantek> inverse of 0 squared = infinity [11:20:01.0000] <tantek> ergo the most accurate metadata is to not have separate metadata at all and just use the data as metadata. [11:23:00.0000] <hober> right. given the 'people are lazy' thing, there's no reason to expect people to update content separated at all from the content they /have to/ update. [11:29:00.0000] <Philip`> tantek: In that case it's not metadata at all, it's just data :-p [11:30:00.0000] <gsnedders> /me thought the inverse of 0^2 was undefined :P [11:32:00.0000] <Philip`> Depends on how you define it [11:33:00.0000] <Philip`> If you define it to be defined, then it's not undefined [11:44:00.0000] <tantek> gsnedders - you can define it via Ruby's formula and limits. Calculus 101. [11:57:00.0000] <gsnedders> "Ruby's formula"? Googling it comes up with nothing useful… [11:58:00.0000] <jgraham> gsnedders: The one above... [11:58:01.0000] <gsnedders> Oh, just that [11:59:00.0000] <jgraham> (Although I would really be impressed o see evidence for the squared number rather than just an assertion :) ) [12:00:00.0000] <jgraham> (if people are going to take it seriously) [12:08:00.0000] <tantek> jgraham , indeed, which is specifically why I chose the word "hypotheses" for this blog post which also mentions inverse (geometric) proportionality: http://tantek.com/log/2007/02.html#d19t1813 [12:26:00.0000] <takkaria> gsnedders: uh, I'm not impressed with manchester uni [12:26:01.0000] <takkaria> gsnedders: depends what you're studying, mind [12:26:02.0000] <takkaria> gsnedders: life and physical sciences are excellent [12:26:03.0000] <takkaria> gsnedders: everything else is dross [12:50:00.0000] <Wolfman2000> Afternoon. I'd like confirmation please: there is officially no such concept as frames in HTML5, right? [12:51:00.0000] <Hixie> <iframe> is allowed in HTML5 [12:51:01.0000] <Hixie> <frame> and <frameset> is not [12:51:02.0000] <Wolfman2000> ...wonder why <iframe> is allowed when <object> should be able to replace it [12:51:03.0000] <Hixie> (browsers will continue to support them of course) [12:51:04.0000] <Dashiva> Iframe does more [12:51:05.0000] <Wolfman2000> more of what is my question [12:51:06.0000] <Hixie> <iframe> always has a nested browsing context; <object> might not [12:52:00.0000] <Wolfman2000> ...then I've got an interesting situation. [12:55:00.0000] <Wolfman2000> http://i42.tinypic.com/2w3bcz9.png I'm nearly done re-building/preserving a website for someone with HTML5. Originally, this page was in three frames: the top, the nav, and the rest. I'm trying to keep roughly the same layout as previously, with the disadvantage that once new content is loaded, the scroll bar goes back to the top. I don't want to break accessibility, but......well, I'm a little unsure where to proceed from he [12:55:01.0000] <Wolfman2000> re. [12:57:00.0000] <Hixie> which scrollbar? [12:57:01.0000] <Wolfman2000> the common scrollbar [12:57:02.0000] <Wolfman2000> basically: when you click on a kanji symbol, information shows up on top [12:58:00.0000] <Hixie> oh i see [12:58:01.0000] <Hixie> just use an iframe [12:58:02.0000] <Hixie> :-) [12:58:03.0000] <Wolfman2000> doesn't that also break accessibility? [12:59:00.0000] <othermaciej> I don't know of any reason iframes would break accessibility [12:59:01.0000] <othermaciej> my understanding is that screen readers know how to traverse into iframes [13:00:00.0000] <Wolfman2000> ...I am going to feel awkward using an iframe, but......I can understand the advantages. The question becomes...how do I adapt my framework properly? [13:02:00.0000] <Wolfman2000> /me is using CodeIgniter, PHP, with semi-dynamic content stored in a database. Little unsure how to keep everything working at this point. [13:28:00.0000] <dbaron> Hixie, for what it's worth, I removed support for fallback background colors from Gecko (since the WG dropped the current approach to the feature), so the change to Acid2 is no longer needed [13:30:00.0000] <gsnedders> /me grumbles at Adobe again [13:36:00.0000] <annevk42> Hixie, the other day there was a comment in IRC regarding elements with a content model of "Text"; apparently it does not link to a definition and is only defined for "the HTML syntax" [13:39:00.0000] <Hixie> dbaron: excellent, thanks [13:40:00.0000] <Hixie> annevk42: what would it link to? [13:40:01.0000] <Hixie> annevk42: text is... text [13:41:00.0000] <annevk42> 1) apparently this is not clear to people; the concept of "Text" is also not explained in the section that explains content models 2) text is defined in the HTML writing section confusing matters [13:42:00.0000] <Hixie> annevk42: ah, fair enough. file a bug? [13:42:01.0000] <Hixie> pretty please :-) [13:42:02.0000] <annevk42> sugar on top? [13:42:03.0000] <Hixie> coarse cane sugar [13:44:00.0000] <annevk42> done :) [13:46:00.0000] <Hixie> thanks [14:18:00.0000] <Philip`> jgraham: I presume the squared is so that metadata isn't considered negatively accurate if it comes after the data instead of before [15:22:00.0000] <remysharp> please excuse the newbish question, but where does, if at all, querySelectorAll fit in with HTML5? From the specs I'm reading, it's not - is that correct? [15:22:01.0000] <othermaciej> remysharp: it's a separate spec, not part of HTML5 [15:23:00.0000] <remysharp> you don't have a link handy do you - or at least where I go ahuntin'? [15:23:01.0000] <othermaciej> remysharp: although, sometimes people use the term "HTML5" loosely to describe the whole current generation of client-side Web standards [15:23:02.0000] <othermaciej> http://www.w3.org/TR/selectors-api/ [15:23:03.0000] <remysharp> othermaciej: that's a lot of what I'm seeing [15:23:04.0000] <remysharp> perfect - thanks. [15:24:00.0000] <remysharp> on that topic - of the umbrella term of HTML5 - would you say that Web Sockets are part of HTML5 or not? [15:24:01.0000] <remysharp> given that they've been split out too? [15:24:02.0000] <gsnedders> They're still in the same source file even if the created docs are seperate :P [15:25:00.0000] <remysharp> I'm currently kinda putting it under the same umbrella, but I don't want to be misleading people I speak to [15:25:01.0000] <othermaciej> remysharp: WebSocket is officially a separate specificiation (actually, it will likely be more than one separate spec) [15:25:02.0000] <othermaciej> remysharp: it was once part of HTML5 proper [15:25:03.0000] <othermaciej> people do sometimes include it in the HTML5 blanket term [15:26:00.0000] <remysharp> do you think that's right / is there an opinion? [15:26:01.0000] <othermaciej> do I think what is right? [15:26:02.0000] <remysharp> putting it all under the blanket term [15:26:03.0000] <gsnedders> If you want to wrap it all up and keep it all warm, then it's right :P [15:27:00.0000] <remysharp> gsnedders: then that's what I'll do ;-) [15:27:01.0000] <gsnedders> remysharp: I, personally, would only call HTML 5 what is in the HTML 5 spec :) [15:28:00.0000] <remysharp> you're one of the peeps working on the spec (right?) [15:28:01.0000] <gsnedders> Hixie is the sole author/editor [15:28:02.0000] <remysharp> cripes - I didn't realise it was a one man job! [15:28:03.0000] <remysharp> fair enough. [15:28:04.0000] <gsnedders> He is employed to work on it full time though :) [15:28:05.0000] <gsnedders> (And has been for several years) [15:29:00.0000] <remysharp> I guess what I'm trying to check, is that I'm going to be speaking to a few people about the JS HTML5 APIs - [15:29:01.0000] <remysharp> I don't want to give the wrong information (or too much at least) [15:29:02.0000] <remysharp> and querySelectorAll is something I wanted to point out - [15:29:03.0000] <remysharp> but if it's not part of HTML5 - should people know or understand that? [15:29:04.0000] <remysharp> people == html/js authors [15:29:05.0000] <remysharp> (or the people I'm speaking to certainly!) [15:30:00.0000] <gsnedders> I'd just call it all stuff that's new to browsers :P [15:30:01.0000] <gavin_> /me wonders whether he is related to remysharp [15:30:02.0000] <remysharp> My gut feeling is that I should just put it under the broad term of HTML5 - but I wanted to sound you chaps out [15:31:00.0000] <beowulf> i say they're seperate, otherwise people go looking for them in the html5 spec [15:31:01.0000] <remysharp> which is what I was doing in fact. [15:32:00.0000] <remysharp> erm ... [15:32:01.0000] <remysharp> that all said though - [15:32:02.0000] <remysharp> http://dev.w3.org/html5/ - why is web sockets sitting in there? [15:33:00.0000] <remysharp> is it just a legacy location? [15:43:00.0000] <othermaciej> remysharp: I would prefer to have a separate term for the latest generation of Web standards, one that could encompass the many recent important Web API specs, ES5, and perhaps some of the interesting new CSS3 modules [15:44:00.0000] <othermaciej> remysharp: but I think it would not be a good use of time to correct everyone using the term HTML5 in a broad way [15:44:01.0000] <remysharp> perhaps it's another comet or ajax term [15:44:02.0000] <othermaciej> we've discussed ideas like "Web 5.0" or just "Web 5" to use as a buzzword [15:44:03.0000] <remysharp> though html5 isn't that sexy I guess.... [15:44:04.0000] <othermaciej> but someone has to actually use it first [15:44:05.0000] <remysharp> or the acronym isn't... :) [15:45:00.0000] <remysharp> web 5! my client's heads would explode!!! [15:45:01.0000] <remysharp> they're seeing blogs and referring to them as "two dot oh" [15:45:02.0000] <remysharp> lol [15:45:03.0000] <remysharp> /me can I lol at my own statements...I think not... [15:47:00.0000] <remysharp> so, backpeddling a bit - am I right in saying the storage api is also not* part of HTML5 [15:50:00.0000] <othermaciej> technically it's not part of the actual HTML5 spec [15:50:01.0000] <othermaciej> it was once [15:50:02.0000] <othermaciej> now it is separate [15:50:03.0000] <othermaciej> though it is still part of the blanket term many people use [15:50:04.0000] <othermaciej> for example Google [15:51:00.0000] <remysharp> okay, good - sorry, it's just I'm going through the specs trying to wrap my head around what is and isn't HTML5. [15:51:01.0000] <remysharp> My eyes are opening now - cheers! 8-) [15:51:02.0000] <remysharp> /me that wasn't supposed to be shades... [15:52:00.0000] <remysharp> one last thing before I leave you guys be for another day? I've got a drag and drop question - is that cool for this channel? [15:57:00.0000] <gsnedders> remysharp: Yeah, sure, whatever. It's really just a social channel, where most of the people involved in HTML 5 hangout and discuss stuff, including HTML 5 [15:57:01.0000] <remysharp> cheers - http://jsbin.com/ohuzi [15:57:02.0000] <remysharp> needs Firefox 3.5b4 - with DnD support [15:57:03.0000] <remysharp> and a console to see what's going on [15:58:00.0000] <remysharp> The 'drop' event is never triggering and I can't for the life of me work out why... [16:00:00.0000] <remysharp> This is the working example I based my code from - which works, but I don't understand why some events we return false, and why we have to setData on the transfer prop: http://jsbin.com/uzima [16:05:00.0000] <tantek> hober, regarding metadata, and inaccuracy, here is an illustrative photo: http://www.flickr.com/photos/tantek/430429916/ [16:09:00.0000] <hober> tantek: excellent [16:10:00.0000] <remysharp> okay, so if there's anyone left active in this room - if I've got an authoring question (which is hitting walls here), do I shoot it to help⊙wo - is that the right mailing list? Cheers [16:12:00.0000] <Hixie> that's the right list, whether you'll get an answer depends on if anyone knows the answer :-) [16:13:00.0000] <Hixie> for the drag and drop events i really need to write a better intro [16:13:01.0000] <gsnedders> Hixie: Yeah, learn to write! [16:13:02.0000] <gsnedders> /me still can't decide what to do at uni… [16:13:03.0000] <gsnedders> /me is quite often tempted by English [16:13:04.0000] <remysharp> It's not so much the intro - it just needs a quick demo of code [16:14:00.0000] <remysharp> that doesn't use inline handlers [16:14:01.0000] <remysharp> (based on the only one I could find) [16:15:00.0000] <Hixie> that's what i mean by intro :-) [16:17:00.0000] <remysharp> Ah, then yeah - that would help a ton :-) 2009-06-17 [19:22:00.0000] <JoePeck> I'm reading about localStorage at http://dev.w3.org/html5/webstorage/#introduction [19:22:01.0000] <JoePeck> It says "The localStorage DOM attribute is used to access a page's local storage area." [19:22:02.0000] <JoePeck> by "page" does it mean "domain" ? [19:23:00.0000] <JoePeck> I played around in WebKit and it seems to mean domain. I think using the word "domain" would be more clear, but I haven't read too much of the specs. Is it common to use the word "page" and mean domain? [19:24:00.0000] <JoePeck> I haven't read too many of the specs. [19:59:00.0000] <othermaciej> JoePeck: it means "page", but all pages on the same domain share the same ultimate back end for storage [20:00:00.0000] <JoePeck> othermaciej: I think the use of "page" in the example is okay, but the line before the example I think is misleading [20:01:00.0000] <JoePeck> albeit I'm still reading through the document, and I haven't gotten down to the technical documentation of localStorage yet =) [20:45:00.0000] <JoePeck> Is there a legend describing the many different styles of content used in the specifications. For instance, what does it mean when something is boldy underlined in this thick blue-gray line? http://tr.im/oK6j?grabup [20:46:00.0000] <JoePeck> wow, I don't see my entire line (using Client Linkinus) so I'll just repeat the last part [20:46:01.0000] <JoePeck> For instance a thick blue-gray underline [20:46:02.0000] <JoePeck> http://tr.im/oK6j?grabup [20:47:00.0000] <JoePeck> The style rule that applies that underline seems to explain, any paragraph or span without a title gets the underline [20:47:01.0000] <JoePeck> so its probably an indicator for something missing an explanation? [20:53:00.0000] <JoePeck> I'm guessing IRC channel is probably busier during normal work hours? =) [00:07:00.0000] <Hixie> http://www.w3.org/mid/21B8BC99-055B-47AF-9515-6CE680B149BA⊙gc [00:08:00.0000] <Hixie> (follows this e-mail from the same thread: http://www.w3.org/mid/AC39CC88-2B2F-41F2-B69E-5DF02A3F0E5E⊙gc ) [00:26:00.0000] <othermaciej> I had to review a poorly written spec today which seemed to be badly attempting to imitate HTML5 style [00:26:01.0000] <othermaciej> it was full of MUSTs with no obvious referent [00:28:00.0000] <othermaciej> and stepwise algorithms like: "(1) If A, do B, then go to step 3. (2) If !A, do nothing, then go to step 3. (3) Terminate this algorithm." [00:29:00.0000] <Hixie> heh [00:29:01.0000] <Hixie> the only algorithms i have that are that bad are the ones that started much more complicated and that i didn't realise got simple enough to rewrite sanely [00:29:02.0000] <Hixie> sadly there are some i'm sure :-( [01:10:00.0000] <Hixie> woo, got the number of XXXs below 200 [01:10:01.0000] <Hixie> which was my goal for the quarter [01:11:00.0000] <annevk42> CSSOM is done in the CSSWG fwiw [01:11:01.0000] <annevk42> as for where to log it, maybe in the CSSWG issue tracker? [01:13:00.0000] <Hixie> how do i add things to that? i looked in bugzilla but didn't find a css component [01:14:00.0000] <Hixie> (and i didn't think sending anything to www-style would be conducive to anything getting done :-) ) [01:14:01.0000] <Hixie> ah, found http://www.w3.org/Style/CSS/Tracker/ [01:14:02.0000] <Hixie> wow, that's awesome [01:14:03.0000] <Hixie> almost every action is overdue [01:16:00.0000] <Hixie> annevk42: ok added, thanks [01:16:01.0000] <Hixie> ok bed time [01:16:02.0000] <Hixie> nn [01:21:00.0000] <annevk42> /me "fixes" due dates for his actions [01:27:00.0000] <annevk42> nn [01:54:00.0000] <remysharp> morning - did I read somewhere that class manipulation (as in class attributes) is covered somewhere in HTML5/web apps api? [01:55:00.0000] <annevk42> search for classList and DOMTokenList [01:56:00.0000] <remysharp> cheer - exactly what I was after. [01:56:01.0000] <remysharp> cheers [01:59:00.0000] <remysharp> darn - none of the browsers have it yet though :-( [01:59:01.0000] <annevk42> maybe if you write a compelling test suite ;) [02:00:00.0000] <remysharp> would that spur on development for the classList.add functionality then? [02:01:00.0000] <annevk42> having a good test suite is one of the main concerns when implementing new features [02:02:00.0000] <annevk42> I can't promise anything of course, but it will increase the odds :) [02:02:01.0000] <remysharp> :) [02:02:02.0000] <remysharp> Is there any guidelines on test suites? [02:02:03.0000] <remysharp> not that I've got oodles of time on my hands, but if it's something I can help with, there's no reason I shouldn't. [02:04:00.0000] <jgraham> My new mantra for testsuites is: don't make the harness more complex than the tests :) [02:05:00.0000] <jgraham> (but there are also probably useful guidelines somewhere) [02:05:01.0000] <Philip`> I think I tend to violate that mantra :-( [02:05:02.0000] <annevk42> remysharp, we don't have published guidelines, but in general being able to run them automated would be nice and when opening the test it would be nice if it was directly clear whether it passed or failed [02:05:03.0000] <annevk42> remysharp, e.g. http://tc.labs.opera.com/apis/getElementsByClassName/001.htm is clear on whether it passes or fails (it's not integrated into a test harness though atm) [02:06:00.0000] <remysharp> okay, that'll do for a starting point for me. [02:06:01.0000] <Philip`> /me has seven hundred lines of Python for his canvas tests [02:06:02.0000] <remysharp> John Resig has an arse load of tests for jQuery and class manipulation, any reason why I shouldn't glean tests from there? [02:07:00.0000] <jgraham> Philip`: Actually having lots of external code is OK. It's when there are 10,000 lines of javascript in the harness for 1 line tests that yo have a problem [02:09:00.0000] <remysharp> annevk2: seeing as you've appeared to self.clone(true) yourself, did you catch my last question on taking a lead from the jQuery tests for addClass et al? [02:09:01.0000] <annevk42> remysharp, can't think of any [02:09:02.0000] <remysharp> cool [02:09:03.0000] <Philip`> jgraham: Oh, good, I've only got ~250 lines of JS in the harness, plus a couple of dozen lines of HTML boilerplate in each test case [02:09:04.0000] <annevk42> remysharp, that's just my other laptop waking up to play music :) [02:09:05.0000] <remysharp> ah, fair enough, if only there was a self.clone function ... [02:09:06.0000] <remysharp> /me dreaming on [02:09:07.0000] <annevk42> :) [02:16:00.0000] <jgraham> remysharp: Parts of http://www.w3.org/Style/CSS/Test/guidelines.html are relevant to HTML tests [02:16:01.0000] <remysharp> cool - cheers. [02:18:00.0000] <jgraham> (specifically all of section 2, and much of section 4) [02:38:00.0000] <annevk42> heh, http://www.youtube.com/watch?v=o4MwTvtyrUQ is great [02:43:00.0000] <hsivonen> annevk42: the sad part is that over here a person with that level of conceptual clarity gets to censor DNS entries [02:52:00.0000] <annevk42> I suppose there's only so many smart people that want to be in governing [02:52:01.0000] <annevk42> s/smart/informed/ [02:53:00.0000] <jgraham> Goos substitution :) [02:53:01.0000] <jgraham> Good [02:53:02.0000] <jgraham> /me should give up [02:53:03.0000] <annevk42> maybe take a day off to get over that :p [02:56:00.0000] <hsivonen> well, having to look at the kind of stuff that may qualify for censoring doesn't seem like a job one would want to have [03:01:00.0000] <hsivonen> contains-p? how did that name come about? [03:01:01.0000] <hsivonen> hmm. maybe I misunderstood the diff [03:03:00.0000] <hsivonen> it scares me that one day I or Someone Else will have to review the Gecko script execution choreography against HTML5 [03:03:01.0000] <hsivonen> now I'm using the old script execution dance [03:05:00.0000] <annevk42> in theory you can reveal bugs by tests [03:06:00.0000] <annevk42> I guess fixing the bugs will be the interesting bit [03:13:00.0000] <Philip`> Writing the tests will be 'interesting' too [03:14:00.0000] <hsivonen> reviewing existing test cases vs. spec will be 'interesting', too [05:27:00.0000] <annevk42> I wonder why I'm still surprised into what detail the TAG has studied HTML5: http://lists.w3.org/Archives/Public/www-archive/2009Jun/0092.html [05:32:00.0000] <myakura> Is there something which defines XML syntax for XQuery? [05:32:01.0000] <myakura> /me found and kinda amazed there exists such: http://www.w3.org/TR/xqueryx/ [05:33:00.0000] <annevk42> it's pretty crazy [05:34:00.0000] <annevk42> especially when you compare the two [05:34:01.0000] <myakura> http://www.w3.org/TR/xqueryx/#Examples [05:35:00.0000] <annevk42> hmm, in Nov 2004 I said "I once read XQueryX and I'm glad that one didn't made it." [05:35:01.0000] <annevk42> apparently it did in Jan 2007! [05:36:00.0000] <JohnResig> oh Jesus, that's some crazy syntax [05:36:01.0000] <myakura> /me is surprised once again, thought it was just a W3C Note or such [05:36:02.0000] <JohnResig> it's not often that you see "here's 3 lines of simple code translated into this 150 line pile of mess, impressed?" [05:37:00.0000] <annevk42> gives "tag soup" a whole new meaning [05:39:00.0000] <myakura> 'The [XML Query 1.0 Requirements] states that "The XML Query Language MAY have more than one syntax binding. One query language syntax MUST be convenient for humans to read and write. One query language syntax MUST be expressed in XML in a way that reflects the underlying structure of the query."' [05:43:00.0000] <jgraham> That is a pisstake, right? [05:47:00.0000] <myakura> /me tries to finde an impl. report for that spec. [06:31:00.0000] <Philip`> "The ideal situation: you can take any HTML5 document, convert it to some XML-based language designed for the purpose (not necessarily XHTML), convert it back, and get a semantically equivalent HTML5 document." [06:31:01.0000] <Philip`> That seems trivial - just output <html5><![CDATA[<!doctype html><title>foo</title>...]]></html5> [06:31:02.0000] <Dashiva> Brillant! [06:32:00.0000] <Dashiva> You could also print it out and take a picture of it, and store the image as XML [06:33:00.0000] <Philip`> http://thedailywtf.com/Articles/Interesting-Bitmap.aspx [06:33:01.0000] <jgraham> You mean like <pixel><red>1</red><green>1</green><blue>1</blue></pixel> [06:45:00.0000] <Philip`> jgraham: That's not very extensible [06:45:01.0000] <Philip`> You need to be able to cope with multiple bit-depths [06:46:00.0000] <Philip`> and floating-point pixel values too [06:46:01.0000] <Philip`> and that needs to be configurable per pixel, so you can losslessly merge multiple images into one [06:48:00.0000] <annevk42> <!doctype html><title></title>]]> mwaha [06:51:00.0000] <Philip`> annevk42: Obviously any characters like ']' or '&' or U+0000 etc in the input will get converted to '&#nnn;' before you put the <html5><![CDATA[ wrapper around it [06:53:00.0000] <annevk42> oooh :p [07:00:00.0000] <jgraham> Philip`: To be fair I was assuming pure floating point pixels :) [07:01:00.0000] <Philip`> jgraham: That's not going to work if I want a pixel to be precisely a third of maximum brightness [07:02:00.0000] <jgraham> What if I want a pixel to be exactly 1/pi of the maximum brightness? [07:03:00.0000] <Philip`> You would write a schema for your pixel brightnesses, then use its URI as the namespace for your component elements [07:06:00.0000] <Philip`> (The schema would look like <pixel:schema xmlns:pixel="..."><pixel:type id="eightbit"><pixel:max inclusive="1">255</pixel:max><pixel:min inclusive="1">0</pixel:min><pixel:descr>Linear mapping from value onto pre-gamma brightness</pixel:descr></pixel:type></pixel:schema> with an equivalent bit for your 1/pi scale) [07:23:00.0000] <Lachy> http://www.darkreading.com/security/encryption/showArticle.jhtml?articleID=217801293 [07:24:00.0000] <Lachy> the article says they can create a darket using a page written using PHP and an "HTML 5-based browser" [07:24:01.0000] <doublec_> I guessing they don't say which one [07:24:02.0000] <Lachy> I'm curious what feature of HTML5 they've actually used to that, or whether "HTML5" has become a replacement buzzword for Ajax, and they're simply using some new scripting techniques [07:30:00.0000] <Philip`> Maybe they're using local storage [07:32:00.0000] <Lachy> I'm not sure how localstorage would help you connect to a darknet [07:33:00.0000] <Philip`> It would help you store data that you could share through the darknet [07:33:01.0000] <Lachy> maybe, but I'm more curious about the connection technique [07:33:02.0000] <Philip`> except apparently they don't persist any data after you shut down the browser [07:34:00.0000] <Lachy> I expect they're using XHR to connect to the server, and it's the server itself that handles the connection to other distributed servers [07:34:01.0000] <Lachy> which would explain why they said it needed PHP [07:35:00.0000] <Philip`> http://www.forbes.com/2009/06/15/darknet-hewlett-packard-technology-security-darknet.html - "Someone navigates to a Web site that serves up some JavaScript code that runs in the user's browser. That code uses the local storage capacity built into the latest version of browsers like Google ( GOOG - news - people ) Chrome and Internet Explorer." [07:35:01.0000] <Lachy> so it's not entirely anonymous, but it would be difficult to work out which server in the network a given user is connected to, assuming there are enough servers to make a process of elimination improbable [07:36:00.0000] <Philip`> "It's more like a peer-to-peer network where the computer nodes can't talk to each other directly." - so, it's like a client-server network, and not like a peer-to-peer network at all? [07:38:00.0000] <Lachy> yeah, that makes sense [07:40:00.0000] <Dashiva> http://www.facebook.no/ [07:40:01.0000] <Dashiva> Those are some interesting tooltips you get when hovering Facebook, © and 2009 in the lower corner [07:48:00.0000] <gsnedders> Dashiva: Translate? [07:48:01.0000] <gsnedders> Oh, there [07:48:02.0000] <gsnedders> I looked at the Facebook in top left :P [07:49:00.0000] <gsnedders> Dashiva: The "20" and "09" are different too [07:49:01.0000] <Philip`> The title text on the © is different to what's in the source, too [08:04:00.0000] <Lachy> Philip`, what? [08:06:00.0000] <Lachy> oh, the title is being modified by a script for some reason [10:27:00.0000] <annevk2> http://docs.google.com/View?id=dcgd8hk6_2g7c6zzc6 [10:32:00.0000] <annevk2> https://lists.webkit.org/pipermail/webkit-dev/2009-June/thread.html#8066 also suggests the display types are not actually supported [10:32:01.0000] <annevk2> hmm [10:35:00.0000] <JoePeck> is there a guide/legend for the different types of styles/syntax in the html5 specifications? [10:36:00.0000] <annevk2> JoePeck, section 1.8.2. does not help? [10:38:00.0000] <JoePeck> annevk2: For the most part, and thats what I was looking for, but it didn't help me with the 1 syntax I was actually wondering about. [10:38:01.0000] <JoePeck> annevk2: thick blue underlines: http://tr.im/oOv2?grabup [10:39:00.0000] <JoePeck> although the CSS rule makes it pretty clear its for a span without a title... maybe when the spec is completed these should have been eliminated? [10:39:01.0000] <annevk2> ah, thick blue underlines are an unimplemented feature [10:39:02.0000] <annevk2> they are references to a term defined in another specification [10:40:00.0000] <annevk2> they should really be links, but the preprocessor is not smart enough just yet [10:40:01.0000] <JoePeck> annevk2: ahh, I had that feeling. Thanks! =) [10:44:00.0000] <Philip`> annevk2: It would be nice if that document said who wrote it and who to send feedback to [10:44:01.0000] <Philip`> (Not that I want to send feedback myself) [10:45:00.0000] <annevk2> the second link I provided indicates the author [10:46:00.0000] <Philip`> annevk2: Seems like a more general failing of Google Docs that it gives no indication of who's responsible for a document [10:47:00.0000] <Philip`> At least with normal web documents you can look around the domain it's hosted on to find someone who probably knows about it [11:56:00.0000] <JoePeck> http://downforeveryoneorjustme.com/http://lists.whatwg.org/ [11:58:00.0000] <annevk2> weird, nothing on http://twitter.com/dhstatus [12:01:00.0000] <Hixie> wow did rigo really use "think of the children" as an argument [12:05:00.0000] <annevk2> he was right about P3P [12:05:01.0000] <annevk2> and to a certain extent it is being done again with Powder afaict [12:08:00.0000] <Hixie> not really, one of the bugs he quoted had a comment from someone basically saying the same as i did, and maciej replied saying that the same applied for him [12:08:01.0000] <Hixie> and i can assure you the same applies for chrome [14:39:00.0000] <expensivelesbian> hello. Are questions about seeming quirks in the HTML5 validator welcome here? I seem to be getting an odd response using multiple source elems inside a media elem [14:39:01.0000] <annevk2> there's a good chance hsivonen will pick it up [14:40:00.0000] <annevk2> alternatively you could use http://bugzilla.validator.nu/ [14:40:01.0000] <expensivelesbian> thanks [14:49:00.0000] <billyjackass> expensivelesbian: about validating multiple source elems inside a media elem, what's the question? do have your source or a test case online? [14:50:00.0000] <gsnedders> /me doesn't know what to say about a billyjackass and an expensivelesbian in #whatwg… There has to be a joke in there. [14:51:00.0000] <krijnh> :P [14:51:01.0000] <krijnh> /me adds a new subtitle [14:51:02.0000] <billyjackass> heh [14:51:03.0000] <expensivelesbian> billyjackass, sure, ignoring doctype stuff, the basic meat of the media elem is quite simple [14:51:04.0000] <gsnedders> krijnh: Do I have to reload the page until I find it? [14:51:05.0000] <expensivelesbian> safe to paste here? [14:52:00.0000] <krijnh> gsnedders: no, why do people keep asking that :) [14:52:01.0000] <gsnedders> krijnh: Where then!? [14:52:02.0000] <billyjackass> expensivelesbian: pastebin.com or somewhere is better, probably [14:52:03.0000] <expensivelesbian> k [14:53:00.0000] <annevk2> krijnh, nobody remembers the URL [14:53:01.0000] <krijnh> annevk2: that's a feature [14:53:02.0000] <gsnedders> krijnh: What is the URL!? [14:54:00.0000] <krijnh> I don't know either :$ [14:54:01.0000] <expensivelesbian> billyjackass http://pastie.org/515647 [14:55:00.0000] <gsnedders> krijnh: You must find it and tell us! We demand it of you! [14:55:01.0000] <expensivelesbian> billyjackass running that through the HTML5 validator produces an error, but I'm not exactly sure why, as it seems like it's following the spec [14:55:02.0000] <krijnh> Then I really have to come up with a new subtitle as well [14:55:03.0000] <expensivelesbian> it doesn't like the source elem for some reason [14:58:00.0000] <billyjackass> expensivelesbian: OK, I'll take a look now [15:29:00.0000] <billyjackass> expensivelesbian: sorry, got distracted. looking at the source for the schema that v.nu uses, it seems that it is requiring that if you have multiple source elements there, the first source in document order must have a type attribute [15:30:00.0000] <billyjackass> I don't know if that's per-spec or a schema bug, looking now [15:30:01.0000] <annevk2> hmm, sees like either type or media in that case would make sense [15:31:00.0000] <billyjackass> annevk2: yeah, that's what the schema requires actually [15:31:01.0000] <expensivelesbian> right, phew. That's a webkit quirk, which I raised a ticket about a while back. [15:31:02.0000] <annevk2> HTML5 however does not seem to make any requirements [15:31:03.0000] <expensivelesbian> thanks for your insights into that [15:31:04.0000] <billyjackass> source.attrs = common.attrs & src & ((type & media?) | media) [15:32:00.0000] <expensivelesbian> if I remember, I switched them around to try and coax iPhone <audio> support into life. That's still an ongoing project [15:32:01.0000] <annevk2> it does require src [15:32:02.0000] <annevk2> billyjackass, if you move the ? to after the ) would it work? [15:34:00.0000] <billyjackass> annevk2: yeah, it'd work but would mean something different of course. I'm looking at the spec now to see what it really should be [15:34:01.0000] <annevk2> well, with work I mean matching the spec :) [15:37:00.0000] <expensivelesbian> I tried adding the type attr to the audio elem for various Safari specific encodes, and it breaks it, hence why I removed it. It was this that I raised as a ticket with the Webkit people. But this is slightly off topic for here [15:38:00.0000] <billyjackass> expensivelesbian: well, as far as the spec goes, the current spec does not require it there (as annevk2 points out) [15:38:01.0000] <billyjackass> maybe some previous draft did, but the current one does not seem to [15:38:02.0000] <billyjackass> so it's a bug in v.nu [15:40:00.0000] <billyjackass> annevk2: it seems like this would match the spec too: source.attrs = common.attrs & src & type? & media? [15:41:00.0000] <expensivelesbian> ok, that's interesting to know. I'd assumed the type was a requirement, and webkit was being naughty. However, the general gist of the ticket is still for them to fix, as they are (by their own design, if I remember) supposed to be ignoring this attr. I can't remember the history of this now, but some dude at Apple said "yeah, it's a bug" [15:41:01.0000] <billyjackass> ok [15:42:00.0000] <annevk2> i think it changed because the spec now actually fetches the resource to determine the type as well [15:42:01.0000] <billyjackass> ah [15:42:02.0000] <annevk2> loading media resources changed like lots of times [15:42:03.0000] <billyjackass> yeah [16:27:00.0000] <Hixie> annevk42: where is the credentials flag initialised in xhr2? 2009-06-18 [17:13:00.0000] <gsnedders> /me wonders why Google.com is being so slow [17:14:00.0000] <jcranmer> gsnedders: a lot of networks just went down? [17:15:00.0000] <gsnedders> I mean, taking a minute to connect to "www.google.com" is wrong. [00:06:00.0000] <heycam> Hixie, ping? [00:07:00.0000] <Hixie> here [00:08:00.0000] <heycam> so i'm thinking about your [XXX]es that are meant to be like "partial" interfaces [00:08:01.0000] <heycam> and i see that they all only have attributes on them, no operations [00:09:00.0000] <heycam> so that means what you really want is to avoid having a new object in the prototype chain for objects implementing these interfaces [00:09:01.0000] <heycam> e.g. you don't want a new, empty prototype object in the chain that corresponds to the interface with the applets attribute on it [00:10:00.0000] <heycam> but, i don't think "extending" the interfaces, at the IDL level, is the right thing to do [00:10:01.0000] <heycam> you just want to control the JS aspects of it [00:10:02.0000] <heycam> so e.g. if these extended interfaces did have operations, then they would go on to the HTMLDocument prototype object (in this case) [00:13:00.0000] <heycam> well actually i see you have two kinds of XXXes now [00:13:01.0000] <heycam> on WindowModal, where you separate it because sometimes you want this interface to be implemented and sometimes not, but if it is, then there's no extra object in the prototype chain for WindowModal [00:14:00.0000] <heycam> and for the things in the Obsolete section, where they're separate just so you don't see them in the earlier sections [00:14:01.0000] <heycam> but they are actually part of the same interface [00:16:00.0000] <heycam> ok maybe i'm talking myself into allowing "partial" interface definitions [00:16:01.0000] <heycam> for that second case [00:16:02.0000] <heycam> even though it's only because of how you want to organise the spec [00:16:03.0000] <heycam> but for the WindowModal case i think your XXX does need to be something like NoPrototypeObject [00:19:00.0000] <heycam> maybe partial should also exist to support adding new constants from different specs, like DOMException codes [00:22:00.0000] <heycam> also, i'm thinking of moving a bunch of things from extended attributes into the "proper" syntax [00:22:01.0000] <heycam> like optional arguments, variadic operations [00:23:00.0000] <heycam> i'm gonna give up pretending web idl is a superset of omg idl [00:26:00.0000] <Hixie> (still here, hold on) [00:27:00.0000] <Hixie> heycam: basically my desire is to not have the spec separation affect the prototype chain [00:28:00.0000] <Hixie> heycam: so we should be able to take, e.g., HTMLDocument, or Window, and split it amongst multiple specs and have zero effect on implementation requirements [00:28:01.0000] <Hixie> heycam: basically i'm trying to avoid the problem hsivonen mentions regularly where you end up splitting the tech into silos per wg [00:30:00.0000] <heycam> so, having a [NoPrototypeObject] could do that [00:30:01.0000] <heycam> especially if the interface only has attributes, since they never go in prototype objects [00:30:02.0000] <Hixie> they'll have methods sometimes [00:30:03.0000] <Hixie> i'm surprised that they don't already [00:31:00.0000] <heycam> maybe something like [AugmentPrototype=Window] instead [00:31:01.0000] <Hixie> yes [00:31:02.0000] <Hixie> that would work [00:31:03.0000] <Hixie> and then the "actual" interface name wouldn't matter? [00:32:00.0000] <Hixie> i.e. it'd be an editorial concern only? [00:32:01.0000] <heycam> yes, except for other language bindings [00:32:02.0000] <heycam> which don't have prototypes [00:32:03.0000] <Hixie> well for bindings that have classes i'd want the same thing [00:32:04.0000] <heycam> they'd be proper, separate interfaces still [00:32:05.0000] <heycam> yeah? [00:32:06.0000] <Hixie> the idea is for there to be zero differences [00:32:07.0000] <heycam> how should it matter? [00:32:08.0000] <heycam> if you say "all HTMLDocument objects also implement HTMLDocumentObselete" [00:33:00.0000] <heycam> oh casting would matter [00:33:01.0000] <Hixie> don't you have to explicitly cast from one to the other? [00:33:02.0000] <Hixie> right [00:33:03.0000] <heycam> i mean, you'd need to cast [00:33:04.0000] <heycam> hmm [00:33:05.0000] <heycam> so it sounds like you exactly want a "partial" modifier on interface [00:34:00.0000] <Hixie> the point is we should be able to take HTML5, 10 years from now, when it's all perfectly implemented (hah), and make 15 specs out of it and be able to release that as an Edited Recommendation that has zero changes required to any test suites in any languages [00:34:01.0000] <Hixie> with all the same code working unchanged [00:34:02.0000] <Hixie> that's basically what i'm doing, btw -- taking DOM2 HTML, and splitting it into the "obsolete" and "today" sections in html5 [00:35:00.0000] <Hixie> 10 years after DOM2 HTML was published [00:36:00.0000] <heycam> ok i'm gonna go for a "partial" keyword then, see how it goes [00:36:01.0000] <Hixie> cool [00:36:02.0000] <heycam> partial interface HTMLDocument : ... { / now stuff / }; partial interface HTMLDocument { / old stuff / }; [00:37:00.0000] <Hixie> do i put it on all interfaces? [00:37:01.0000] <heycam> hmm [00:37:02.0000] <heycam> you want any interface to be extendable? [00:37:03.0000] <Hixie> i'm sure there will be specs that add to Window if there aren't already [00:37:04.0000] <heycam> right [00:37:05.0000] <Hixie> and it would have been nice to be able to add to HTMLElement instead of having the side interfaces, for some cases [00:38:00.0000] <Hixie> but i dunno [00:38:01.0000] <heycam> yeah i'm not sure [00:38:02.0000] <Hixie> i mean, if we have to have it to extend, you're basically saying it's the opposite of "final" [00:38:03.0000] <heycam> well, not really [00:38:04.0000] <heycam> you can extend with ":" [00:38:05.0000] <heycam> but it's a different kind of extending [00:38:06.0000] <Hixie> yeah [00:39:00.0000] <Hixie> ":" and also with "side" interfaces [00:39:01.0000] <Hixie> like how EventTarget is on all Nodes [00:39:02.0000] <heycam> right [00:39:03.0000] <heycam> depends on how you want the prototype chain to look [00:39:04.0000] <Hixie> but for some cases like Window that seems... wrong [00:39:05.0000] <Hixie> yeah [00:39:06.0000] <heycam> yeah Window is a bit special in that regard, in that you think of it as just a pool of properties that you add to [00:39:07.0000] <Hixie> i guess in practice people would just say that the interface they want to extend had "partial" on it [00:39:08.0000] <Hixie> even if it didn't [00:40:00.0000] <heycam> heh perhaps [00:40:01.0000] <Hixie> so in practice even if you say we have to put it on both, it won't matter [00:40:02.0000] <Hixie> it's just nice to annotate one of them (the "not primary" one, there's usually a clear one that is the "primary" one) as being an extention of another [00:42:00.0000] <heycam> it's kinda weird for languages like java to have an interface whose definition changes depending on which specs you decide to implement [00:43:00.0000] <Hixie> why? [00:43:01.0000] <Hixie> how is that different from when you decide how much of one spec to implement? [00:43:02.0000] <Hixie> it's not like spec boundaries reflect actual implementation boundaries, like, ever [00:44:00.0000] <heycam> it's different if you want to use two libraries together, say [00:44:01.0000] <heycam> let's say a spec first defines interface A [00:44:02.0000] <heycam> and a later spec defines an extension to A [00:44:03.0000] <heycam> and both of these specs used this "partial" keyword to allow this [00:45:00.0000] <heycam> one library was compiled with the first A, and another library was compiled with the extended A [00:45:01.0000] <heycam> now both libraries have different A.class files [00:45:02.0000] <heycam> whichever one the class loader finds first is the one that will be usable [00:45:03.0000] <heycam> you need to be sure to get the extended one loaded first somehow [00:46:00.0000] <heycam> but then it's impossible if there were two extensions to A, and each library chose the different extended one [00:46:01.0000] <heycam> you couldn't get both libraries working together in the one program [00:48:00.0000] <Hixie> what about if you use two libraries, and they both implement a part of the same spec? [00:48:01.0000] <Hixie> same problem [00:48:02.0000] <Hixie> literally the same problem, because the next day that spec might be split into two [00:49:00.0000] <Hixie> and the day after that the specs might be merged together again [00:49:01.0000] <heycam> you could at least have the same all-in-one interface implemented by both libraries, but some of those methods just throw a NotImplementedException or something [00:49:02.0000] <heycam> in fact you would need to, to get it to compile [00:50:00.0000] <Hixie> i guess in java maybe. in javascript we encourage implementations that don't support something to not expose it, so that you can test to see if it exists [00:50:01.0000] <heycam> yes [00:51:00.0000] <heycam> it's a benefit of not having to implement a whole interface i suppose [00:51:01.0000] <Hixie> well it's a benefit of easy reification [00:52:00.0000] <heycam> benefit of duck typing (sorta)? [00:53:00.0000] <Hixie> vaguely [00:53:01.0000] <Hixie> not exclusively though [00:53:02.0000] <Hixie> it's a benefit of late binding and easy reification [00:54:00.0000] <Hixie> sorry, not reification, reflection [00:54:01.0000] <Hixie> late binding and reflection [00:54:02.0000] <Hixie> because you can compile code that uses something without it needing to be supported, and then you can test to see if that code would work, and if it doesn't, you can skip it [00:57:00.0000] <heycam> so for something like WindowTimers, would you want to do that with partial instead? [01:01:00.0000] <Hixie> yeah [01:01:01.0000] <Hixie> actually no [01:01:02.0000] <Hixie> well [01:01:03.0000] <Hixie> gah, i dunno [01:02:00.0000] <Hixie> WindowTimers is also implemented by the WorkerUtils object [01:02:01.0000] <heycam> ah [01:02:02.0000] <Hixie> but i'd want it to appear on the prototype of WorkerUtils (which itself should just appear on the prototype of two other interfaces), and i'd want it to appear as if it was part of Window also [01:02:03.0000] <Hixie> and it's specced in the same spec as Window but not the same spec as WorkerUtils [01:03:00.0000] <heycam> do any of these use ":"? [01:03:01.0000] <heycam> or is it [ImplementedOn] [01:04:00.0000] <Hixie> none of these use : [01:04:01.0000] <heycam> btw i think i like the "Node implements EventTarget;" kind of syntax [01:04:02.0000] <Hixie> oh actually that's a lie [01:04:03.0000] <Hixie> the WorkerUtils thing is implemented on interfaces that do use inheritance [01:05:00.0000] <heycam> [AugmentPrototype] Window implements WindowTimers; [01:05:01.0000] <heycam> or if you're doing something with ":" [01:06:00.0000] <heycam> interface ExtendedWindow : [AugmentPrototype] Window ... [01:06:01.0000] <heycam> but have Java interfaces change depending on which specs you choose to implement seems bad to me [01:07:00.0000] <Hixie> well i don't really mind how you end up doing things for java really [01:07:01.0000] <Hixie> i'm mildly professionally interested in how what you do ends up for C++, because of Native Client's use of WebIDL [01:07:02.0000] <Hixie> but really my interest lies in JS [01:09:00.0000] <heycam> having c++ classes with different members seems as bad as with java [01:09:01.0000] <heycam> mixin abstract classes (just like different java interfaces) seems ok though [01:12:00.0000] <hsivonen> /me raises eyebrows at www-tag [01:15:00.0000] <hsivonen> s/www-tag/www-archive- [01:20:00.0000] <heycam> Hixie, i'll think about partial/[AugmentPrototype] a bit more, gotta head home now [01:51:00.0000] <Hixie> heycam: cool, thanks [04:40:00.0000] <zcorpan_> gsnedders: http://gsnedders.html5.org/outliner/process.py?url=http%3A%2F%2Fsimon.html5.org%2Fsandbox%2Fhtml%2Fgoogle-chrome-comic seems wrong [07:40:00.0000] <annevk42> http://troogle.wordpress.com/2009/06/17/google-is-trying-to-set-us-up-breaking-news-the-new-web-project-revealed-before-its-release/ wtf [07:40:01.0000] <annevk42> especially since http://whois.domaintools.com/thenewwebproject.com points back to them [07:41:00.0000] <annevk42> some kind of elaborate joke that did not cover its tracks? [07:42:00.0000] <Dashiva> "we believe the GOVERNMENT is involved too" [07:42:01.0000] <Philip`> "Google has plans to work together with Microsoft ..." - clearly this whole thing is just a work of fiction [07:43:00.0000] <Dashiva> It looks like a parody of that British guy who walked around with visible secret papers [07:44:00.0000] <annevk42> http://cocaman.ch/wp/2009/06/the-new-web-project-its-a-movie-and-plot-against-google/ [07:46:00.0000] <annevk42> they could have at least used a proxy for the whois [08:55:00.0000] <Philip`> http://asyourworldchanges.wordpress.com/2009/05/31/hear-me-stumble-around-white-house-recovery-and-data-gov-web-sites/ looks potentially informative [13:08:00.0000] <ojan> I know arun's not around, but maybe others here know the answer to my question...why does File API need getAsDataURI and getAsText? [13:09:00.0000] <ojan> as in, can't you replace getAsDataURI with a few lines of code wrapping getAsText? [13:18:00.0000] <annevk2> i guess it's just a convenience thing [13:18:01.0000] <annevk2> not sure it's needed if we get the getAsLocalURL() [13:19:00.0000] <annevk2> ojan, seems best to comment on public-webapps [13:20:00.0000] <ojan> annevk2: will do. just wanted to make sure i wasn't missing something obvious [13:20:01.0000] <ojan> annevk2: yeah, getAsLocalURL does seem very useful to me [13:20:02.0000] <annevk2> oh wait, I guess PNG and such won't get you text but will get you data URLs [13:21:00.0000] <annevk2> /me wonders why he did not think of that [13:21:01.0000] <annevk2> but then getAsText might work anyway, it's not very well defined [13:22:00.0000] <ojan> i was thinking that if you passed base64 as the encoding to getAsText it ought to work for binary content [13:24:00.0000] <annevk2> base64 is not a character encoding, however [13:24:01.0000] <annevk2> but then getAsText is underdefined [14:26:00.0000] <annevk5> oh yay, some guy is advocating software patents on www-style [14:54:00.0000] <sicking> Hixie, ping [15:15:00.0000] <Hixie> sicking: here [15:15:01.0000] <sicking> Hixie, is there a draft for File drag-n-drop? [15:16:00.0000] <sicking> Hixie, we have an intern that's going to start working on it pretty much today [15:16:01.0000] <sicking> Hixie, we can do a lot without a spec for now I think, but it'd be nice to have a draft [15:16:02.0000] <annevk5> there's a proposal on the WHATWG list [15:17:00.0000] <Hixie> i'm waiting for arun's draft to stabalise [15:17:01.0000] <Hixie> but basically i expect datatransfer to get some new api for files [15:18:00.0000] <Hixie> maybe make getData and setData take |any| instead of DOMString [15:18:01.0000] <Hixie> and have them return either DOMStrings for Files [15:18:02.0000] <Hixie> or alternatively, have the files be totally separate [15:18:03.0000] <sicking> Hixie, ok [15:18:04.0000] <Hixie> just a .files or something on DataTransfer [15:19:00.0000] <gsnedders> /me hits more bugs in PHP. ergh. [15:19:01.0000] <Hixie> oh hey, arun is mozilla too, i forgot [15:19:02.0000] <Hixie> go light a fire under his chair :-) [15:19:03.0000] <Hixie> i think we'll go with just .files [15:20:00.0000] <gsnedders> Hixie: But that'll mean he'll have to move away from his computer, thus making him take longer to write the spec [15:20:01.0000] <Hixie> gsnedders: it is implied that he'd be tied down first and only released once the spec was done, thus increasing the motivation to work on it :-) [15:20:02.0000] <Hixie> sicking: yeah so i imagine we'll add a .files that returns a sequence<File> [15:22:00.0000] <sicking> Hixie, ok [15:22:01.0000] <sicking> Hixie, ugh, we'll have to add support for sequence<> to gecko... [15:22:02.0000] <sicking> Hixie, the idea is that that'll map to a JS array, right? [15:24:00.0000] <sicking> annevk5, where is the proposal on the whatwg list? What's the subject line? [15:25:00.0000] <virtuelv> /me thinks "file" should be the atom here, whether that is a file or collection of files [15:26:00.0000] <virtuelv> and treat, at least, zip files as if they were collections of files [15:26:01.0000] <othermaciej> sicking: fwiw we have a patch to implement it for WebKit, and we proposed an API on WHATWG [15:26:02.0000] <virtuelv> that way, an application could easily process openoffice (and similar) documents like epub [15:27:00.0000] <othermaciej> sicking: our version is fairly agnostic with respect to details of Arun's draft [15:27:01.0000] <sicking> othermaciej, cool [15:27:02.0000] <sicking> othermaciej, do you have a link to the draft you proposed? [15:27:03.0000] <gsnedders> Hixie: But then you risk him dying before he finishes the spec, which means you need to find a new editor [15:28:00.0000] <othermaciej> http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2009-June/020302.html [15:28:01.0000] <othermaciej> sicking: ^ [15:30:00.0000] <Hixie> that's basically what i just described :-) [15:30:01.0000] <Hixie> sicking: sequence<T> is just shorthand for an interface with a length property and an item() property or something like that, i think [15:30:02.0000] <Hixie> sicking: basically i just mean your regular DOM list of things [15:32:00.0000] <gsnedders> /me gives up trying to obtain Lightroom, seeming Adobe keep refusing to sell it to me [15:32:01.0000] <gsnedders> Oh well, their lost sale. [15:34:00.0000] <sicking> Hixie, ok [15:44:00.0000] <sicking> othermaciej, btw, have you had a look at the API that Arun proposed for File? [15:45:00.0000] <sicking> othermaciej, I know FileDialog is still controversial, but it'd be nice to hear how you feel about File [15:45:01.0000] <othermaciej> sicking: not yet [15:45:02.0000] <othermaciej> sicking: on my TODO list of spec review [15:46:00.0000] <othermaciej> I just did a careful review of ForceTLS so I'm in a reviewing mood [15:46:01.0000] <sicking> hehe [15:46:02.0000] <sicking> sweet [15:46:03.0000] <othermaciej> sicking: I expect to have issues with it if there's still synchronous file I/O, which I suspect there is [15:47:00.0000] <sicking> othermaciej, nope, asynch only [15:47:01.0000] <othermaciej> then maybe not so much! [15:47:02.0000] <sicking> sweet! [15:47:03.0000] <sicking> i know you guys had concerns about googles Blob proposal, which was async only [15:47:04.0000] <othermaciej> we're obviously interested in making file access better, and it would be nice to have fine DnD and our <input type="file" multi> fuly spec'd [15:48:00.0000] <sicking> not sure what that feedback was though, so not sure if it applies to File [15:48:01.0000] <othermaciej> I dunno [15:48:02.0000] <Hixie> i'll spec both of those within days of arun's proposal being stable [15:48:03.0000] <sicking> othermaciej, Agreed. I think the only controversial part is FileDialog, assuming you're fine with file I/O [15:48:04.0000] <othermaciej> right now we're just using the File interface as a dumb handle [15:56:00.0000] <Hixie> sicking: oh i got a response back from the gmail guys about the filedialog thing -- if you can get arun here that'd be great to discuss it [15:57:00.0000] <sicking> Hixie, you should mail arun and me so we can figure out a day that works [16:00:00.0000] <sicking> Hixie, what about storing File in localStorage? [16:00:01.0000] <sicking> Hixie, I'd imagine that that's needed for offline-gmail [16:07:00.0000] <Hixie> sicking: well i'm free whenever [16:07:01.0000] <annevk5> sicking, would that do a copy or reference the file on disk? [16:07:02.0000] <Hixie> sicking: gmail uses the sql database stuff for their offline gmail [16:08:00.0000] <Hixie> sicking: but yes, they would like a way to store references to Files while waiting for an opportunity to upload them [16:08:01.0000] <Hixie> sicking: i was more talking about FileDialog though [16:08:02.0000] <Hixie> sicking: turns out they didn't use flash to avoid using <input type=file> at all -- in fact they didn't even list that as a reason for using flash until i specifically asked them about it [16:08:03.0000] <Hixie> sicking: and then they dismissed that as not a big deal [16:08:04.0000] <annevk5> if you just wanna store references the local:<id> thing would work I think, you can just store those in localStorage [16:09:00.0000] <sicking> annevk5, how its stored seems like an implementation detail. I'd imagine sqlite has blob support [16:09:01.0000] <annevk5> Hixie, did they use it for async upload and multiple files? [16:10:00.0000] <Hixie> sicking: i wouldn't expect the local:<id> thing to survive page load, but i guess that depends on how we define it [16:10:01.0000] <sicking> Hixie, yeah, my question about localStorage was orthogonal to the meeting [16:10:02.0000] <Hixie> er [16:10:03.0000] <Hixie> s/sicking/anne/ [16:10:04.0000] <annevk5> sicking, it's not an implementation detail, a reference to the file is very different from copying it [16:11:00.0000] <annevk5> e.g. the user could modify the file later on [16:11:01.0000] <Hixie> annevk5: the two things they said were multiselection of photos, and file upload progress. [16:11:02.0000] <sicking> annevk5, ah, yes, we should define that it's the contents of the file at the time of the call to store it in localStore [16:11:03.0000] <annevk5> Hixie, yay for me guessing why they used Flash [16:12:00.0000] <annevk5> and boo for sicking not believing me :p [16:12:01.0000] <Hixie> annevk5: and then they added dnd of files, local display of files while uploading, and selecting or uploading a whole folder at once [16:12:02.0000] <sicking> annevk5, i prefer data over guessing [16:12:03.0000] <Hixie> and finally they mentioned the <input accept=""> feature (though not by name) [16:12:04.0000] <sicking> annevk5, except when I'm the one guessing [16:12:05.0000] <sicking> annevk5, :) [16:12:06.0000] <annevk5> sicking, so it appears :) [16:12:07.0000] <Hixie> oh and resuming uploads after browser crash [16:13:00.0000] <Hixie> though that one was more theoretical [16:13:01.0000] <sicking> Hixie, flash supports that? [16:13:02.0000] <annevk5> i always wonder whether selecting a folder should be a UI feature or some semantic thing like <input type=folder> [16:13:03.0000] <Hixie> i think that was after they transitioned from why they used flash to what they wanted most of all [16:14:00.0000] <sicking> Hixie, localStorage being able to store files would take care of that [16:14:01.0000] <Hixie> anyway their answer to "so, uh, styling of <input type=file> isn't an issue?" was basically "oh i guess that would be nice too" but they basically said it wasn't a big deal compared to the rest [16:14:02.0000] <Hixie> so i vote for dropping FileDialog for now [16:14:03.0000] <sicking> we'll see [16:14:04.0000] <Hixie> sicking: or the sql database, yeah [16:15:00.0000] <sicking> Hixie, yeah, though i'm not sure what our timeline is for sql database. I think we still want to see some other API [16:15:01.0000] <sicking> Hixie, have had some discussions about that internally recently [16:15:02.0000] <Hixie> yeah, we were talking about it in #developers recently too [16:15:03.0000] <sicking> ok [16:16:00.0000] <Hixie> basically i'm gonna do nothing on the sql front until october and then i'm gonna suggest moving to LC [16:16:01.0000] <sicking> mostly due to the pains of standardizing a SQL dialect, and then having to implement a sql parser that maps that dialect to sqlite [16:16:02.0000] <Hixie> and after that look into speccing teh actual dialect [16:16:03.0000] <Hixie> so you have until then to come up with a better solution :-) [16:16:04.0000] <sicking> i don't see how we could move into LC without a dialect [16:16:05.0000] <Hixie> which i'd likely support since i have no desire to actual spec out the dialect :-) [16:17:00.0000] <sicking> heh [16:17:01.0000] <sicking> hence i vote for File in localStorage for now :) [16:17:02.0000] <Hixie> well the dialect would be its own spec i think [16:17:03.0000] <sicking> and across postMessage [16:17:04.0000] <Hixie> i don't know what file in localStorage would really be [16:17:05.0000] <Hixie> postMessage should definitely support it [16:18:00.0000] <sicking> Hixie, what do you mean, what it would be? [16:18:01.0000] <gsnedders> Oh dear… [16:18:02.0000] <Hixie> sicking: well localStorage up to now has only been unstructured name-value string pairs [16:18:03.0000] <sicking> Hixie, can't we just make the value <string-or-file>? [16:18:04.0000] <gsnedders> If you follow the rules for comparing charsets in UTS #22, you end up with iso-ir-9-1 and iso-ir-91 being identical, but they shouldn't be. [16:19:00.0000] <Hixie> sicking: i don't think gmail wants an unstructured answer here; and extending DOMString to DOMString-or-File seems weird when you don't allow anything else [16:19:01.0000] <sicking> nono, not touch DOMString [16:19:02.0000] <Hixie> gsnedders: do browsers support those? [16:19:03.0000] <sicking> that would be, um, crazy [16:19:04.0000] <gsnedders> Hixie: No idea. [16:19:05.0000] <Hixie> sicking: that's what i'm saying :-P [16:19:06.0000] <Hixie> sicking: anyway, send mail [16:19:07.0000] <Hixie> sicking: i'm swamped right now [16:19:08.0000] <sicking> ok [16:20:00.0000] <Hixie> sicking: and waiting for arun to finish his draft before i think of this [16:20:01.0000] <gsnedders> Hixie: On an HTML 5 note, it'd be nice if the charset override table used the same case as the IANA registry (i.e., lowercase "w" for "windows-") [16:20:02.0000] <sicking> ok, i'll try to spend some time with him touching up the current draft [16:21:00.0000] <sicking> it needs a lot of editing, but I'm not sure that the actual APIs will be much different from what we have now [16:21:01.0000] <sicking> apart from Annes bikeshedding ;) [16:21:02.0000] <Hixie> gsnedders: send mail [16:21:03.0000] <sicking> and possibly FileDialog since you're all a bunch of pansies ;) [16:21:04.0000] <gsnedders> /me was expecting that answer :) [16:21:05.0000] <othermaciej> Hixie: the reasons cited for using Flash to upload sound just like what the MobileMe people told us [16:22:00.0000] <othermaciej> which is why we added multi-file selection and file upload progress (via XHR) to Safari [16:22:01.0000] <annevk5> sicking, arun already accepted all my comments :) [16:22:02.0000] <annevk5> sicking, well, apart from FileDialog, but it seems I'm not alone [16:23:00.0000] <sicking> annevk5, see above re 'pansy' ;) [16:24:00.0000] <annevk5> sicking, says the man who doesn't wanna do redirects :p [16:25:00.0000] <sicking> annevk5, redirects are the root of all evil [16:25:01.0000] <sicking> annevk5, together with money [16:25:02.0000] <sicking> annevk5, and premature optimizations [16:25:03.0000] <Hixie> scripting modal dialogs are the root of more evil than redirects :-P [16:26:00.0000] <sicking> there's going to be a dialog no matter what (in the non-dnd case) [16:26:01.0000] <Hixie> a user-invoked one is not a big deal [16:26:02.0000] <sicking> The difference between having it triggered by an arbitrary click, and on a JS call, is very small [16:26:03.0000] <sicking> Hixie, UAs are free to warn instead of putting up a dialog if the JS-call happens outside a click [16:27:00.0000] <sicking> Hixie, a'la popup blockers [16:28:00.0000] <Hixie> sicking: any feature that is admitted to need popup-blocker-like treatment _at the design stage_ is imho a non-starter [16:35:00.0000] <sicking> Hixie, i don't think it needs a popup blocker [16:35:01.0000] <sicking> Hixie, IE has had this API for ages. Granted, that's not the best indicator ever [16:36:00.0000] <sicking> Hixie, also, geoid is doing the same thing [16:39:00.0000] <Hixie> IE doing something is not a good indicator that it is secure. [16:39:01.0000] <sicking> i know, but i've never heard or people being annoyed by it [16:39:02.0000] <Hixie> and the geoloc stuff isn't a modal dialog, it pops up an asynchronous UI that can then later cause the script to get a callback [16:39:03.0000] <sicking> or attacked by it [16:39:04.0000] <Hixie> anyway [16:39:05.0000] <sicking> FileDialog is async too [16:39:06.0000] <Hixie> we don't have a need for this, and people aren't really happy about it [16:40:00.0000] <Hixie> so why do it [16:54:00.0000] <Hixie> hsivonen, others: i'm writing a section for the parser spec talking about some error handling cases and basically explaining how they end up as they do [16:54:01.0000] <Hixie> if you have any specific error cases you would like me to discuss in particular, let mek now 2009-06-19 [19:26:00.0000] <jruderman> Hixie: is http://dev.w3.org/html5/spec/Overview.html#dom-window-close intended to be vague on what browsers are allowed to do in non-auxilary cases? [19:26:01.0000] <jruderman> (e.g. show a dialog, or employ heuristics mentioned in https://bugzilla.mozilla.org/show_bug.cgi?id=190515#c70) [20:38:00.0000] <MikeSmith> sayrer: you around? [20:38:01.0000] <sayrer> MikeSmith: hi [20:38:02.0000] <MikeSmith> hey man [20:39:00.0000] <MikeSmith> about your message http://www.w3.org/mid/4A3AF8B9.8040700⊙mc [20:39:01.0000] <sayrer> ah, yes [20:39:02.0000] <MikeSmith> you are encouraged to open bugs in the W3C bugzilla [20:39:03.0000] <sayrer> but not reopen them? [20:39:04.0000] <MikeSmith> there is intentionally no process for it [20:39:05.0000] <MikeSmith> no, you can reopen them too [20:40:00.0000] <sayrer> oh, I wonder if I just missed the button [20:40:01.0000] <MikeSmith> it is intended to be a relative free-for-all in that regard [20:40:02.0000] <MikeSmith> intended by me, I mean, when I set it up for the group [20:40:03.0000] <sayrer> mmm, no, I can't reopen things [20:41:00.0000] <sayrer> http://www.w3.org/Bugs/Public/show_bug.cgi?id=7034 [20:41:01.0000] <sayrer> no buttons to do so [20:41:02.0000] <MikeSmith> hmm, OK [20:41:03.0000] <MikeSmith> hang on for a couple minutes, I will take a look at it [20:41:04.0000] <sayrer> I can file things [20:41:05.0000] <MikeSmith> I am pretty sure I can change your individual perms [20:42:00.0000] <sayrer> hope this HTML WG is the right HTML WG [20:42:01.0000] <sayrer> :) [20:42:02.0000] <MikeSmith> heh [20:42:03.0000] <MikeSmith> I think we are just using the default bugzilla perms, but anyway, lemme take a look [20:43:00.0000] <MikeSmith> sayrer: so, to be clear, when you are logged in an go to http://www.w3.org/Bugs/Public/show_bug.cgi?id=7034 [20:44:00.0000] <MikeSmith> there is no "Reopen bug" radio button [20:44:01.0000] <MikeSmith> under the Additional Comments textarea? [20:44:02.0000] <sayrer> where should I look? [20:44:03.0000] <sayrer> nope [20:44:04.0000] <sayrer> not there [20:45:00.0000] <MikeSmith> OK [20:48:00.0000] <MikeSmith> sayrer: please try again now [20:48:01.0000] <sayrer> tada [20:50:00.0000] <MikeSmith> I just gave you bugzilla "editbugs" perms. which I guess in bugzilla terms adds you to the editbugs "group" [20:50:01.0000] <MikeSmith> = "Can edit all aspects of any bug." [20:50:02.0000] <MikeSmith> now, what I would like to do is have that be the default for all users of our bugzilla [20:52:00.0000] <MikeSmith> but unfortunately, I reckon that trying to make it so would involve a relatively time-consuming discussion with the powers that be [20:52:01.0000] <MikeSmith> so I guess for now I will just deal with it case by case [20:52:02.0000] <MikeSmith> I wish bugzilla allowed those perms to be set per-Product [20:54:00.0000] <MikeSmith> sayrer: anyway, if you run into w3c bugzilla problems or have questions about our specific instance of it, please feel free to ping me directly and I'll check asap [20:54:01.0000] <MikeSmith> rather than on public-html, I mean [20:55:00.0000] <sayrer> oh, I wouldn't mail /only/ about bugzilla problems [20:55:01.0000] <sayrer> issue seemed worth raising on the list anyway [20:55:02.0000] <MikeSmith> yeah, the issue does [20:56:00.0000] <MikeSmith> I just mean not administrivia-like stuff on public-html [20:56:01.0000] <sayrer> ok [20:56:02.0000] <MikeSmith> just in the interest of trying to keep traffic volume down on the list [20:57:00.0000] <MikeSmith> so, about your actual statement in that message, regarding the topic of "content model" [20:57:01.0000] <sayrer> yeah [20:57:02.0000] <MikeSmith> "I can see why this might be tempting to specify for theoretical reasons, but I can't think of a use case." [20:58:00.0000] <sayrer> http://www.whatwg.org/specs/web-apps/current-work/images/content-venn.svg [20:58:01.0000] <sayrer> why does that exist? [20:58:02.0000] <MikeSmith> oh, that [20:58:03.0000] <sayrer> there's also this business of forming outlines [20:58:04.0000] <MikeSmith> it's not clear from you message that that's what you meant [20:59:00.0000] <sayrer> that is the first illustration under "3.4 Content models" [20:59:01.0000] <sayrer> "There are also certain elements that are sectioning roots. These are distinct from sectioning content, but they can also have an outline." [20:59:02.0000] <MikeSmith> yeah, I thought you had brought that up before -- the question of the appropriateness of the outline algorithm, I mean [20:59:03.0000] <sayrer> like... what is this... [21:00:00.0000] <sayrer> well, I'm sure whoever answered it thought they did a good job addressing the issue [21:00:01.0000] <sayrer> I should pay more attention [21:01:00.0000] <MikeSmith> sayrer: I think the term "sectioning roots" is Hixie coining a language primitives that he needs to reference in multiple definitions elsewhere [21:02:00.0000] <MikeSmith> I think "sectioning content" is fairly unambiguous [21:02:01.0000] <MikeSmith> or at least a familiar concept [21:03:00.0000] <MikeSmith> sayrer: anyway, I would personally advise/prefer that you raise those kinds of questions individual bugzilla issues -- if/when you have time [21:04:00.0000] <sayrer> hmm yeah [21:05:00.0000] <MikeSmith> in some cases it might be useful to post them directly to public-html too, but I have the bugzilla set up so that an automated notification gets sent to the list each time any bug is first opened or resolved [21:06:00.0000] <MikeSmith> so that way, people who are actually interested enough to want to take time to follow the discussion further can take time to cc themselves to the bugzilla issue [21:07:00.0000] <sayrer> gtg [21:37:00.0000] <zcorpan_> "Font family names containing more than one word should be quoted." - i've never understood why it should be quoted [21:37:01.0000] <zcorpan_> http://www.w3.org/TR/2009/WD-css3-fonts-20090618/ [21:57:00.0000] <othermaciej> MikeSmith: I have sympathy with sayrer's position (I am not sure authoring conformance requirements are actually useful) but I think his bug report is ridiculous and insulting [21:57:01.0000] <othermaciej> MikeSmith: and should be closed [21:57:02.0000] <othermaciej> MikeSmith: snark and sarcasm are not appropriate ways to express differences of opinion in a bug tracker [21:58:00.0000] <MikeSmith> othermaciej: well, I'm the one who raised it initially, with that summary (though in jest) [21:58:01.0000] <MikeSmith> I have learned my lesson [21:59:00.0000] <othermaciej> MikeSmith: in that case - I think your bug report is not an appropriate way to raise an issue [21:59:01.0000] <zcorpan_> gsnedders: http://krijnhoetmer.nl/irc-logs/whatwg/20090618#l-356 [21:59:02.0000] <othermaciej> MikeSmith: I think it should be closed and you and/or Rob should raise sincere issues [21:59:03.0000] <othermaciej> such as "HTML5 should have no authoring conformance requirements" [22:00:00.0000] <othermaciej> or suggesting specific changes to the authoring conformance requirements [22:00:01.0000] <othermaciej> it's not clear to me which the bug is really asking for - I think it's obvious that literally replacing the text "conformance checker" with "loyalty checker" everywhere is not something anyone sincerely wants [22:00:02.0000] <zcorpan_> i think the requirements are very useful for authors [22:00:03.0000] <MikeSmith> othermaciej: yeah, of course [22:00:04.0000] <MikeSmith> othermaciej: but I will let Rob do what he wants we the issue, and let anybody else who uses bugzilla do anything they want with it [22:01:00.0000] <othermaciej> since I haven't used the bug tracker much, I don't feel it's my place to close it [22:01:01.0000] <MikeSmith> I am not going to let that bugzilla become yet another place where we get mired in discussions about what's appropriate and what the process should be [22:01:02.0000] <zcorpan_> isn't the bug closed already? [22:01:03.0000] <othermaciej> but if someone filed a bug like that in WebKit bugzilla, it would be INVALID city, and if people tried to reopen it, they would get a talking to about responsible use of bugzilla [22:02:00.0000] <MikeSmith> zcorpan_: I closed it, but Rob said he wanted to reopen it, and I said, go ahead [22:02:01.0000] <MikeSmith> othermaciej: our bugzilla is relatively lightly used now -- nothing on the scale of Webkit bugzilla -- and we do not have a general problem [22:03:00.0000] <MikeSmith> if this develops in a bigger problem, then I will do something about it [22:03:01.0000] <othermaciej> sure, but filing sarcastic bug reports and reopening them are still abusive things to do [22:03:02.0000] <othermaciej> you in particular should set a better example [22:03:03.0000] <MikeSmith> othermaciej: add it to the list of many other things in the group that are considered abusive by some members of the group [22:04:00.0000] <MikeSmith> othermaciej: christ [22:04:01.0000] <MikeSmith> I hope you mean that "you in particular should set a better example" sarcastically [22:04:02.0000] <MikeSmith> I can't tell [22:04:03.0000] <othermaciej> fine, I don't want to get mired in a postmodern discussion of whether courtesy is an illusion [22:05:00.0000] <MikeSmith> good [22:05:01.0000] <MikeSmith> I don't want to either [22:06:00.0000] <MikeSmith> but please, seriously, don't schoolmarm me with "you in particular should set a better example" nonsense [22:06:01.0000] <MikeSmith> I get plenty enough of that already [22:06:02.0000] <MikeSmith> believe me [22:09:00.0000] <zcorpan_> i wonder why jd doesn't put quotes around ECMAScript 4 [22:11:00.0000] <MikeSmith> zcorpan_: because he's not afraid of ECMAScript 4 [22:11:01.0000] <MikeSmith> he only uses scare quotes to indicate stuff that he's afraid of [22:13:00.0000] <othermaciej> they were scare quotes of disdain [22:14:00.0000] <othermaciej> ECMAScript 4 was based on ActionScript so it would be hard for him to consistently disdain it [22:14:01.0000] <othermaciej> (ES4 also had a lot more hype than reality, somewhat undermining his point...) [22:19:00.0000] <othermaciej> the comments here are entertaining: http://news.ycombinator.com/item?id=662620 [22:50:00.0000] <sayrer> entertaining, or insulting? [22:50:01.0000] <othermaciej> both! [22:50:02.0000] <sayrer> regarding the bug report: I did not file it, but I do think it raises a real issue [22:51:00.0000] <sayrer> and the title is only insulting to the extent that it is poignant [22:51:01.0000] <othermaciej> sayrer: I think a sincere bug that suggests removing conformance requirements would be useful [22:51:02.0000] <sayrer> there is no swearing or personal attacks [22:51:03.0000] <othermaciej> or a bug that suggested specific changes to removing conformance requirements [22:52:00.0000] <sayrer> so I think politeness police kind of stuff is bikeshedding [22:52:01.0000] <othermaciej> I think sarcastically requesting a change that is not what is actually desired, without making it clear what is desired, is a bad use of the bug tracker [22:52:02.0000] <othermaciej> filing the bug in the first place could be seen as comedy, reopening it is abusive [22:52:03.0000] <sayrer> yes, I think your point, while possibly valid, is unimportant [22:52:04.0000] <sayrer> reopening the bug is not abusive [22:53:00.0000] <MikeSmith> sayrer: I agree with othermaciej that we would be better off if you changed the title of the bug [22:53:01.0000] <sayrer> I went and asked for permision to reopen it from the person who filed it [22:53:02.0000] <othermaciej> chain reopening sarcastic bugs is a waste of people's time, and not likely to lead to any productive outcome [22:53:03.0000] <MikeSmith> sayrer: I mean better off in that if you actually want people to take the issue seriously [22:53:04.0000] <sayrer> Larry and Maciej seem to want to change the title... I don't care what it is [22:54:00.0000] <MikeSmith> sayrer: if you don't care what it is, I will go ahead and change it [22:54:01.0000] <sayrer> go for it [22:54:02.0000] <sayrer> of course, now it's been defaced by maciej lecturing :) [22:56:00.0000] <othermaciej> I think my point was valid, and if it was unimportant, then surely debating its unimportance is an even greater waste of time [23:02:00.0000] <MikeSmith> I have re-learnedd the lesson that attempts at humor are most successful among a group of people who are reasonably familiar with each, but less successful in a group a people who have a tendency to be distrustful of each others and who are prone to suspect (or even openly accuse) each other of having ulterior motives [23:02:01.0000] <MikeSmith> (reasonably familiar with each other) [23:21:00.0000] <othermaciej> MikeSmith: I get that you were trying to be funny, and I don't want to say humor in a bug tracker is always wrong, it's just that mixing it up with a substantive issue of controversy and combined with reopening it starts to be less funny [23:22:00.0000] <MikeSmith> yep, point taken [23:22:01.0000] <othermaciej> I regret being heavy-handed in my complaining about it [23:23:00.0000] <MikeSmith> othermaciej: np. I regret having filed it as bug to begin with [23:41:00.0000] <zcorpan_> /me wonders why http://www.w3.org/TR/2009/WD-css3-fonts-20090618/ refers to "XHTML headings" rather than "HTML headings" [23:56:00.0000] <Hixie> jruderman: send mail, sorry can't reply right now [23:57:00.0000] <zcorpan_> i wonder if 'unicode-range' is a good feature to have [23:57:01.0000] <zcorpan_> does anyone implement it? [23:59:00.0000] <zcorpan_> if an author wants to optimize bandwidth, it seems Philip`'s approach is more effective [00:00:00.0000] <zcorpan_> i mean, in the example, a 4.5 MB font is downloaded just to render an arrow [00:21:00.0000] <zcorpan_> "Matching occurs in a well-defined order to insure ..." [00:39:00.0000] <heycam> Hixie, are you happy with the statement in html5 about document.all booleanifying to false remaining there? or do you think it needs to web idl? [00:39:01.0000] <heycam> i suppose whatever spec defines style.filter needs it [00:40:00.0000] <heycam> then again, it's just one sentence of prose, so it's not that much of a burden to really need an [Undetectable] in web idl [00:48:00.0000] <zcorpan_> /me updates https://developer.mozilla.org/En/Using_audio_and_video_in_Firefox to use <object> instead of <applet> [00:49:00.0000] <zcorpan_> heycam: typeof document.all should return 'undefined', iirc [00:49:01.0000] <heycam> zcorpan_, ah. at least that's within the realms of conformance to the ecmascript spec. [00:50:00.0000] <heycam> i suppose Hixie should add that then (as long as [Undetectable] isn't in web idl) [00:50:01.0000] <Hixie> heycam: i've no idea. i couldn't get browser vendors to tell me anything useful. [00:52:00.0000] <zcorpan_> http://software.hixie.ch/utilities/js/live-dom-viewer/saved/146 [00:54:00.0000] <zcorpan_> http://software.hixie.ch/utilities/js/live-dom-viewer/saved/147 - firefox false, webkit and opera true [00:54:01.0000] <zcorpan_> do instanceof doesn't seem to need anything special [00:54:02.0000] <zcorpan_> but typeof does [00:56:00.0000] <zcorpan_> firefox stringifies to 'undefined' too, but opera and webkit don't [01:01:00.0000] <zcorpan_> it seems firefox tries to make it as undetectable as possible, while opera and webkit try to do as little as necessary to be compatible with web content [01:22:00.0000] <othermaciej> I can tell you guys what we do to make document.all undetectable [01:23:00.0000] <othermaciej> it's more than just the typeof thing [01:23:01.0000] <othermaciej> does HTML5 intend to spec document.all? [01:23:02.0000] <othermaciej> (I couldn't find it in the IDL but maybe it's spec'd elsewhere?) [01:23:03.0000] <Hixie> it does spec it already [01:23:04.0000] <zcorpan_> othermaciej: you can reply to my email on public-html :) [01:24:00.0000] <Hixie> zcorpan_: i did the reverse engineering; the question was what was actually needed and what were browser vendors willing to change to doing [01:24:01.0000] <othermaciej> the HTMLDocument interface definition doesn't seem to list an "all" property [01:25:00.0000] <othermaciej> where is it spec'd? [01:25:01.0000] <zcorpan_> http://www.whatwg.org/specs/web-apps/current-work/multipage/the-xhtml-syntax.html#dom-document-all [01:26:00.0000] <zcorpan_> Hixie: someone at opera said typeof was required for web compat [01:26:01.0000] <othermaciej> I see [01:27:00.0000] <othermaciej> splitting the interfaces is a bit unwieldy [01:27:01.0000] <othermaciej> though I can understand it as a way to hide the bad stuff [01:28:00.0000] <zcorpan_> Hixie: maybe you should include the obsolete stuff together with the other things but hide it behind class=impl (and have // obsolete in the IDL)? [01:28:01.0000] <zcorpan_> although i guess having it in a separate section is better for authors and implementors will see the obsolete section anyway [01:28:02.0000] <Philip`> zcorpan_: It'd be great if someone made a font format that could just download the glyphs that were needed on demand, so you wouldn't have to download the entire font or subset it in advance [01:29:00.0000] <zcorpan_> Philip`: indeed [01:29:01.0000] <Philip`> but it seems everyone wants to invent new formats to add DRM capabilities, not to add any actually useful features [01:29:02.0000] <heycam> Hixie, did you in the end need something from Web IDL for the split window "this" issue? [01:29:03.0000] <zcorpan_> Philip`: still, 'unicode-range' is not it [01:29:04.0000] <heycam> i.e. where "this" means the WindowProxy instead of the Window? [01:29:05.0000] <Hixie> heycam: no [01:29:06.0000] <heycam> ok cool [01:29:07.0000] <heycam> /me is going through some old notes [01:29:08.0000] <Hixie> heycam: search HTML5 for WindowProxy if you want to see how i ended up doing it [01:32:00.0000] <zcorpan_> /me should get better at quoting the spec in his spec feedback emails [01:34:00.0000] <othermaciej> I sent what I think is an exhaustive list of what we do [01:34:01.0000] <othermaciej> zcorpan_: I'm an implementor and I didn't see it til it was pointed out [01:34:02.0000] <othermaciej> though I guess I will know to look for them now [01:34:03.0000] <zcorpan_> Hixie: maybe you should have <span class=impl>// further members are specified <span>xref</span></span> in the main IDL [01:35:00.0000] <Hixie> zcorpan_: that's a good idea; file a bug for me? [01:35:01.0000] <othermaciej> basically what we do is make sure document.all compares as false in boolean comparisons, has a type of "undefined", and compares equal to undefined, null and false [01:35:02.0000] <othermaciej> (bot not strict equal) [01:35:03.0000] <Hixie> othermaciej: i saw all kinds of weird per-op-code behaviour when i was looking at jsc [01:36:00.0000] <othermaciej> Hixie: I believe all of that behavior implements what I described above and nothing more [01:36:01.0000] <othermaciej> I may have missed a spot [01:36:02.0000] <othermaciej> but I don't think so [01:36:03.0000] <Hixie> cool [01:37:00.0000] <zcorpan_> opera doesn't do the equals thing it seems [01:37:01.0000] <othermaciej> I believe we apply the exact same set of hacks to style.filter [01:37:02.0000] <heycam> Hixie, at one point you asked about an open issues list for Web IDL. That list should just be the red notes in the spec, now. [01:38:00.0000] <Hixie> heycam: cool [01:38:01.0000] <Hixie> heycam: good to know [01:38:02.0000] <heycam> so if there is something you requested that hasn't been handled but doesn't have a red note, let me know [01:38:03.0000] <Hixie> roger [01:38:04.0000] <othermaciej> (btw I should note that, at least for us, undetectability belongs to the object implementing document.all, not to the property) [01:38:05.0000] <othermaciej> (so if you do var failAPI = document.all, then failAPI will have all the same forms of undetectability) [01:40:00.0000] <zcorpan_> that's the same in opera i think [01:40:01.0000] <othermaciej> I think Mozilla does something elaborate and weird that may be specific to the context of an if statement [01:41:00.0000] <Hixie> not sure how to define that things are == null and == undefined [01:41:01.0000] <Hixie> but i'm sure i'll figure out a way [01:42:00.0000] <othermaciej> what's hard about it? [01:42:01.0000] <zcorpan_> Hixie: opera doesn't do that, so it might not be needed for compat [01:42:02.0000] <othermaciej> can't you just literally list some of the ECMAScript expressions that MUST evaluate to certain values, despite what the ES spec says for normal objects? [01:43:00.0000] <Hixie> i don't know [01:43:01.0000] <Hixie> i'm not really sure how to define the expression [01:43:02.0000] <othermaciej> I can't remember if we did it for consistency or because we thought some sites really check that way [01:43:03.0000] <Hixie> is |document.all == foo| true if |foo === null|? [01:44:00.0000] <othermaciej> if (document.all != undefined) seems like a plausible check [01:44:01.0000] <othermaciej> that's the case in WebKit [01:44:02.0000] <Hixie> what about things like isNaN, and |'all' in document|, and so on? [01:44:03.0000] <othermaciej> it does it based on the value being null, the expression doesn't have to contain the null literall [01:45:00.0000] <Hixie> so is it a property of the == operator or of the document.all object? [01:45:01.0000] <Hixie> or both? [01:45:02.0000] <Hixie> or something else? [01:45:03.0000] <Hixie> anyway like i said i'm sure i'll find a way to express it [01:45:04.0000] <othermaciej> alert('all' in document) reports true in Safari [01:45:05.0000] <Hixie> i just haven't look at it in detail yet [01:45:06.0000] <othermaciej> it's a property of the combination of the object and the == operator [01:46:00.0000] <Hixie> so if there is some other language mechanism to do comparisons, then it isn't affected by this magic? [01:46:01.0000] <Hixie> and if, say, JS added operator overloading, an overloaded '==' wouldn't see magic? [01:46:02.0000] <othermaciej> there is the === operator [01:46:03.0000] <othermaciej> which is not affected [01:47:00.0000] <Hixie> basically my point is just that the wording has to be very carefully thought out; i don't think it'll be especially hard or anything [01:47:01.0000] <othermaciej> I can't predict how it would interact with hypothetical future features [01:47:02.0000] <zcorpan_> if (document.all === undefined) seems like an equally plausible check as if (document.all == undefined), although i guess would be a lot less common [01:47:03.0000] <Hixie> ideally if we define it clearly (i.e. not by example), it will be well-defined [01:49:00.0000] <othermaciej> hmmm [01:49:01.0000] <othermaciej> I can't find signs of document.all existing at all in Firefox [01:49:02.0000] <zcorpan_> othermaciej: try quirks mode [01:49:03.0000] <othermaciej> does the page need to have some special property for document.all to appear even in hidden form? [01:49:04.0000] <Hixie> has to be in quirks mode in firefox [01:49:05.0000] <othermaciej> oh hey, google.com has a docype now [01:50:00.0000] <Hixie> i'm trying :-) [01:50:01.0000] <othermaciej> there goes my favorite quirks mode test page! [01:50:02.0000] <Hixie> my influence is slow, but non-zero :-) [01:50:03.0000] <Hixie> jens meiert has been helping a lot too [01:50:04.0000] <Hixie> he even set up an internal instance of henri's validator [01:50:05.0000] <zcorpan_> doesn't validate though :( [01:51:00.0000] <Hixie> not yet :-) [01:51:01.0000] <Hixie> google.com is an impressively deceptive page in that while it looks simple, it is insanely complicated [01:51:02.0000] <othermaciej> Firefox makes document.all == to undefined and null [01:51:03.0000] <othermaciej> which is probably why we did that [01:51:04.0000] <Hixie> so many domains, so many languages, so many locales, so many things to expose differently in different countries... [01:52:00.0000] <zcorpan_> /me is waiting for Hixie to respond to emails from 2005 where people ask Hixie to make google.com front page validate [01:52:01.0000] <Hixie> i doubt i have those around still [01:52:02.0000] <Hixie> oh did i mention that the web elements guys fixed the problems ou raised? [01:52:03.0000] <Hixie> or were going to, anyway? [01:53:00.0000] <Hixie> dunno if they actually did [01:53:01.0000] <Hixie> let me know in a few months if you still find them broken [01:53:02.0000] <zcorpan_> cool [01:53:03.0000] <Hixie> they apparently were already intending to do it [01:53:04.0000] <Hixie> and had introduced a bunch of errors at the last minute trying to fix things up in other browsers [02:31:00.0000] <Philip`> Woah, Chrome can run Canvex almost full-screen with almost acceptable performance :-o [02:31:01.0000] <Philip`> (Maybe other modern browsers can too, but I've not tried them) [02:31:02.0000] <Philip`> I need to work on it again to make it slower [02:51:00.0000] <gsnedders> zcorpan_: Send email. I probably won't get around to doing anything about it till in Sweden anyway though. :) [02:52:00.0000] <gsnedders> zcorpan_: What's wrong about the output? [03:24:00.0000] <Lachy> JohnResig, yt? [03:25:00.0000] <Lachy> JohnResig, I wanted to discuss getting the selectors api test suite completed, as mentioned in my last mail to public-webapps [03:33:00.0000] <zcorpan_> gsnedders: the h2s should be sub sections of the h1 [03:37:00.0000] <gsnedders> /me remembers the header/hgroup change. that's the bug :P [03:45:00.0000] <Philip`> http://www.guardian.co.uk/politics/gallery/2009/jun/18/mps-expenses-houseofcommons?picture=349036998 - noooo! Comic Sans :-( [05:03:00.0000] <roc> Philip`: download glyphs on demand? that's crazy talk [05:03:01.0000] <roc> hmm [05:03:02.0000] <roc> why is jd proud of the fact that Flash only checks for updates once a month? [05:05:00.0000] <annevk5> /me thinks jd is best ignored at this point [05:19:00.0000] <Lachy> who's jd? [05:27:00.0000] <roc> http://blogs.adobe.com/jd/ [05:34:00.0000] <JohnResig> Lachy: "IE lacks support for many of the CSS3 selectors tested in the test suite, but to be fair, these failure should be ignored. [05:35:00.0000] <JohnResig> that's bullshit :-( [05:35:01.0000] <JohnResig> do we really have to remove those tests? [05:35:02.0000] <JohnResig> ok, hang on, I should probably keep reading [05:35:03.0000] <gsnedders> JohnResig: Is CSS 3 selectors really need to be a prerequisite for the selectors test suite, though? [05:36:00.0000] <JohnResig> gsnedders: for a fully interoperable implementation, yep! [05:36:01.0000] <gsnedders> JohnResig: How so? What in selectors requires CSS 3 selectors to be tested? [05:38:00.0000] <JohnResig> gsnedders: I'm saying that if we care about having interoperability - and making sure that all of the possible selectors work correctly - we need to test the full gamut of CSS selectors. A number of bugs were caught in the implementations of WebKit, Firefox and Opera since these tests existed - stripping them out will mean 1) There will be regressions 2) That IE will implement more bugs in their implementation and 3) That the implementations will be more li [05:39:00.0000] <gsnedders> JohnResig: Is that not an issue for the CSS 3 selectors test suite, though? [05:39:01.0000] <JohnResig> gsnedders: no - it's simply assumed that ass CSS selectors work identically as they do in the browser - it's very likely that they don't [05:39:02.0000] <annevk5> gsnedders, that test suite does not test them in the context of the Selectors API [05:40:00.0000] <JohnResig> there were a number of selectors that failed in the suite but "passed" in the official CSS selector test suite [05:42:00.0000] <Lachy> JohnResig, no, I didn't say we should remove them. [05:42:01.0000] <JohnResig> Lachy: heh, ok, I'm still reading [05:42:02.0000] <JohnResig> Lachy: just got a little flustered ;) [05:43:00.0000] <Lachy> Just that counting them when working out IE's level of conformance isn't really fair, because not supporting CSS3 doesn't affect their conformance with the requirements in Selectors API [05:43:01.0000] <gsnedders> /me would nevertheless argue that unless the Selectors API requires support for CSS 3 Selectors (normatively) its test suite should just check against CSS 2.1, and optionally CSS 3 Selectors [05:44:00.0000] <JohnResig> Lachy: I see, yeah - and the fact that there's been 3 nearly-conformant implementations (as you pointed out) kind of removes the need for IE to be fully conformant [05:44:01.0000] <JohnResig> regardless [05:45:00.0000] <JohnResig> Lachy: I'm reading over the list on the bottom now [05:45:01.0000] <Philip`> roc: Why so crazy? It's just like downloading seekable media resources, except with lots more seeking, so you could add some kind of batching, and then it'd be trivial! :-) [05:45:02.0000] <JohnResig> Lachy: the test suite was actually, originally, xhtml - and included SVG inside of it [05:45:03.0000] <JohnResig> Lachy: I removed it because it broke in IE and we wanted the suite to run in IE :-/ [05:46:00.0000] <Lachy> as far as selectors api conformance is concerned, for interoperability, we only need to ensure that the selectors that the implementations do support work correctly in the API. Requiring that they also work with Selectors that the implementation don't even claim support is silly [05:47:00.0000] <Lachy> ok. In that case, it's probably easiest to just make a new version that tests XHTML/SVG specific stuff [05:48:00.0000] <JohnResig> the first batch of hixie tests are pretty painless [05:50:00.0000] <JohnResig> woah, Hixie's 002-001 test is pretty nuts [05:52:00.0000] <Lachy> do you mean checking for "[object NodeList]" [05:52:01.0000] <Lachy> ? [05:52:02.0000] <JohnResig> Lachy: yeah - checking that the resulting NodeList is == to "[object NodeList]" [05:52:03.0000] <Lachy> I'm fine wil leaving the whole of Hixie's 002 tests out [05:53:00.0000] <JohnResig> Lachy: we already test for 002-003 [05:53:01.0000] <JohnResig> and isn't 002-002 part of a different spec? [05:53:02.0000] <JohnResig> ok, anyway, I'll merge in 001 [05:53:03.0000] <Lachy> well, the IDL says NoInterfaceObject [05:54:00.0000] <Lachy> so technically, 002-002 is testing a selectors api requirement, though it depends on a relatively unstable spec 2009-06-20 [17:02:00.0000] <Hixie> http://www.w3.org/mid/OF73CC06FC.50275ECF-ON852575DA.0072E82C-852575DA.00734691⊙lc [17:02:01.0000] <Hixie> given all the amazing APIs we're inventing these days, i wonder if noah has a particular reason to prefer the 2DContext one, or if he just doesn't know about the others [17:03:00.0000] <Hixie> i mean, WebSocket, File, and Geolocation would all, imho, rate as more important APIs than canvas' [17:05:00.0000] <Hixie> hsivonen: any comments on http://www.whatwg.org/specs/web-apps/current-work/#an-introduction-to-error-handling-in-the-parser ? [17:06:00.0000] <Hixie> Philip`: same to you ^ [17:06:01.0000] <Hixie> or annevk5 [17:06:02.0000] <Hixie> or anyone who's worked on the parser :-) [17:08:00.0000] <Philip`> Hixie: "errorneous" [17:10:00.0000] <Philip`> Hixie: The element names in the example trees should probably be lowercase, I'd imagine [17:11:00.0000] <Hixie> yeah, i guess [17:11:01.0000] <Hixie> (fixed the typo, thanks) [17:12:00.0000] <Philip`> Some of the text nodes in the example trees get thick green underlines [17:13:00.0000] <Philip`> "The interesting part is with the end tag token with the tag name "b" is parsed." - s/with/when/ [17:13:01.0000] <Philip`> or 'where' or whatever [17:15:00.0000] <Hixie> ok reload, fixed all that [17:15:01.0000] <Hixie> (and thanks) [17:16:00.0000] <Hixie> bbiab, meeting [17:17:00.0000] <Philip`> I'm too lazy to check if all the details of the algorithms are correct, but I don't see anything blatantly wrong, and I think these sections seem useful as a way of giving people some intuition as to what the algorithms are trying to do [17:17:01.0000] <Hixie> cool [17:17:02.0000] <Hixie> i'll add some more then! [17:18:00.0000] <Philip`> But I could be wrong, and maybe they don't help at all for someone who's not already familiar with it [17:18:01.0000] <Philip`> I've been corrupted by years of blog posts and discussions and tests and implementations of this stuff, so I can't approach it with a fresh mind :-( [17:19:00.0000] <Philip`> (That's not to say I actually understand it, just that I'm vaguely familiar with the general concepts) [19:27:00.0000] <annevk5> Hixie, look ok, but also unfinished [20:20:00.0000] <Hixie> annevk5: yeah, wip [21:21:00.0000] <Hixie> why do i get garbage when i go to mpilgrim's site? [21:21:01.0000] <Hixie> how odd [04:13:00.0000] <annevk5> http://ejohn.org/blog/unimpressed-by-nodeiterator/ should we nuke it again (including from Acid3)... [04:16:00.0000] <jgraham_> annevk5: I guess acid3 is the only reason Mozilla implemented it. But it doesn't suck more than the other similar DOM apis, really [04:18:00.0000] <jgraham_> I'm not sure that his proposed replacement is really that valuble [04:21:00.0000] <jgraham_> Particularly I'm not sure how valuable it is to be able to select all the comment nodes or text nodes in the document and it is already rather easy to get all the element nodes [04:22:00.0000] <annevk5> I wasn't saying we should support the replacement [04:22:01.0000] <annevk5> just to nuke APIs that are not obviously helpful [16:56:00.0000] <Midler> will there be a target="_tab" in html5? [16:56:01.0000] <Midler> Forcing the browser to open a link in a tab [16:56:02.0000] <Midler> :P [16:59:00.0000] <Dashiva> No 2009-06-21 [17:00:00.0000] <Midler> ok [17:03:00.0000] <Philip`> It'd seem silly five years from now when the new UI paradigm doesn't have the concept of tabs [17:03:01.0000] <Midler> ye, i know [17:04:00.0000] <Midler> All browsers nowadays has tabs [17:04:01.0000] <Midler> However, u might want for some reason (that i cant get right now), open a page in a new window [17:05:00.0000] <Midler> Example, instead of javascript window.open [17:05:01.0000] <Philip`> Like target="_blank"? [17:05:02.0000] <Midler> yes [17:05:03.0000] <Midler> Somethimes _blank [17:05:04.0000] <Midler> and (most of the thime) _tab [17:06:00.0000] <Midler> but know when i think of it, target seems useless [17:06:01.0000] <Philip`> I think Opera already loads target=_blank into a new tab by default [17:06:02.0000] <Midler> most browser does by default [17:06:03.0000] <Midler> but thats the thing [17:07:00.0000] <Midler> If you by some reason dont want to open a page up in a new tab, then what? [17:07:01.0000] <Midler> I cant come to think of any of these moments [17:07:02.0000] <Midler> right now* [17:09:00.0000] <Philip`> /me knows this has been discussed before but can't remember the relevant arguments, and so goes to bed instead :-) [17:10:00.0000] <Midler> Hm, well [17:10:01.0000] <Midler> google had some javascript examples somewhere [17:10:02.0000] <othermaciej> hi everyone [17:10:03.0000] <Midler> And one of the example was about moving a ball thru different windows [17:11:00.0000] <Midler> Cant remember the page [17:12:00.0000] <Midler> Multiple windows interacting with each other* [17:14:00.0000] <Philip`> http://www.chromeexperiments.com/detail/browser-ball/ ? [17:15:00.0000] <Midler> jup [17:15:01.0000] <Midler> thats the one [17:16:00.0000] <Dashiva> Midler: The user has the choice to open in window and tab, that's all you need [17:36:00.0000] <Midler> Is there anyone here that have some time to spend and listen to my stupid questions =) [17:36:01.0000] <Midler> ? [17:37:00.0000] <annevk2> just write them down [17:37:01.0000] <annevk2> stuff is logged, maybe some day you'll get a reply :) [17:38:00.0000] <Midler> ^^ [17:38:01.0000] <Midler> Well, for example [17:38:02.0000] <Midler> typing  all the time to write a comment is quite...well [17:38:03.0000] <Midler> /me is a lazy guy [17:39:00.0000] <Midler> And second of all (i dont know if its within the standardisation) [17:39:01.0000] <Midler> But favicon [17:39:02.0000] <Midler> I would like to be able to have favicons in .png and .svg [17:39:03.0000] <Midler> And in bigger size for a reason [17:40:00.0000] <Midler> There is ppl that for example saves the homepage as a link on their desktop [17:40:01.0000] <Midler> And usually it will be desplayed with the broswer loggo [17:40:02.0000] <annevk2> you can do that by using the sizes attribute on <link> [17:41:00.0000] <Midler> well, but right now favicon has a maxsize doesnt it? [17:42:00.0000] <Midler> And also, as more stuffs is getting on the web. A link with the page logos instead of browser logo could be something to consider, but that means bigger logos [18:01:00.0000] <Midler> hmm [18:02:00.0000] <Midler> target is also implemented in area [18:04:00.0000] <Midler> If you take the example from w3c [18:04:01.0000] <Midler> http://www.w3schools.com/tags/html5_map.asp [18:04:02.0000] <Midler> Would you like that to open in a new tab, or blank? [18:04:03.0000] <Midler> /me would prefer window.open [18:04:04.0000] <Midler> so i answered myself ofcourse :P [05:21:00.0000] <takkaria> /me is now in Sweden [05:36:00.0000] <Philip`> /me offers condolences [05:37:00.0000] <takkaria> :) [05:49:00.0000] <MikeSmith> takkaria: you know yet what you'll be working on? [06:03:00.0000] <Philip`> gsnedders: http://gsnedders.html5.org/outliner/process.py?url=--%3E%3Cscript%3Ealert('haxxored')%3C/script%3E [06:35:00.0000] <MikeSmith> hmm, the spec's not particularly consistent about making it clear whether a particular attribute is required or not [06:36:00.0000] <MikeSmith> e.g., for the source element [06:36:01.0000] <MikeSmith> http://www.whatwg.org/specs/web-apps/current-work/multipage/video.html#the-source-element [06:36:02.0000] <MikeSmith> the spec doesn't say whether the media element is optional or required [06:37:00.0000] <MikeSmith> I guess readers are meant to infer that if the spec doesn't explicitly say a particular attributed is "required" or "must be present", it's optional [06:38:00.0000] <takkaria> MikeSmith: I'm on the opera Mini team apparently [06:43:00.0000] <MikeSmith> takkaria: really, on Opera Mini team? [06:43:01.0000] <MikeSmith> great team [06:43:02.0000] <MikeSmith> I reckon you're going to have a chance to do some pretty interesting stuff [06:43:03.0000] <MikeSmith> you're going to be coding, right? [06:46:00.0000] <MikeSmith> cool to see somebody actually implementing <datagrid> [06:46:01.0000] <MikeSmith> hey Rik` [06:47:00.0000] <Rik`> hey MikeSmith [06:49:00.0000] <hsivonen> whoa! who is implementing datagrid? [06:49:01.0000] <hsivonen> with all the copious CSS pseudos within? [06:49:02.0000] <MikeSmith> hsivonen: webkit [06:49:03.0000] <MikeSmith> dhyatt [06:49:04.0000] <MikeSmith> https://bugs.webkit.org/show_bug.cgi?id=26545#c1 [06:49:05.0000] <hsivonen> interesting [06:50:00.0000] <MikeSmith> [[ [06:50:01.0000] <MikeSmith> This implementation may end up being very different from what's in the spec. [06:50:02.0000] <MikeSmith> The goal is to create a simpler implementation that can help improve the spec. [06:50:03.0000] <MikeSmith> ]] [07:01:00.0000] <jgraham> /me wonders if hsivonen was expecting <datagrid> to be quietly dropped :) [07:07:00.0000] <hsivonen> jgraham: I was expecting it to get dropped. not necessarily quietly [13:07:00.0000] <Philip`> Hixie: You might want to remove sven⊙zc from the WHATWG list, else he'll probably send an "I'm not here" email once a day for the next week [16:51:00.0000] <Philip`> gsnedders: http://krijnhoetmer.nl/irc-logs/whatwg/20090621#l-155 in case you missed it earlier 2009-06-22 [19:49:00.0000] <MikeSmith> hendry: 404 for http://natalian.org/archives/tag/mobile/feed/atom/ [23:49:00.0000] <billyjackass> is [Reflect] not part of WebIDL? [23:49:01.0000] <annevk5> not part of the current draft [23:49:02.0000] <annevk5> what does it do? [23:50:00.0000] <olliej> annevk2: check darin's recent email to webkit-dev [23:50:01.0000] <olliej> annevk2: i would guess billyjackass is referring to that [23:50:02.0000] <billyjackass> yeah [23:51:00.0000] <billyjackass> https://lists.webkit.org/pipermail/webkit-dev/2009-June/008458.html [23:53:00.0000] <MikeSmith> annevk5: see IDL at bottom of https://bugs.webkit.org/attachment.cgi?id=31632 [23:53:01.0000] <MikeSmith> would [Reflect] be a useful addition to Web IDL? [23:54:00.0000] <othermaciej> maybe! [23:54:01.0000] <othermaciej> it would save spec writers time [23:54:02.0000] <othermaciej> but it's kind of a layering violation [23:54:03.0000] <othermaciej> since it's so specific to DOM elements [23:54:04.0000] <MikeSmith> I see [23:55:00.0000] <MikeSmith> no heycam around.. would be interesting to hear his thoughts when he gets on [23:55:01.0000] <othermaciej> (assuming you mean the same Reflect we have added or are planning to add to WebKit's IDL) [23:55:02.0000] <annevk5> ah, so that's why Hixie uses the term "reflect" [23:55:03.0000] <othermaciej> no, we copied it from Hixie [23:55:04.0000] <annevk5> oh :) [23:55:05.0000] <othermaciej> we just didn't feel like writing the boilerplate code in C++ for every trivial reflection DOM wrapper [23:55:06.0000] <MikeSmith> othermaciej: yeah, the message from Darin to webkit-dev was what got me asking [23:56:00.0000] <annevk5> MikeSmith, that email, referenced bug, and patch referenced from the bug, do not mention reflect [23:56:01.0000] <annevk5> ah, but https://lists.webkit.org/pipermail/webkit-dev/2009-June/008457.html does [23:57:00.0000] <othermaciej> yeah, Darin and I came up with that design [23:57:01.0000] <MikeSmith> annevk5: sorry, I guess the link I pasted was to the wrong e-mail message [23:57:02.0000] <othermaciej> it does seem like it would be a timesaver for markup language specs [23:57:03.0000] <othermaciej> to avoid the need to mention reflection in prose every time [23:58:00.0000] <othermaciej> but we are doing it to reduce the amount of handwritten code [23:58:01.0000] <MikeSmith> seems like other UA implementors might could find it useful for that too [23:59:00.0000] <MikeSmith> and would save words in the HTML5 spec at least [23:59:01.0000] <MikeSmith> maybe for SVG too? [23:59:02.0000] <annevk5> a few sentences [23:59:03.0000] <annevk5> though it would be neat if all specs used the same definitions, agreed [00:00:00.0000] <othermaciej> in general we'd like to get our IDL as close to Web IDL as possible [00:00:01.0000] <othermaciej> we may still need a few extra extended attributes for truly implementation-specific things [00:03:00.0000] <gsnedders> /me comes back to life [00:08:00.0000] <MikeSmith> / iht.com relies on id returning the empty string when no id is present. [00:08:01.0000] <MikeSmith> is an interesting comment [00:08:02.0000] <MikeSmith> heycam! [00:09:00.0000] <heycam> hi MikeSmith [00:10:00.0000] <MikeSmith> heycam: any thoughts on https://lists.webkit.org/pipermail/webkit-dev/2009-June/008457.html ? [00:10:01.0000] <MikeSmith> as far as whether it'd be useful to add a standard mechanism for that to WebIDL? [00:13:00.0000] <heycam> i suppose Web IDL doesn't really deal with elements [00:14:00.0000] <MikeSmith> yeah [00:14:01.0000] <MikeSmith> still, it's handy [00:14:02.0000] <othermaciej> when Darin originally proposed doing this in WebKit, I was hesitant because it seemed kind of layer-violating [00:14:03.0000] <heycam> yeah. maybe html 5 could mint that [Reflect] extended attribute. [00:14:04.0000] <othermaciej> on the other hand, it's going to be really convenient, allowing us to remove a lot of handwritten code [00:14:05.0000] <heycam> or Web DOM Core [00:15:00.0000] <othermaciej> it's less convenient from the spec perspective [00:15:01.0000] <othermaciej> Web DOM Core might be a good place if Web IDL is defined to be extensible in that way [00:15:02.0000] <gsnedders> /me curses at olliej for spoiling the race for him on Twitter :P [00:15:03.0000] <MikeSmith> maybe a "WebIDL extensions for markup-language specifications" mini-spec? [00:15:04.0000] <heycam> Web IDL currently says "Extensions to language binding requirements can be specified using extended attributes that do not conflict with those defined in this document." [00:16:00.0000] <heycam> in its extensibility section [00:16:01.0000] <MikeSmith> ah, good [00:16:02.0000] <heycam> so i think such extensions are sanctioned [00:17:00.0000] <heycam> i guess it would just save a sentence per attribute ("The foo attribute reflects the foo content attribute" or whatever.) [00:18:00.0000] <othermaciej> it could also save multiple specs having to define what it means to reflect a content attribute [00:18:01.0000] <othermaciej> if, for instance, SVG ever cares to use Web IDL [00:18:02.0000] <heycam> yep [00:18:03.0000] <heycam> though many of SVG's reflections are more complex SVGAnimatedBlah things [00:19:00.0000] <othermaciej> SVG could probably use its own extended attribute for that [00:19:01.0000] <othermaciej> being reminded of the SVGAnimated interfaces makes me sad though [00:19:02.0000] <othermaciej> now I need a drink before bed [00:20:00.0000] <heycam> haha [00:20:01.0000] <heycam> we were thinking of making them easier to use in SVG 2 [00:20:02.0000] <heycam> with some [PutForwards] and other things on them [00:20:03.0000] <heycam> so that you could do circleElt.cx = 100; and similar simpler things [00:20:04.0000] <othermaciej> I hate not only the API but the ridiculousness of implementing them [00:21:00.0000] <othermaciej> though fixing the API would certainly be a positive step [00:21:01.0000] <heycam> we discussed them at the f2f the week before last, though we came to the conclusion that there's probably enough content to avoid replacing those APIs completely. [00:22:00.0000] <heycam> so we're going to investigate those [PutForwards] kind of improvements, so that the cumbersome API still works too [00:22:01.0000] <othermaciej> sure, I'm not asking to remove the old bad API [00:22:02.0000] <othermaciej> it's just one of many legacy things that make me sad [00:23:00.0000] <heycam> makes me feel #00f too [00:23:01.0000] <heycam> (that's a web joke) [00:24:00.0000] <olliej> gsnedders: ? [00:24:01.0000] <gsnedders> olliej: http://twitter.com/ohunt/status/2261936217 [00:24:02.0000] <gsnedders> olliej: I not even started watching yet :P [00:25:00.0000] <olliej> gsnedders: top sites doomed me [00:25:01.0000] <gsnedders> olliej: :) [00:25:02.0000] <olliej> gsnedders: as i saw the podium in the updated snapshot [00:25:03.0000] <olliej> it made me sad [00:25:04.0000] <olliej> gsnedders: but yes, sorry for posting that [00:25:05.0000] <gsnedders> olliej: Snapshot from what? [00:25:06.0000] <olliej> gsnedders: formula1.com is in my topsites [00:25:07.0000] <gsnedders> olliej: Ah, yes, that has photos. [00:26:00.0000] <heycam> i find i have to turn off my twitter client before just popular TV show finales air, until i've watched them [00:27:00.0000] <shepazu> Pantone 292! [00:28:00.0000] <gsnedders> /me virtually always watches it live so normally doesn't have this issue [00:29:00.0000] <hsivonen> /me thinks the Web should use royalty-free colors [00:31:00.0000] <othermaciej> hippie [00:36:00.0000] <shepazu> hsivonen: so, no purple? [00:37:00.0000] <heycam> no cadbury purple, anyway [00:37:01.0000] <shepazu> don't they make eggs? [00:37:02.0000] <gsnedders> Yes [00:37:03.0000] <gsnedders> Of the chocolate variety [00:38:00.0000] <MikeSmith> http://suika.fam.cx/gate/2007/cssom/viewer [00:39:00.0000] <shepazu> /me notes that there's 2 jokes hidden in the "royalty-free purple" [00:39:01.0000] <hsivonen> shepazu: only purple made of r, g and b [00:40:00.0000] <gsnedders> hsivonen: Even if those r, g and b are identical to Pantone 292? [00:40:01.0000] <hsivonen> gsnedders: yeah. Independent implementations FTW! [00:40:02.0000] <gsnedders> /me wishes his shoulder didn't hurt [01:12:00.0000] <gsnedders> Huh, Adobe charge more _for_ download than shipping. [03:12:00.0000] <zcorpan_> /me adds a note about [Reflect] to web dom core [06:37:00.0000] <hsivonen> is <object> fallback known to suck in IE6? [06:50:00.0000] <hsivonen> Does IE6 ever show <object> fallback? [07:00:00.0000] <Lachy> In the Window interface, why is the window attribute defined as WindowProxy? It only seems to be IE that exhibits that weird behaviour of making (window === window.window) return false. [07:01:00.0000] <Lachy> though it still returns true when just == is used instead of === [07:03:00.0000] <hsivonen> Does Opera still have a feature that shows the document outline? [07:03:01.0000] <hsivonen> I think it did in the Opera 5 or 7 days [07:03:02.0000] <hsivonen> but now I can't find it in 10 [07:04:00.0000] <Lachy> I've never seen that feature [07:04:01.0000] <hsivonen> maybe my memory fails me [07:05:00.0000] <Lachy> I never used Opera back in the version 5 days though, so it's possible that it did [07:06:00.0000] <jgraham> Lachy: AIUI all browsers have something like WindowProxy [07:08:00.0000] <jgraham> hsivonen: View -> Style -> Table of Contents, maybe? [07:08:01.0000] <hsivonen> jgraham: I think that's not it. [07:41:00.0000] <rubys> hsivonen: do you develop on / have ready access to a Linux box? [07:42:00.0000] <hsivonen> rubys: I don't develop on Linux. I have a VM in the office and a real but slow Linux box at home. [07:44:00.0000] <rubys> http://svn.versiondude.net/whattf/htmlparser/trunk/ruby-gcj/test/domencoding.rb [07:45:00.0000] <rubys> doesn't do much yet, but is wrapping a Java Document in Ruby and exposing a method on that Ruby object that extracts data from the underlying Java object. [07:46:00.0000] <hsivonen> is encoding the data? [07:46:01.0000] <rubys> yes: String encoding = jdoc->getXmlEncoding(); [07:48:00.0000] <rubys> I've been slowed down by two Ubuntu/gcj/jaxp bugs... but if I'm past them, it is now simply a matter of adding more objects and more methods. [07:48:01.0000] <rubys> I've only tested it so far on Ubuntu and Fedora; don't know what it would take to reproduce on OSX or Windows. [08:04:00.0000] <MikeSmith> rubys: [08:04:01.0000] <rubys> /me listening [08:05:00.0000] <MikeSmith> gcjh -force -o headers/DomUtils.h -cp /usr/share/java/libgcj-4.3.jar:classes DomUtils [08:05:01.0000] <MikeSmith> rake aborted! [08:05:02.0000] <MikeSmith> Command failed with status (127): [gcjh -force -o headers/DomUtils.h -cp /usr...] [08:05:03.0000] <rubys> os? [08:06:00.0000] <MikeSmith> debian linux [08:06:01.0000] <MikeSmith> testing/unstable [08:06:02.0000] <MikeSmith> rubys: I just need to run "rake" in the build dir? [08:06:03.0000] <MikeSmith> or is there a configure step? [08:06:04.0000] <MikeSmith> /me opens the README [08:06:05.0000] <rubys> rake runs extconf... checking. [08:06:06.0000] <MikeSmith> maybe I need ruby-dev [08:07:00.0000] <MikeSmith> not sure I have it [08:07:01.0000] <MikeSmith> yet [08:07:02.0000] <rubys> that shouldn't cause gcjh to fail [08:08:00.0000] <MikeSmith> rubys: actually, I think I had no gcj [08:08:01.0000] <MikeSmith> because I removed it [08:08:02.0000] <rubys> that would cause gcjh to fail :-) [08:09:00.0000] <MikeSmith> gcj package pulls in dependencies like that fastjar thing, that screw with existing tools [08:09:01.0000] <MikeSmith> in that they don't behave like the stuff they are supposed to replace [08:11:00.0000] <MikeSmith> anyway, now installing fastjar gcj-4.3 libecj-java libecj-java-gcj libgcj9-dev [08:14:00.0000] <jgraham> hsivonen: BTW I looked in Opera 6 + 7 and couldn't see an outline toolbar [08:15:00.0000] <jgraham> s/toolbar/tool/ [08:15:01.0000] <jgraham> (but it is quite likely I missed something) [08:18:00.0000] <gsnedders> hsivonen: Netscape used to have it [08:18:01.0000] <gsnedders> hsivonen: And maybe early Moz suite too? [08:18:02.0000] <MikeSmith> rubys: coool. I now have a validator.so [08:18:03.0000] <MikeSmith> and libnu- stuff [08:19:00.0000] <rubys> :-) [08:19:01.0000] <rubys> doesn't do much yet, but the rest should be cookie-cutter [08:19:02.0000] <MikeSmith> /me now tries "rake test" [08:20:00.0000] <MikeSmith> works [09:39:00.0000] <jgraham> gsnedders: Which version? [09:39:01.0000] <gsnedders> jgraham: Dunno. [09:39:02.0000] <gsnedders> jgraham: I certainly remember it existing, and I'm pretty sure it was Netscape [09:40:00.0000] <gsnedders> jgraham: I want to say the early (pre-1.0) moz releases though [09:40:01.0000] <gsnedders> Which was what, 6? [09:40:02.0000] <gsnedders> Was 7 as well? [09:40:03.0000] <jgraham> gsnedders: I can't see it in 7 [09:41:00.0000] <jgraham> Although it makes my eyes bleed to look at [09:41:01.0000] <jgraham> so I can't search that well... [09:41:02.0000] <gsnedders> I certainly remember some old Mac OS browser having it. [09:42:00.0000] <jgraham> gsnedders: iCab? [09:42:01.0000] <gsnedders> jgraham: No [09:42:02.0000] <gsnedders> jgraham: IE/NN I think [09:42:03.0000] <gsnedders> /me opens IE 5.2 [09:43:00.0000] <gsnedders> Not IE [09:45:00.0000] <jgraham> gsnedders: It seems possible that someone developed an outline view for the sidebar thing in NS6/7 [09:45:01.0000] <gsnedders> jgraham: It was a default thing in sidebar for some browser [09:46:00.0000] <jgraham> gsnedders: Not afaict. Could you have had an ISP provided "enhanced" browser or something? [09:46:01.0000] <gsnedders> jgraham: No [09:47:00.0000] <gsnedders> Anyhow, I head off to pack [10:19:00.0000] <MikeSmith> rubys: don't make the switch, man [10:19:01.0000] <rubys> so... what do you know that I don't know :-) [10:20:00.0000] <MikeSmith> nothing... I know nothing. [10:20:01.0000] <MikeSmith> (like the dude on Hogan's Heroes) [10:20:02.0000] <Philip`> Hmm, shipping versions of Fedora have beta versions of Firefox? That doesn't sound like an excellent idea for stability [10:20:03.0000] <rubys> Ubuntu did that for Firefox 2 IIRC. [10:21:00.0000] <Philip`> That doesn't sound like an excellent idea for stability either [10:21:01.0000] <Philip`> though I thought it was for FF3? [10:21:02.0000] <MikeSmith> Philip`: fedora doesn't seem to care much about stability [10:21:03.0000] <Philip`> which had some nasty fsync bug on Linux, that all the Ubuntu users were exposed to [10:21:04.0000] <rubys> /me might have been 3 [10:21:05.0000] <Philip`> or something like that [10:22:00.0000] <Philip`> /me is happy to stick with Gentoo, where there's no temptation to mark a package as stable prematurely just to avoid having to delay it a year until the next release [10:24:00.0000] <rubys> releases are every six months for Ubuntu (and apparently for Fedora) [10:24:01.0000] <Philip`> Ah [10:24:02.0000] <Philip`> s/a year/six months/ then :-) [10:28:00.0000] <gsnedders> /me grumbles at finding the suitcase he was going to use has a mass of 12kg, which with Ryanair's limit of 15kg is problematic [10:30:00.0000] <Philip`> Fortunately 12 < 15 [10:30:01.0000] <beowulf> is your suitcase made of wood? [10:30:02.0000] <beowulf> somewhat like a clog, carved from a single scots pine [10:31:00.0000] <Philip`> You could fill the suitcase with helium to reduce its weight (which is what really matters) [10:32:00.0000] <Dashiva> I don't think gas containers are allowed [10:33:00.0000] <Philip`> I think it's kind of hard to have a container that doesn't contain gas, unless it contains a vacuum [10:33:01.0000] <Philip`> (or a liquid, but that's certainly not allowed) [10:34:00.0000] <Dashiva> I don't think you'll get through airport security with dictionary definitions [10:34:01.0000] <jgraham> No, seriously, how do you have a 12kg suitcase? [10:34:02.0000] <jgraham> Or does it have stuff in? [10:34:03.0000] <beowulf> perhaps it has hidden compartments stuffed full of narcotics [10:35:00.0000] <jgraham> In hich case you haven't deined the problem well enough [10:35:01.0000] <beowulf> but i like to imagine it's made of wood [10:35:02.0000] <Dashiva> It's lined with osmium [10:36:00.0000] <Philip`> Dashiva: It won't matter anyway because after your bag has been weighed, the security people can open it up to search for illicit contents and the helium gas will disperse harmlessly and invisibly [10:36:01.0000] <Philip`> although the security people might start speaking in high-pitched voices which would give the game away [10:36:02.0000] <Dashiva> Yes [10:36:03.0000] <Dashiva> And then you'd get sued for emotional distress too [10:38:00.0000] <gsnedders> jgraham: I can't work that out. [10:38:01.0000] <gsnedders> jgraham: There's nothing in it. [10:38:02.0000] <Philip`> Fill it with hydrogen, and then it'll be much safer [10:38:03.0000] <gsnedders> :P [10:38:04.0000] <Philip`> since that has less obvious effects than helium [10:38:05.0000] <gsnedders> What happens if the security people smoke? [10:38:06.0000] <Philip`> unless the security people are smokers, in which case they clearly deserve what they get [10:38:07.0000] <jgraham> gsnedders: You have a suitcase that, when empty, has a mass of 12kg? [10:38:08.0000] <gsnedders> jgraham: Yes. [10:39:00.0000] <Philip`> gsnedders: Oh, you beat me :-( [10:39:01.0000] <jgraham> That is insane [10:39:02.0000] <gsnedders> Philip`: Is smoking in a public place indoors illegal in England nowadays? [10:39:03.0000] <jgraham> Or at least not designed for air travel [10:39:04.0000] <jgraham> gsnedders: Yes [10:40:00.0000] <jgraham> gsnedders: WTF does said suitcase date from, and wtf was its intended purpose? [10:40:01.0000] <gsnedders> jgraham: A couple of years ago, dunno. [10:40:02.0000] <Philip`> Perhaps it was designed for carrying radioactive devices? [10:41:00.0000] <gsnedders> Philip`: Well, it'd leak a lot of radiation even with its weight [10:41:01.0000] <jgraham> Philip`: Sounds like a reasonable theory [10:41:02.0000] <gsnedders> (As it doesn't appear to be made of lead, though I expect it would stop alpha and beta radiation… but I expect most suitcases would) [10:44:00.0000] <jgraham> Seriously? 13kg? [10:44:01.0000] <Philip`> jgraham: No, 12kg [10:45:00.0000] <jgraham> Oh [10:45:01.0000] <jgraham> Well that's alright then [10:45:02.0000] <jgraham> 3kg should be more than enough for anyone [10:49:00.0000] <takkaria> 12kg for a suitcase is indeed impressive [10:50:00.0000] <Dashiva> Most laptops only weigh 2kg (or less) anyhow [10:52:00.0000] <Philip`> You could remove the keys that aren't very useful, like ` and F6, to save more weight [10:53:00.0000] <takkaria> my macbook is 2kg and it's pretty light [10:54:00.0000] <gsnedders> Philip`: That would be relevant if I had a keyboard in the hold. [10:54:01.0000] <gsnedders> /me finds a slightly larger case weighing 8kg [10:54:02.0000] <jgraham> takkaria: You still upstairs or you have got internet wherever you are residing now? [10:55:00.0000] <jgraham> gsnedders: Surely you can find a bag that weighs < 2kg? [10:55:01.0000] <takkaria> jgraham: I got to my apartment a couple hours ago, and internet is here [10:55:02.0000] <jgraham> takkaria: Oh cool [10:56:00.0000] <gsnedders> jgraham: Seemingly not. [10:56:01.0000] <gsnedders> /me has internet in his apartment [10:56:02.0000] <jgraham> gsnedders: Well obviously the blame lies with you [10:56:03.0000] <Midler> takkaria: If i sat down on ur macbook, it would break [10:56:04.0000] <Midler> =) [10:57:00.0000] <gsnedders> Midler: If I sat down on your computer, it would likely break. [10:57:01.0000] <Midler> ok, no more jokes for me in here... [10:58:00.0000] <takkaria> gsnedders: whereabouts are you living/when areyou gettinghere/what team are you on? [10:58:01.0000] <gsnedders> jgraham: Suitcases all appear to be surprisingly heavy [10:58:02.0000] <takkaria> my suitcase is 5kg [10:59:00.0000] <gsnedders> takkaria: A street whose name I cannot spell and which I have no idea how to pronounce, number 12. Friday. Dunno. [10:59:01.0000] <takkaria> /me grins, sounds about right [10:59:02.0000] <gsnedders> (I'm going down on sleeper to London tonight, and I fly from STN on Friday) [11:00:00.0000] <gsnedders> (And I come back from Linköping via Berlin and London, so I don't get home for around two weeks after my internship ends) [11:00:01.0000] <gsnedders> (So I need to pack for until then) [11:01:00.0000] <gsnedders> takkaria: How big is your suitcase? [11:01:01.0000] <takkaria> about so big [11:01:02.0000] <takkaria> /me gestures towards the suitcase [11:01:03.0000] <gsnedders> :) [11:02:00.0000] <gsnedders> takkaria: Where are you staying? [11:02:01.0000] <gsnedders> /what team are you on? [11:02:02.0000] <takkaria> I'm in Gottfriedsberg which is southwest of the office, about 20 minutes walk, on a road named Bjälbogatan and I'm on the Mini team [11:03:00.0000] <takkaria> we haven't quite figured out what I'm actually going to be doing on the Mini team yet though [11:03:01.0000] <gsnedders> hah. awesome. [11:05:00.0000] <gsnedders> takkaria: I'm basically north of you then [11:08:00.0000] <gsnedders> Weee… less humid in Linköping than it is here [11:09:00.0000] <takkaria> jgraham: amusingly we haven't been able to get me a login yet, so my first day was mostly chatting [11:09:01.0000] <gsnedders> takkaria: Are there other interns around? [11:09:02.0000] <takkaria> not AFAIK [11:09:03.0000] <gsnedders> Then I guess that means they'll probably fix it for all interns :( [11:10:00.0000] <takkaria> fix what? [11:10:01.0000] <gsnedders> takkaria: Fix insofar as get logins [11:11:00.0000] <takkaria> oh, right [11:12:00.0000] <takkaria> hopefully I'll have one tomorrow [11:13:00.0000] <gsnedders> /me wonders how much clothing to take [11:13:01.0000] <Midler> How come you are wisiting sweden? [11:13:02.0000] <Midler> visiting [11:13:03.0000] <gsnedders> Midler: I'm not; both myself and takkaria are interns at Opera's Linköping office this summer [11:14:00.0000] <Midler> aha, ok [11:15:00.0000] <takkaria> gsnedders: not much, I'd say [11:16:00.0000] <gsnedders> /me is far too much of a girl to accept that :P [11:16:01.0000] <takkaria> a week's worth of clothes, basically is what I went for [11:17:00.0000] <gsnedders> I mean, it's hardly as if one of my friends was telling some of the younger people from my school on the Hadrian's Wall trip at the weekend that I was their "gay best friend in female form" :P [11:17:01.0000] <Midler> Atleast the weather is getting nice now =) Enjoy your visit! [11:18:00.0000] <gsnedders> (Does that make me a lesbian?) [11:19:00.0000] <gsnedders> (It's certainly more fun to claim to be a lesbian, as that just confuses people) [11:20:00.0000] <takkaria> do you wear short, spiky hair and chequed shirts? :) [11:20:01.0000] <gsnedders> No, long hair in pony tail [11:20:02.0000] <gsnedders> :P [11:20:03.0000] <gsnedders> Though I do own a chequed shirt :) [11:21:00.0000] <gsnedders> takkaria: http://www.flickr.com/photos/gsnedders/3331908722/ [11:22:00.0000] <takkaria> ah, I used to have long brown hair [11:22:01.0000] <takkaria> then I cut it short and dyed it blonde [11:27:00.0000] <gsnedders> There's a photo of me on Facebook with green hair [11:28:00.0000] <gsnedders> http://www.flickr.com/photos/gsnedders/3494620206/ — one of the very few occasions I've had my hair down in around six months [12:08:00.0000] <jgraham> gsnedders: There are a couple of interns in core already [12:09:00.0000] <jgraham> (so I guess it is just takkaria who is having difficulties) [12:54:00.0000] <Dashiva> http://www.cssquirrel.com/2009/06/22/comic-update-who-really-is-the-wizard-of-html5/ [12:54:01.0000] <Dashiva> "If the data-harvesting Ian performs can’t be independently verified" -- Doesn't Philip` do this most of the time already? [12:54:02.0000] <Philip`> Yes [12:55:00.0000] <Philip`> I posted a comment there a while ago, but I guess it's being moderated or something [12:55:01.0000] <gsnedders> Dashiva: Lies! [12:56:00.0000] <gsnedders> /me grumbles at it being virtually impossible to do anything on his computer with IO at limit on all disks [12:56:01.0000] <Dashiva> So since there are no actual complaints about the contents, I suppose everything is okay again then [13:02:00.0000] <jgraham> I just posted a comment saying basically what Philip` was presumably going to say and it seems to have got through [13:04:00.0000] <Philip`> I tried saying what I'd said again but it said I'd already said it [13:29:00.0000] <Dashiva> So did the RDFa spec writers really expect that nobody would care enough to use it in HTML? [13:32:00.0000] <jgraham> Dashiva: Are you wondering for a reason? [13:34:00.0000] <Dashiva> It's easy to imagine reasoning from ignorance (HTML is irrelevant) and malice (grandfather it in later), but I don't think either of those really was the case. Surely they had some other reasoning. [13:34:01.0000] <Philip`> They were chartered to write a spec for RDFa in XHTML, so what else could they do? [13:35:00.0000] <Dashiva> They could make something that would be easy to extend into HTML :) [13:36:00.0000] <Dashiva> After all, most of the XHTML on the web is text/html anyhow. It's not like it makes a difference. [13:39:00.0000] <gsnedders> Dashiva: See /topic [13:39:01.0000] <Dashiva> But they aren't whatwg :P [13:40:00.0000] <jgraham> On the basis that the people doing the work now were likely to have had some influence on the charter, I would guess that the most likely reason is that they believed all the hype about HTML being dead and XHTML being the future. Indeed I would guess that they were generating that hype [14:05:00.0000] <gsnedders> jgraham: Here's how I think a suitcase can weigh 12kg: it actually weighs 12 lbs (e.g., 5.4 kg) [14:19:00.0000] <takkaria> ah, a unit mismatch. that's one of the classic blunders 2009-06-23 [18:17:00.0000] <Hixie> hsivonen: http://identi.ca/notice/5639842 :-( ? [21:45:00.0000] <MikeSmith> /me smirks about "spelling errors occurred" [21:49:00.0000] <MikeSmith> description of r3304 says "Write some explanatory text around the HTML parser." but the actual change just looks like some additions to the embedded CSS stylesheet [21:53:00.0000] <MikeSmith> ah, I see now [23:34:00.0000] <hsivonen> Hixie: V.nu whines about cellspacing and cellpadding for me. Steps to reproduce? [23:51:00.0000] <Hixie> hsivonen: no idea, just saw the twitter [01:40:00.0000] <hsivonen> /me wishes cssquirrel studied the mailing lists enough to call other people on practicing the bible of handling people [01:47:00.0000] <MikeSmith> I wish cssquirrel comics has more sex and violence [01:48:00.0000] <MikeSmith> because there are few things in this world that are funnier than depictions of cartoon characters having sex with each other [02:05:00.0000] <Dashiva> I'm not sure I get the cssquirrel argument [02:06:00.0000] <Dashiva> It's been shown that these data studies can be independently verified, and his response is that we should throw away the useful data just because... it has to be verified? [02:07:00.0000] <Dashiva> Next up, dot.org is revealed to be the real mastermind behind this [02:07:01.0000] <hsivonen> Dashiva: it's the vast Google-wing conspiracy! independent verifications doesn't change that [02:08:00.0000] <Dashiva> It does seem to reduce to "Data is evil, let's just make up opinions instead" [02:17:00.0000] <Philip`> "[using non-Google data] helps take the wind out of any Google influence concerns that may arise (such as mine)" - I'm not sure that's necessarily the case, because we already do that stuff and he has those concerns anyway based on the misconception that we don't [02:19:00.0000] <Philip`> so I guess the way to alleviate those concerns is to make it clearer to concerned persons what data we use, rather than actually doing anything much differently [02:24:00.0000] <Dashiva> Data is evil, I tell you [02:25:00.0000] <Philip`> Only Google data is evil, because the editor works for Google [02:26:00.0000] <Philip`> Also, data collected by humans is evil, because the editor is a human [02:26:01.0000] <Philip`> My data's okay though [02:26:02.0000] <hsivonen> Philip`: why is your data okay? is it an expert opinion? [02:26:03.0000] <Philip`> even though all the actual processing is done in secret with non-public code and it could be full of bugs or I could be making it all up [02:28:00.0000] <Philip`> hsivonen: My expert opinion is that I'm not an expert and it would not generally be sensible to trust anything I say :-) [02:30:00.0000] <Dashiva> So you admit you're making it all up? [02:30:01.0000] <hsivonen> fortunately, real Web data is so crazy you can't make it up [02:31:00.0000] <Philip`> Dashiva: I can neither confirm nor deny such rumours [02:32:00.0000] <Dashiva> Google's data verifies that you're making it up [02:33:00.0000] <Philip`> http://www.google.com/search?q=Am+I+making+it+all+up%3F - apparently I am not making this up [02:34:00.0000] <Dashiva> That's a trope [02:34:01.0000] <Dashiva> You obviously wouldn't fall for that [02:34:02.0000] <Dashiva> So you are making it up! [02:40:00.0000] <Philip`> Dashiva: Hmph! Foiled again [02:48:00.0000] <jgraham> Philip`: It would be insane not to use Google data where it is useful. [02:49:00.0000] <jgraham> So it seems like the CSSquirrel guy should have said "oh, my whole article was based on the misconception that no one was trying to reproduce these results. Sorry!" [02:49:01.0000] <jgraham> Rather than saying "oh well you shouldn't use that readily avaliable source of data because it could lead to accusations like the one I just made" [02:50:00.0000] <jgraham> "even though people have taken reasonable care to apply rigour and reproducability to the analayses" [03:13:00.0000] <hsivonen> /me wonders what happened to markp's 'this week' posts [03:25:00.0000] <Lachy> regarding this test http://software.hixie.ch/utilities/js/live-dom-viewer/saved/150 it's not clear to me whether the UA can wait till the loop has finished executing before firing the events, or whether it can fire them synchronously each time setItem() is invoked [03:26:00.0000] <Lachy> Minefield seems to fire the events synchronously, but our internal Opera build waits until the loop is done and then fires all events together. [03:27:00.0000] <Lachy> (that test freezes the latest WebKit though) [03:27:01.0000] <hsivonen> does this mean Chrome 1.x will be gone as soon as instances autoupdate selves? http://googlechromereleases.blogspot.com/2009/06/stable-beta-update-security-fix.html [03:29:00.0000] <annevk5> Chrome 1.x was already gone [03:29:01.0000] <abarth> which is A Good Thing (tm) [03:30:00.0000] <Lachy> does Chrome automatically update itself without asking the user? [03:30:01.0000] <annevk5> hey abarth, I've been meaning to think you for taking care of the CORS thread [03:30:02.0000] <abarth> although i talked to someone the other day who still had Chrome 0.2 [03:30:03.0000] <annevk5> oops [03:30:04.0000] <abarth> he had hacked the omaha client to stop updates [03:30:05.0000] <abarth> annevk5: no problem. hopefully tyler isn't too mad at me [03:31:00.0000] <annevk5> Lachy, storage fires async [03:31:01.0000] <annevk5> Lachy, though maybe not yet in Gecko [03:32:00.0000] <Lachy> annevk5, ok [03:33:00.0000] <abarth> Lachy: yes. although you can control the updates with registry keys [03:33:01.0000] <hsivonen> it seems that Chrome makes it hard for Web authors to keep old versions around, which I guess is a good thing [03:33:02.0000] <Lachy> then I wonder what would happen if I let the loop run for a few million loops and let the task queue fill up [03:34:00.0000] <abarth> Lachy: http://code.google.com/p/omaha/wiki/UpdateControls [03:34:01.0000] <abarth> hsivonen: why do you want old versions? 99% of users are on the latest version [03:35:00.0000] <hsivonen> abarth: I don't. I said I guess it's a good thing that it's hard to keep old versions around. [03:35:01.0000] <abarth> hsivonen: ah, misread :) [03:35:02.0000] <annevk5> /me wonders how this will work if Chrome is adopted by corporations [03:35:03.0000] <Philip`> Old versions are interesting as a historical curiosity [03:36:00.0000] <abarth> yeah [03:36:01.0000] <abarth> we tried running some super old version of navigator [03:36:02.0000] <abarth> they were super crashy [03:36:03.0000] <abarth> but fun [03:36:04.0000] <hsivonen> maybe I should subscribe my IE6 VM to Chrome stable, IE7 VM to Chrome beta and IE8 VM to Chrome dev [03:36:05.0000] <abarth> haha [03:37:00.0000] <abarth> the dev channel is too unstable for me [03:37:01.0000] <abarth> they're trying to make it better though [03:37:02.0000] <Philip`> (http://browsers.evolt.org/ is great, and lots of them still work in Windows or Wine) [03:37:03.0000] <abarth> with more automated testing [03:37:04.0000] <Philip`> (though some old versions of Netscape are presumably failing to send the Host header or something, and so are incapable of viewing almost every web site on the planet) [03:38:00.0000] <abarth> haha [03:38:01.0000] <Philip`> (or at least Google and Wikipedia, which I consider to be almost every web site on the planet) [03:38:02.0000] <abarth> we were looking at the lock icon [03:38:03.0000] <jgraham> Hmm it seems like the autoupdating + corporate users incompatibility would seriously damage the abaility of chrome to remove IE marketshare [03:38:04.0000] <abarth> i think the number of teeth on the lock represented the strength of the encryption [03:39:00.0000] <abarth> jgraham: chrome isn't optimized for enterprises yet [03:39:01.0000] <jgraham> (which I thought was the stated goal) [03:39:02.0000] <abarth> jgraham: the update controls are a first step in that direction [03:39:03.0000] <gsnedders> jgraham: I thought the stated goal was to help drive competition and hence development of all browsers [03:40:00.0000] <Philip`> /me once talked to someone who seemingly had a better-than-average understanding of web security, because he knew "if the thing at the top says something like HPTPS then it's safe" [03:41:00.0000] <jgraham> gsnedders: Maybe [03:41:01.0000] <Philip`> /me wouldn't really fancy having to explain how subtle changes in an icon represent encryption strength [03:41:02.0000] <abarth> yeah, it's just an interesting historical oddity [03:42:00.0000] <abarth> although it gives you some insight into their frame of mind at the time [03:42:01.0000] <Philip`> gsnedders: I thought the stated goal was world domination [03:43:00.0000] <Philip`> abarth: The userbase for web browsers was a bit different at that time too, so maybe it was not quite as unreasonable as it seems today [03:44:00.0000] <gsnedders> Philip`: What's the weather like there? [03:45:00.0000] <Philip`> gsnedders: I don't know, the weather monitoring station has been taken down for maintenance [03:45:01.0000] <gsnedders> :) [03:45:02.0000] <jgraham> gsnedders: It's like the siberian tundra with only slight changes to the temperature, precipitation and pressure [03:46:00.0000] <annevk5> Hixie, in the new Selector chapter you want to define "case-insensitive" as "ASCII case-insensitive" and link it and such [04:28:00.0000] <MikeSmith> /me wonders where zcorpan might be [04:29:00.0000] <MikeSmith> can't remember if Simon was around when we were discussing the [Reflect] thing yesterday [04:30:00.0000] <hsivonen> MikeSmith: he added a note to Web DOM Core later [04:30:01.0000] <MikeSmith> oh [04:30:02.0000] <gsnedders> MikeSmith: Linköping [04:31:00.0000] <MikeSmith> hsivonen: aha, I see it now [04:31:01.0000] <jgraham> gsnedders: He might be, he might not. It's hard to tell [04:31:02.0000] <MikeSmith> gsnedders: gesundheit [04:34:00.0000] <gsnedders> jgraham: Do you know whether you are right about where you thought I would be sitting? [04:35:00.0000] <jgraham> Yes. I know. [04:37:00.0000] <gsnedders> jgraham: Were you write? [04:37:01.0000] <gsnedders> right [04:38:00.0000] <takkaria> oo, where? [04:38:01.0000] <jgraham> As I am handed [04:38:02.0000] <gsnedders> /me sighs [04:39:00.0000] <gsnedders> jgraham: Are the people I'll be around evil? :P [04:41:00.0000] <jgraham> Do you consider "overendowed with natural programming ability" to be a sign of evil? [04:41:01.0000] <gsnedders> jgraham: Possibly. [04:42:00.0000] <Dashiva> Depends on the language [04:43:00.0000] <gsnedders> If Perl, certainly. [04:43:01.0000] <Philip`> There's nothing natural about the ability to program Perl [04:44:00.0000] <jgraham> Hixie and Philip` both have a certian fondness for Perl. I guess that explains a lot [04:46:00.0000] <jgraham> Maybe the CSSquirrel is right [04:46:01.0000] <Philip`> /me hasn't written any Perl more complex than "while (($_ = <$authz>) =~ /\S/) {" today, and that's quite readable :-( [04:46:02.0000] <hsivonen> maybe it's a vast Perl-wing conspiracy instead of a Google conspiracy. Google uses Python. [04:50:00.0000] <Philip`> Perl users are too pathologically eclectic to form a conspiracy [04:55:00.0000] <hsivonen> looks like IBM now has an Orbeon-like product: http://www.ibm.com/developerworks/blogs/page/JohnBoyer?entry=xforms_xfdl_and_lotus_domino?ca=drs-bl [05:00:00.0000] <hsivonen> why isn't webapps killing mutation events altogether? [05:46:00.0000] <takkaria> I've not found anywhere in Sweden yet which sells fresh potatoes [05:46:01.0000] <takkaria> this saddens me [05:46:02.0000] <takkaria> I like potatoes [05:48:00.0000] <zcorpan_> /me has eaten swedish fresh potatoes several times this year [05:49:00.0000] <jgraham> takkaria: Did you try a shop? [05:49:01.0000] <takkaria> I have [05:50:00.0000] <jgraham> /me has no idea what the difference is beteen fresh and non-fresh potatoes but has had no difficulty in aquiring potatoes [05:50:01.0000] <takkaria> well, I can find jars of salted potatoes [05:50:02.0000] <takkaria> I couldn't find any in Hemk�p, anyway [05:51:00.0000] <jgraham> Odd [05:51:01.0000] <jgraham> I am reasonably sure I have bought potatoes there [05:52:00.0000] <jgraham> I mean potatoes, in all their forms, seem to make up about 9/10 of the national diet [05:52:01.0000] <takkaria> maybe they've just not been then when I have [05:53:00.0000] <takkaria> s/then/there/ [06:12:00.0000] <beowulf> takkaria: ask gsnedders to bring some potatoes in his suitcase... [06:24:00.0000] <Philip`> He could smuggle them across the border disguised as Mr Potato Heads [06:24:01.0000] <Philip`> to avoid having to pay import duty [06:51:00.0000] <zcorpan_> Hixie: did you look at the emails when writing the latest checkin? [07:32:00.0000] <drcode> hi all [08:19:00.0000] <zcorpan_> Philip` says "I don’t remember Hixie quoting any private data about that issue." - http://www.cssquirrel.com/2009/06/22/comic-update-who-really-is-the-wizard-of-html5/ [08:19:01.0000] <zcorpan_> the response? "Philip, then Ian needs to use these external data sources from now on, rather than falling back on Google’s index." [08:21:00.0000] <zcorpan_> interesting that people complain about Hixie using google's index when he has only been looking at other people's data [08:22:00.0000] <jgraham> zcorpan_: I'm not sure interesting is quite the right word [08:23:00.0000] <Philip`> /me 's memory could easily be wrong, but he'll stick with it until someone provides evidence demonstrating otherwise [08:26:00.0000] <MikeSmith> that side of the discussion sorta seems to be missing the basic point that some data is at least better than no data [08:27:00.0000] <zcorpan_> MikeSmith: that's not what i see. i see complaints about hixie basing decisions on data from google's index [08:28:00.0000] <zcorpan_> MikeSmith: while in fact he's basing decisions on data provided by other people [08:28:01.0000] <MikeSmith> that is, if person/group P don't like the data that's been given by person/group H, then P is free to collect/publish other data [08:28:02.0000] <MikeSmith> zcorpan_: yeah, understood [08:29:00.0000] <Philip`> MikeSmith: Some data is no worse than no data, because you can always choose to ignore the data; but the complaints seem to be about people not ignoring the data [08:29:01.0000] <jgraham> I think that, at a more fundamental level, that side of the discussion is missing the fact that language design is fundamentally about HCI design. [08:29:02.0000] <MikeSmith> but my point about this is that it's perhaps not about the particular data, it's instead a kind of general fear of data for this case [08:30:00.0000] <jgraham> I really mean to write a blog post about that but I don't know how to get going [08:30:01.0000] <MikeSmith> that is, a realization that no matter where the data comes from for this case, it's going to support the conclusion that's been made so far after observing it [08:31:00.0000] <MikeSmith> so some choose to argue against the particular data-collection methods [08:31:01.0000] <MikeSmith> but they do so to get people to ignore the data [08:32:00.0000] <takkaria> quite a brilliant method, really [08:32:01.0000] <takkaria> I believe it's outlined in Hixie's Bible. :) [08:32:02.0000] <MikeSmith> so cssquirrel comic illustrates that pretty well [08:32:03.0000] <MikeSmith> ignore the data behind the curtain [08:32:04.0000] <MikeSmith> look over here where I'm waving my hands instead [08:33:00.0000] <Philip`> MikeSmith: The CSSquirrel thing doesn't say what specific case it's referring to, which makes it harder to argue against using specifics, but I guess it's really about accessibility again [08:33:01.0000] <MikeSmith> Philip`: any time nothing is mentioned specifically, it's usually about accessibility [08:33:02.0000] <jgraham> /me is still clinging on to the hope that people aren't malicious but misguided [08:34:00.0000] <MikeSmith> because unfortunately many people don't want to talk about it explicitly [08:34:01.0000] <MikeSmith> not talk about it in any kind of negative way, at least [08:35:00.0000] <gsnedders> Is there any way to resubmit a form in Opera? [08:36:00.0000] <gsnedders> Or do I really need to re-write what I just wrote? [08:36:01.0000] <Dashiva> ? [08:36:02.0000] <MikeSmith> jgraham: I don't think anybody involved in this is actually maliciously or even that they are particularly misguided [08:36:03.0000] <Philip`> /me thinks it's hard to avoid being misguided, because there's so much information about HTML5 development that it's pretty much impossible to keep up with it all, and so most people only notice the few specific areas where someone shouts loudly [08:36:04.0000] <Dashiva> gsnedders: When you go back, the form contents should still be there? [08:36:05.0000] <gsnedders> Dashiva: It isn't. It reloads the page with the original content. [08:36:06.0000] <Philip`> gsnedders: The back button works for me, except when the site is stupid and tries to do something clever [08:37:00.0000] <MikeSmith> jgraham: seems like most everybody has an investment/agenda they want to protect, even when they don't admit it or call it that [08:38:00.0000] <jgraham> MikeSmith: Maybe [08:39:00.0000] <MikeSmith> jgraham: and I guess some people are more easily able to cut their losses on a what turns out might have been a not-so-great investment, and some people find it a lot harder [08:40:00.0000] <MikeSmith> (how's that for talking around things?) [08:40:01.0000] <MikeSmith> I should shut up now [08:41:00.0000] <beowulf> i don't see how it helps that people are scared to talk about accessibility, if i can put it that way [08:41:01.0000] <beowulf> but what do i know? [08:41:02.0000] <Philip`> You shouldn't shut up - you should try saying more and more outrageous things, and see how long people keep thinking everything you say in IRC should be taken seriously [08:41:03.0000] <MikeSmith> heh [08:41:04.0000] <MikeSmith> I may be past that point already [08:42:00.0000] <MikeSmith> Philip`: and I've already succeeded in tainting our bugzilla with that kind of ambiquity and confusing now to [08:42:01.0000] <gsnedders_> /me contemplates going back to using Safari to write blog posts, because he never lost data there [08:42:02.0000] <gsnedders_> /me sighs [08:43:00.0000] <Philip`> gsnedders_: You should write them in Notepad and then paste them into the browser [08:43:01.0000] <beowulf> gsnedders_: use vim for that :) [08:43:02.0000] <zcorpan_> takkaria: do you have your irc client set to utf-8? [08:43:03.0000] <takkaria> zcorpan_: I think so [08:43:04.0000] <takkaria> gsnedders_: text editor first, paste into form later? [08:43:05.0000] <gsnedders_> Reloading the submitted form page (which was just "connection failed" gives me a blank page) [08:43:06.0000] <zcorpan_> takkaria: http://krijnhoetmer.nl/irc-logs/whatwg/20090623#l-321 [08:44:00.0000] <takkaria> oh, damn, not one but two people got there first [08:44:01.0000] <gsnedders_> My browser should not cause me to lose data. [08:44:02.0000] <takkaria> zcorpan_: hm, I'll tweak [08:44:03.0000] <gsnedders_> /me wonders if there is anyway to get at the data I lost [08:44:04.0000] <MikeSmith> beowulf: on the subject of humor, you riff about gsnedders_ suitcase yesterday was solid gold [08:44:05.0000] <MikeSmith> "carved out of single Scottish oak" or something like that [08:45:00.0000] <Philip`> Solid gold? That would explain why it was so heavy [08:45:01.0000] <beowulf> finally i make a contribution, where's my cv... [08:45:02.0000] <takkaria> gsnedders_: depends [08:45:03.0000] <takkaria> gsnedders_: do a memory dump to disk and grep [08:45:04.0000] <takkaria> worked for me once [08:45:05.0000] <MikeSmith> the helium and stuff from Philip` and others after that that was gold-ish too [08:45:06.0000] <takkaria> took me about an hour [08:45:07.0000] <jgraham> Anyway, like I said, the big philosopical divide I see is beteen people who view writing a language as basically an engineering problem (we have to provide features x y and z so we will have markup corresponding to x y and z) and people wwho view it as a HCI problem (we have to maximise the chance that users are exposed to the features x y and z and that they are used in the right way) [08:47:00.0000] <gsnedders_> /me wonders how to get a memory dump [08:48:00.0000] <Philip`> 'cat /dev/mem' if you have a sensible OS [08:48:01.0000] <takkaria> as root [08:50:00.0000] <gsnedders_> There is no /dev/mem [08:50:01.0000] <Philip`> Install a sensible OS first, then [08:50:02.0000] <Philip`> being careful not to reboot as it may disrupt your RAM content [09:01:00.0000] <MikeSmith> heh [09:02:00.0000] <MikeSmith> (about "being careful not to reboot as it may disrupt your RAM content") [09:03:00.0000] <takkaria> another way would be to write a program which malloc()s progressively more memory without freeing it and saving each chunk todisk [09:03:01.0000] <takkaria> eventually it crashes and you have a dump of free memoryspace [09:04:00.0000] <Philip`> Except that the OS will zero any pages before giving them to your process [09:04:01.0000] <Philip`> (otherwise it'd be a horrendous security hole) [09:04:02.0000] <takkaria> yeah, that's a point, really [09:04:03.0000] <takkaria> it would have worked on RISC OS [09:05:00.0000] <takkaria> that is all [09:06:00.0000] <Philip`> Would you need to malloc at all in that case, rather than just doing fwrite((void)0, 1, 0xffffffff, f)"? [09:06:01.0000] <Philip`> s/"// [09:08:00.0000] <takkaria> hmm [09:08:01.0000] <takkaria> I don't know if RISC OS would let you do that [09:23:00.0000] <MikeSmith> hmm, speaking of fear, I predict the jd-style practice of people putting scare quotes around particular words, an indicator that they fear them, will spread [09:23:01.0000] <MikeSmith> or has already [09:27:00.0000] <beowulf> i'd really like a dummies guide to the @summary issue [09:28:00.0000] <Philip`> MikeSmith: Hmm, I wish more people published indexes from their web "spiders" [09:36:00.0000] <Philip`> beowulf: That would be a nice thing to have, but I'm not aware of any that have been written with biases I agree with :-( [09:44:00.0000] <Midler> whats is it called again, the specs that make it possible to save data localy? [09:44:01.0000] <Midler> (i really sucks when typing in english) [09:46:00.0000] <Philip`> Midler: Do you mean http://dev.w3.org/html5/webstorage/ ? [09:46:01.0000] <takkaria> I don't know if RISC OS would let you do that [09:46:02.0000] <takkaria> /me rms that last line [09:47:00.0000] <Philip`> Someone needs to make an IRC client with an "undo send" feature [09:48:00.0000] <jgraham> /me wonders what rms has to do with it [09:48:01.0000] <takkaria> i guess like gmail's but with a shorter gap before sending [09:49:00.0000] <Midler> Philip`: yup, thars the one. Localstorage [09:54:00.0000] <beowulf> Philip`: then that means people will make their minds up based on which personality they happen to like, or which football team they support, i guess [09:59:00.0000] <Philip`> beowulf: I don't think that's a great solution, since I don't support any football team [10:00:00.0000] <Philip`> (but I don't care enough about accessibility to try to find or write a suitable document describing the issues) [10:00:01.0000] <Philip`> (so I'll let other people worry about it instead) [10:01:00.0000] <beowulf> /me will be more careful in choosing his metaphors in future [10:03:00.0000] <takkaria> yes, Shelley, "data is ... used to undercut those with experience and expertise in accessibility" [10:03:01.0000] <Philip`> You should choose a metaphor like you choose a wine [10:03:02.0000] <takkaria> that's the only reason people use data [10:03:03.0000] <takkaria> is to undercut other people's expertise [10:03:04.0000] <takkaria> not because it's an acutal grounds for making decisions on or anything [10:04:00.0000] <takkaria> /me rages a bit [10:04:01.0000] <Midler> How does firefox save localstore data? [10:04:02.0000] <Midler> in localstore.rdf? [10:04:03.0000] <takkaria> it uses sqlite [10:04:04.0000] <jgraham> Philip`: I don't drink wine. Does that mean I can't use metaphors? [10:04:05.0000] <Midler> ok [10:04:06.0000] <takkaria> Philip`: what if you just buy cheap red all the time? [10:05:00.0000] <beowulf> Philip`: I don't drink wine, you don't support a football team, we're doomed, doomed I say [10:06:00.0000] <Philip`> This all proves my point about metaphors! [10:06:01.0000] <Philip`> I don't know what my point was, but it's proved nonetheless [10:07:00.0000] <takkaria> in much the same way that bread is proved, I'm assuming [10:08:00.0000] <jgraham> By covering it in clingfilm and leaving it in a warm place? Or are we talking sourdough? [10:08:01.0000] <takkaria> I did that with some curry by accident today [10:09:00.0000] <Philip`> 10g yeast, 500g flour, 295ml warm water, QED? [10:09:01.0000] <takkaria> forgot to put it in the fridge [10:10:00.0000] <jgraham> Philip`: Depends on the type of yeast and the type of flour [10:12:00.0000] <beowulf> are we still talking in code about accessibility? [10:12:01.0000] <beowulf> i'm only slightly less lost than the actual threads on the wg list, if that's the case [10:12:02.0000] <jgraham> beowulf: I was just talking about bread making. What were you talking about? [10:12:03.0000] <jgraham> :) [10:12:04.0000] <jgraham> s/:/;/ [10:13:00.0000] <Philip`> http://www09.wolframalpha.com/input/?i=1+hogshead+of+bread - 2705% of the recommended daily amount of fat? That can't be good [10:13:01.0000] <beowulf> mmm, 2705% rda of fat [10:16:00.0000] <Philip`> /me continues his quest to avoid productive discussion [10:20:00.0000] <takkaria> 2705% rdfa of fat would make some people happier [10:31:00.0000] <Philip`> /me finds a program which appears to have a self-defeating anti-debugging mechanism [10:31:01.0000] <Philip`> If you run an expired trial version, it (presumably) checks some files or registry entries (on Windows) to decide it's expired, and then complains [10:32:00.0000] <Philip`> but if you run Process Monitor, to attempt to see which files/entries it's checking, it appears to skip the checks entirely and acts as if it's non-expired [10:32:01.0000] <Philip`> (presumably to prevent you easily reverse-engineering it and working out what it's checking) [10:33:00.0000] <Philip`> so all you need to do is keeping Process Monitor running in the background all the time [10:34:00.0000] <MikeSmith> x [10:34:01.0000] <MikeSmith> sweet [10:37:00.0000] <Midler> Or, like my friend did with one app.. Just changed the date for the computer [10:37:01.0000] <Midler> And if u are wondering, its still 1994 for him ;) [10:39:00.0000] <zcorpan_> hsivonen, MikeSmith: "Warning: Attribute with the local name href is not serializable as XML 1.0." - http://html5.validator.nu/?doc=http%3A%2F%2Fhacks.mozilla.org%2F2009%2F06%2Fhtml5-video-fallbacks-markup%2F [10:40:00.0000] <zcorpan_> also, "Error: The internal character encoding declaration must be the first child of the head element." [10:42:00.0000] <zcorpan_> it seems wordpress inserts <br /> for each new line in the <video> markup [10:43:00.0000] <MikeSmith> great [10:43:01.0000] <zcorpan_> the algorithm to select a <source> copes, though [10:43:02.0000] <zcorpan_> if it would insert a <p>, then it would not work [10:46:00.0000] <zcorpan_> i think "Error: Attribute href not allowed here." is supposed to say xmlns:dc, too [10:51:00.0000] <zcorpan_> http://bugzilla.validator.nu/show_bug.cgi?id=584 [10:55:00.0000] <MikeSmith> /me reads through hsivonen treebuilder code, feels sense of intimidation setting in [10:56:00.0000] <MikeSmith> zcorpan_: I don't get the "Error: The internal character encoding declaration must be the first child of the head element." message [10:57:00.0000] <MikeSmith> zcorpan_: I'm hoping the xmlns:dc problem is just an easy-to-fix off-by-one error [11:02:00.0000] <MikeSmith> zcorpan_: oh, actually, I do [11:02:01.0000] <MikeSmith> but only if I use html5.validator.nu [11:02:02.0000] <MikeSmith> not if I use validator.nu [11:04:00.0000] <MikeSmith> d'oh! [11:05:00.0000] <MikeSmith> (hadn't manually set validator.nu select option to html5) [11:13:00.0000] <zcorpan_> /me filed a bug on wordpress http://core.trac.wordpress.org/ticket/10247 [11:33:00.0000] <beowulf> would <span> be allowed in a summary element? [11:34:00.0000] <beowulf> or <p>? [11:35:00.0000] <MikeSmith> zcorpan_: fix for "encoding declaration must be the first child" problem checked in [12:29:00.0000] <zcorpan_> MikeSmith: nice [12:29:01.0000] <zcorpan_> /me notes that there would be much less resistance to just allow summary="" in html5 [12:31:00.0000] <Dashiva> zcorpan_: What design principle is that? :) [12:32:00.0000] <zcorpan_> Dashiva: it's not a design principle [12:33:00.0000] <zcorpan_> "Seems to me that this working group's underlying practice is more to take the path of least resistance," -- http://www.w3.org/mid/643cc0270906230930j1c37cf7dv46e58637d5dda46⊙mgc [12:57:00.0000] <Hixie> zcorpan_: there would be less resistance from one particular group of people, yes [12:57:01.0000] <Hixie> zcorpan_: my goal though is not to make those peopel happy but to actually measurably improve accessibility [13:08:00.0000] <jgraham> Expertise in the HTMLWG is like penises. Everyone likes to brag about how much they have but when it comes time to do some measurements they become remarkably reluctant to put it on the table. [13:08:01.0000] <Hixie> lol [13:09:00.0000] <takkaria> /me grins [13:43:00.0000] <annevk2> so summary="" is by default not revealed to end users? [13:45:00.0000] <Hixie> ? [13:45:01.0000] <annevk2> Hixie, at some point earlier today I pointed out that case-insensitive in the new Selectors section was not linked to ASCII case-insensitive [13:46:00.0000] <Hixie> yeah that was somewhat intentional... i was hoping css defined it [13:46:01.0000] <annevk2> Hixie, (the summary="" was about http://lists.w3.org/Archives/Public/public-html/2009Jun/0604.html and zcorpan's reply) [13:46:02.0000] <annevk2> Hixie, CSS wants the markup language to define it [13:46:03.0000] <Hixie> oh i'll get to public-html in a few fours [13:46:04.0000] <Hixie> swamped with www-style right now [13:47:00.0000] <annevk2> "The case sensitivity of document language element names, attribute names, and attribute values in selectors depends on the document language." -- http://www.w3.org/TR/css3-selectors/#casesens [13:49:00.0000] <Hixie> ah, bummer [13:49:01.0000] <Hixie> ok [13:53:00.0000] <Hixie> zcorpan_: technically <video hidden> means "this isn't relevant" and so we should probably make it mute the audio as well [13:53:01.0000] <annevk2> it's sort of nice as it means we could make all attribute values case-sensitive (which I think we should do) [13:54:00.0000] <Hixie> well we can't make all of them case-sensitive [13:54:01.0000] <Hixie> the ones in the spec are case-insensitive already in at least two browsers [13:54:02.0000] <annevk2> IE until IE7 had everything case-sensitive [13:54:03.0000] <Hixie> IE until IE7 didn't have attribute selectors, did it? [13:54:04.0000] <annevk2> including IE7, I should say [14:02:00.0000] <Hixie> shelley thinks i'm an expert? even _i_ don't think i'm an expert. [14:04:00.0000] <Philip`> An expert is just a retired pert [14:14:00.0000] <atwilson> I have a question about built-in constructors, and what the behavior should be if you pass a constructor to a different window or frame. [14:16:00.0000] <atwilson> Let's say I pass the window.XMLHttpRequest constructor into a child frame. That child frame uses the constructor to instantiate an XHR object. Does the XHR object get the origin from the originating window (the window that the XHR constructor came from) or from the calling scope (the child frame)? [14:16:01.0000] <atwilson> I have a similar question about Worker objects, but I figure there's probably a general principle at work here but I don't know where in the spec I might look for it. [14:18:00.0000] <ap> atwilson: I don't remember the rules, but the current behavior is a result of many tweaks and fixes - so there's chance that it's correct [14:19:00.0000] <ap> atwilson: sorry, didn't notice where we are :) [14:19:01.0000] <ap> atwilson: I was talking about WebKit [14:19:02.0000] <atwilson> ap: The reason I ask is the WebKit behavior for XHR differs from the WebKit behavior for Workers. [14:19:03.0000] <atwilson> ap: so I wanted to see the actual spec. [14:20:00.0000] <atwilson> ap: (per the recent webkit-dev thread) [14:21:00.0000] <ap> atwilson: like <http://dev.w3.org/2006/webapi/XMLHttpRequest-2/#origin-and-base-url>? [14:21:01.0000] <ap> atwilson: I think that annevk2 was going to update that though [14:22:00.0000] <atwilson> ap: yes, but I'm not sure how to parse that spec language. [14:23:00.0000] <atwilson> ap: how does one determine 'which is the Document object associated with the Window object for which the XMLHttpRequest interface was created'? [14:24:00.0000] <ap> atwilson: I think we should wait for Anne to weigh in on whether this text in in flux indeed [14:24:01.0000] <atwilson> Just look up the scope chain to find the global scope, and use the Window object associated with that? Because if that's the spec, then we aren't doing that in WebKit. [14:24:02.0000] <atwilson> Agreed. And I'm only using XHR as an example - the same issue applies to Workers. [14:25:00.0000] <atwilson> ap: Worker lifecycle depends on what document a given Worker is associated with, so making sure we associate with the correct Window is important. [14:32:00.0000] <othermaciej_> atwilson: XMLHttpRequest is totally clear on what to do [14:33:00.0000] <othermaciej_> (in my opinion) [14:33:01.0000] <othermaciej> and I am quite sure WebKit matches the intended spec behavior [14:33:02.0000] <othermaciej> Web Workers is a bit less clear [14:33:03.0000] <othermaciej> the only use of an original Window or Document would be for resolving a relative script URL against a base [14:33:04.0000] <othermaciej> it says to resolve the URL but it's not very clear about the base [14:34:00.0000] <othermaciej> my assumption is it should work like Location, which would mean the calling function's global object's current document is right [14:35:00.0000] <othermaciej> atwilson: the intent of the spec (as I understand it) is that whenever the XMLHttpRequest object is first instantiated by retrieving it from a particular Window object, it should record that Window object's current Document [14:36:00.0000] <othermaciej> WebKit records the Window instead of the Document [14:36:01.0000] <othermaciej> I can't remember if this makes an observable difference [14:36:02.0000] <othermaciej> there is definitely no looking up the scope chain involved [14:40:00.0000] <ap> othermaciej: I think the difference is that we record the window in constructor, not just in XMLHttpRequest object [14:40:01.0000] <ap> othermaciej: an I think that Anne was going to update the spec in this regard [14:47:00.0000] <othermaciej> ap: I see, the spec is probably wrong then [14:47:01.0000] <othermaciej> ap: I find its use of "the XMLHttpRequest object" a little confusing [14:47:02.0000] <othermaciej> hard to tell if it means an instance, or the constructor [14:48:00.0000] <othermaciej> it's "Window object for which the XMLHttpRequest interface object was created" that is unclear I guess [15:00:00.0000] <atwilson> othermaciej: yeah, if the spec is updated to reflect that the Window is bound to the constructor rather than to any given instance, then the spec is clear (and matches WebKit). [15:01:00.0000] <zcorpan_> aren't interface objects and constructors different things? [15:01:01.0000] <zcorpan_> window.XMLHttpRequest - interface object [15:01:02.0000] <zcorpan_> new XMLHttpRequest() - constructor [15:02:00.0000] <atwilson> Hixie: Do you have an opinion about how WebWorkers should work (if I pass window.Worker into a child frame, and then the child frame instantiates a Worker using that passed-in constructor, which document object ends up in that Worker's document set)? [15:04:00.0000] <atwilson> zcorpan_: Quite possibly. I'm trying to find out where 'interface object' is defined in the spec now. I can't find that string in the HTML5 spec when I search for it, although it is referenced in the XHR spec. [15:04:01.0000] <zcorpan_> atwilson: it's defined in webidl [15:04:02.0000] <zcorpan_> constructors too [15:05:00.0000] <atwilson> awesome, I'll look there. Thanks for the pointer. [15:06:00.0000] <zcorpan_> (though maybe a constructor is always also an interface object in implementations?) [15:08:00.0000] <zcorpan_> "An ECMAScript implementation supporting these interfaces would have a [[Construct]] property on the Circle interface object which would return a new object that implements the interface." [15:09:00.0000] <atwilson> The XHR spec is much clearer now, since it's clear that the constructor is bound to the originating window (i.e. var f = window.XMLHttpRequest; var xhr = new f(); yields an XHR object bound to the original window, regardless of what context new f() was invoked in) [15:12:00.0000] <atwilson> It sounds like Workers operate differently, though - the owning document seems to be tied to a "script's global scope". [15:12:01.0000] <atwilson> (section 4.5 of the Web Workers spec) [15:17:00.0000] <zcorpan_> it seems webidl allows [NoInterfaceObject] [Constructor] interface Foo {} [15:17:01.0000] <zcorpan_> which makes no sense in ecmascript [15:35:00.0000] <Hixie> atwilson: looking at "if I pass window.Worker into a child frame, and then the child frame instantiates a Worker using that passed-in constructor, which document object ends up in that Worker's document set" now... [15:38:00.0000] <Hixie> atwilson: web workers doesn't say how to resolve the argument to the constructor. will fix that now. [15:54:00.0000] <annevk2> so I did have a pass at XHR for those wanting to know [15:54:01.0000] <annevk2> currently base URL and origin are fixed when invoking open() [15:54:02.0000] <annevk2> I'm wondering if that's entirely correct or not [15:55:00.0000] <annevk2> it seems implementations differ a bit on the details and I've to look at all relevant testcases again (plus maybe make some more) to figure out if anything needs to change [16:02:00.0000] <othermaciej_> zcorpan_: that's not how I understood "interface object" but I could be confused [16:02:01.0000] <othermaciej_> zcorpan_: the object used for window.XMLHttpRequest is the same as the one used when invoing new XMLHttpRequest() [16:02:02.0000] <othermaciej_> zcorpan_: but different from the object returned* by XMLHttpRequest [16:04:00.0000] <Hixie> atwilson: ok, made it so that the url is resolved relative to the /first script/'s /script's base URL/ (see HTML5 for meanings of those terms), so all Worker constructors work identically [16:05:00.0000] <annevk2> othermaciej_, window.XMLHttpRequest is the interface object [16:06:00.0000] <annevk2> othermaciej_, though what you're saying there is very interesting and something I'd like to learn about [16:10:00.0000] <annevk2> specifically, the relation between the various objects [16:10:01.0000] <annevk2> but I should go to bed [16:11:00.0000] <annevk2> Hixie, in Web Workers you also need to define the relevant bits for XMLHttpRequest [16:12:00.0000] <annevk2> Hixie, I introduced the concepts "XMLHttpRequest base URL" and "XMLHttpRequest origin" [16:12:01.0000] <annevk2> Hixie, for when there's no such thing as "XMLHttpRequest Document" [16:15:00.0000] <atwilson> Hixie: OK, so it sounds like your intent is, indeed, to use the current script's global scope to determine the Document to insert in the set. So if I pass window.Worker to another frame, it doesn't carry with it any intrinsic relationship to that original window or its document. Correct? [16:16:00.0000] <zcorpan_> othermaciej: yeah, the constructor is the interface object [16:17:00.0000] <othermaciej> in that case the XHR spec is fairly clear and correct, I think [16:18:00.0000] <Hixie> atwilson: correct, all Worker constructors are created equal [16:18:01.0000] <Hixie> annevk2: can you drop me a mail about that? [16:18:02.0000] <atwilson> Hixie: thx. [16:25:00.0000] <jgraham> /me thought he had got over his "write a long essay" phase of email composition. Sadly not [16:26:00.0000] <jgraham> Also we should totally have a HTMLWG metup in Linkoping. Me + zcorpan_ + gsedders + takkaria for starters :) [16:28:00.0000] <annevk2> I have this plan to come over to Sweden [16:28:01.0000] <annevk2> maybe end of July? [16:28:02.0000] <annevk2> /me is really going to bed now; has a flight around 10ish [16:30:00.0000] <jgraham> For the record of the logs I am away at the end of July [16:34:00.0000] <othermaciej> jgraham: yeah, that was kind of a wall of text [16:43:00.0000] <zcorpan_> http://twitter.com/tieleman/statuses/2292450418 [16:49:00.0000] <Hixie> is there an up-to-date implementation of the <th scope> algorithm in a web tool i can poke at? [16:50:00.0000] <Hixie> is http://james.html5.org/tables/table_inspector.html up to date? [16:50:01.0000] <Hixie> (jgraham?) [16:52:00.0000] <zcorpan_> Hixie: it's up to jan 2009 [16:55:00.0000] <Hixie> k [16:57:00.0000] <zcorpan_> "Did you know that HTML5 editor Ian Hickson stated that HTML5 won’t fully be implemented in all browsers until 2022?" - http://www.davidtucker.net/2009/06/19/an-honest-open-discussion-on-web-standards-and-html-5/ [16:57:01.0000] <zcorpan_> i can't remember Hixie having stated that [16:57:02.0000] <Hixie> i already commented 2009-06-24 [17:03:00.0000] <zcorpan_> http://validator.nu/?doc=http%3A%2F%2Fwww.mindfly.com%2F [17:04:00.0000] <othermaciej> hmm [17:04:01.0000] <othermaciej> all of his "Did you know..." questions seem false [17:04:02.0000] <othermaciej> in their premise [17:04:03.0000] <othermaciej> (except maybe Google Wave?) [17:05:00.0000] <zcorpan_> othermaciej: see Hixie's comment [17:05:01.0000] <Hixie> even his google wave one is wrong; see my comment [17:06:00.0000] <Hixie> (and btw the use of gears in wave will become a non-issue as soon as arun's draft is ready and browsers have drag-and-drop File objects) [17:06:01.0000] <Midler> isnt that the year w3c is having it to be a standard? [17:06:02.0000] <Midler> 2022? [17:06:03.0000] <Hixie> see my comment on the blog entry :-) [17:07:00.0000] <ojan> and FWIW, eric's patch to add drag-and-drop files to webkit came out of the Wave team's request so they would not have to rely on Gears [17:10:00.0000] <Philip`> jgraham: You should include some ASCII-art diagrams to break up the flow of your long emails [17:11:00.0000] <Hixie> /me looks through ben millard's interesting tables collection to get ideas for examples for the spec [17:13:00.0000] <Hixie> i need a table that has a scope=rowgroup header cell in a non-left-most cell [17:15:00.0000] <othermaciej> I feel like this guy doesn't understand how HTML works [17:20:00.0000] <Midler> iam still reading it [17:21:00.0000] <Midler> "The truth is that plugins can ‘upgrade the web‘ in under a year." [17:21:01.0000] <Midler> That doesnt sounds that fantastic [17:21:02.0000] <zcorpan_> Hixie: use google's index to find it [17:21:03.0000] <Midler> Like we will be using flash forever [17:23:00.0000] <Hixie> zcorpan_: hah [17:23:01.0000] <Midler> I must ask how much the possibility for huge changes in the html5 document is? [17:24:00.0000] <Hixie> Midler: if someone finds a huge problem, pretty high :-) [17:24:01.0000] <Midler> Ok [17:24:02.0000] <Midler> I read comments concerning that [17:24:03.0000] <Midler> On Swedish forums [17:24:04.0000] <zcorpan_> Midler: pointer? [17:24:05.0000] <Midler> Its like: Ok, mozilla firefox add html 5 support [17:24:06.0000] <Philip`> Hixie: I found a huge problem in the spec [17:25:00.0000] <Philip`> Hixie: It doesn't rhyme at all [17:25:01.0000] <Midler> but you would never know if that support will be left 5 years from now [17:25:02.0000] <Hixie> Philip`: good lord, you're right. i'll make it rhyme right away. [17:27:00.0000] <zcorpan_> Midler: features that are shipped in multiple browsers and used on a wide scale on the web are very unlikely to change in incompatible ways [17:27:01.0000] <Midler> "very unlikely" is the concern...you can never know [17:29:00.0000] <Hixie> who knows [17:29:01.0000] <Hixie> tomorrow maybe the entire internet infrastructure will collapse [17:29:02.0000] <Hixie> and there will be no web [17:29:03.0000] <Hixie> or maybe adobe will go bankrupt, and there will be no flash [17:29:04.0000] <Hixie> you can never know :-) [17:29:05.0000] <Philip`> I think Google should buy Adobe [17:30:00.0000] <beowulf> which group does Shelley Powers refer to when she describes them as 'not diverse'? [17:30:01.0000] <zcorpan_> Midler: do you have a link to said forums? [17:31:00.0000] <zcorpan_> Hixie: i guess there would still be flash if adobe went bankrupt [17:31:01.0000] <Midler> zcorpan_: http://www.idg.se/2.1085/1.236210/se-upp-flash--html-5-vill-vara-med-och-leka?articleRenderMode=listpostings#a19.651633 [17:32:00.0000] <Midler> Its not forum, bit comments [17:32:01.0000] <othermaciej> did Shelley re-un-quit? [17:33:00.0000] <Dashiva> Yes [17:35:00.0000] <Midler> zcorpan_: Swedish? [17:38:00.0000] <Midler> There is comment that says this: "the problem with W3C is that the whole standard in theory could be rewritten entirely, which would make that they do not have as good implementation of the longer." <-- slightly modified [17:38:01.0000] <Midler> a google translate* [17:42:00.0000] <Dashiva> Not quite [17:43:00.0000] <Dashiva> "Of course, it only applies to the candidate [spec of] CSS 2.1; the problem with w3c being so senselessly slow is that the entire standard could in theory be rewritten, which would make it so they don't have as good an implementation any more" [17:50:00.0000] <Midler> oh i see. Wrong of me [17:51:00.0000] <Midler> but reading the discussion show a concern about the time it takes for it to be a standard [17:52:00.0000] <Midler> "The web is evolving faster than w3c makes something a standard" [17:54:00.0000] <othermaciej> Midler: in theory it could be, but in practice it won't [17:54:01.0000] <Midler> =) [17:54:02.0000] <Midler> Even i understand that [17:54:03.0000] <othermaciej> Midler: once a spec is at the Candidate Recommendation phase, any significant changes require a lot of bureaucracy [17:55:00.0000] <othermaciej> the basic idea is not to lock it down until we have two interoperable implementations so we're sure it works [17:55:01.0000] <othermaciej> it's similar to the IET [17:55:02.0000] <othermaciej> IETF I mean [17:55:03.0000] <othermaciej> where in theory, a "Draft Standard" could be changed completely before becoming an "Internet Standard", but in practice it won't be [17:56:00.0000] <jcranmer> beowulf: ancient Saxony? [18:23:00.0000] <Hixie> jgraham: i crashed your table inspector :-( [18:26:00.0000] <Hixie> jgraham: http://junkyard.damowmow.com/391 [18:27:00.0000] <Hixie> jgraham: http://james.html5.org/tables/table_inspector.py?algorithm=html5c&input_type=type_uri&uri=http%3A%2F%2Fjunkyard.damowmow.com%2F391&source=%3Ctable%3E%3Cthead%3E%3Ctr%3E%3Cth%3E%3Cth%3E%3Ctbody%3E%3Ctr%3E%3Cth+scope%3Drowgroup%3E [18:27:01.0000] <Hixie> ->crash [18:27:02.0000] <Hixie> it works if i add the two missing cells that hsivonen's validator complains about [18:37:00.0000] <zcorpan_> Hixie: maybe you should have an example in the spec where a script moves the element that the parser is inserting to [18:37:01.0000] <Hixie> send mail :-) [18:38:00.0000] <Hixie> i'm busy trying to make exampels for tables [18:54:00.0000] <zcorpan_> hmm, i tried to annotate "Parsing HTML documents" but it added a box to #comments instead :-/ [18:56:00.0000] <Hixie> yeah the annotation thingy is broken [18:56:01.0000] <Hixie> not sure what the bug is [18:56:02.0000] <Hixie> haven't looked [18:56:03.0000] <Hixie> i think it has to do with the "impl" annotations [19:28:00.0000] <heycam> zcorpan_, fixed the [NoInterfaceObject] / [Constructor] thing, thanks [19:33:00.0000] <zcorpan_> heycam: nice [19:35:00.0000] <zcorpan_> Hixie: the parts of html5 that were split out to w3c docs don't have the same permissive license as html5 [19:36:00.0000] <zcorpan_> Hixie: so it's harder to fork web storage [19:39:00.0000] <tantek> /me prefers open standards specs that use a public domain declaration in order to be truly open. [19:43:00.0000] <zcorpan_> Hixie: you should have a third color for legs and tails [19:44:00.0000] <zcorpan_> Hixie: at first i thought the green arrows were all from Cats and English speakers [20:23:00.0000] <othermaciej> tantek: in practice, I don't think a public domain declaration is better than, say, a BSD-style copyright license or rough equivalent [21:18:00.0000] <Hixie> zcorpan: the source is still in the whatwg svn under that license :-) [00:33:00.0000] <tantek> othermaciej re: PD vs BSD etc. - in general I would agree, however, my experience with book publishers specifically for example is that they know/understand/like content that is cleared as PD but few have heard of BSD, and unknown = fear/uncertainty/doubt. [00:50:00.0000] <Mrmil> What is more correct? Open <code>bloodyscript.rb</code>, open <var>bloodyscript.rb</var> or open <bloodysomethingelse?>bloodyscript.rb</bloodysomethingelse?> [00:55:00.0000] <jgraham> So. Much. Mail. [00:58:00.0000] <othermaciej> Mrmil: I think <samp> would be most correct [00:59:00.0000] <Mrmil> othermaciej: Ook, thanks [00:59:01.0000] <othermaciej> no, I guess not, maybe <kbd> [00:59:02.0000] <othermaciej> HTML has too many elements to set things in monospace [00:59:03.0000] <Mrmil> right [01:00:00.0000] <Mrmil> <kbd>/var/www/test/script.rb</kbd>? [01:00:01.0000] <Mrmil> it's not too right though, it's more a path then input... but then you can't follow path without an input [01:02:00.0000] <jgraham> Mrmil: It is irelevent. No UA will ever care. People are really unlikely to screen scrape for it and if you care about people trying to you could always use <code class=path> or something [01:02:01.0000] <Mrmil> jgraham: ok, I generally use <code> anyways. I just wanted to satisfy my curiosity. :) [01:13:00.0000] <sayrer> Mrmil: if you want something to be monospace, you can use <font> [01:15:00.0000] <Mrmil> sayrer: I use css for that, I was curious if there is a way to correctly annotate files and paths, looks like there is not just one solution [01:16:00.0000] <sayrer> I see, I'm sure the annotation is important ;) [01:18:00.0000] <Mrmil> sayrer: :P [01:30:00.0000] <hsivonen> hmm. I've lost the URL of Hixie's <meta charset> tests again [01:31:00.0000] <hsivonen> http://hixie.ch/tests/adhoc/html/parsing/encoding/ [02:18:00.0000] <MikeSmith> Hixie: the hide/highlight feature doesn't work at all in webkit-based browsers, afaict [02:18:01.0000] <Hixie> wfm [02:18:02.0000] <MikeSmith> hmm [02:18:03.0000] <Hixie> what happens for you? [02:18:04.0000] <Hixie> i only tested it in webkit trunk [02:18:05.0000] <Hixie> but it doesn't do anything really fancy [02:19:00.0000] <MikeSmith> I'm testing in webkit nightly [02:20:00.0000] <Hixie> what does it do? [02:20:01.0000] <Hixie> what are you doing? [02:20:02.0000] <MikeSmith> the buttons work as expected, but the URLs don't seem to for me [02:20:03.0000] <Hixie> paste one here? [02:20:04.0000] <MikeSmith> http://www.whatwg.org/specs/web-apps/current-work/?style=author [02:20:05.0000] <MikeSmith> if I go to that URL, it still shows me the impl stuff [02:21:00.0000] <Hixie> hm, yeah [02:21:01.0000] <Hixie> the right radio button is checked though right? [02:22:00.0000] <MikeSmith> Hixie: yeah [02:23:00.0000] <Hixie> weird [02:23:01.0000] <Hixie> the styles are being disabled properly [02:23:02.0000] <gsnedders> /me wonders how long it'll take to buy tickets at King's Cross [02:23:03.0000] <MikeSmith> maybe webkit bug [02:24:00.0000] <Hixie> must be [02:24:01.0000] <Hixie> don't see what else it could be [02:25:00.0000] <MikeSmith> Hixie: want me to ask or on #webkit, or do you? [02:26:00.0000] <Hixie> go ahead [02:27:00.0000] <MikeSmith> k [02:27:01.0000] <Hixie> if the right radio is being checked, i don't see how it could fail [02:31:00.0000] <zcorpan_> hixie: multipage/styler.js is a 404 [02:31:01.0000] <Hixie> oops [02:32:00.0000] <Hixie> ok, updater.js and styler.js will be symlinked properly next time i regen [02:36:00.0000] <Hixie> heycam: did you come up with an attribute name for merging interfaces? [02:49:00.0000] <heycam> Hixie, haven't got to that one yet [02:50:00.0000] <Hixie> heycam: let's make one up for now, so i don't have to check in about six bazillion [XXX]s [02:50:01.0000] <Hixie> (i'm merging in the old DOM2 HTML deprecated stuff) [02:51:00.0000] <heycam> sure, just make something up and i'll let you know when i get to it [02:51:01.0000] <Hixie> ok [02:53:00.0000] <Hixie> i'll use [Supplemental] for now [02:56:00.0000] <heycam> righto [03:33:00.0000] <gsnedders> Philip`: ping [03:36:00.0000] <gsnedders> Philip`: You want to meet up at 17:00 or so today? [03:36:01.0000] <gsnedders> /me probably will be gone before you respond, but can probably lurk around at the normal place at that time anyway :P [03:36:02.0000] <gsnedders> (weee… multi-second lag!) [03:37:00.0000] <hsivonen> "In reality, version 3 of the iPhone OS is the only solid platform that has full HTML5 support (as well as some other fixed development platforms)." [03:37:01.0000] <hsivonen> http://www.davidtucker.net/2009/06/19/an-honest-open-discussion-on-web-standards-and-html-5/ [03:38:00.0000] <Hixie> i'm sure apple will be really glad to hear that they don't have to implement any more stuff [03:38:01.0000] <Hixie> still no acknowledgement of my comment on that post [03:39:00.0000] <gsnedders> /me really does have to go now, so maybe see you [03:39:01.0000] <Hixie> later [03:39:02.0000] <gsnedders> (Actually, I think you have my number, so phone me) [03:40:00.0000] <gsnedders> /me heads off to Cambridge [03:58:00.0000] <zcorpan_> Hixie: hmm, why is size DOMString on font bug long on basefont? [03:58:01.0000] <Hixie> dunno, but that's what DOM2 HTML says [04:05:00.0000] <zcorpan_> seems in ie both are long, in opera both are DOMString, while firefox and webkit follow dom2html [04:27:00.0000] <hsivonen> zcorpan_: hmm. Interesting bug about the errors for xmlns:dc being reported for href [04:28:00.0000] <hsivonen> btw, that's pretty certainly a useless use of RDFa [04:28:01.0000] <hsivonen> pure metacrap [04:31:00.0000] <zcorpan_> hsivonen: it's useful because it found a bug in your validator :) [04:33:00.0000] <hsivonen> :-) [04:40:00.0000] <zcorpan_> Hixie: the note about dom feature strings seems to be incorrect now [04:41:00.0000] <zcorpan_> Hixie: maybe you should just use "HTML", "2.0" [04:42:00.0000] <Hixie> send [04:42:01.0000] <Hixie> mail [04:42:02.0000] <Hixie> :-P [04:42:03.0000] <zcorpan_> working on it [04:42:04.0000] <Hixie> :-) [04:44:00.0000] <hsivonen> whoa. lots of email on public-html. [04:44:01.0000] <zcorpan_> hsivonen: probably partially my fault :( [04:44:02.0000] <hsivonen> /me has been happily looking away from public-html (towards the Mozilla land instead) today [05:15:00.0000] <Lachy> yeah, there's way too much e-mail on public-html [05:16:00.0000] <Lachy> I've got through about a third of this week's discussion, still have just over 100 to go [05:32:00.0000] <Dashiva> Last week was almost empty, though. Strange how it fluctuates. [05:49:00.0000] <Lachy> http://www.contemporis.com/2009/06/translation-from-adobe-pr-speak-to-english/ [05:50:00.0000] <zcorpan_> Lachy: that's copied from somewhere else [05:51:00.0000] <zcorpan_> http://news.ycombinator.com/item?id=662620 [05:52:00.0000] <Hixie> oops, marked my massive DOM2 HTML checkin as editorial! [05:52:01.0000] <Hixie> oh well [05:53:00.0000] <Hixie> i think that is it as far as the obsolete section goes, so if anyone knows of anything that should be in there, please file a bug or send mail [05:53:01.0000] <Lachy> zcorpan_, oh well. Wherever it originally came from, it's still somewhat funny. [05:54:00.0000] <Lachy> although, I'm curious about the claim "iPhone helped to radically increase the number of phones with Flash support" since the iPhone doesn't support Flash [05:54:01.0000] <Hixie> ok bed time [05:54:02.0000] <Hixie> nn [06:01:00.0000] <MikeSmith> zcorpan_, hsivonen - you guys around? [06:02:00.0000] <zcorpan_> MikeSmith: yes [06:13:00.0000] <jgraham> The local rate of public html emails is about 1 per 10 minutes. So making the reasonable assumption of a constant rate we should have about 600,000 emails between now and last call [06:13:01.0000] <MikeSmith> zcorpan_: sorry, got distracted [06:13:02.0000] <MikeSmith> zcorpan_: about http://bugzilla.validator.nu/show_bug.cgi?id=586 from hsivonen [06:14:00.0000] <MikeSmith> zcorpan_: is that not the same as http://bugzilla.validator.nu/show_bug.cgi?id=584 ? [06:15:00.0000] <zcorpan_> MikeSmith: seems so [06:16:00.0000] <MikeSmith> OK [06:16:01.0000] <MikeSmith> I thought at first that Henri was filing a new bug caused by my checkin [06:16:02.0000] <MikeSmith> zcorpan_: as far as I can tell, my checkin to the treebuilder code fixes that problem, without any regression [06:17:00.0000] <zcorpan_> MikeSmith: cool [06:18:00.0000] <MikeSmith> and checked in a change for the encoding-must-be-first-child thing yesterday too, so that should be fixed too, next time hsivonen deploys [06:18:01.0000] <MikeSmith> zcorpan_: it was just those two problems you found yesterday, right? [06:20:00.0000] <zcorpan_> MikeSmith: yeah [06:20:01.0000] <MikeSmith> great [06:20:02.0000] <zcorpan_> MikeSmith: in v.nu anyway [06:21:00.0000] <MikeSmith> zcorpan_: where else would it need to be fixed? [06:21:01.0000] <zcorpan_> i found and reported problems in the blog post itself, in wordpress, in mozilla, and in mozilla documentation [06:23:00.0000] <zcorpan_> actually, not mozilla [06:26:00.0000] <zcorpan_> sadly i didn't find problems in the spec [06:37:00.0000] <MikeSmith> zcorpan_: I see [06:38:00.0000] <MikeSmith> have we ever had any success in getting wordpress to make fixes? [06:38:01.0000] <zcorpan_> i noticed wordpress was being patched for <video> support already [06:39:00.0000] <zcorpan_> or at least to let it through their filter of whitelisted elements and attributes [06:41:00.0000] <zcorpan_> http://core.trac.wordpress.org/ticket/10151 [06:42:00.0000] <zcorpan_> http://core.trac.wordpress.org/ticket/9437 - case-insensitive svg in html ++ [06:42:01.0000] <Dashiva> <video src="spam-ad.ogg" autoplay /> [06:42:02.0000] <Philip`> zcorpan_: Hopefully not the whitelist for comments? [06:43:00.0000] <Philip`> Dashiva: That would be bad, because it would give browsers a disincentive to implement Ogg support [06:43:01.0000] <zcorpan_> Philip`: there were concerns about that in the bug [06:43:02.0000] <Dashiva> Sorry, spam-ad.mp4 of course [07:51:00.0000] <Philip`> Licensing is hard :-( [07:53:00.0000] <Philip`> I have some project that includes C++ source files, which will be GPL; and some art data (bitmap textures etc) which will be CC-by-sa; but then there's stuff like XML files that tie together the art data, so they're sort of art data but since they're XML they're sort of like code too, and it gets a bit confusing [07:54:00.0000] <jgraham> Philip`: Lets go shopping [07:54:01.0000] <jgraham> Or, if you prefer I will ask you a more interesting question [07:54:02.0000] <jgraham> Actually I will anywway [07:55:00.0000] <jgraham> (you can always stick your fingers in your eyes and hum) [07:55:01.0000] <Philip`> (That might hurt my eyes) [07:55:02.0000] <jgraham> Is there a good/standard way of empirically determining the big-O of a function? [07:56:00.0000] <Philip`> What do you count as "empirical"? [07:56:01.0000] <jgraham> At least under some reasonable set of assumptions e.g. the function has to be drawn from a finite set [07:56:02.0000] <jgraham> s/function/big O function/ [07:57:00.0000] <zcorpan_> http://forums.whatwg.org/viewtopic.php?t=4101 - spam or not? [07:57:01.0000] <jgraham> Philip`: Say I have an implementation of a "sort" function and I want to know if it is quadratic or subquadratic purely by black box measurement [07:58:00.0000] <jgraham> In a way that is reasonably robust against the environment [07:58:01.0000] <Lachy> JohnResig, I just sent a revised copy of the selectors api exit criteria to public-webapps, which also describes what should be done with the testsuite. Let me know what you think, and if and when you'll be able to make those changes [07:59:00.0000] <jgraham> zcorpan_: spam [07:59:01.0000] <Philip`> jgraham: It's not possible to determine a theoretically correct answer - the function could have cost n^1.5 + n^2/1000000000, and it would be O(n^2) but all your black box measurements would make you think it was O(n^1.5) [08:00:00.0000] <Philip`> (The definition of big-O notation is that it's all about what happens in the limit, and you can't perform measurements in the limit) [08:01:00.0000] <jgraham> Philip`: Indeed. So lets say that we don't care so much about the true asymptotic behaviour but the behaviour for some reasonable range of input values [08:01:01.0000] <jgraham> Which isn't the theoretically pure case but is more practically interesting [08:02:00.0000] <Philip`> Then you can just run it for a range of n values, and pass it into some statistical analysis package (e.g. Excel's graph plotter) and get it to tell you the polynomial equation of the curve [08:02:01.0000] <jgraham> Philip`: Right but does that seem likely to be robust in the face of environmental factors [08:03:00.0000] <jgraham> like, say, GC pauses or whatever [08:03:01.0000] <jgraham> and there are practical concerns about the range of input values that should be chosen [08:03:02.0000] <Philip`> (There's also the issue that big-O is technically about worst case behaviour, and e.g. quicksort has terrible worst-case behaviour but you'll never notice that except on carefully-chosen inputs) [08:04:00.0000] <jgraham> (Yeah the worst-case behaviour thing is a reasonable concern but we'll ignore that for now) [08:04:01.0000] <Philip`> jgraham: It seems likely to be sufficiently robust that the difference between n log n and n^2 would be blatantly obvious [08:05:00.0000] <jgraham> Philip`: OK but you are not aware of any useful prior art on doing this kind of meaurement? [08:05:01.0000] <Philip`> I don't see why it's a complicated enough issue that one would need prior art, rather than just measuring some values and looking to see what the curve is like :-) [08:06:00.0000] <Philip`> or looking at the implementation and working out what it's doing [08:06:01.0000] <jgraham> Philip`: I have been told that this is harder in practice than it sounds in theory :) [08:07:00.0000] <Philip`> (In the latter case, see e.g. Knuth for more information than you ever wanted to know about sorting algorithms) [08:07:01.0000] <Philip`> jgraham: I guess it depends on how precise an answer you want [08:07:02.0000] <Philip`> e.g. n log n vs n^2 is easy to detect, but n^1.99 vs n^2 would be much harder [08:09:00.0000] <Philip`> I suppose the problem is I don't understand what problem you're trying to solve [08:11:00.0000] <jgraham> Philip`: Fair enough :) [08:12:00.0000] <jgraham> /me isn't quite sure either [08:12:01.0000] <Philip`> Easy solution: Just use heap sort [08:12:02.0000] <Philip`> Uh, merge sort [08:13:00.0000] <Philip`> I think that's the one I mean [08:15:00.0000] <zcorpan_> smylers++ on UAs generating a description of the table structure [08:16:00.0000] <jgraham> Philip`: It isn't really a sort-specific question [08:17:00.0000] <Philip`> jgraham: Oh [08:24:00.0000] <Philip`> jgraham: Anyway, I think in conclusion I don't have any useful non-obvious knowledge that would help you do whatever it is you're trying to do [08:27:00.0000] <JohnResig> Lachy: k - I'll probably be able to look at it/make the changes early July [08:35:00.0000] <Lachy> ok [08:37:00.0000] <zcorpan_> http://blog.benward.me/post/128294973 [08:41:00.0000] <jgraham> I think it's sweet of Adobe to talk about HTML5 at every possible oppertunity. [08:41:01.0000] <jgraham> It's hard to imagine what more they could do to promote it [08:44:00.0000] <Philip`> They could make Flash buggy and crash-prone, and not port it to new platforms and devices [08:45:00.0000] <jgraham> Philip`: Like it said it's hard to imagine what more they could do [08:45:01.0000] <jgraham> s/it/I/ [08:59:00.0000] <zcorpan_> jgraham: do you think it would be possible to hack together a demo implementation of generating a description of a table's structure? [09:00:00.0000] <jgraham> zcorpan_: All things are possible [09:00:01.0000] <zcorpan_> jgraham: i mean, do you want to do it? [09:00:02.0000] <jgraham> zcorpan_: I can have a go [09:00:03.0000] <zcorpan_> cool [09:02:00.0000] <jgraham> zcorpan_: Another interesting project would be to see how well the Mozilla layout table detector works [09:02:01.0000] <zcorpan_> yeah [09:02:02.0000] <zcorpan_> i wonder how mozilla's algorithm came into being [09:02:03.0000] <Lachy> are you show that it's possible for UAs to determine for itself what would otherwise be provided by a summary attribute (if one is supplied and used properly)? [09:03:00.0000] <jgraham> If it works reasonably well hsivonen could add it to the validator [09:03:01.0000] <jgraham> Lachy: No. But it seems worth finding out [09:03:02.0000] <zcorpan_> is it just a bunch of guesses? or is it based on research? [09:03:03.0000] <jgraham> /me guesses it is a bunch of guesses [09:03:04.0000] <Lachy> jgraham, certainly, if we could show that summary was redundant, then it would end the debate [09:04:00.0000] <Philip`> Lachy: I assume that was a joke [09:04:01.0000] <jgraham> Lachy: Hah ha [09:05:00.0000] <Lachy> no, it wasn't a joke. I'm serious. If everything an author is supposed to provide with a summary attribute can be determined automatically from the table structure, then it is redundant [09:05:01.0000] <zcorpan_> jgraham: look at ben's collection of tables for inspiration [09:06:00.0000] <jgraham> Lachy: I doubt it will be possible to determine "everything an author is supposed to provide" [09:06:01.0000] <jgraham> Even if it is I think people will still claim @summary is needed [09:07:00.0000] <jgraham> (for legacy UAs or for super-complex corner cases or something) [09:07:01.0000] <Philip`> Is summary="Calendar" considered a useful use of summary? That's something that's very hard to determine algorithmically, and hard for someone using a non-graphical browser to determine, but trivial for someone using a graphical browser [09:08:00.0000] <jgraham> Philip`: That's not what @summary is for, apparently [09:08:01.0000] <Lachy> Philip`, I don't think so [09:08:02.0000] <jgraham> (my current understanding is that it is _only_ supposed to be used for describing the structure of the table) [09:08:03.0000] <Philip`> jgraham: It doesn't really matter what it's apparently for, it matters whether it's beneficial to users [09:09:00.0000] <zcorpan_> would be interesting to see if it's possible to determine whether a given table is a calendar or not [09:09:01.0000] <zcorpan_> calendars usually have a specific format [09:09:02.0000] <Lachy> "This attribute provides a summary of the table's purpose and structure ..." - For most common simple cases, it's possible to determine the structure simply by looking at the headings. The purpose should be indicated by the caption or surrounding prose. [09:15:00.0000] <beowulf> if the ua can't provide a summary, allow the author to use @summary or whatever, no? [09:16:00.0000] <Lachy> beowulf, even if the UA can't, there are better proposals than the @summary [09:42:00.0000] <remysharp> hi - I've been playing with the manifest for offline apps in Safari, and checking the docs because Safari won't reload a server side change - is this being handled correctly? [09:42:01.0000] <remysharp> i.e. the page is loaded and told to cache xyz assets [09:42:02.0000] <remysharp> I make a change to the page, like the h1 text [09:42:03.0000] <remysharp> and Safari doesn't reflect the change [09:42:04.0000] <remysharp> I can only get it to load the new page if I cache bust the URL manually. [09:43:00.0000] <ap> remysharp: you need to change the manifest for resources to be refreshed [09:43:01.0000] <remysharp> but it was the actual page that was being loaded, i.e. it's not specified in the manifest [09:44:00.0000] <ap> remysharp: it goes to the manifest implicitly (as a master resource) [09:44:01.0000] <remysharp> aye, that's what I figured - [09:44:02.0000] <remysharp> are you saying if I 'touch' the manifest, it should update it? [09:45:00.0000] <remysharp> or - something else, how should I be able to tell the browser the master resource has been updated? [09:45:01.0000] <ap> remysharp: its content should change, not just modification date [09:45:02.0000] <ap> remysharp: its == manifest [09:45:03.0000] <remysharp> but I've got nothing to add though - it's just a change to the html [09:45:04.0000] <ap> remysharp: you can add a version in a comment [09:45:05.0000] <remysharp> ah - right. [09:45:06.0000] <remysharp> so comments can trigger an update then? [09:46:00.0000] <ap> remysharp: any change to the manifest content will [09:46:01.0000] <remysharp> hm - just tried it - it didn't. [09:46:02.0000] <remysharp> but I guess that's an implementation bug, right? [09:46:03.0000] <ap> remysharp: why do you think so? [09:47:00.0000] <remysharp> because I changed the manifest, as you suggested [09:47:01.0000] <remysharp> and the rendered html didn't update in safari [09:47:02.0000] <ap> remysharp: so far, so good [09:47:03.0000] <ap> remysharp: changing the manifest and reloading should start an update process [09:47:04.0000] <remysharp> yeah, but it didn't. [09:47:05.0000] <remysharp> at all. [09:48:00.0000] <ap> remysharp: once it's done, and the cache is updated, your page gest an event [09:48:01.0000] <remysharp> it's not though [09:48:02.0000] <ap> remysharp: is is_not_ magically reloaded again [09:48:03.0000] <remysharp> I'm listening for all the events [09:48:04.0000] <remysharp> and I've also triggered manual update() [09:48:05.0000] <ap> remysharp: what event listeners do you have installed? [09:48:06.0000] <remysharp> and swap [09:48:07.0000] <remysharp> one sec [09:48:08.0000] <remysharp> checking,error,noupdate,downloading,progress,updateready,cached,obsolete [09:48:09.0000] <remysharp> that lot! [09:48:10.0000] <ap> remysharp: which ones did fire? [09:49:00.0000] <remysharp> currently - after manually refreshing - it's on 'downloading' [09:49:01.0000] <remysharp> but it's stuck on downloading, and the master resource just isn't updating so far. [09:50:00.0000] <ap> remysharp: sounds strange. it is possible that there's a bug, but we do have tests verifying the behavior [09:50:01.0000] <remysharp> hmm - and now Safari crashed! [09:50:02.0000] <remysharp> are these the ones that drive the test data on the whatwg specs? [09:50:03.0000] <ap> remysharp: all right, that's a bug :) [09:50:04.0000] <remysharp> because I've been writing a bunch of simple tests myself, and a few of them definitely don't marry up. [09:51:00.0000] <zcorpan_> hsivonen, MikeSmith: maybe all obsolete features in html should be in the html5 schema (including framesets), with assertions saying that it's obsolete and what to use instead [09:51:01.0000] <remysharp> for instance, drag and drop, from what I can see in testing - is only supported in Firefox - and not Webkit [09:51:02.0000] <ap> remysharp: WebKit tests are just for WebKit regression testing (we're getting offtopic here, could move to #webkit) [09:51:03.0000] <remysharp> ap: aye, sorry. [09:52:00.0000] <remysharp> ap: so, to trigger an update to the application cache, from the server, I should be updating the manifest - that's the approach [09:52:01.0000] <ap> right [09:52:02.0000] <remysharp> ap: is that in the spec - I didn't come across it when I was trying to see if I had to manually trigger it somehow. [09:52:03.0000] <remysharp> (I mean, I assume it is, and I missed it) [09:54:00.0000] <ap> remysharp: see step 7 in 6.9.4 [09:54:01.0000] <MikeSmith> zcorpan_: I have a patch to add frameset written up already [09:54:02.0000] <zcorpan_> MikeSmith: sweet [09:55:00.0000] <ap> remysharp: update process is invoked in many circumstances (e.g. each time a new page opens), and usually quickly finishes due to the manifest being unchanged [09:55:01.0000] <MikeSmith> frameset and basefont are the only obsolete things that aren't already in teh v.nu schemas [09:55:02.0000] <ap> remysharp: please file bugs at bugs.webkit.org about the crash, and about other issues you may find [09:55:03.0000] <remysharp> ap: yeah, the new page opens - that's what I was figuring on - but it doesn't seem to happen in webkit - re: bugs, will do - if I can consistently create them. [09:56:00.0000] <remysharp> ap: re: 6.9.4 is that on www, or dev.w3 ? I'm on www and offline is in 5.x [09:56:01.0000] <zcorpan_> MikeSmith: is marquee in? [09:56:02.0000] <ap> remysharp: I was looking at whatwg.org [09:56:03.0000] <remysharp> ap: ah! [09:56:04.0000] <zcorpan_> MikeSmith: spacer? [09:57:00.0000] <MikeSmith> zcorpan_: nope. I meant elements that were valid in HTML4 but that are not valid in the HTML5 draft [09:57:01.0000] <zcorpan_> MikeSmith: ok. i'm looking at the obsolete section in html5 [10:06:00.0000] <zcorpan_> MikeSmith: is it possible to generate messages that say what to use instead for obsolete attributes? [10:12:00.0000] <MikeSmith> zcorpan_: yeah, it's possible to do with the assertions-checking code [10:12:01.0000] <remysharp> ap: cheers for the help. [10:12:02.0000] <MikeSmith> zcorpan_: anything is possible with that. we can make the error messages say whatever we want [10:13:00.0000] <zcorpan_> MikeSmith: i think it would be useful to rip off what Hixie wrote in the obsolete section [10:14:00.0000] <zcorpan_> MikeSmith: hmm, actually [10:14:01.0000] <zcorpan_> MikeSmith: some messages already have the content model stuff [10:15:00.0000] <MikeSmith> zcorpan_: yeah, for some cases [10:15:01.0000] <zcorpan_> MikeSmith: maybe it would make sense to include a separate field in a similar manner for advice on what to use instead [10:15:02.0000] <zcorpan_> MikeSmith: instead of adding everything to the schema and assertions [10:17:00.0000] <MikeSmith> we can't add arbitrary messages to the grammar-based schema checking anyway. we are stuck with what Jing reports there, for the most part [10:17:01.0000] <MikeSmith> the content-model stuff that hsivonen has it emitting now is something that he hooked in with custom code [10:18:00.0000] <MikeSmith> it just copies that from a copy of the spec, verbatim [10:18:01.0000] <zcorpan_> MikeSmith: can't the advice be implemented the same as the content model stuff? [10:19:00.0000] <MikeSmith> it could be, but it would be more trouble than just adding it to the assertions checking [10:19:01.0000] <zcorpan_> ok [10:19:02.0000] <MikeSmith> the assertions checking is now implemented as custom Java code anyway [10:19:03.0000] <MikeSmith> so we can do anything we want with that [10:19:04.0000] <MikeSmith> there is still the assertions.sch file, but v.nu doesn't actually use that [10:20:00.0000] <zcorpan_> aha [10:20:01.0000] <MikeSmith> anyway, v.nu has an mechanism already for emitting Info-class messages (as opposed to Error or Warning) [10:20:02.0000] <MikeSmith> so that part's already in there [10:21:00.0000] <zcorpan_> i was having in mind including the advice in the same error message [10:21:01.0000] <zcorpan_> like content model stuff is tied to a an error message [10:22:00.0000] <MikeSmith> I think it would need to be Info so that users can suppress it if they want [10:22:01.0000] <MikeSmith> IMHO, that's how the content-model stuff should be too [10:22:02.0000] <MikeSmith> they can still be tied together [10:23:00.0000] <MikeSmith> could just think of them as different fields of the same message [10:24:00.0000] <MikeSmith> anyway, I chatted with hsivonen a while back of using the assertions-checking backend for more cases, particularly cases where the error messages from Jing aren't very useful [10:24:01.0000] <MikeSmith> which is actually a lot of cases [10:24:02.0000] <zcorpan_> MikeSmith: some parse errors already have advice on what to do instead [10:25:00.0000] <MikeSmith> yeah, the error messages for parse errors is another separate thing [10:27:00.0000] <MikeSmith> as far as the schema-checking part goes, I think one lesson from this is that you can't ever expect to get useful messages for all cases from a grammar-based generic checker like Jing or any other RelaxNG or XSD checker [10:28:00.0000] <MikeSmith> if you want good error messages for more cases, it has to be supplemented with other kinds of checking [10:28:01.0000] <MikeSmith> schematron assertion-based checking is actually a really good way [10:29:00.0000] <MikeSmith> but the performance sucks [10:30:00.0000] <MikeSmith> and it is based on XPath, and not everybody likes XPath, and writing XPath expressions for some cases is a PITA [10:31:00.0000] <jgraham> /me onders how the fraction of custom java code in validator.nu has changed as a function of time [10:31:01.0000] <jgraham> +w [10:31:02.0000] <zcorpan_> so the conclusion is to use relax ng for anything that doesn't need useful messages and custom code for the rest? [10:32:00.0000] <MikeSmith> jgraham: the custom code for assertion checking has been going down recently because Hixie has relaxed or removed some constraints [10:32:01.0000] <MikeSmith> like the one about encoding declaration needing to be first child of head [10:32:02.0000] <MikeSmith> or around datagrid content model [10:33:00.0000] <MikeSmith> the datagrid content-model checking was a big chunk of the assertions code [10:33:01.0000] <MikeSmith> zcorpan_: no, not exactly -- not IMHO at least [10:33:02.0000] <MikeSmith> though that is basically what v.nu is doing now [10:34:00.0000] <MikeSmith> well, not exactly [10:34:01.0000] <MikeSmith> lemme rephrase that [10:35:00.0000] <MikeSmith> v.nu now is mostly using relax ng for anything that can be expressed practically in relax ng, and then using the assertions checking for stuff that can't, as then other custom code for stuff that can't be done with assertions checking (e.g., overlapping table cells) [10:36:00.0000] <MikeSmith> so relax ng does emit useful messages for a lot of cases [10:36:01.0000] <MikeSmith> but there are some others where it doesn't -- like cases that Rick Jelliffe mentioned [10:37:00.0000] <MikeSmith> e.g., where the constraint is around allowed sequences of elements [10:47:00.0000] <webben_> /me notes that one major advantage of DTD-based validation is developers can hack the DTD to change the constraints. I don't think that offsets the way DTDs limit what you can check, but it is a major advantage nevertheless. [10:48:00.0000] <webben_> e.g. at Yahoo!, we're on route to using a custom DTD internally for validating the sub- and super-sets of HTML4 we want to be using. [10:48:01.0000] <brale> hi all. I need some help, an idea. I need to extract only (printable) text from some page. Is there an easy way to do this using html5lib or some other library? [10:50:00.0000] <zcorpan> webben_: you can hack a relax ng schema, too [10:51:00.0000] <webben_> zcorpan: Oh absolutely, and I much prefer that. [10:52:00.0000] <webben_> The contrast is more with arbitrary code vs declarative rulesets. [10:52:01.0000] <webben_> so we're on route to a custom DTD, but the amount of effort it would require to customize the CSS validator is much higher [10:53:00.0000] <webben_> so there our immediate approach is to try to comment out bits of CSS we know won't validate [10:53:01.0000] <webben_> which is problematic in itself. [10:58:00.0000] <Philip`> webben_: Might it be easier to filter the output of the CSS validator, to remove the errors you don't care about? [10:59:00.0000] <webben_> Philip`: There are some problems with that. [10:59:01.0000] <webben_> some errors will cause the validator to stop parsing, or (iirc) parse differently than the parsing specified in CSS2.1 [10:59:02.0000] <Philip`> Oh, that doesn't sound good [11:00:00.0000] <webben_> other errors that would work for - e.g. I think that would work for unrecognized properties. [11:00:01.0000] <atwilson> Hixie: Is it our intent to eventually expose WebSockets to workers? [11:04:00.0000] <webben_> Philip`: contrast the output for a{height:50px;min-height:50px;}a{color:red;} and a{height:50px;min-height:50px;}a{color:red;} [11:05:00.0000] <webben_> Philip`: in the first case, W3C validator implies there's no "Valid CSS information" [11:05:01.0000] <webben_> it recovers from a{foo:50px;min-height:50px;}a{color:red;} rather differently [11:11:00.0000] <jgraham> brale: Yes that is possible [11:11:01.0000] <jgraham> How easy it is depends on exactly what you want [11:13:00.0000] <jgraham> e.g. import html5lib;import urllib2;from lxml import etree;print etree.tostring(html5lib.parse(urllib2.urlopen(url)), method="text") [11:13:01.0000] <jgraham> er [11:13:02.0000] <jgraham> import html5lib;import urllib2;from lxml import etree;print etree.tostring(html5lib.parse(urllib2.urlopen(url), treebuilder="lxml"), method="text") [11:14:00.0000] <jgraham> /me guesses he made a dozen other mistakes too [11:20:00.0000] <brale> jgraham: thanks. that was something I was thinking about (only i would use text_content). The thing is that I need only text that would actually be shown on the screen [11:21:00.0000] <brale> jgraham: so I am hoping that there is a more elegant solution to my problem. The main problem with lxml is that i will get text that's inside <script></script>, etc. [11:23:00.0000] <jgraham> brale: Yeah you would need to blacklist some elements. If you also care about the style rules that are in effect hings become much more complex [11:24:00.0000] <zcorpan> brale: hmm, <script> is something that http://lists.w3.org/Archives/Public/public-html/2009Jun/0682.html maybe should deal with too [11:24:01.0000] <brale> zcorpan: i think it would be smart [11:25:00.0000] <zcorpan> brale: what's your use case? [11:25:01.0000] <brale> jgraham: I don't care about style, I'm only interested in text [11:26:00.0000] <brale> zcorpan: text-mining, statistical models [11:27:00.0000] <zcorpan> brale: other than <img>, <script> and <style>, do you know which elements should be special-cased? [11:29:00.0000] <brale> zcorpan: <video> and some other html5 tags [11:29:01.0000] <zcorpan> how would you special case <video>? [11:30:00.0000] <brale> zcorpan: well I dont need anything thats inside [11:30:01.0000] <brale> as it would not give my any text on the screen [11:32:00.0000] <zcorpan> brale: would you want the text in <img alt="...">? [11:32:01.0000] <brale> zcorpan: I suppose not [11:34:00.0000] <brale> maybe I should stop searching and start coding with html5lib/lxml and deal with problems as they come :D [11:41:00.0000] <zcorpan> smylers++ [11:42:00.0000] <brale> zcorpan, jgraham: thanks for help [11:48:00.0000] <jgraham> smylers keeps posting good stuff today. I hope it has more success than my attempt to explain why there might be disagreement [11:50:00.0000] <jgraham> Which seemed to fail. I'm not sure if it's because a) it was actually not that insightful b) it was not as well reasoned as I had hoped c) everything I say is regarded as tainted or d) other [11:53:00.0000] <takkaria> I suspect it's one of them [11:55:00.0000] <Philip`> jgraham: You could empirically test option c by adopting a fake name when posting to the list, so nobody will realise it's really you [12:13:00.0000] <brale> jgraham: do you now anything about lxml/html5lib error "Invalid attribute name u'xml:lang'" ? [12:17:00.0000] <zcorpan> brale: is it an lxml error or html5lib error? [12:24:00.0000] <brale> zcorpan, jgraham: Now I'me using html5lib from mercurial and it seems to work :D [12:38:00.0000] <brale> jgraham, zcorpan: as I see there is a small bug in current mercurial version in html5lib/treewalkers/_base.py [12:38:01.0000] <brale> jgraham, zcorpan: in function endTag, missing "," [12:39:00.0000] <jgraham> brale: The treewalkers stuff on head is probably slightly broken [12:39:01.0000] <zcorpan> jgraham: is the foreign branch still more up to date? [12:40:00.0000] <jgraham> zcorpan: No they have been merged [12:40:01.0000] <zcorpan> ok [12:40:02.0000] <jgraham> brale: I will check if I have any local patches that I should push out later but I am going out for a while now [12:42:00.0000] <brale> jgraham: ok. ( missing "," isn't that big ;) ) [12:42:01.0000] <jgraham> brale: Patches welcome :) [12:43:00.0000] <brale> jgraham: will html5lib.HTMLParser be fine for using? [13:03:00.0000] <Dashiva> /me wonders if debate about accessibility is much different from debate about religion [13:07:00.0000] <Philip`> Dashiva: It is, because accessibility has some underlying quantitative facts (like how many people can read how many sites), and the debate might lead to a better understanding of those facts [13:16:00.0000] <Dashiva> Philip`: This is true. But religion also has some underlying facts (history, written works, etc) that could be argued objectively. The problem is the people who turn every discussion into rhetorics. [13:19:00.0000] <tantek> Dashiva, I think the bigger problem is people who say "this is the way it is because I say so and I'm a [self-proclaimed] expert" - in other words, people that ask you to accept their assertions on faith. [15:07:00.0000] <Hixie> atwilson: yes [15:07:01.0000] <atwilson> OK, thx. [15:24:00.0000] <zcorpan> marginheight="" seems to be a way for pages to do cross-document communication, at least in one direction [15:25:00.0000] <zcorpan> cross-origin even [15:26:00.0000] <Hixie> ? [15:26:01.0000] <Hixie> oh, yeah, i guess [15:26:02.0000] <Hixie> heh [15:36:00.0000] <zcorpan> woot, <pre width> is not in the rendering section [15:38:00.0000] <zcorpan> though it seems only opera and firefox support it [15:47:00.0000] <Hixie> for those of you who were worried about paying for tpac using paypal, btw, apparently they will also accept checks made payable to MIT [16:29:00.0000] <zcorpan> http://wiki.whatwg.org/wiki/Presentational_elements_and_attributes [16:33:00.0000] <zcorpan> /me realizes that he uses border="1" on that page, but it seemed that class="wikitable" did not result in anything 2009-06-25 [17:47:00.0000] <zcorpan> http://zope.cetis.ac.uk/members/scott/blogview?entry=20090624222327 [18:13:00.0000] <Dashiva> So about those summaries... has anyone extracted a list of _useful_ summaries from that list Philip` made? [18:14:00.0000] <zcorpan> Dashiva: what's a useful summary? [18:15:00.0000] <Dashiva> There was that argument "Most of those summaries would never be used since they're on layout tables" [18:16:00.0000] <Dashiva> But that doesn't imply that the summaries that were on data tables are good. So I'm wondering if anyone has actually found some examples. [18:17:00.0000] <zcorpan> where's the data again? [18:17:01.0000] <Dashiva> http://philip.html5.org/data/table-summary-values-dotbot.html [18:17:02.0000] <Dashiva> (It seems reasonable to assume that a good, descriptive summary would not come from a layout table) [18:22:00.0000] <zcorpan> A table of four, single-cell rows. Cell one contains the trail to this section from the document home. Cell two contains a navigation bar to the pages within the current section of the overall document. Cell three contains the body text of the document. Cell four contains navigation to reference information for this web site www.mf-ct.upc.es/bathfreesim_doc/bfs_pra17.html [18:22:01.0000] <zcorpan> that's the first i found [18:23:00.0000] <zcorpan> not clear to me whether it's a layout table or data table [18:23:01.0000] <zcorpan> but it's describing the structure of the table, which is the stated purpose of summary="" [18:23:02.0000] <Dashiva> That actually looks like a layout table, yeah. [18:25:00.0000] <zcorpan> the actual data table on that page does not* have a summary [18:25:01.0000] <zcorpan> while all layout tables have elaborate summaries [18:26:00.0000] <zcorpan> (the data table on that page seems simple enough to not needing a summary, though) [18:28:00.0000] <zcorpan> "A table with two columns listing title, author, date, source, subject headings and comments for selected newspaper articles" [18:28:01.0000] <zcorpan> http://www.santacruzpl.org/history/clippingfile/cliplist.php?page=3&subjhead2=621&newspaper=&datelimit= [18:29:00.0000] <Dashiva> By the way, I was only asking if such a list existed. You don't have to make it for my sake :) [18:31:00.0000] <zcorpan> i'm not making a list, i'm just having a peek [18:45:00.0000] <zcorpan> actually, that was the only occurrence i found [18:46:00.0000] <zcorpan> obviously i skipped any that were in languages i don't understand [18:47:00.0000] <heycam> Hixie, I think some of the interfaces that were added to the obsolete section need to inherit from HTMLElement [18:47:01.0000] <zcorpan> that table is also simple enough to not actually need a summary [02:16:00.0000] <zcorpan> http://code.google.com/intl/es/speed/articles/optimizing-html.html [02:47:00.0000] <othermaciej> zcorpan: I wonder how effective those kinds of changes are post-gzip [02:48:00.0000] <zcorpan> othermaciej: might have no effect at all [02:49:00.0000] <othermaciej> that would be my gues [02:50:00.0000] <othermaciej> I also can't make much sense of this article: http://code.google.com/intl/es/speed/articles/reflow.html [02:50:01.0000] <Rik`> there's a lot of small perf improvements for big trouble reading code afterwards in those google articles [02:52:00.0000] <othermaciej> what exactly does it think "reflow" is? [02:52:01.0000] <Philip`> Adding </li>s and </p>s (with any necessary whitespace to http://www.google.com/intl/en/privacy.html makes the file 251 bytes larger, and makes the gzipped file 7 bytes larger [02:52:02.0000] <othermaciej> if it's a relayout, how can they say padding or width changes won't cause one in some browsers? [02:52:03.0000] <Philip`> s/whitespace/whitesspace)/ [02:53:00.0000] <othermaciej> or that changing the background will cause one in Firefox? (Gecko can't possibly be that stupid) [02:53:01.0000] <Philip`> so it's not no effect at all [02:53:02.0000] <Philip`> s/whitesspace/whitespace/ [02:54:00.0000] <Rik`> even if it has effects, the research done by the Bing team claims it has no effect on users [02:54:01.0000] <Rik`> http://radar.oreilly.com/2009/06/bing-and-google-agree-slow-pag.html last paragraph [02:54:02.0000] <jgraham> Philip`: How does it compare to removing all whitespace bwetween elements? I guess that saves more bytes but feer gzip bytes [02:54:03.0000] <jgraham> fewer [02:56:00.0000] <Philip`> jgraham: Removing whitespace from privacy.html seems to save 696 bytes, or 76 bytes in the gzipped version [02:58:00.0000] <zcorpan> maybe they should have focused on gzip and caching etc, and skip the minor optimizations that affect workflow and maintainability [02:59:00.0000] <zcorpan> the organization of css files is another thing that has close to no effect once gzipped [02:59:01.0000] <Philip`> But micro-optimisations are more fun :-( [02:59:02.0000] <Rik`> look at php otimizations, they are even more fun [03:00:00.0000] <othermaciej> what they should have done is told you that you how to measure the results of optimizations and not bother with ones that don't help [03:00:01.0000] <zcorpan> Rik`: like use single quotes instead of double quotes when the string has no variables? [03:00:02.0000] <Rik`> zcorpan: that and "don't copy variables !" [03:00:03.0000] <Philip`> "Sometimes PHP novices attempt to make their code "cleaner" by copying predefined variables to variables with shorter names." - stupid novices, trying to make their code cleaner [03:01:00.0000] <Rik`> I don't wanna see internal google code [03:01:01.0000] <hsivonen> /me thinks jgraham's points were good but the "engineering" and "HCI" labels weren't quite apt given than people with an engineering background have been in the "HCI" mindset and vice versa [03:01:02.0000] <othermaciej> or better yet, how to tell where your page load timte is going so you can tell if it is even worth the effort to work on size of your HTML document, compared to other things you could do [03:02:00.0000] <othermaciej> hsivonen: I think "scientist" vs "empath" was in some ways a more apt comparison, though it highlights a different kind of divide [03:02:01.0000] <Philip`> The Page Speed tool is useful, since it does actual measurements and lets you see what's slow in practice [03:02:02.0000] <othermaciej> i.e. valuing measurable results vs. wanting to signal that you care [03:04:00.0000] <Hixie> i was trying to explain the summary="" thing to a non-tech person today, trying to make it sound like the pro-summary="" position was the one i was supporting [03:04:01.0000] <Hixie> and as soon as i mentioned the data we collected, they basically were like "oh well then you should come up with something better" [03:05:00.0000] <Hixie> made me feel better that we're not the only ones who come to that conclusion [03:05:01.0000] <jgraham> hsivonen: Yeah, I wondered about that. But the labels seem to reflect reality quite well even if they don't reflect preconceptions about how reality should be [03:06:00.0000] <jgraham> Maybe labels in general are bad. I could have just gone with "approach 1" and "approach 2" [03:07:00.0000] <jgraham> But I really wanted to get across the fact that language design is a HCI problem [03:07:01.0000] <Hixie> jgraham: the problem is that labelling the two "camps", or explaining the assumptions or approaches those "camps" are taking, doesn't help resolve the issue [03:08:00.0000] <hsivonen> aargh. I've managed to break the parser so that it puts <html> inside <body> [03:08:01.0000] <Hixie> which is that fundamentally, the pro-summary="" advocates have, as far as i can tell, refused to acknowledge that the data we've collected shows that summary="" hasn't worked [03:08:02.0000] <hsivonen> also, I think the data pretty well debunks the notions that [03:09:00.0000] <hsivonen> 1) summary is long [03:09:01.0000] <hsivonen> 2) summary is different from caption [03:09:02.0000] <jgraham> Hixie: The idea was to sho that reasonable people with different appraoches could come to different conclusions [03:09:03.0000] <jgraham> +w [03:09:04.0000] <Philip`> http://code.google.com/speed/articles/prefetching.html - "Profile your changes" - a revolutionary idea! Maybe the PHP article should talk more about profiling and less about changing double quotes to single quotes [03:09:05.0000] <Hixie> jgraham: sure [03:09:06.0000] <jgraham> Hixie: A great deal of friction has arisen because people have assumed that anyone with a different opinion is being unreasonable [03:10:00.0000] <Hixie> jgraham: i'm not sure that anyone who was assuming that before, if there are any, will think any differently now [03:10:01.0000] <jgraham> Hixie: Yeah. [03:11:00.0000] <jgraham> I guess it was a waste of time. Hopefully it wasn't also harmful [03:12:00.0000] <Philip`> It could still help convince people who aren't firmly on any side of the discussion [03:13:00.0000] <Hixie> fundamentally, if two groups have contradictory goals (in this case "make it possible to write fully-accessible pages even if the overall quality of pages is lower than with other approaches" and "improve the overall quality of pages so that more are accessible even if the theoretical maximum accessibility is lower than with other approaches"), they won't come to an agreement [03:13:01.0000] <jgraham> Hixie: To be fair there has been some acknoledgement that it hasn't worked. It is just that they consider that it is important to have it even though it hasn't worked in general; either because they think it might work well in the future or because they think experts can use it and only experts matter [03:13:02.0000] <Hixie> jgraham: that's what i meant by "hasn't worked" [03:14:00.0000] <Hixie> jgraham: ("would only work if used by experts") [03:14:01.0000] <jgraham> Hixie: Right but reasonable people could disagree with that definition of "hasn't worked" [03:14:02.0000] <Hixie> sure [03:15:00.0000] <Hixie> anyway i don't see how i can reconcile the two goals here [03:15:01.0000] <Hixie> they seem to be contradictory [03:15:02.0000] <jgraham> Particularly if those "reasonable people" consider themselves to be the sort of experts who will use it [03:17:00.0000] <jgraham> Hixie: Allow flow content in <caption>, promote <caption><details> as a solution for cases where an extended summary is needed. It's not sure to work but it is a different approach that might be worth trying; soesn't seem to have potential for harm and might resolve the deadlock :) [03:17:01.0000] <Hixie> <caption><details> really isn't what i would recommend imho [03:17:02.0000] <Hixie> a disclosure triangle in a table caption would be really silly-looking [03:18:00.0000] <Hixie> and it's not clear to me that it is necessary anyway [03:19:00.0000] <Philip`> jgraham: Wouldn't any solution that graphically hides some text by default be subject to the same harm of people putting useless/incorrect/unreadable strings in it? [03:19:01.0000] <jgraham> Yeah, I worry about how it would look. It would be good if it could be easilly styled [03:19:02.0000] <Hixie> i think <details> would be far less spammed than summary="" [03:19:03.0000] <tantek> /me wonders how many hours of how many individuals' time has been spent on the summary attribute debate, time that could have perhaps been directed at designing and user-testing alternative solutions with an eye towards developing something that user-tests better with typical (or even "modern") HTML authors. [03:19:04.0000] <jgraham> Hixie: In general not allowing flow content in <caption> makes no sense to me [03:20:00.0000] <jgraham> Why can't I have multi-paragraph captions [03:20:01.0000] <jgraham> ? [03:20:02.0000] <Hixie> jgraham: sure, but that's another issue [03:21:00.0000] <Hixie> assuming you've sent (non-summary-related) mail on it i'm sure i'll get around to it and fix that relatively soon [03:21:01.0000] <zcorpan> jgraham: i think the idea was to be consistent with other heading-like elements like h1, legend, dt [03:21:02.0000] <jgraham> Hixie: I'm not sure if I did [03:21:03.0000] <Hixie> jgraham: ah well then file a bug or something [03:21:04.0000] <zcorpan> jgraham: although it seems reasonable to be able to put two paragraphs or a list in a caption [03:22:00.0000] <jgraham> But anyway once you get flow content you get <details> for free so it doesn't seem unreasonable to suggest that for the cases where a complex table needs an extended summary [03:23:00.0000] <Hixie> if i'm going to suggestion something for complex tables that are so complex they need an extended summary, what i would suggest is "make your table simpler" [03:23:01.0000] <Hixie> not "hide your explanation so that people have to trigger a button to see it" [03:23:02.0000] <zcorpan> jgraham: maybe <legend> should allow blocks too in figures [03:24:00.0000] <Hixie> (are multiparagraph captions something one sees in print at all?) [03:25:00.0000] <jgraham> Hixie: right, that would be my first point of advice too. But pragmatically there are some cases where you might be constrained to not do that [03:25:01.0000] <Hixie> i don't know that i'd ever tell people to use <details> even then [03:25:02.0000] <Hixie> but that's mostly academic imho [03:26:00.0000] <Hixie> the question is how do we proceed on this issue given that as far as i can tell, summary="" advocates won't accept anything short of including summary="" [03:27:00.0000] <Hixie> and given that the chairs are apparently unwilling to make any sort of statements regarding a decision [03:27:01.0000] <jgraham> Hixie: I wwould be interested to see if they would accept using <details>. It seems to be basically equivalent to using a <summary> element which they have advocated as a long term solution [03:27:02.0000] <annevk2> we could take the path of least resistance as shelley thinks we already do, but I'm not sure that'd be good [03:27:03.0000] <Hixie> annevk2: what's the path of least resistance? [03:27:04.0000] <Hixie> jgraham: have they not responded to those proposals? [03:28:00.0000] <jgraham> Hixie: Who is "they"? [03:28:01.0000] <jgraham> Oh I see what you mean [03:28:02.0000] <Hixie> jgraham: have summary="" advocates and <summary> advocates not commented on the <details> proposals? [03:29:00.0000] <jgraham> IIRC the response was more or less "we like the name 'summary' better" [03:29:01.0000] <Hixie> so they're ok with it? [03:29:02.0000] <Hixie> wow, i somehow missed that [03:30:00.0000] <jgraham> Hixie: I think they didn't say that :) [03:30:01.0000] <Hixie> what did they say then? [03:30:02.0000] <zcorpan> i don't understand the <summary> proposal [03:30:03.0000] <jgraham> AFAICT proposals are only taken seriously when you or Sam expresses an opinion on them [03:30:04.0000] <annevk2> Hixie, include it? [03:30:05.0000] <zcorpan> would <summary> be visible or not? [03:31:00.0000] <Hixie> annevk2: that's not the path of least resistance; i, for one, would object strongly to that solution as it has been shown to actually harm users [03:31:01.0000] <jgraham> zcorpan: Leif had a somewhat more coherent version of the proposal onthe list [03:31:02.0000] <zcorpan> <summary> would have to go inside the <caption> for legacy parsing reasons [03:32:00.0000] <jgraham> I think <summary> would have to work like <table><caption>My shory summary <summary>My description of the table structure</summary></caption> [03:32:01.0000] <jgraham> and <sumamry> would be hidden by default and get a disclosure triangle [03:33:00.0000] <Hixie> i think that'd be exactly as bad as <details> [03:33:01.0000] <jgraham> (btw I find it counterintuituve that <details> uses <legend> around the bit that is visible by default rather than the bit that is hidden) [03:34:00.0000] <Hixie> <details> is like <fieldset> except the contents are hidden [03:34:01.0000] <Hixie> anyway i should sleep [03:34:02.0000] <Hixie> i guess i wish the chairs would at least give some guidance on how to proceed [03:34:03.0000] <jgraham> Hixie: Agreed. The technical difference between the proposals is minimal. The differences are mainly about whether it is seen to be Designed for the Blind (tm) [03:35:00.0000] <Hixie> (as i have asked for several times so far, with no response) [03:35:01.0000] <Hixie> nn [03:35:02.0000] <jgraham> gn [03:37:00.0000] <jgraham> (Actually the original proposal has <summary> as a child of <table> which is a technical difference but has really bad legcay behaviour so I was kind of overlooking it) [03:45:00.0000] <hsivonen> <summary> is unworkable for legacy reasons. There's really no point in even discussing it. [03:45:01.0000] <hsivonen> if anyone wants a child element of <table>, it has to be called <caption> no matter what its semantics are defined to be [03:45:02.0000] <jgraham> hsivonen: Which legacy reasons? [03:46:00.0000] <hsivonen> jgraham: fosterparenting [03:46:01.0000] <jgraham> hsivonen: Ah OK. I wondered if there is something else I overlooked [03:46:02.0000] <jgraham> As I said, it would have to be a child of <caption> [03:47:00.0000] <Lachy> there's no reason the summary needs to be a child of the table. All that's needed is some kind of association between the summary and the table, and there are many ways to do that [03:48:00.0000] <jgraham> Lachy: I can think of three (caption, figure and aria-describedby) which is not really "many" [03:49:00.0000] <hsivonen> jgraham: <summary> as sibling of <table> (quite unconventional, yes) [03:49:01.0000] <zcorpan> putting it as child of caption has the advantage that the association works in legacy AT [03:51:00.0000] <Philip`> hsivonen: Before or after the <table>? Ignoring whitespace? Ignoring comments? [03:52:00.0000] <jgraham> Philip`: sibling is quite eel defined, no? [03:52:01.0000] <jgraham> s/eel/well/ [03:52:02.0000] <Philip`> Oh, I'm probably misinterpreting it as "adjacent sibling" [03:52:03.0000] <Philip`> You could write <table><summary>... and then it'd be a sibling of the table :-) [03:52:04.0000] <Lachy> I'm not convinced of the need to have such explicit assocation. Just summarise the table in the preceding paragraph [03:53:00.0000] <Lachy> so <p>This is a summary of the following table...</p><table>...</table> [03:54:00.0000] <jgraham> Lachy: If people who use AT argue it is helpful for them to have such an association it is quite hard to argue it is not helpful (I'm not sure if people actually did argue this for table summaries) [03:54:01.0000] <jgraham> (but it seems plausible at least because you can e.g. reread the summary eailly if you get lost in the middle of the table) [03:55:00.0000] <Philip`> Sibling would be bad because people would write <!doctype html><p>Blah blah <summary>Blah</summary> <table> and then it wouldn't actually be a sibling [03:57:00.0000] <Lachy> I'm sure they did. That doesn't mean they're right. I haven't seen any usability studies demonstrating how effective or ineffective a summary would be if it doesn't have an explicit association in the markup [03:59:00.0000] <Philip`> In the absence of any usability studies (because presumably the WG doesn't care enough about accessibility to actually do any), the best information we have about user behaviour is what users claim [04:05:00.0000] <Lachy> Philip`, sure, but anyone in the field of usability knows you should avoid giving too much weight to what the user says, as their behaviour can often completely contradict that [04:07:00.0000] <Philip`> Lachy: Indeed, but hypothesising about other people's behaviour is likely to be even less accurate, and we don't currently have any other sources of information [04:08:00.0000] <Lachy> I havent' hypothesised about it. I just said I wasn't convinced of the need for explicit association [04:09:00.0000] <Lachy> and in the interst of finding the simplest solution possible, I think simply including the summary in the surrounding prose may be sufficient [04:10:00.0000] <Philip`> You're hypothesising that it's sufficient for users [04:10:01.0000] <zcorpan> http://simon.html5.org/articles/mobile-results first used summary="", then moved it to a <p> before the table, and now includes it in the <caption> [04:10:02.0000] <Lachy> well, in that case, you're hypothesising that it isn't, and neither one of us have any data to test it [04:11:00.0000] <Philip`> Lachy: I'm not hypothesising anything myself [04:12:00.0000] <Lachy> you're hypothesising just as much as I am, which is none. But it is you that claimed I was [04:12:01.0000] <Philip`> You said "I think [stuff] may be sufficient" which sounds like a hypothesis to me :-) [04:13:00.0000] <jgraham> If you knew whether users were right about their own needs more often than they were wrong then you could use that to pick the more likely hypothesis [04:14:00.0000] <jgraham> /me notes that ouldn't actually work for all sorts of reasons [04:14:01.0000] <jgraham> wouldn't [04:16:00.0000] <zcorpan> Lachy: i think bug 7035 is questioning the use case of the content model concept, not the diagram itself [04:21:00.0000] <Philip`> On the subject of page speed: if you use mod_deflate in Apache 2.2, then it breaks caching (by generating invalid ETags) [04:21:01.0000] <Philip`> I would have hoped this kind of thing would be more of a solved problem by now [04:26:00.0000] <Dashiva> Oh look, antoher reference to the bible on public-html [04:26:01.0000] <hsivonen> Philip`: before, ignoring comments and whitespace [04:27:00.0000] <Lachy> zcorpan, it clearly said: [04:27:01.0000] <Lachy> <sayrer> [04:27:02.0000] <Lachy> http://www.whatwg.org/specs/web-apps/current-work/images/content-venn.svg [04:27:03.0000] <Lachy> <sayrer> why does that exist? [04:28:00.0000] <Lachy> and the reason content models exist should be fairly obvious to anyone [04:30:00.0000] <jgraham> Lachy: sayrer is a HTML anarchist. He doesn't believe in any sort of central authority over what is allowed in a document [04:30:01.0000] <MikeSmith> I think Rob may be questioning whether the use cases for implementors needing to care anything about the content classes that the spec defines -- interactive, sectioning, etc. [04:31:00.0000] <Lachy> MikeSmith, conformance checkers do. Browser's probably don't [04:32:00.0000] <jgraham> MikeSmith: I failed to parse that sentence [04:32:01.0000] <jgraham> +exist maybe? [04:32:02.0000] <MikeSmith> jgraham: remove "whether", I think [04:32:03.0000] <MikeSmith> anyway, it's fun to speculate about Rob thinks [04:32:04.0000] <MikeSmith> we should play this game more often [04:32:05.0000] <MikeSmith> for other stuff [04:33:00.0000] <Lachy> MikeSmith, but it was you who filed the bug. Why couldn't Rob do it? [04:33:01.0000] <MikeSmith> like, instead of What Would Jesus Do?, we could ask, What Would Rob Think? [04:33:02.0000] <MikeSmith> Lachy: I filed that bug solely for humor purposes. [04:33:03.0000] <Lachy> ok [04:36:00.0000] <takkaria> so the bugtracker is the place to explore possibility for entertainment value but the mailing list isn't? :) [04:36:01.0000] <beowulf> takkaria: irc is no place for humour [04:37:00.0000] <MikeSmith> takkaria: well, I already got dressed down by a range of partisans for using the bugzilla for entertainment purposes [04:37:01.0000] <MikeSmith> so I suppose I won't do it again [04:37:02.0000] <beowulf> basg.org is not acceptable evidence as bash.org is not an expert on irc [04:37:03.0000] <MikeSmith> unless/until I forget the lesson I learned from this [04:37:04.0000] <beowulf> s/basg/bash [04:38:00.0000] <MikeSmith> which should take at least 2 weeks for me to forget [04:38:01.0000] <MikeSmith> anyway, in other news, fwiw to anybody on here who might be running their own instance of validator.nu -- I just checked in an new "--local" option that lets you run v.nu without it needing to grab the IANA language registry and Wiki stuff over the network remotely [04:39:00.0000] <zcorpan> speaking of wiki and v.nu... [04:39:01.0000] <MikeSmith> so it will help you to say, prevent yourself from looking likely an total asshead (as I did) when you go somewhere and try to do a v.nu demo and find you have no network connection [04:39:02.0000] <zcorpan> MikeSmith: do you think it would be useful for v.nu to point to http://wiki.whatwg.org/wiki/Presentational_elements_and_attributes when complaining about certain obsolete elements? [04:40:00.0000] <MikeSmith> zcorpan: yeah, that could be added to the assertions-checking code, I suppose [04:41:00.0000] <MikeSmith> like I said, I think there's a lot more that could be done yet with putting out advice for obsolete elements and other stuff [04:41:01.0000] <MikeSmith> zcorpan: there was something I was going to ask you about earlier today, but I forgot [04:43:00.0000] <zcorpan> MikeSmith: you should ask it even if i'm not here, so that you can point me to the logs later :) [04:44:00.0000] <MikeSmith> true that [04:48:00.0000] <MikeSmith> speaking of lessons learned, I got this amateur schoolmarm who wants to learn me about how to look smarter in my blog postings [04:48:01.0000] <MikeSmith> http://sideshowbarker.net/2009/06/22/webkit-sniffer/#comments [04:48:02.0000] <MikeSmith> he mistakes me for somebody who cares about appearing to be articulate [04:49:00.0000] <Dashiva> There's nothing wrong with Realpolitik [04:50:00.0000] <MikeSmith> Dashiva: I spelled it wrong, apparently [04:50:01.0000] <MikeSmith> Realpolitick [04:50:02.0000] <MikeSmith> with a "ck" [04:50:03.0000] <takkaria> I noticed that when I read it but cba pointing it out [04:50:04.0000] <MikeSmith> I call that the "tribute to 17th-century spelling conventions" spelling [04:51:00.0000] <MikeSmith> anyway, the few blog postings I manage to get around doing are pretty much the equivalent of scribblings on cocktail napkins [04:51:01.0000] <MikeSmith> or on toilet paper [04:52:00.0000] <MikeSmith> anybody who expects me to spell words in them right is going to be disappointed [04:53:00.0000] <takkaria> since befriending an extremely dyslexic person and talking to them online a fair bit, my tolerance for spelling errors has gone up a lot [04:53:01.0000] <Dashiva> MikeSmith: I'm sure those international readers aren't steady on the spelling anyhow [04:54:00.0000] <takkaria> I'm almost slightly annoyed that I'm not annoyed at people not spelling things anymore [04:54:01.0000] <takkaria> s/things/things right/ [04:55:00.0000] <MikeSmith> I think I will reply to his comment with, "Thanks for you pedantrie." [04:55:01.0000] <MikeSmith> and then he can correct me on my misspelling of "pedantrie" [04:56:00.0000] <MikeSmith> that would be great [04:56:01.0000] <Dashiva> If you took each word and randomly jumbled the letters, would that count as 'not spelling' [05:03:00.0000] <hsivonen> /me agrees with zcorpan about hasFeature and "2.0" [05:03:01.0000] <hsivonen> hasFeature is pointless already. Better just freeze it. [05:06:00.0000] <Dashiva> Would it be needed at all if Java had proper capability detection? [05:06:01.0000] <hsivonen> Dashiva: no [05:06:02.0000] <hsivonen> Java has reflection :-) [05:07:00.0000] <Dashiva> Been there, done that. :) [05:07:01.0000] <hsivonen> I wonder how often people actually write non-mobile Java apps that use hasFeature [05:07:02.0000] <hsivonen> one would think that first you target JDK 1.4.2 [05:07:03.0000] <hsivonen> then when you upgrade to targeting JDK 5, you just stop running on 1.4.2 [05:08:00.0000] <hsivonen> and when you upgrade to JDK 6, you stop running on JDK 5 [05:08:01.0000] <hsivonen> and when your aren't ready to increment the platform requirement, you just keep targeting the older JDK [05:09:00.0000] <hsivonen> and if you are using a non-Sun JDK that opts to ship a non-Xerces JAXP back end, you are in the world of hurt anyway [06:05:00.0000] <Lachy> LOL, this is what not to do with your ssh private key :-) http://lists.w3.org/Archives/Public/public-webapps/2009AprJun/1241.html [06:23:00.0000] <Philip`> Lachy: Old news :-p [06:23:01.0000] <Philip`> (See #html-wg) [06:27:00.0000] <Dashiva> Lachy got Warnocked [06:30:00.0000] <hsivonen> does Opera support <script defer>? what about IE? [06:30:01.0000] <Philip`> Dashiva: That seems like an entirely novel use of the term "Warnocked" [06:31:00.0000] <Philip`> (unless I'm misunderstanding the context) [06:31:01.0000] <Dashiva> Philip`: If someone had replied to the email, he would know it's an "old" topic. [06:31:02.0000] <Dashiva> But without replies, maybe he was the first to notice? [06:32:00.0000] <Lachy> Dashiva, how did I get warnocked? what does that mean? [06:34:00.0000] <Dashiva> Well, maybe more like second order warnocking. [06:34:01.0000] <Dashiva> http://en.wikipedia.org/wiki/Warnock's_Dilemma [07:14:00.0000] <zcorpan> hsivonen: opera no, ie yes [07:21:00.0000] <hsivonen> zcorpan: ok. thanks [07:34:00.0000] <MikeSmith> annevk2: while you are in Denmark, if you happen to stop by a CD shop, please buy me "I've Got You on Tape" [09:45:00.0000] <gsnedders> jgraham, zcorpan: Would either of you happen to have a bed sheet I could borrow tomorrow? :P [09:46:00.0000] <jgraham> gsnedders: Sure [09:46:01.0000] <gsnedders> jgraham: thanks [09:46:02.0000] <jgraham> gsnedders: Mail, please [09:47:00.0000] <gsnedders> jgraham: Mail for what? Can I just pick it up at the office seeming I need to get flat key from there? [09:47:01.0000] <gsnedders> jgraham: Can you make do with getting it back when I next see you (which I guess should be no later than Monday) [09:47:02.0000] <jgraham> gsnedders: For the purpose of me remembering to look for it [09:47:03.0000] <gsnedders> jgraham: Oh, email. duh. [09:48:00.0000] <gsnedders> jgraham: I can't promise it will manage to send on this wifi. [09:48:01.0000] <gsnedders> :( [09:48:02.0000] <jgraham> Actually, come to think of it, I wonder if we have two bed sheets spare since we have visitors... I guess it is not impossible [09:48:03.0000] <gsnedders> jgraham: cam or opera email? [09:48:04.0000] <jgraham> gsnedders: opera or hopipolla [09:48:05.0000] <gsnedders> jgraham: I only need one bed sheet! :P [09:48:06.0000] <jgraham> hoppipolla [09:52:00.0000] <gsnedders> jgraham: sent [09:52:01.0000] <gsnedders> (hopefully) [09:52:02.0000] <gsnedders> Still sending… [09:53:00.0000] <gsnedders> Should have sent [13:23:00.0000] <spithash> hey niggers [13:53:00.0000] <Hixie> http://twitter.com/itpastorn/statuses/2329152650 [13:53:01.0000] <Hixie> when two conspiracy theories conflict... [13:55:00.0000] <Dashiva> Nah [13:56:00.0000] <Dashiva> Google doesn't actually like RDFa. Their implementation is intended to subvert existing support, while the spec work ensures it doesn't get any more popularity! [14:26:00.0000] <takkaria> i heard rumours that Sergei actually sits by Hixie's desk and tells him what to specify when he can't make up his mind [14:27:00.0000] <Hixie> actually sergey has never made a suggestion for the spec as far as i recall [14:27:01.0000] <Hixie> the acknowledgements section does list other people who have, though, whose names may be unexpected [14:29:00.0000] <takkaria> you would say that [14:29:01.0000] <takkaria> it must be hard to edit independently with that corporate agenda pushing all your decisions [14:29:02.0000] <takkaria> and the need to keep it quiet so conspiracy theories won't form [14:30:00.0000] <inimino> wow, Rasputin and the Bilderbergs, that /was/ unexpected [14:38:00.0000] <Dashiva> Maybe Hixie doesn't actually exist [14:43:00.0000] <takkaria> Hixie: with regards to disagreeing with stuff in the spec [14:44:00.0000] <takkaria> surely if the arguments for something's inclusion are sound, then your opinion is that that thing should be included? [14:45:00.0000] <beowulf> i read on twitter that hixie is in something called a gene pool [14:45:01.0000] <beowulf> i hope he has an inflatible dinosaur [14:46:00.0000] <Hixie> takkaria: i think the arguments for including the /> syntax are solid, but i still don't _want_ the /> syntax in text/html [14:46:01.0000] <Hixie> takkaria: if that makes sense [14:47:00.0000] <Hixie> takkaria: if i could find a way to get rid of style="", i'd want to do so [14:49:00.0000] <takkaria> hmm, OK [14:49:01.0000] <takkaria> the idealist in you hasn't quiet yet died then :) [14:51:00.0000] <Hixie> right [15:21:00.0000] <gsnedders> Woops. Do I have IRC open? I guess I'll drop offline with this terrible wifi soon [16:01:00.0000] <hober> I guess Michael Jackson missed his opportunity to report a typo in the spec & gain immortality via inclusion in the Acknowledgments section. [16:16:00.0000] <sayrer> jgraham: ftr, I am not an HTML anarchist [16:17:00.0000] <sayrer> but I don't like pointless rules either [16:17:01.0000] <sayrer> and I think "semantic markup" is not worth very much [16:18:00.0000] <sayrer> so disallowing <font> while allowing style="" seems ridiculous [16:19:00.0000] <sayrer> go ahead, just move that presentation stuff into attribute values rather than element and attribute names... then everything is ok! [16:19:01.0000] <sayrer> it's like that monty python skit where everyone changes chairs and says "there, that's much better" [16:39:00.0000] <beowulf> sayrer: pity, i liked the idea of someone being an HTML anarchist [16:41:00.0000] <sayrer> beowulf, prescriptivists have a tough time telling a descriptivist from an anarchist, so there HTML anarchists in the minds of some [16:45:00.0000] <beowulf> sayrer: what is descriptivism in the context of HTML? [16:46:00.0000] <sayrer> beowulf: that would mean writing down what mean, without judging whether they're doing it wrong [16:46:01.0000] <sayrer> what markup means [16:46:02.0000] <sayrer> so for instance, @valign is banned as wrong in Ian's draft [16:47:00.0000] <sayrer> descriptivist approach would be to write it down, and note that the same task can be accomplished other ways (synonyms) [16:49:00.0000] <sayrer> http://en.wikipedia.org/wiki/Ain%27t is a could example [16:49:01.0000] <sayrer> good example [16:49:02.0000] <sayrer> "It is a word that is widely used by many people, but its use is commonly considered to be improper.[1]" [16:50:00.0000] <beowulf> sayrer: ah, i get you [16:51:00.0000] <sayrer> A related concept is the virtue of "semantic markup" [16:51:01.0000] <sayrer> people who obsess about whether they are using <dt> right [16:51:02.0000] <sayrer> my opinion is that the section that most precisely defines the semantics of these elements is section 11, rendering 2009-06-26 [19:22:00.0000] <zcorpan> jgraham: when it comes to generating table descriptions, i think a good first step would be to include the number of rows and columns, and where the table headers are [19:27:00.0000] <zcorpan> "Table with 19 rows and 7 columns. The first two rows contain headers. The first row and the first column contain merged cells." [19:27:01.0000] <zcorpan> http://projectcerbera.com/web/study/2007/tables/clark2006/05-housing/original [19:29:00.0000] <zcorpan> hmm actually that table has merged cells on the first two rows and the first two columns [19:29:01.0000] <zcorpan> but i guess the top left cell should be ignored [19:30:00.0000] <zcorpan> or maybe empty cells should be ignored [19:31:00.0000] <zcorpan> maybe it's not so useful to say where there are merged cells [19:31:01.0000] <zcorpan> maybe just stating that it contains merged cells [19:35:00.0000] <zcorpan> /me looks at the forming a table algorithm in the spec [19:47:00.0000] <zcorpan> might also be good to announce when there are more than one tbody or when there's a tfoot [20:57:00.0000] <MikeSmith> zcorpan: you still around? [21:01:00.0000] <MikeSmith> sayrer: you got time to chat? [21:07:00.0000] <zcorpan> MikeSmith: yes [21:08:00.0000] <MikeSmith> zcorpan: do you have a local workspace for testing v.nu changes? [21:09:00.0000] <zcorpan> MikeSmith: local workspace? [21:10:00.0000] <MikeSmith> zcorpan: I mean, an svn workspace, so that you can sync up to latest svn revision and then re-run a local v.nu to test changes? [21:11:00.0000] <zcorpan> MikeSmith: ah. no [21:12:00.0000] <zcorpan> i checked out the source on my windows laptop but didn't get it to run [21:12:01.0000] <zcorpan> i probably don't have JDK on my mac [21:12:02.0000] <MikeSmith> ok, no problem [21:13:00.0000] <zcorpan> is there something you'd like me to test? [21:13:01.0000] <MikeSmith> I'm adding frameset to the actual schema [21:13:02.0000] <MikeSmith> and wanted to make sure I've not muffed it up [21:14:00.0000] <MikeSmith> I have my own simple tests for it, but wanted to double-check [21:16:00.0000] <zcorpan> ping me when it's live in v.nu and i can try to break it [21:16:01.0000] <MikeSmith> OK [21:20:00.0000] <MikeSmith> zcorpan: anyway, the only difference my change will introduce is that when you check a document that has <frameset>, the "Error: Element frameset not allowed as child of element html in this context. (Suppressing further errors from this subtree.)" message will be suppressed [21:21:00.0000] <MikeSmith> oh, and it will actually report any errors in the subtree [21:22:00.0000] <zcorpan> MikeSmith: will it complain about required children missing (i.e. the <body>)? [21:25:00.0000] <zcorpan> /me guesses not [21:28:00.0000] <MikeSmith> zcorpan: you guess right, it won't [21:28:01.0000] <zcorpan> MikeSmith: have you tested noframes before and after the frameset? [21:29:00.0000] <MikeSmith> zcorpan: nope [21:30:00.0000] <MikeSmith> but if you give a test, I will add it and test it before I check in [21:31:00.0000] <zcorpan> html4 only allows noframes inside frameset, but the html5 parser inserts noframes to head or html for before and after respectively [21:32:00.0000] <zcorpan> <!doctype html><title>noframes</title><noframes></noframes><frameset><noframes></noframes></frameset><noframes></noframes> [21:32:01.0000] <zcorpan> i guess it should whine about the first and last [21:34:00.0000] <zcorpan> and missing frame or frameset child [21:37:00.0000] <MikeSmith> zcorpan: OK, this is what I get: [21:37:01.0000] <MikeSmith> http://pastebin.com/m4ba0c569 [21:37:02.0000] <MikeSmith> so I guess I need to add <noframes> as a valid child of <head> [21:40:00.0000] <zcorpan> MikeSmith: why? [21:40:01.0000] <zcorpan> MikeSmith: html4 doesn't allow noframes in head [21:41:00.0000] <MikeSmith> ah, OK [21:41:01.0000] <MikeSmith> so what should it be reporting for that case? [21:43:00.0000] <zcorpan> MikeSmith: what you pasted seems correct [21:45:00.0000] <MikeSmith> zcorpan: OK [21:46:00.0000] <zcorpan> MikeSmith: maybe it's not so useful to assert that the elements other than the top-most frameset being obsolete [21:58:00.0000] <MikeSmith> dglazkov: great to finally have http basic auth working in chrome on linux [22:06:00.0000] <MikeSmith> zcorpan: btw, I have an open bugzilla issue for keeping track of ideas around how to do better reporting for obsolete elements [22:07:00.0000] <MikeSmith> and obsolete attributes [22:07:01.0000] <MikeSmith> http://bugzilla.validator.nu/show_bug.cgi?id=481 [22:08:00.0000] <dglazkov> MikeSmith: I take full credit for this. [22:08:01.0000] <dglazkov> MikeSmith: what wasn't working again? [22:08:02.0000] <dglazkov> :) [22:08:03.0000] <MikeSmith> heh [22:08:04.0000] <MikeSmith> dglazkov: I think it just wasn't implemented at all until 190 [22:08:05.0000] <MikeSmith> anyway, it's still only implemented so far for linux, not Mac [22:09:00.0000] <MikeSmith> dglazkov: hey, btw and just fyi - http://code.google.com/p/chromium/issues/detail?id=15422 [22:11:00.0000] <othermaciej> MikeSmith: so would the validator report attributes or elements as obsolete instead of unknown? [22:11:01.0000] <othermaciej> that would be neat [22:12:00.0000] <MikeSmith> othermaciej: yeah, it does that already for a number of elements -- just elements that were in HTML4, but not in HTML5 [22:12:01.0000] <MikeSmith> so, it doesn't do it for <marquee>, but does for <big>, etc. [22:13:00.0000] <MikeSmith> except we didn't have <frameset> and <noframes> and <basefont> in there [22:13:01.0000] <sayrer> "warning <big> is far too short, please use a style attribute" [22:13:02.0000] <MikeSmith> sayrer: heh [22:13:03.0000] <MikeSmith> othermaciej: so all that I'm changing now is just to add those [22:14:00.0000] <MikeSmith> sayrer: so is it your position that validators should not report anything as obsolete? [22:14:01.0000] <MikeSmith> or that we shouldn't have validators at all? [22:14:02.0000] <sayrer> I am totally in favor of lint tools [22:14:03.0000] <sayrer> I'm not so sure about a hard line between valid/invalid [22:15:00.0000] <sayrer> I gather I'm in good company on that one [22:15:01.0000] <MikeSmith> sayrer: well, you are in company, but not sure if it's necessarily "good" company (scare quotes intentional) [22:15:02.0000] <sayrer> I was thinking of a tbl presentation ;) [22:16:00.0000] <othermaciej> MikeSmith: I'm not sure <marquee> is "obsolete" in quite the same way [22:16:01.0000] <MikeSmith> (scare quotes seem to be the trendy thing to do, so I'm trying to use them more frequently) [22:16:02.0000] <MikeSmith> othermaciej: yeah, true [22:16:03.0000] <othermaciej> MikeSmith: don't you mean "scare" quotes? [22:16:04.0000] <MikeSmith> heh [22:16:05.0000] <othermaciej> sayrer: I'm dubious about the value of a normative spec defining the behavior of lint tools [22:16:06.0000] <sayrer> I think that's what he "means" [22:17:00.0000] <sayrer> othermaciej: me too, that's why I don't like the current author conformance requirements [22:17:01.0000] <othermaciej> sayrer: however, hsivonen made the plausible argument that having a certain common level of baseline rules is an interoperability benefit for authoring tools and validators [22:17:02.0000] <sayrer> since there are many rules that are basically lint tool kind of things [22:17:03.0000] <MikeSmith> we should have double scare quotes -- e,g., ""working group"" -- to indicate an extra amount of disdain [22:17:04.0000] <othermaciej> that's the only really convincing argument I have heard for why such things make sense as conformance requirements [22:18:00.0000] <othermaciej> sayrer: do you think the spec should contain any authoring conformace requirements? [22:18:01.0000] <Dashiva> MikeSmith: Shouldn't you use a different quote set for one of them, to avoid it parsing as "", working group, ""? [22:18:02.0000] <MikeSmith> heh [22:18:03.0000] <othermaciej> sayrer: it's not totally clear to me if you doubt the value of all of them or only some? [22:18:04.0000] <othermaciej> MikeSmith: I would think nested scare quotes would neutralize [22:19:00.0000] <MikeSmith> ah yeah, true [22:19:01.0000] <othermaciej> MikeSmith: since in general "foo" means 'people call it foo, but I don't really think it is, I'm just using their term under protest' [22:19:02.0000] <sayrer> othermaciej: some rules seem to be in place because the markup is "presentational" [22:19:03.0000] <MikeSmith> you can use them to quote somebody else's use of scare quotes -- to show disdain for the fact that they are using scare quotes [22:20:00.0000] <othermaciej> sayrer: are those the only kinds of rules you don't like? [22:20:01.0000] <MikeSmith> sayrer: very true, and not everybody is in agreement about that class of constraints [22:20:02.0000] <MikeSmith> e.g, the presentational attributes on <table> [22:20:03.0000] <othermaciej> "make presentational markup conforming" is a pretty clear position and worth discussing on that basis [22:21:00.0000] <sayrer> oh, I think concrete spec text would make it easier to discuss [22:21:01.0000] <othermaciej> I have an easier time reading a few sentences than a whole alternate spec [22:21:02.0000] <sayrer> then there are those things that seem kind of vestigial or ineffective [22:22:00.0000] <sayrer> like @profile and @summary [22:22:01.0000] <MikeSmith> the bulk of the kind of conformance checking that v.nu is doing is simply checks like, "you can't have a <dd> as a child of a <ul>" [22:22:02.0000] <othermaciej> I suspect many others are in that position [22:22:03.0000] <sayrer> those vestigial things hardly seem worth bannign [22:23:00.0000] <sayrer> or declaring invalid [22:23:01.0000] <othermaciej> sayrer: do you think rules like what MikeSmith cited (can't have a <dd> as a child of a <ul>) are worthwhile? [22:23:02.0000] <othermaciej> or more specifically, are they worthwhile as conformance requirements? [22:23:03.0000] <sayrer> yes, that is the class of things that I have a hard time characterizing. a lint tool would clearly want to report "<b><i>foo</b></i>" [22:24:00.0000] <sayrer> to use a really brain dead example [22:24:01.0000] <othermaciej> misnested tags are in some sense a different level of error than structurally sound markup that happens to violate content model rules [22:25:00.0000] <MikeSmith> yeah, <b><i>foo</b></i> gets reported by the parser, not validator code [22:25:01.0000] <othermaciej> I think some people are interested in avoiding presentational markup in their output, and others are not [22:25:02.0000] <sayrer> for sure [22:25:03.0000] <othermaciej> right now the spec is a compromise and therefore doesn't fully address either need [22:26:00.0000] <othermaciej> I wonder if it would make sense to have two classes of conformance to address this [22:26:01.0000] <MikeSmith> "Error: End tag b violates nesting rules." [22:26:02.0000] <sayrer> I am in favor of an authoring guide [22:27:00.0000] <MikeSmith> I think it would make sense to have user-tunable validators [22:28:00.0000] <othermaciej> if we accept Henri's premise that there needs to be a shared set of rules, so that different combinations of content generators and validators can interoperate without large switching costs, then an authoring guide may be insufficient [22:28:01.0000] <othermaciej> I'm not sure if that premise is right [22:29:00.0000] <MikeSmith> there's a market for validators, just like there is a market for browsers [22:29:01.0000] <othermaciej> but it sounds plausible [22:29:02.0000] <MikeSmith> and people place value on validators [22:30:00.0000] <othermaciej> the idea being, my CMS generates markup that is error-free in Validator A, but I decide I like the user interface of Validator B better [22:30:01.0000] <othermaciej> can I switch without being flooded with errors? [22:30:02.0000] <MikeSmith> right [22:30:03.0000] <othermaciej> that seems to be a case where interoperability matters [22:31:00.0000] <othermaciej> so that makes me think that there has to be some shared definition of what counts as content errors [22:31:01.0000] <othermaciej> so it's not a viable position for the WG to completely fail to define authoring conformance [22:31:02.0000] <sayrer> I don't think any of that follows [22:32:00.0000] <sayrer> in particular, I am not suggesting the WG fail to define anything [22:32:01.0000] <othermaciej> I'm not saying you are suggesting that [22:32:02.0000] <sayrer> ok [22:32:03.0000] <MikeSmith> anyway, it does at least seem that there's a fundamental baseline of document-conformance rules that we could get agreement on [22:32:04.0000] <othermaciej> it is now clear to me that you think there should be some definition of what constitutes an error [22:32:05.0000] <othermaciej> and that at the very least, you don't think presentational markup or obsolete but harmless attributes should be in that category [22:32:06.0000] <sayrer> well, there are things that I would like to have attention brought to, as an author [22:33:00.0000] <sayrer> "this may not mean what you think it means" is something I want to know about [22:33:01.0000] <MikeSmith> right [22:33:02.0000] <MikeSmith> e.g., a <dd> as a child of <ul> [22:34:00.0000] <sayrer> well, what is the error you would report for that? [22:34:01.0000] <othermaciej> it's not clear to me what you think should be in that definition, or whether you think those rules should be labeled something other than "conformance", or how you feel about whether they need to be in a separate document from browser conformance rules [22:35:00.0000] <sayrer> 1.) not sure what should be there, probably requires user testing [22:35:01.0000] <sayrer> 2.) don't want the labeled as conformance requirements in the HTML document, somewhere else is fine [22:35:02.0000] <sayrer> 3.) probably separate [22:36:00.0000] <sayrer> MikeSmith: what is the ideal error for <dd> in <ul> ? [22:36:01.0000] <othermaciej> so it's ok to label the rules conformance requirements, but not if they are in the same document as browser conformance requirements? [22:36:02.0000] <sayrer> yes [22:36:03.0000] <othermaciej> I can understand the rest of your position, but not that part [22:37:00.0000] <sayrer> the other conformance requirements in the document have interop ramifications [22:37:01.0000] <MikeSmith> sayrer: I don't know what would be ideal, but at a minimum, the checker should report, "<dd> is not allowed as a child of <ul>" [22:37:02.0000] <sayrer> well, it is allowed [22:37:03.0000] <MikeSmith> and then report what is allowed as a child of <ul> [22:38:00.0000] <othermaciej> hsivonen would argue that validity rules have interop ramifications (between different validators and different content generators) [22:38:01.0000] <sayrer> I find that position pretty circular [22:38:02.0000] <othermaciej> I don't see how it's circular [22:39:00.0000] <othermaciej> the idea is that you can switch content generating tools without changing validators, or vice versa [22:39:01.0000] <othermaciej> I think it's a much less important form of interoperability than browser <--> content [22:40:00.0000] <othermaciej> but being able to switch out pieces independently is pretty much what interoperability means [22:40:01.0000] <MikeSmith> sayrer: OK, if I get your point, the error might more precisely be, "using <dd> as a child of <ul> is probably not going to have the effect you might be expecting" [22:40:02.0000] <MikeSmith> sayrer: "<dd> as a child of <ul> is probably not something you want to do" [22:41:00.0000] <sayrer> "we need this rule so the things that check rules will have a rule" [22:42:00.0000] <othermaciej> I think a more fair presentation of the position would be "we need rules about what is an error so that content producers and tools that check for content errors can agree on what is an error" [22:43:00.0000] <othermaciej> you could say "tools that check for content errors" should not label anything they report as errors, but the same reasoning applies if you call them "errors" or "conditions that indicate the author may not have said what he meant" or whatever [22:44:00.0000] <othermaciej> or "warnings" [22:47:00.0000] <sayrer> validator interop is not something the w3c is chartered to [22:48:00.0000] <sayrer> to do [22:49:00.0000] <sayrer> I am not even sure it's desirable, but if it is, every validator implementor will rush to become Rec FooValidator2015 or whatever [22:49:01.0000] <sayrer> I suppose there's a subtext of everyone being paranoid about rogue additions to the language [22:49:02.0000] <sayrer> but I don't think these rules are going to help that [22:52:00.0000] <othermaciej> I could buy that validator interop is not an important goal [22:53:00.0000] <sayrer> it would certainly be easier to suss out if the author conformance rules had clear, uniform motivations [22:53:01.0000] <othermaciej> so far, the one validator developer who has spoken up thinks it is [22:53:02.0000] <othermaciej> they do seem to have different motivations [22:54:00.0000] <MikeSmith> sayrer: it might also help if we used the term "document conformance" instead of "authoring conformance" [22:54:01.0000] <othermaciej> there's surface syntax errors where the motivation seems to be that they are likely to indicate an error, and may result in divergent behavior in older browsers [22:55:00.0000] <othermaciej> there's nesting rules where violating them results in parser madness (like <p> not allowed in <table>) or because some elements have very fixed behavior and other children don't make sense (like <div> not being allowed in <select>) [22:55:01.0000] <MikeSmith> othermaciej: right [22:56:00.0000] <MikeSmith> there are also cases like, you can't nest <a href>s if you want them to have an expected effect [22:57:00.0000] <MikeSmith> or, as a class, you can't nest interactive elements [22:57:01.0000] <othermaciej> there's nesting rules that are based on what is believed to be proper semantic structure, but which are not in any way enforced and won't cause interop problems if you violate them (like <b> not allowed immediately inside <ul>) [22:58:00.0000] <othermaciej> there's elements or attribute that have an effect but are banned because they are presentational [22:58:01.0000] <MikeSmith> yeah [22:58:02.0000] <Hixie> <b> inside <ul> will cause problems to any tool that tries to do anything useful with list items in a list [22:58:03.0000] <othermaciej> there's elements or attributes that have no effect, and may have been in older specs, but are banned because they are useless [22:58:04.0000] <Hixie> (e.g. scripts that resort lists) [22:59:00.0000] <othermaciej> there's elements or attributes that have no current effect and were never in a spec (the general ban on unknown elements or attributes) [22:59:01.0000] <heycam> Hixie, "but the novice authors is cautioned" in the new "A quick introduction to HTML" section [22:59:02.0000] <othermaciej> there's rules for attribute values that disallow things that can't have an effect because they don't satisfy the basic syntax of the attribute (that would be in the "likely to be an error" category) [22:59:03.0000] <Hixie> heycam: thanks [23:00:00.0000] <othermaciej> I'm not sure if that list of categories was exhaustive [23:00:01.0000] <sayrer> elements that have an effect but were never in a spec [23:01:00.0000] <sayrer> that's all I can think of [23:02:00.0000] <othermaciej> there's also mandatory content, like the fact that documents must have a <title> or that <img> elements must have a src attribute [23:05:00.0000] <sayrer> those last two seem different to me [23:05:01.0000] <sayrer> maybe it's not important [23:06:00.0000] <othermaciej> I see your point; not sure how to express the difference [23:06:01.0000] <othermaciej> some mandatory content is mandatory because the associated container won't work otherwise [23:07:00.0000] <othermaciej> many elements are also required to have " at least one descendant text node that is not inter-element whitespace, or at least one descendant element node that is embedded content" [23:07:01.0000] <othermaciej> which effectively rules out being empty [23:08:00.0000] <othermaciej> I'm not sure if that's in the "otherwise the element doesn't work" category of mandatory content, or the "because it seems right to require this" category [23:08:01.0000] <sayrer> one class of content nesting rules would seem to concern mark up where the DOM is quite different from the lexical nesting of the markup [23:08:02.0000] <sayrer> parser madness [23:11:00.0000] <othermaciej> if I'm reading the spec right, it seems like it is conforming for <table> or <ul> to be empty, but not for <p> or <b> to be empty [23:12:00.0000] <sayrer> interesting. people put empty <p> elements in to space things out quite a lot [23:12:01.0000] <sayrer> heretical [23:13:00.0000] <othermaciej> empty <p> elements collapse away [23:14:00.0000] <othermaciej> (at least by default; you could presumably style them not to) [23:15:00.0000] <Hixie> you must be reading an old version of the spec [23:15:01.0000] <Hixie> either that or i missed some bit [23:15:02.0000] <Hixie> because empty <p> is explicitly allowed (though mildly discouraged) to enable scripts to fill them in later [23:15:03.0000] <Hixie> and to allow templates to be filled in later yet still be validated [23:17:00.0000] <sayrer> oh right [23:17:01.0000] <sayrer> must be <br>s or something when they do that [23:19:00.0000] <sayrer> anyway, we have quite a few axes for these rules [23:19:01.0000] <Hixie> <br>s? [23:19:02.0000] <othermaciej> I see, "flow content" non-emptiness is a SHOULD now [23:44:00.0000] <Hixie> hsivonen: yt? [23:44:01.0000] <Hixie> hsivonen: i'm still baffled by http://www.w3.org/Bugs/Public/show_bug.cgi?id=6776 [23:46:00.0000] <hsivonen> Hixie: what's unclear? [23:49:00.0000] <Hixie> should i add something to the xhtml syntax section saying "XSLT 1.0 processors, when the method is 'html', must violate the XSLT 1.0 spec in the following way:"? [23:49:01.0000] <Hixie> or some other section? [23:50:00.0000] <Hixie> basically i don't understand where to put this [23:50:01.0000] <hsivonen> I'd put it near the text/html DOM API differences [23:50:02.0000] <Hixie> (and i am concerned about getting shot by the xslt working group for doing something that i really don't care about :-) ) [23:50:03.0000] <Hixie> the case sensitivity stuff? [23:50:04.0000] <hsivonen> it would be confusing to put this in XHTML section [23:50:05.0000] <hsivonen> yeah [23:51:00.0000] <hsivonen> for wg concerns, see the linked bug [23:51:01.0000] <Hixie> i don't see how the section you're indicating has anything to do with xslt whatsoever [23:51:02.0000] <Hixie> it's all about the DOM [23:51:03.0000] <hsivonen> this is about the DOM, too [23:53:00.0000] <hsivonen> XSLT processors that output to a character or byte stream are unaffected [23:55:00.0000] <Hixie> hmmm [23:55:01.0000] <Hixie> ok [00:26:00.0000] <Hixie> hsivonen: ok, how is http://www.whatwg.org/specs/web-apps/current-work/#dom-based-xslt-1.0-processors [00:31:00.0000] <Hixie> well i checked it in [00:31:01.0000] <Hixie> let me know how it is [01:12:00.0000] <hsivonen> Hixie: looks good [01:12:01.0000] <Hixie> woo [01:13:00.0000] <hsivonen> thanks [02:26:00.0000] <Hixie> hsivonen: still here? [02:28:00.0000] <Hixie> hsivonen: http://www.whatwg.org/specs/web-apps/current-work/#interactions-with-xpath-and-xslt [02:28:01.0000] <Hixie> checked in [02:44:00.0000] <hsivonen> Hixie: looks good. thanks [02:45:00.0000] <hsivonen> Hixie: can the XPath bit be annotated as implemented in Gecko and WebKit and the XSLT bit be annotated as implemented in Gecko? [02:47:00.0000] <hsivonen> hmm. does WebKit output to DOM from XSLT? [02:48:00.0000] <othermaciej> WebKit's XSLT will serialize and reparse [02:48:01.0000] <hsivonen> othermaciej: thanks [02:49:00.0000] <Hixie> hsivonen: reload and there should be separate IDs for the two parts of that section [02:49:01.0000] <othermaciej> I believe that makes the given conformance requirement inapplicable [02:49:02.0000] <hsivonen> othermaciej: it does, yes [02:50:00.0000] <hsivonen> I annotated Gecko and WebKit nightlies as supporting the section [02:50:01.0000] <hsivonen> since the last part doesn't apply to WebKit [02:53:00.0000] <MikeSmith> Hixie: can you conceive of any possible workaround for the Webkit bug that prevents get-author-view-onload of the spec from working? [02:54:00.0000] <othermaciej> MikeSmith: what's the bug? [02:54:01.0000] <Hixie> MikeSmith: i wrote a workaround earlier today [02:55:00.0000] <Hixie> othermaciej: .disabled doesn't seem to work on <link rel=stylesheet> in webkit the first time it's set, or something [02:55:01.0000] <Hixie> hsivonen: cool, thanks [02:55:02.0000] <othermaciej> weird [02:56:00.0000] <Hixie> othermaciej: i haven't created a minimal test case, so i don't know exactly what the issue is [02:57:00.0000] <MikeSmith> othermaciej: https://bugs.webkit.org/show_bug.cgi?id=26673 [02:58:00.0000] <othermaciej> sounds plausible [02:58:01.0000] <othermaciej> maybe a kind volunteer will make a reduction [03:03:00.0000] <takkaria> hey, has anyone got a recentish IE to hand? [03:03:01.0000] <jgraham> othermaciej: The thing that we are more likely to have to spend time on is how aria interacts with HTML semantically e.g. what does aria-labelledby mean on a table cell if it does/does not point to a table header [03:03:02.0000] <jgraham> takkaria: Install virtualbox :) [03:03:03.0000] <jgraham> (or: yes) [03:04:00.0000] <takkaria> I'm wondering what IE does if you go to "http://spaces.ru/fff/6789004718042409/0/427562/Gameloft_Asphalt_4_Elite_Racing_3D_HD_v1.0.6_S60v3_SymbianOS9.sisx", mainly whether it tries to display the page or whether it pops up a download box [03:04:01.0000] <othermaciej> jgraham: I could also imagine additional validity rules preventing nonsensical combinations [03:04:02.0000] <othermaciej> like <input type="text" role="checkbox"> [03:05:00.0000] <Hixie> currently aria actually says we're not allowed to disallow that [03:05:01.0000] <jgraham> othermaciej: So can I [03:05:02.0000] <Hixie> that was one of my last call comments [03:05:03.0000] <Hixie> are there any content models that use "Text" other than <option>, <title>, and <textarea>? [03:06:00.0000] <othermaciej> Hixie: what's the specific part that says that? [03:06:01.0000] <Hixie> othermaciej: i forget the exact section off-hand [03:07:00.0000] <othermaciej> does it have conformance requirements for markup language specifications? [03:08:00.0000] <othermaciej> (they don't seem to have a clear statement of conformance classes) [03:08:01.0000] <Hixie> http://www.w3.org/WAI/PF/comments/details?comment_id=267 [03:08:02.0000] <Hixie> see section 6.1.1 [03:09:00.0000] <Hixie> http://www.w3.org/TR/2009/WD-wai-aria-20090224/#host_general_role [03:09:01.0000] <othermaciej> I see, that does seem like a problem [03:10:00.0000] <othermaciej> I wonder if that is intentionally meant to prevent banning of crazy combinations, or just an attempt to say a language can't ban all use of an ARIA role wholesale, without considering the issue [03:12:00.0000] <jgraham> othermaciej: I have heard various aria types insist that aria should always override host semantics so <input type=text role=checkbox> would be considered fine and indicative of a checkbox [03:13:00.0000] <othermaciej> jgraham: it's kind of crazy to consider it fine [03:13:01.0000] <othermaciej> but one thing I didn't realize before is that overriding host semantics has some associated implementation challenge [03:14:00.0000] <othermaciej> because any time an ARIA-related attribute is set at all, you have to stop doing all or most of the built-in accessibility behavior [03:14:01.0000] <Philip`> takkaria: I get a download prompt in IE8 on Vista [03:14:02.0000] <othermaciej> or at least, it seems like the output of assistive technologies would be nonsense if something was both a checkbox and a text box, or (as in Hixie's example) both checked and unchecked [03:16:00.0000] <takkaria> Philip`: thanks [03:25:00.0000] <Hixie> i just got my first ironic error message from hsivonen's validator [03:26:00.0000] <jgraham> The validator does irony now? [03:27:00.0000] <Hixie> while validating the spec after changing it to allow <p> in <caption>, which included a change to an example to add an actual <p> to an actual <caption>, the validator told me that that was wrong. [03:27:01.0000] <takkaria> you need a "I am the spec author and what I say goes" box on v.nu [03:29:00.0000] <Hixie> actually adding the validator step to my generator script has caught so many typos and errors that it's awesome [03:41:00.0000] <MikeSmith> jgraham, Hixie: browsers generally behave as expected for multiple pargraphs or lists in a <caption>? [03:43:00.0000] <jgraham> MikeSmith: AFAICT yes [03:45:00.0000] <hsivonen> how many WebKit ports does Nokia have and how many of them are active? [03:45:01.0000] <MikeSmith> hsivonen: as far as I know of it least, they have one: the native S60 port [03:46:00.0000] <othermaciej> they have the Qt port [03:46:01.0000] <MikeSmith> and it is not being actively developed, and has not been for a long time [03:46:02.0000] <othermaciej> as far as I know, that's the only port that Nokia actually maintains upstream [03:46:03.0000] <MikeSmith> othermaciej: they have an S60 QtWebKit port now [03:47:00.0000] <othermaciej> I don't know what if anything they do in private trees [03:47:01.0000] <MikeSmith> a "pre-release" one - http://sideshowbarker.net/2009/06/26/s60-qtwebkit/ [03:48:00.0000] <hsivonen> MikeSmith: the native S60 port that shipped with S60r3.1 is dead, right? so is QtWebKit what they are on track shipping when they next refresh the bundled browser on S60? [03:48:01.0000] <MikeSmith> hsivonen: as far as I can tell, yeah, the native S60 port is dead. [03:48:02.0000] <hsivonen> ok [03:49:00.0000] <MikeSmith> as far as how QtWebKit fits into their product plans, they have said nothing publicly [03:49:01.0000] <othermaciej> they don't maintain the webkit.org copy of it that lives on a very old branch [03:49:02.0000] <othermaciej> (the native S60 port) [03:50:00.0000] <MikeSmith> http://sideshowbarker.net/2008/04/11/s60-webkit-dead/ [03:52:00.0000] <MikeSmith> that was more than 1 year ago, they closed all open bugs on it en masse [03:52:01.0000] <MikeSmith> after there had at that time not been any checkins to the branch for 8 months [03:52:02.0000] <othermaciej> I would guess the native S60 port is abandoned but there's not been a definitive statement [03:52:03.0000] <othermaciej> it's possible they did all sorts of things in a private tree, or maybe not [03:53:00.0000] <othermaciej> who knows? [03:54:00.0000] <MikeSmith> the work QtWebKit port for S60 is being done in public, I think [03:54:01.0000] <MikeSmith> I mean, I think they have a publicly readable repository [03:54:02.0000] <MikeSmith> heh [03:54:03.0000] <MikeSmith> this is what dude wrote at the time in response to my post - [03:54:04.0000] <MikeSmith> http://blogs.s60.com/2008/04/whoa_there [03:55:00.0000] <othermaciej> sure, it's clear that QtWebKit for S60 is being worked on [03:55:01.0000] <othermaciej> and one could certainly guess they want this to be part of their product strategy [03:55:02.0000] <MikeSmith> "We are not backing away from Webkit no matter what you might have read on one of those new-fangled “website” things ... No, no, no — it’s just that we’re focusing on the newer Leopard branch" [03:56:00.0000] <MikeSmith> anyway, ancient history now [03:56:01.0000] <othermaciej> I believe they tried to bring their port up to the Safari 3.0 version of WebKit and failed [03:58:00.0000] <Hixie> ok time to go to bed [03:58:01.0000] <Hixie> actually long past time to go to bed [03:58:02.0000] <Hixie> nn [04:16:00.0000] <Romme> is here a right place to ask about html5lib? i've googled, but still haven't found the answer [04:17:00.0000] <takkaria> sure [04:20:00.0000] <jgraham> Romme: Ask away [04:21:00.0000] <hsivonen> Nokia also has a port of WebKit for Maemo, right? A variant of the GTK port. [04:21:01.0000] <Romme> jgraham: how do i serialize an element without including it's start and end tags? [04:24:00.0000] <Romme> i've tried ''.join(serializer.HTMLSerializer().serialize(walker(lxmletree.FragmentRoot(content)), 'UTF-8')), where content is my element [04:25:00.0000] <Romme> but it then serializes everything until my document ends [04:28:00.0000] <Romme> and i don't quite understand what FragmentWrapper and FragmentRoot are for [04:31:00.0000] <jgraham> Romme: Of the top of my head I don't think we support that in an easy way but it does seen like a reasonable feature request. I'm pretty sure it is possible to hack [04:31:01.0000] <jgraham> (sorry I got distracted by rl) [04:32:00.0000] <Romme> i've heard that if i parse my data with parseFragment, then i will get a fragment when serializing as well, but the problem is, i'm not parsing a fragment, but a document instead [04:34:00.0000] <jgraham> Romme: Which tree backend are you using? [04:34:01.0000] <Romme> jgraham: lxml [04:36:00.0000] <takkaria> are there any tests for abarth's content-type sniffing draft? [04:36:01.0000] <jgraham> Romme: Give me a moment I will just check something [04:36:02.0000] <jgraham> takkaria: IIRC gsnedders has some [04:37:00.0000] <takkaria> hm, I guess I'll wait til he comes online again [04:38:00.0000] <jgraham> takkaria: You may see him in person before then :) [04:39:00.0000] <takkaria> that is a point :) [04:41:00.0000] <jgraham> Romme: t = html5lib.parse("<div><a>b</a> abc <b>c</b></div>", treebuilder="lxml") [04:41:01.0000] <jgraham> tw = html5lib.treewalkers.getTreeWalker("lxml") [04:41:02.0000] <jgraham> div = t.find("//div") [04:41:03.0000] <jgraham> for item in tw(list(div)):print item [04:42:00.0000] <jgraham> gives all the children of the <div> but not the <div> itself. So I guess that will serialize OK [04:42:01.0000] <Romme> i would have just serialized it with lxml then :P [04:43:00.0000] <Romme> i need the features of html5lib, like omitting optional tags or quotes in attribute names [04:43:01.0000] <Romme> oops [04:43:02.0000] <Romme> sorry [04:44:00.0000] <Romme> didn't notice the "tw" [04:44:01.0000] <jgraham> s = html5lib.serializer.HTMLSerializer() [04:44:02.0000] <jgraham> s.render(tw(list(div))) [04:44:03.0000] <jgraham> u'<a>b</a> abc <b>c</b>' [04:46:00.0000] <Romme> i wonder why render() converts it to a list instead of just iterating [04:48:00.0000] <Romme> jgraham: thanks, but that serialized it until my document ended [04:49:00.0000] <jgraham> Romme: I don't actually know, but it might be a theory that doing join on a list is faster than joing it on an iterator. I have no idea if this is true or not [04:49:01.0000] <Romme> with end tags for element which haven't started [04:49:02.0000] <jgraham> Romme: Can you give an example of something that doesn't work? [04:50:00.0000] <Romme> jgraham: wait a moment, i'll create a testcase [04:52:00.0000] <jgraham> /me notes that using list() or not seems to make no speed difference [04:55:00.0000] <Romme> a few more minutes [05:05:00.0000] <Romme> jgraham: http://files.getdropbox.com/u/268483/testcase.tar.bz2 [05:15:00.0000] <Midler1> In case someone of you have missed (its not about html5...but still) http://blogs.msdn.com/outlook/archive/2009/06/24/the-power-of-word-in-outlook.aspx#commentform [05:16:00.0000] <Midler1> its like they will stay in the 90:s forever [05:24:00.0000] <Romme> well, at least they're not dicks about letting us hide that thing out of sight. i'm quite tired of minimizing evolution and moving it to the second workspace [05:26:00.0000] <jgraham> Romme: AFAICT your testcase works. At least the textcontent of the serialized fragment seems to be the same as that of the div element [05:27:00.0000] <jgraham> Possibly the div just doesn't get closed? [05:28:00.0000] <Romme> jgraham: Vim confirms that the div is indeed closed [05:28:01.0000] <Romme> hmmph, maybe it's just my version of html5lib, i'll try installing the newest one [05:28:02.0000] <Romme> what version are you using? [05:29:00.0000] <jgraham> Romme: Some random development version. I don't recommend pulling the head at the moment though; things are not really in a perfect state [05:29:01.0000] <jgraham> This is unlikely to have changed [05:34:00.0000] <jgraham> Oh wait, I am probably wrong [05:58:00.0000] <jgraham> Romme: When I said I was wrong, I meant about the testcase working for me. I was comparing the wrong things. I think the solution is more complex than I first thought [05:58:01.0000] <jgraham> One approach would be to write a filter that dropped the first tag and everything after the matching end tag [06:01:00.0000] <jgraham> In theory using copy.deepcopy to get the nodes into a list could work (because afaik it breaks parent pointers) but I think it doesn't actually work right now [06:01:01.0000] <jgraham> File a feature request [06:04:00.0000] <Philip`> /me decides that writing a GUI half in C++ and half in JS is actually quite a pain [06:04:01.0000] <Philip`> so I suppose I should just rewrite it all in JS [06:04:02.0000] <hsivonen> Philip`: Firefox? [06:05:00.0000] <Philip`> hsivonen: No, something using wxWidgets [06:10:00.0000] <Philip`> Hmm, I currently have three separate sections of code viewing and controlling the same piece of data (GUI written in JS; GUI written in C++; game engine in C++ in another thread) - I wonder if there's meant to be some magic design pattern that makes it all stay in sync trivially... [06:12:00.0000] <Philip`> s/makes/would make/ [06:22:00.0000] <hendry> Philip`: when in doubt, write in C [06:50:00.0000] <Philip`> hendry: I'd generally prefer to make things easier for myself, not harder :-) [16:07:00.0000] <gsnedders> /me waves [16:08:00.0000] <gsnedders> (From Sweden) [16:13:00.0000] <Midler> lol [16:13:01.0000] <Midler> comhem.. [16:13:02.0000] <Midler> gsnedders: I can see that 2009-06-27 [17:25:00.0000] <gsnedders> ergh. I just realized I left jacket on the bus from the airport. [17:25:01.0000] <Midler> less funny [17:25:02.0000] <Midler> try to call the buscompany [17:26:00.0000] <gsnedders> Between 8 and 23 [17:26:01.0000] <gsnedders> /me tries the email option :P [17:27:00.0000] <eighty4> not like you'll need it this week :D [17:27:01.0000] <gsnedders> Indeed. [17:27:02.0000] <gsnedders> But the hoodie would be nice as a pillow :p [17:27:03.0000] <eighty4> 28-29* all week [17:27:04.0000] <eighty4> /me goes wheee [17:27:05.0000] <Midler> yea, for once its hot WEEEE [17:27:06.0000] <Midler> =D [17:27:07.0000] <eighty4> you don't have a pillow? [17:27:08.0000] <gsnedders> No [17:27:09.0000] <gsnedders> I only have bed linen at all because I asked James for some yesterday :P [17:28:00.0000] <Midler> eighty4: he has some hair as ive understood it [17:28:01.0000] <eighty4> Midler: and I have to work like crazy all week [17:28:02.0000] <eighty4> gsnedders: you need ikea [17:28:03.0000] <gsnedders> :P [17:28:04.0000] <gsnedders> I think everything else I need is supermarket stuff [17:28:05.0000] <eighty4> ever been to ikea? [17:28:06.0000] <gsnedders> I'll think of more and more stuff tomorrow morning [17:28:07.0000] <gsnedders> Yes [17:28:08.0000] <eighty4> they have everything :) [17:29:00.0000] <gsnedders> eighty4: Really everything in Sweden? :P [17:29:01.0000] <eighty4> everything you need at least [17:29:02.0000] <eighty4> and if you get hungry they have hotdogs for 5kr [17:29:03.0000] <gsnedders> Power for laundry? [17:29:04.0000] <gsnedders> :P [17:30:00.0000] <eighty4> no, but you don't need that. Not really [00:45:00.0000] <Hixie> jesus christ, took an hour and a half for comcast to solve my problem [00:45:01.0000] <Hixie> including them asking for my MAC address _twice_ despite the support call being an IM session WITH A CHAT LOG [01:01:00.0000] <eighty4> :) [01:01:01.0000] <eighty4> Not surprising [01:11:00.0000] <Hixie> not surprising, but still tiresome :-) [01:13:00.0000] <gsnedders> /me waves tiredly [01:17:00.0000] <gsnedders> /me would head out and wonder, but thinks it is too likely he'll get lost [01:20:00.0000] <Hixie> MikeSmith: yt? [01:45:00.0000] <gsnedders> /me could do with some breakfast… [01:45:01.0000] <gsnedders> /me is somewhat limited by the emptiness of this flat [03:19:00.0000] <cyclist> what's the status on HTML comment tags? [03:20:00.0000] <cyclist> For example, in a page i was recently it had: [03:20:01.0000] <cyclist> ", the text appeared [05:20:00.0000] <annevk2> http://wiki.whatwg.org/wiki/HowToWriteASpec is nice; hopefully someone will use it :) [05:41:00.0000] <Philip`> http://itpastorn.blogspot.com/2009/06/validation-and-doctype-myths-and.html - <!DOCTYPE html PUBLIC "null" "null"> - oops [05:46:00.0000] <annevk2> ouch, that page argues for using window.open rather than target=_blank [10:06:00.0000] <takkaria> wow [10:06:01.0000] <takkaria> "Ian, I don't think that it is fair to imply or assert that longdesc and [10:06:02.0000] <takkaria> summary [10:06:03.0000] <takkaria> do not work in practice." [10:06:04.0000] <takkaria> (http://lists.w3.org/Archives/Public/public-html/2009Jun/0767.html) [10:06:05.0000] <takkaria> I'm not quite sure who it would be unfair to to assert that longdesc doesn't work in practice [10:20:00.0000] <webben> takkaria: I guess the argument is: UA vendors are consulted on whether features work. In the case of accessibility features, AT vendors are a key class of UA vendors who should be consulted. [10:20:01.0000] <webben> (it somewhat sidesteps the problematic nature of actually consulting AT vendors) [10:21:00.0000] <webben> takkaria: The point about many "bad" summary values not being presented to the user in practice could be taken as an example of the benefits of such consultation. [10:22:00.0000] <takkaria> I was more heading at the longdesc bit [10:22:01.0000] <takkaria> I mean, there is no data anywhere that supports the idea that longdesc is a useful part of the web platform [10:22:02.0000] <webben> The same may be true of longdesc; the longdesc lottery article didn't (IIRC) test or consult as to what longdesc values are or aren't presented to users. [10:23:00.0000] <Wolfman2000> Morning/afternoon. Is there some way to make this page split into smaller sections? Firefox lags in general when this page is opened. http://www.whatwg.org/specs/web-apps/current-work/ [10:23:01.0000] <webben> (There may be additional problems with the longdesc/summary values /actually/ presented to users, of course.) [10:23:02.0000] <takkaria> Wolfman2000: add multipage/ on the end of the url [10:23:03.0000] <Philip`> Wolfman2000: http://www.whatwg.org/specs/web-apps/current-work/multipage/ [10:23:04.0000] <takkaria> indeed not, but I did some of the original research for that article, and I really don't know how you can conclude that longdesc works in practice [10:23:05.0000] <Wolfman2000> That works. Thanks. [10:25:00.0000] <takkaria> you can't just go around saying "I don't think it's fair to assert that x" when everything speaks for that assertion and nothing against it [10:26:00.0000] <Wolfman2000> Hmm...Safari 4 seems to handle the singlepage/ fine [10:26:01.0000] <webben> /me has never really bought into the evidence brought to bear in that article, but there we go. [10:28:00.0000] <takkaria> I'm just annoyed at people casting aspersions without giving good reasons for casting them [10:29:00.0000] <webben> I think the comment should have acknowledged that AT vendors are at liberty to join the discussion; and that some AT vendors are already present in the discussion. [10:30:00.0000] <takkaria> I think there's also a big difference here between adding new features and evaluating existing features [10:30:01.0000] <takkaria> if we introduce new a11n features, AT vendors should be consulted [10:30:02.0000] <takkaria> but longdesc/summary are about seeing if an existing experiment has given good results or not [10:31:00.0000] <webben> I don't think there's any difference really. [10:31:01.0000] <webben> As we can see from summary, consultation can change one's interpretation of existing features. [10:33:00.0000] <takkaria> Murray's argument just seem to be suggesting to me that the AT people should say what a11n features we should have, and that's that [10:35:00.0000] <webben> takkaria: I'm not sure that is what Murray is arguing. [10:35:01.0000] <takkaria> maybe I'm wilfully misinterpreting given my own biases :) [10:36:00.0000] <webben> takkaria: Seems to me his argument is: if browser vendors are the ultimate arbiters of features, then why are AT arbiters not the ultimate arbiters of a11y features. [10:36:01.0000] <webben> (earlier in the email he rejects the idea that browser vendors can be trusted to get things right for their users) [10:36:02.0000] <takkaria> well, they are [10:37:00.0000] <takkaria> but the AT vendors tend to refuse to take part in the HTMLWG, so there's not much can be done about that... [10:37:01.0000] <webben> some AT vendors. [10:37:02.0000] <webben> there are certainly options for external consultation [10:38:00.0000] <webben> particularly with opensource projects like NVDA and Orca. [10:39:00.0000] <webben> And to some extent that's happening, if perhaps only rather indirectly - I know e.g. jgraham has been chatting with Marco Zehe (Mozilla a11y, ex-Freedom Scientific guy) in here. [10:40:00.0000] <takkaria> mm [10:40:01.0000] <takkaria> I'm just not sure we need AT people to be there to see that longdesc has failed on the web [10:42:00.0000] <webben> Personally, I find the idea of putting a11y features into a box and putting it at the feet of AT vendors as problematic as it would be to exclude the later from discussions of supposedly non-a11y features. [10:42:01.0000] <webben> there's a lot of accessibility that has little to do with AT. [10:42:02.0000] <webben> but is provided by authoring practice or ordinary browser features. [10:43:00.0000] <webben> I think the argument about the utility of "summary" to sighted users (however we choose to judge on that issue) touches directly on that. [10:44:00.0000] <Dashiva> webben: It also reinforces the view of accessibility as a bolt-on thing [10:45:00.0000] <webben> Dashiva: well the idea of "a11y feature" is itself the idea of a bolt-on thing. [10:46:00.0000] <webben> I don't think the notion "a11y features" might be the province of AT vendors adds much bolt-on-ed-ness. [10:47:00.0000] <Dashiva> The idea itself is author bolt-on, but leaving it to AT makes it a vendor bolt-on too. [10:47:01.0000] <webben> I see what you mean. [10:48:00.0000] <Dashiva> Thinking mainly of ARIA, but there might be other cases [10:49:00.0000] <webben> ARIA, whatever its merits, is indeed an example of AT- rather than accessibility-driven design. But then it was basically conceived as a bolt-on to Dojo widgets so that's not surprising. [16:09:00.0000] <sayrer> huh, I see that inputmode was removed [16:10:00.0000] <sayrer> in favor of -waf-format-mode or some such [16:10:01.0000] <sayrer> is there an HTML5 way to do predictOn/predictOff ? [16:18:00.0000] <Philip`> http://thevideobay.org/ - "TVB aims to use the new HTML5 features, more specificly the <video> and <audio> tags with the ogg/theora video and audio formats" [16:30:00.0000] <annevk2> sayrer, it wasn't removed in favor of anything [16:31:00.0000] <annevk2> sayrer, iirc it was removed because the sole reason we had it was because of XForms [16:33:00.0000] <sayrer> annevk2: the checkin comment says something else [16:34:00.0000] <sayrer> http://lists.whatwg.org/pipermail/commit-watchers-whatwg.org/2008/001395.html [16:35:00.0000] <sayrer> I found that the carrington mobile wordpress theme uses it [16:36:00.0000] <annevk2> oh, interesting [16:36:01.0000] <sayrer> looks like webkit supports it [16:36:02.0000] <sayrer> or maybe only android? [16:36:03.0000] <sayrer> hmm [16:37:00.0000] <sayrer> I guess I should try predictOff [16:41:00.0000] <sayrer> don't see any resolved webkit bugs for it [16:41:01.0000] <sayrer> but it is in XHTML Basic [16:42:00.0000] <Midler> Philip`: ofcourse [16:42:01.0000] <sayrer> maybe I should ask alex king why he put it in there [16:47:00.0000] <sayrer> ah, looks like they added it to satisfy the mobileOk validator [16:47:01.0000] <sayrer> wonder if any real handsets pay attention to it 2009-06-28 [17:53:00.0000] <Hixie> it's so hot here that my macbook pro's right-hand-side usb port doesn't provide enough power to power my keyboard [17:53:01.0000] <Hixie> i don't even know how that makes sense [17:55:00.0000] <gsnedders> So if you put a fan next to it it works? [17:56:00.0000] <Hixie> dunno [17:56:01.0000] <Hixie> but the port was working fine earlier before it was so hot [17:56:02.0000] <gsnedders> Are you sure it's not just coincidence? [17:57:00.0000] <Hixie> the heat has also caused my cats to crash, one of the thinks the wall is the floor and is sleeping with his four feet against the wall [17:57:01.0000] <Hixie> it's very cute [17:57:02.0000] <gsnedders> My father had EyeTV stop working after upgrading to a new version. It turned out, after a week or two, that a cable had fallen out. [17:57:03.0000] <Hixie> i expect what's going on is that the heat has caused some wire to not connect properly [17:57:04.0000] <Hixie> something similar is going on on my mac mini [17:58:00.0000] <Hixie> on hot days, one of my drobos disconnects [17:58:01.0000] <gsnedders> Why would the heat cause that? Surely it should cause it to expand? But if the case expands and takes the two things further apart than the metal expands… [17:58:02.0000] <Hixie> (doesn't appear to be anything wrong with the drobo, but it could be that too of course) [17:58:03.0000] <gsnedders> Yeah, I guess I can make some sense out of it [17:58:04.0000] <Hixie> gsnedders: different metals can expand in different ways [17:58:05.0000] <Hixie> look up how an old-style thermostat works :-) [17:58:06.0000] <gsnedders> Hixie: Hey, this is cem. I don't claim to know anything :P [17:59:00.0000] <gsnedders> /me hasn't done chem. in his last four years of school [17:59:01.0000] <gsnedders> (I somewhat regret that, but by the time I regretting not taking it, it was too late) [18:00:00.0000] <gsnedders> Hixie: Also, we have a old-style thermostat at home :P [18:01:00.0000] <Hixie> well then [18:01:01.0000] <gsnedders> /me ought to sleep seeming it is 2am [20:08:00.0000] <Hixie> http://lists.whatwg.org/pipermail/help-whatwg.org/2009-June/000258.html ...? [22:10:00.0000] <Hixie> anyone got any short but compelling examples for me to put in the <div> section? [22:10:01.0000] <Hixie> showing how it's not a sectioning element... and not showing something that can be done any other way [02:37:00.0000] <Hixie> can anyone think of a variable-size encoding that is a superset of ASCII and that has 0x3C as the second character of a two- (or more) byte sequence? [02:38:00.0000] <Hixie> (Shift_JIS doesn't ever use 0x3C in the second byte apparently) [03:11:00.0000] <Hixie> MikeSmith: yt? [03:11:01.0000] <MikeSmith> Hixie: yep [03:25:00.0000] <Hixie> ok _one_ of these modes in 2022 has to be able to have a byte sequence that contains a 3C without that being a "<" somehow [03:27:00.0000] <MikeSmith> Hixie: need somebody like Junshik Shin [03:27:01.0000] <MikeSmith> encodings guru [03:29:00.0000] <MikeSmith> /me remembers he meant to file a bug about definition of "sectioning content" [03:30:00.0000] <annevk2> Hixie, why in 2022? I'd hope we have less encodings by then, or at least the same amount as now [03:31:00.0000] <svl> annevk2: ISO-2022-x [03:31:01.0000] <Hixie> MikeSmith: i was talking to smontagu about this earlier [03:32:00.0000] <MikeSmith> Hixie: well, if he doesn't know either, it's time to worry [03:48:00.0000] <Hixie> hsivonen: yt? [03:49:00.0000] <gsnedders> /me feels weak. [03:49:01.0000] <gsnedders> I can't open a jam jar :( [03:53:00.0000] <eighty4> use a knife to "pop" it [03:55:00.0000] <gsnedders> eighty4: The jam jar is stronger than the knife [03:55:01.0000] <takkaria> you've tried using a towel for extra grip? [03:56:00.0000] <gsnedders> takkaria: yeah [03:56:01.0000] <Hixie> you've tried using hot and cold water to expand the top and shrink the glass? [03:56:02.0000] <eighty4> hot water? [03:56:03.0000] <eighty4> :) [03:56:04.0000] <gsnedders> Hixie: yes [03:56:05.0000] <Hixie> are you sure it's actually openable? [03:56:06.0000] <eighty4> yeah it is [03:57:00.0000] <gsnedders> Hixie: No, but I'd hope it is so I can eat it :P [03:57:01.0000] <eighty4> I've managed with the same kind. It's just crazy hard .) [03:57:02.0000] <Hixie> have you tried stabbing the top with something sharp to equalise the pressure? [03:57:03.0000] <gsnedders> (seriously though, it's a normal jam jar) [03:57:04.0000] <Hixie> (of course you'll have to finish all today if you do that) [03:57:05.0000] <gsnedders> Hixie: I don't really have anything sharp enough, and I wouldn't finish it that quickly. [03:57:06.0000] <Hixie> i recommend sticking it in the freezer for a bit and trying the hot water trick again [03:58:00.0000] <eighty4> does cold water actually shrink the glass? [03:58:01.0000] <gsnedders> /me notices he's broken his nail trying to open this :P [03:59:00.0000] <eighty4> I've never used that. Only hot water and it's worked before. Hot water will expand the tiny amount of air inside... [03:59:01.0000] <takkaria> gsnedders: well, if you pierce the top, you can clingfilm it afterward [04:00:00.0000] <gsnedders> takkaria: I need clingfilm for that [04:00:01.0000] <Hixie> eighty4: hot water on its own works because it expands the glass and the top at different rates [04:01:00.0000] <Hixie> eighty4: i don't think it affects the pressure inside any measurable amount [04:01:01.0000] <hsivonen> Hixie: I'm here briefly [04:01:02.0000] <Hixie> eighty4: cold water on one part and hot on the other just makes the size change more drastic [04:01:03.0000] <Hixie> hsivonen: no worries, i figured it out [04:02:00.0000] <eighty4> Hixie: Sounds reasonable [04:02:01.0000] <Hixie> hsivonen: basically i was going to ask what you wanted to do about charset sniffing with encodings that are ascii supersets but have "modes" where they become multibyte and then can have ascii-like bytes that aren't ascii [04:02:02.0000] <gsnedders> Hixie: The hot is on the metal lid, right? [04:03:00.0000] <Hixie> hsivonen: but i convinced myself that it's not possible to make a real polyglot doc that is valid in two different encodings with different <meta charset>s applying [04:04:00.0000] <Hixie> gsnedders: generally, yeah, unless you have a weird situation where the problem is that the glass is pulling on the lid somehow keeping it shut [04:04:01.0000] <Hixie> gsnedders: which is possible i guess [04:04:02.0000] <gsnedders> /me normally goes for the get-father-to-open-it solution [04:05:00.0000] <Hixie> i used to use the get-sister-to-open-it solution for shrinkwraped products [04:05:01.0000] <Hixie> that stopped working when i left the country too [04:06:00.0000] <eighty4> that proves that one shouldn't leave the country :) [04:06:01.0000] <gsnedders> I guess the right solution is to not leave the country. [04:07:00.0000] <Hixie> eighty4: i later left the continent too :-P [04:08:00.0000] <Hixie> Philip`: please checkin r3331 and try to come up with some encoding combination in which that's possible [04:08:01.0000] <Hixie> Philip`: er, please see checkin r3331 and..., rather [04:09:00.0000] <gsnedders> /me still can't open it [04:10:00.0000] <jgraham> gsnedders: You are quite lame [04:10:01.0000] <Hixie> send it home [04:11:00.0000] <gsnedders> jgraham: certainly [04:11:01.0000] <jgraham> gsnedders: You have seriously tried running the top under hot water for a minute and then using a towel to get a reasonable amount of grip? [04:11:02.0000] <jgraham> And it didn't work? [04:12:00.0000] <gsnedders> w00t! [04:12:01.0000] <annevk2> svl, hehe, confusing that it happens to collide with a certain year [04:12:02.0000] <gsnedders> Finally, after combining it with the freezer, it works [04:16:00.0000] <eighty4> :) [04:16:01.0000] <eighty4> nice [04:21:00.0000] <MikeSmith> gsnedders: making a reputation for yourself [04:21:01.0000] <MikeSmith> "the guy who's such a wussy, he can't even open a jam jar" [04:21:02.0000] <Philip`> Hixie: I don't understand what "a superset of US-ASCII ... ignoring cases where those bytes would be part of multibyte sequences" actually means [04:22:00.0000] <Hixie> not sure how else to phrase it [04:22:01.0000] <Hixie> i mean like 8859-1 and Shift_JIS [04:22:02.0000] <Hixie> but not GSM 03.38 or EBCDIC [04:26:00.0000] <Philip`> I'd probably be happier if it said something like "bytes in the set ..., ignoring cases where they are not the first byte of a multibyte sequence, decode into the same character as in US-ASCII" or something like that [04:26:01.0000] <Philip`> unless it's meant to be about encoding rather than decoding [04:32:00.0000] <Hixie> Philip`: is the new text better? (regenning now, give it 10 seconds or so) [04:36:00.0000] <jgraham> /me laments how hard it is to get computed style of HTML elements [04:37:00.0000] <jgraham> (outside of a browser environment) [04:40:00.0000] <MikeSmith> god help those who need to reproduce/match browser handling of standards outside of an actual browser [04:51:00.0000] <Philip`> Hixie: Oh, it's the whole "superset ..." thing that I don't understand [04:51:01.0000] <Hixie> ah [04:51:02.0000] <MikeSmith> are there any attributes whose values aren't PCDATA? e.g., any that if you put a character reference into their value, it will be handled as a literal string, instead of being dereferenced? [04:52:00.0000] <takkaria> no [04:52:01.0000] <MikeSmith> OK [04:53:00.0000] <Hixie> yes [04:53:01.0000] <Hixie> well [04:53:02.0000] <Hixie> no [04:53:03.0000] <Hixie> but there's an attribute where entities aren't allowed [04:53:04.0000] <Hixie> two, actually [04:53:05.0000] <Hixie> <meta charset> and in certain equivalent cases <meta content> [04:54:00.0000] <Hixie> and they cause trouble if used there [04:54:01.0000] <MikeSmith> OK [04:54:02.0000] <MikeSmith> /me goes to re-read parsing algorithm [04:56:00.0000] <Philip`> (Argh, lost connection) [04:56:01.0000] <Philip`> Hixie: e.g.: Does it mean 0x3C (followed by any other byte sequence, not preceded by a partial multibyte sequence) must decode to U+003C? Or U+003C must always be encoded to 0x3C? [04:56:02.0000] <Hixie> is there any encoding where that distinction is relevant? [04:56:03.0000] <Philip`> (If you could encode U+003C into two different sequences, one of which was 0x3C, surely that would still be a superset) [04:59:00.0000] <Hixie> i rephrased it again, is it better now? (let me know if you're on your pennies-per-byte connection and i'll paste it in here) [04:59:01.0000] <Philip`> http://www.tbray.org/ongoing/When/200x/2003/10/17/UTF8-plus - UTF-8+names probably lets you encode U+003C as 0x26 0x6C 0x74 0x3B as well as as 0x3C [05:00:00.0000] <Philip`> It's pennies-per-megabyte, which isn't quite so bad, but I am on it anyway :-( [05:01:00.0000] <gsnedders> /me wonders where laundry is in this building [05:02:00.0000] <Philip`> gsnedders: I would guess that it's either dumped on your floor, or in a bag [05:02:01.0000] <gsnedders> s/laundry/washing machines/ [05:02:02.0000] <gsnedders> s/is/are/ [05:03:00.0000] <Philip`> Ah [05:10:00.0000] <Philip`> MikeSmith: There's doctypes too, where the things that look quoted attribute strings can't contain entities [05:10:01.0000] <MikeSmith> OK [05:14:00.0000] <Philip`> Hixie: If I had an error-correcting version of UTF16-BE, which mapped the one-byte string 0x3C (and similarly the two-byte string 0x3C 0x00) onto U+003C, would that be ASCII-compatible? [05:22:00.0000] <MikeSmith> so if I wanted to try to have a term that unambiguously describes the contents of <script> and <style> elements, would "unparsed character data" be inaccurate? [05:24:00.0000] <Hixie> Philip`: i... guess? [05:24:01.0000] <Hixie> Philip`: how would it know that 0x3C was an error though? [05:34:00.0000] <Philip`> Hixie: Because it would reach the end of the file/string and flush all unfinished characters [05:43:00.0000] <Philip`> Hixie: "a document that is valid as two different encodings at once, with different <meta charset> elements applying in each case" - depending on what's meant by "valid", how about <meta charset="utf-8 +names"> ? [05:44:00.0000] <Philip`> which has charset equivalent to utf-8 if you decode as utf-8, and equivalent to utf-8+names if you decode as utf-8+names [05:44:01.0000] <MikeSmith> hendry: thanks for dwm+dmenu tip. I'll give it a try [05:44:02.0000] <Philip`> (based on http://www.tbray.org/tag/utf-8+names.html) [05:56:00.0000] <Philip`> I suppose what's really needed is that if a document saying <meta charset="X"> is encoded as X and then decoded as ASCII, it will still be seen to say <meta charset="X">, so the requirement is that U+003C etc must encode to 0x3C, and nothing else must encode to any sequence containing 0x3C etc [05:57:00.0000] <Philip`> s/\encode\/always encode/ [06:25:00.0000] <gsnedders> /me is trying to get hoodie and jacket back from the bus he left it on :( [06:28:00.0000] <gsnedders> /me has been told to ring another number [07:16:00.0000] <gsnedders> Oh yay. [07:16:01.0000] <gsnedders> Ringing that number I just got told to ring the number I got told to call originally. [07:36:00.0000] <Phae> annoying newb question: is @profile still non-html5? [07:36:01.0000] <gsnedders> Yes [07:37:00.0000] <Phae> hokeykokey. just checking. [09:49:00.0000] <gsnedders> jgraham: I have a problem with returning clean bed sheets tomorrow: I can't find washing machine. [10:05:00.0000] <jgraham> gsnedders: Don't worry, we can wash them. [10:05:01.0000] <jgraham> gsnedders: Did you get your jacket back? [10:05:02.0000] <gsnedders> /me does need to find the washing machine though [10:05:03.0000] <gsnedders> jgraham: no :( [10:09:00.0000] <jgraham> gsnedders: Have you looked in the basement for the washing machine? [10:09:01.0000] <gsnedders> jgraham: If there is a basement, how do I get there? :P [10:10:00.0000] <jgraham> gsnedders: Down some stairs? [10:10:01.0000] <gsnedders> /me was actually expecting that answer [10:10:02.0000] <gsnedders> jgraham: I can't find any such stairs. [10:11:00.0000] <eighty4> gsnedders: a friend of mine lives in a similar building. I'll ask him when he gets home. [10:11:01.0000] <jgraham> gsnedders: There may not be a basement. It's just that in the small sample of all Swedish apartments I have lived in a) there is a basement and b) the washing machine is in the basement [10:11:02.0000] <gsnedders> jgraham: thx [10:12:00.0000] <eighty4> jgraham: I think it's true for more then 90% of all swedish houses. [10:15:00.0000] <gsnedders> jgraham: Is there anything I should bring to work with me tomorrow? [10:15:01.0000] <gsnedders> (Apart from your bed sheets) [10:15:02.0000] <jgraham> gsnedders: I would come fully clothed. [10:15:03.0000] <gsnedders> jgraham: OK, apart from the normal social requirements to go outside. [10:16:00.0000] <jgraham> gsnedders: More seriously I guess your passport will help with the tax stuff [10:16:01.0000] <jgraham> You shouldn't really need anything for work [10:18:00.0000] <Midler> jgraham: ususally, i believe the washing machine is inside the toilette room [10:18:01.0000] <Midler> It uses to be it..mostly [10:19:00.0000] <jgraham> gsnedders: Do you have a plan for lunch? Some core people go out regularly (at about 11:30am (!)) so you can join them or you can bring your own stuff [10:19:01.0000] <gsnedders> jgraham: I have no plan. [10:19:02.0000] <gsnedders> jgraham: My plan for tomorrow is about as detailed as go to work some time after 9 and before 10, and to go home sometime. [10:19:03.0000] <jgraham> gsnedders: If you have no food we can go for lunch and you will get to meet some new people [10:19:04.0000] <jgraham> Which is good [10:20:00.0000] <gsnedders> jgraham: I don't have anything to bring food in, more seriously. [10:20:01.0000] <eighty4> Midler: not in rental apartments [10:20:02.0000] <jgraham> gsnedders: OK [10:20:03.0000] <Midler> =/ [10:20:04.0000] <gsnedders> /me notes eighty4 has the advantage of actually having been in my flat :P [10:20:05.0000] <eighty4> at least not the smaller once [10:21:00.0000] <Midler> eighty4: well, thats true [10:21:01.0000] <eighty4> gsnedders: true :) [10:22:00.0000] <eighty4> jgraham: isn't 11.30 a regular lunch time? or was the ! for something else? [10:22:01.0000] <gsnedders> eighty4: For anyone from GB that's absurdly early. [10:22:02.0000] <eighty4> gsnedders: really? I start to get hungry around 11 :) [10:23:00.0000] <gsnedders> eighty4: In Britain we normally start to think about having breakfast at 8, if at all. Then lunch at 12 or 1. [10:23:01.0000] <eighty4> but I eat breakfast around 07 and have 1-2 cups of coffe [10:23:02.0000] <jgraham> eighty4: That is 1-1.5 hours earlier than I would usually eat lunch [10:23:03.0000] <eighty4> gsnedders: that's just crazy [10:24:00.0000] <jgraham> (but then I don't start the day as early as a typical Swede) [10:24:01.0000] <Midler> well...there is a reason we say "eftermiddag" when we mean after 12 a clock [10:24:02.0000] <eighty4> everyone know that breakfast is the most important meal [10:24:03.0000] <eighty4> and lunch is almost as important [10:24:04.0000] <jgraham> /me always has breakfast [10:24:05.0000] <gsnedders> /me never had breakfast while at school, more or less :P [10:25:00.0000] <Midler> Hard translated eftermiddag would mean after dinner [10:25:01.0000] <eighty4> gsnedders: that's no good. You learn better when you've eaten [10:25:02.0000] <eighty4> Midler: no, actually not [10:25:03.0000] <Midler> but it means afternoon in english [10:25:04.0000] <eighty4> it would mean after mid day [10:25:05.0000] <gsnedders> In England, "afternoon" normally means after lunch. [10:25:06.0000] <eighty4> or at least that's what I've learnt :) [10:26:00.0000] <Midler> hmm, and i always thought it meant after dinner =( [10:26:01.0000] <gsnedders> (In Scotland, it normally means after 12:00) [10:26:02.0000] <eighty4> Midler: your translation of it is better then mine :) [10:26:03.0000] <eighty4> lets go with yours [10:26:04.0000] <eighty4> and what's up with everyone being in sweden in here? [10:26:05.0000] <Midler> Sweden rocks? [10:26:06.0000] <Midler> :P [10:26:07.0000] <eighty4> /me votes for transforming this into a swedish channel [10:27:00.0000] <gsnedders> eighty4: This is a WHATWG channel, browser guys are here, and Opera has an office in Sweden? [10:27:01.0000] <Midler> eighty4: well, for once its really hot in sweden [10:27:02.0000] <gsnedders> s/,/,hence / [10:28:00.0000] <eighty4> gsnedders: doesn't Opera have offices in other countries as well? [10:28:01.0000] <gsnedders> eighty4: Transforming it into a Swedish channel probably wouldn't work seeming half of us in Sweden in this channel don't speak Swedish [10:28:02.0000] <eighty4> Midler: and I have to stay inside, it's around 30* in here... [10:28:03.0000] <gsnedders> eighty4: Yeah, sure. But core is mainly in Oslo and Linköping [10:28:04.0000] <eighty4> didn't know :) [10:29:00.0000] <gsnedders> (I think) [10:29:01.0000] <eighty4> so Linköping actually have something [10:29:02.0000] <Midler> opera is a norweigan company as far as i know [10:29:03.0000] <gsnedders> Midler: Yeah [10:29:04.0000] <eighty4> yeah, that's what I've learnt as well [10:29:05.0000] <Midler> But its not strange having an office in sweden, not atleast near linköping [10:29:06.0000] <Midler> as its near a university [10:29:07.0000] <gsnedders> Midler: Just by the railway station in the centre of Linköping :P [10:30:00.0000] <Midler> Iam not familiar with linköping [10:30:01.0000] <gsnedders> /me got here on Friday evening :P [10:31:00.0000] <Midler> gsnedders: Have you bought anything in a store yet? [10:31:01.0000] <gsnedders> Midler: yes [10:31:02.0000] <Midler> Because in norway, you would need to multiply the price by 2 or 3 [10:31:03.0000] <Midler> Than in sweden [10:31:04.0000] <gsnedders> /me doesn't really care about Norway :P [10:31:05.0000] <Midler> /me likes the mountains view in norway [10:31:06.0000] <Midler> But ive only been there once [10:32:00.0000] <Midler> When i was little [10:32:01.0000] <Midler> and it was not that fun to spend 3weeks of my weekly pocket money [10:32:02.0000] <Midler> :P [10:33:00.0000] <gsnedders> /me has never been to Norway [10:35:00.0000] <eighty4> gsnedders: I would put my money on a "classic" swedish basement [10:35:01.0000] <eighty4> you should be able to use your key to get into it [10:35:02.0000] <gsnedders> Yes, I know there is a key for it. [10:35:03.0000] <gsnedders> I don't know where it is though :P [10:35:04.0000] <eighty4> the washer might now be in the same "house" [10:36:00.0000] <eighty4> Midler: Oslo is nice [10:36:01.0000] <eighty4> but it's way more expansive [10:37:00.0000] <Midler> eighty4: as ive say. Ive only been in norway once [10:37:01.0000] <Dashiva> There are so many Swedes working in Norway you might almost think it's in Sweden :P [10:37:02.0000] <Midler> When i was a child [10:37:03.0000] <Midler> The only thing i apreciated was the view [10:37:04.0000] <Midler> However, being a child [10:37:05.0000] <Midler> Norway was really booring [10:37:06.0000] <Midler> It was a sunday [10:37:07.0000] <Midler> and nothing was open [10:38:00.0000] <Midler> And smal thing such icecream ruined me =( [10:38:01.0000] <eighty4> Dashiva: it _was_ sweden once [10:38:02.0000] <Midler> But i guess norway is fine yes [10:38:03.0000] <Midler> eighty4: yeah, decades ago... [10:39:00.0000] <Dashiva> Sweden was just a rebound after getting out of that abuse relationship with Denmark [10:44:00.0000] <eighty4> Midler: yeah yeah :) [10:50:00.0000] <Philip`> Hixie: \x1B$@<meta/\x1B(B charset=utf-8 x="<meta charset=iso-2022-jp>"> [10:50:01.0000] <Philip`> (where "\x1B" is the byte 0x1B) [10:52:00.0000] <Philip`> Decoding as UTF-8 gives charset=utf-8, decoding as ISO-2022-JP gives charset=iso-2022-jp, and ISO-2022-* looks like it would be considered "ASCII compatible" because all the normal characters (everything except 0x1B) are mapped identically to ASCII [10:56:00.0000] <Philip`> The charset johab/cp1361 (http://www.fileformat.info/info/charset/x-Johab/list.htm) might have similar problems, since ß is encoded as 0xDD 0x3C [11:31:00.0000] <Philip`> ISO-2022 is great fun [11:31:01.0000] <Philip`> http://search.yahoo.com/search?p=%E9%B9%BF%E5%A1%A2%E6%AF%80%E8%BB%A2%E9%A6%96%E7%AB%B0%E8%9C%B7%E5%9D%8F%20src=data:text/plain,alert(123)//&ei=UTF-8&eo=iso-2022-jp [11:32:00.0000] <Philip`> View that in a browser that doesn't support iso-2022, or manually switch the encoding to iso-8859-1 [11:32:01.0000] <Philip`> Unfortunately all the browsers I tested do support iso-2022 :-( [11:32:02.0000] <eighty4> "123" [11:32:03.0000] <Philip`> but anyway it's a bit of an XSS hole [11:46:00.0000] <hsivonen> just pushed the HTML5 parser into mozilla-central preffed off. Hoping everything is ok... [12:04:00.0000] <Philip`> http://search.yahoo.com/search?p=%EC%84%9E%E7%87%89%E4%B9%AD%ED%83%91%EC%88%8D%E8%A8%8A%E6%98%B1%E7%A6%BF%E5%88%B7%E5%A5%84%E5%89%94%EA%B5%89%EB%82%91%EC%84%9E%E6%98%8C%E4%BE%84%E5%BE%9E%EC%97%86&ei=UTF-8&eo=iso-2022-kr [12:04:01.0000] <Philip`> View that in Chrome on Windows [12:04:02.0000] <Philip`> and then it runs the script by default [12:05:00.0000] <Philip`> (Chrome 2.0.173.33 in particular) [12:05:01.0000] <Philip`> I hope this counts as a legitimate XSS vulnerability now [12:05:02.0000] <Philip`> only I've got no idea who to report it to [12:05:03.0000] <Philip`> so I'll just complain about it in here instead [12:07:00.0000] <Philip`> Hixie: I guess ISO-2022 should be forbidden/discouraged for the same security reasons as UTF-7 and EBCDIC [12:57:00.0000] <Midler> will it be possible to place for example div tag within video tag? [12:57:01.0000] <takkaria> as fallback for browsers which don't support <video>, yes [12:58:00.0000] <Midler> ok [12:59:00.0000] <Midler> iam just thinking of youtube that has "comments/links" withing the movie. I though of doing the same with div. [13:00:00.0000] <Midler> But its to use layers then [13:00:01.0000] <gsnedders> takkaria: How long did it take you to get to the lake? [13:00:02.0000] <Dashiva> Can you use video with imagemaps? [13:02:00.0000] <takkaria> gsnedders: well, my bike back wheel was flat when I got the canal [13:02:01.0000] <takkaria> so I walked [13:02:02.0000] <takkaria> I have no idea how long it was [13:03:00.0000] <gsnedders> heh [13:03:01.0000] <takkaria> I got back to my bike at 8pm, but I spent a fair bit of time at the lake napping or taking photos on the way [13:04:00.0000] <gsnedders> /me ended up more or less falling asleep at 5pm, oddly [13:04:01.0000] <jgraham> gsnedders: It is about 5km from the town centre to the lake [13:05:00.0000] <gsnedders> jgraham: I know taht [13:05:01.0000] <jgraham> Which is just over an hours walk [13:08:00.0000] <gsnedders> I tend to walk slower in the heat though, so probably a bit more than that for me [13:12:00.0000] <takkaria> apparently today I have taken some exceptionally wonky photos [13:12:01.0000] <jgraham> gsnedders: So does everyone I guess [13:13:00.0000] <gsnedders> takkaria: I seem to have done that over the past week too :( [13:13:01.0000] <jgraham> /me has lots of photos that he should, or should not, upload [13:13:02.0000] <Midler> If i want to get the state of the video with javascript (example: stopped, playing, paused ect). Where can i find it? [13:13:03.0000] <takkaria> and also lots of pictures which are the pictoral qequivalent of me going "wow! big expanse of flat, uninteresting land" [13:13:04.0000] <takkaria> which at the time I was obviously quite excited by, but I'm not so sure why now [13:14:00.0000] <gsnedders> :) [13:14:01.0000] <gsnedders> It is somewhat impressive quite how big the expanses of flat uninteresting land are though. [13:15:00.0000] <Midler> Iam thinking. If i dont like the custom controls or i want to use my own controls. I could do so with placing divs within the video overlayed. And with javascript activate and change blablabal.. [13:15:01.0000] <takkaria> yeah, hard to capture with a 50mm lens though [13:15:02.0000] <takkaria> Midler: http://www.whatwg.org/specs/web-apps/current-work/multipage/video.html#dom-media-paused [13:15:03.0000] <gsnedders> takkaria: You have 350D, right? [13:15:04.0000] <takkaria> gsnedders: aye [13:15:05.0000] <Midler> takkaria: thanks [13:16:00.0000] <gsnedders> takkaria: This is when the 10–22mm comes into use :P [13:16:01.0000] <takkaria> Midler: and the surrounding section [13:16:02.0000] <Midler> mm, i will find [13:16:03.0000] <jgraham> I have lots of photos that are the pictoral equivalent of "I got the wrong camera settings. Again." [13:16:04.0000] <Midler> I just thought of making my own controls [13:16:05.0000] <takkaria> /me grins jgraham [13:17:00.0000] <gsnedders> /me is tempted to get one of the Sigma fisheyes designed for APS-C sized sensor cameras [13:17:01.0000] <jgraham> /me is tempted to get many things but can't afford most of them [13:17:02.0000] <gsnedders> /me tends to be too conservative and boring to run into that often [13:18:00.0000] <Dashiva> Sure is deja vu on public-html lately [13:21:00.0000] <takkaria> isn't it always? [13:21:01.0000] <Philip`> It wasn't the first time each issue was discussed [13:22:00.0000] <Philip`> but after the second and third and fourth etc times, it does get a little tedious [13:24:00.0000] <tantek> Perhaps if the issues were captured and resolved on a wiki with permalinks, they could be quickly dismissed with URLs to those permalinks when brought up the second and third and fourth times (or at least new information could be added instead of rehashing). [13:24:01.0000] <tantek> http://microformats.org/wiki/wiki-better-than-email [13:26:00.0000] <Dashiva> I don't think it's lack of knowledge that's the main problem [13:27:00.0000] <Philip`> /me thinks the volume of communication (email and IRC and blog posts and blog comments and ...) is high enough that it's unreasonable to expect people to follow it all, and one can't expect them to not repeat earlier discussions unless there's a brief summary they can refer to to see that it's been said before [13:27:01.0000] <Dashiva> More of a "They don't agree with me yet, they must've missed it the previous time, so I'll just say it again until it works" [13:28:00.0000] <Dashiva> (From all sides involved) [13:33:00.0000] <Hixie> Philip`: the key is can you make it valid in both encodings. those examples all have at least one conformance error (e.g. unknown attributes) in at least one of the encodings. [13:33:01.0000] <Philip`> Hixie: So it has to be conforming HTML, not just a valid byte sequence in the given encoding? [13:33:02.0000] <Hixie> right [13:34:00.0000] <Hixie> Philip`: the point is to highlight these issues to authors [13:34:01.0000] <Hixie> Philip`: so if the author gets a conformance error, that's enough [13:34:02.0000] <Hixie> imho [13:35:00.0000] <Hixie> i've made 2022 discouraged like utf-32 and ebcdic [13:35:01.0000] <Hixie> i understand it's actually used quite widely, at least for e-mail and maybe for web [13:36:00.0000] <Hixie> Philip`: i also tweaked the text to make utf8+names not ascii-compatible [13:38:00.0000] <Hixie> hsivonen: btw, you probably don't want your classes and so forth to have the "5" in their name [13:39:00.0000] <hsivonen> Hixie: the nsHtml5 prefix is supposed to go away [13:39:01.0000] <Hixie> k [13:39:02.0000] <hsivonen> so as it happened, I was too much of a newbie with hg to do the landing right and now more experienced people are cleaning up the mess I caused [13:40:00.0000] <Hixie> :-/ [13:40:01.0000] <takkaria> oh, you're landing it on trunk? [13:40:02.0000] <takkaria> cool [13:47:00.0000] <Philip`> Hixie: Is <meta charset="utf-8>+names"> conforming (in both encodings)? [13:48:00.0000] <Philip`> It doesn't have any unrecognised attributes now, unless I'm being stupid [13:49:00.0000] <Philip`> http://philip.html5.org/data/charsets.html#charset-iso-2022-jp - hmm, that seems to be used quite a bit [13:49:01.0000] <Hixie> utf-8+names isn't ascii-compatible [13:49:02.0000] <Hixie> so <meta charset can't be used with it [13:49:03.0000] <Hixie> in a conforming page [13:50:00.0000] <Philip`> I thought the idea was to have an example that justifies the inclusion of that ASCII-compatibility requirement [13:51:00.0000] <Philip`> so it just needs to conform to all other requirements [13:51:01.0000] <Hixie> no the idea is to have an example that justifies changing the encoding sniffer to bail as soon as it hits a byte sequence that might screw things up [13:52:00.0000] <Hixie> or changing the ASCII-compatibility requirements to exclude even ascii bytes in later parts of multibyte sequences [14:01:00.0000] <Philip`> Hixie: Hmm, what about <meta charset=\x0Eutf8>x\x0Fiso2022kr> [14:01:01.0000] <Philip`> which is also <meta charset=朮歷없iso2022kr> [14:02:00.0000] <Hixie> how is that valid in 2022? [14:03:00.0000] <Philip`> 0x0E and 0x0F are escape codes in iso-2022-kr (http://tools.ietf.org/html/rfc1557) [14:04:00.0000] <Philip`> Oh, I suppose there's the problem that 朮歷없iso2022kr isn't the preferred name for that encoding [14:04:01.0000] <Hixie> right [14:04:02.0000] <Hixie> or any name for it as far as i can tell [14:05:00.0000] <Philip`> I thought anything [^a-zA-Z0-9] gets stripped when comparing names [14:05:01.0000] <Hixie> hm i guess it would compare correctly, yeah [14:05:02.0000] <Hixie> but it's still not valid [14:05:03.0000] <gsnedders> takkaria: http://www.flickr.com/photos/takkaria/3668544285/ — Now I'm not sure it is worth it. [14:05:04.0000] <Hixie> it's also not valid in UTF-8, is it? [14:06:00.0000] <Hixie> U+000E is a parse error [14:06:01.0000] <Hixie> i'm glad you're having as much trouble as i was with this [14:06:02.0000] <Hixie> makes me feel like i'm not a moron :-) [14:06:03.0000] <Hixie> and btw these encodings are insane [14:07:00.0000] <takkaria> gsnedders: I'm guessing it's more fun if you actually get to the lakeside, but I didn't because I wasn't quite sure how far down the road the lakeside was [14:23:00.0000] <Philip`> Hixie: If you want the document to be valid UTF-8 then it can't contain any non-whitespace bytes < 0x20, so it seems impossible to do anything interesting with any ISO-2022-* encodings, so I suppose that's a lost cause [14:24:00.0000] <Hixie> k [14:24:01.0000] <Philip`> (unless I'm wrong) [14:24:02.0000] <Hixie> that's more or less the conclusion i drew too [14:24:03.0000] <Hixie> i was playing with comments to hide some of the bytes, but yeah, even that wouldn't hide 0x1B. [15:34:00.0000] <gsnedders> Oh well, time to sleep, I think. See some of you tomorrow. [15:35:00.0000] <takkaria> ~ [15:35:01.0000] <gsnedders> takkaria: a tilde? :\ [15:35:02.0000] <takkaria> it's a wave [15:36:00.0000] <gsnedders> Oh, OK 2009-06-29 [18:44:00.0000] <annodomini> Are there any features in HTML5 (other than the different parsing rules) that are currently available only in the HTML syntax and not in the XHTML syntax? [19:04:00.0000] <MikeSmith> annodomini: lack of support for content from arbitrary namespaces [19:12:00.0000] <annodomini> MikeSmith: That's a feature missing from the HTML syntax, right? I'm wondering if I'll be missing anything (other than HTML parsing, and IE compatibility) if I use XHTML. [19:14:00.0000] <MikeSmith> annodomini: lack of support for content from arbitrary namespaces could be considered an actual feature, not a missing feature [19:14:01.0000] <MikeSmith> annodomini: anyway, no, you're not going to be missing anything [19:14:02.0000] <annodomini> Ah, now I see. [19:15:00.0000] <MikeSmith> but a big caveat is that if you are dynamically aggregating content from other places, you need to be very careful to make sure that content doesn't contain any well-formedness errors [19:16:00.0000] <annodomini> Right, that's why I mentioned the HTML parser. [19:20:00.0000] <Hixie> annodomini: <noscript> [19:20:01.0000] <Hixie> annodomini: maybe a few others [19:20:02.0000] <Hixie> annodomini: <svg> and <mathml> without the namespace stuff [19:22:00.0000] <annodomini> Hixie: Sure, you're going to have to use namespaces if you're using XML. I'm curious why <noscript> isn't supported? [19:23:00.0000] <Hixie> <noscript> is implemented as a weird parsing hack [19:23:01.0000] <Hixie> the parser actually acts differently for script-enabled browsers vs script-disabled [19:23:02.0000] <heycam> document.write() is another one [19:23:03.0000] <Hixie> which we can't do in xml [19:23:04.0000] <Hixie> yep, document.write() doesn't work in xml either [19:24:00.0000] <annodomini> Hmm. OK, so there's a non-trivial set of functionality that's not available in the XHTML syntax. Good to know. [19:26:00.0000] <Hixie> nothing good is missing :-) [19:34:00.0000] <annodomini> I'm basically wondering if I want to use XSLT to generate HTML5, if I should be trying to use XML output or HTML output. [19:34:01.0000] <annodomini> If it's just parser hacks like <noscript> and document.write(), I can probably work with XML. [19:35:00.0000] <Hixie> i'd suggest not using xslt, but xml output should be good if you do want to use it (not sure how it works with IE though) [19:36:00.0000] <annodomini> Why would you recommend against xslt? I'm just looking into it now, so I'm curious what the disadvantages are. [19:37:00.0000] <Hixie> i'm just biased against it because i get a headache when i try to use it :-) [19:37:01.0000] <heycam> i heartily recommend xslt! [19:37:02.0000] <heycam> xslt2, that is. xslt1 is slightly painful. [19:38:00.0000] <annodomini> (according to my googling, you can cram xhtml through IE if you lie about its mime type and leave off the XML processing instruction because it triggers quirks mode) [19:39:00.0000] <annodomini> Hixie: Ah, OK. I'll decide on whether it hurts my head once I use it a bit more. At a quick glance it looks fairly reasonable for using as a basic template system; I wouldn't want to use it for much more than that. [19:40:00.0000] <Hixie> i tried to use it to turn an xml file of numbers into an html file of slightly different numbers and that didn't go well [19:40:01.0000] <Hixie> that probably soured my opinion of it [19:40:02.0000] <Hixie> though generally speaking i'm scared of anything that's turing complete using angle brackets in the basic syntax [19:40:03.0000] <Hixie> e.g. also C++ templates :-) [19:44:00.0000] <annodomini> heycam: What are the benefits of xslt2? [19:44:01.0000] <annodomini> And do the browsers support it? I'm using browser xslt support for quickly testing thing out. [19:44:02.0000] <heycam> annodomini, two that i find useful are being able to define your own functions with <xsl:function> [19:45:00.0000] <heycam> and being able to tunnel template parameters [19:45:01.0000] <heycam> so you don't need to explicitly pass them in [19:45:02.0000] <heycam> i don't think any browsers support xslt2 [19:45:03.0000] <heycam> i just use saxon for offline document processing [19:45:04.0000] <MikeSmith> lack of document.write() support in XHTML is a feature [19:46:00.0000] <MikeSmith> support for document.write() in HTML is a bug [19:47:00.0000] <annodomini> Yeah, I don't mind losing that. [19:49:00.0000] <MikeSmith> annodomini: fwiw, I use XSLT1 for generating HTML for some things. It is painful in places, but I guess I've learned what the limitations and problems are and how to deal with them. [19:50:00.0000] <MikeSmith> I've been too lazy to learn XSLT2 [19:50:01.0000] <MikeSmith> when I have time, I plan to learn it by stealing from XSLT2 work that heycam did recently for generating the SVG spec [19:51:00.0000] <annodomini> I'm looking at it only as a basic template system, and not anything fancy, so I think I'll stick with XSLT1 for now. [19:52:00.0000] <heycam> annodomini, when you get to the point where you're writing recursive templates to step through strings a character at a time to do some simple string manipulation, that's when you know it's time to upgrade to xslt2 [19:52:01.0000] <MikeSmith> yeah [19:52:02.0000] <annodomini> Right. Sounds good. [19:52:03.0000] <MikeSmith> limitations on string manipulation in XSLT1 is a joke [19:53:00.0000] <MikeSmith> though there are some EXSLT things that most XSLT processors [19:53:01.0000] <MikeSmith> extensions that make some things easier [19:59:00.0000] <MikeSmith> Hixie: does the spec define where escaping text spans are allowed? [19:59:01.0000] <MikeSmith> are they only allowed in CDATA and RCDATA elements? [19:59:02.0000] <MikeSmith> or are they allowed in normal elements and foreign elements? [20:03:00.0000] <MikeSmith> hmm, I guess the answer is that anything that might look like any escaping text span in normal elements or foreign elements is actually a comment [20:04:00.0000] <MikeSmith> and that anything that looks like a comment in CDATA or RCDATA contents is actually an escaping text span [20:16:00.0000] <MikeSmith> Hixie: I think it's a mistake to use the term "CDATA elements" in the spec [20:16:01.0000] <MikeSmith> It would be worthwhile to try to come up with some better term [20:16:02.0000] <Hixie> file a bug if you come up with a better term :-) [20:16:03.0000] <Hixie> afk [20:17:00.0000] <MikeSmith> OK [20:18:00.0000] <MikeSmith> (problem with the word "CDATA" is that SGML jackasses managed to ensure that there would forever be ambiguity and confusion around it - http://www.flightlab.com/~joe/sgml/cdata.html) [21:28:00.0000] <Dashiva> Is it just me, or is Shelley sounding kinda like a conspiracy theorist lately... [23:40:00.0000] <annevk2> fwiw, UTF-8+names is not an actual encoding [23:40:01.0000] <annevk2> why did it end up in the spec? [23:41:00.0000] <hsivonen> feed autodetection is just a blog post, too :-) [23:42:00.0000] <hsivonen> but UTF-8+names definitely shoudln't be in any spec [23:42:01.0000] <hsivonen> having variants of UTF-8 around is evil [23:42:02.0000] <hsivonen> like CESU-8 [23:46:00.0000] <Dashiva> If they're going to move it into unicode, doesn't it make sense to use the unicode names? [23:48:00.0000] <hsivonen> whose "they" and is UTF-8+names anything but a blog post that went nowhere? [23:48:01.0000] <Dashiva> Apparently there's an i-d [00:16:00.0000] <Philip`> Dashiva: There's a thing written with xml2rfc and linked from a blog post, but that seems to be about it [01:29:00.0000] <takkaria> all these debates about data on htmlwg remind me of a line from Stephen Colbert [01:31:00.0000] <takkaria> "Statistics are just collections of facts, facts about the real world, and as we all know, reality has a liberal bias" [01:32:00.0000] <MikeSmith> heh [01:32:01.0000] <MikeSmith> brilliant [01:32:02.0000] <MikeSmith> takkaria: you quoted that wrong [01:32:03.0000] <MikeSmith> [[ [01:32:04.0000] <takkaria> I'm sure I did, I couldn't find an accurate quote online [01:32:05.0000] <MikeSmith> takkaria: no, not that [01:33:00.0000] <MikeSmith> [[ [01:33:01.0000] <MikeSmith> Statistics are just collections of "facts", "facts" about the real world, and as we all know, "reality" has a liberal bias. [01:33:02.0000] <MikeSmith> ]] [01:33:03.0000] <MikeSmith> never forget your scare quotes, grasshopper [01:33:04.0000] <MikeSmith> have you learned nothing for your Shaolin training? [01:34:00.0000] <takkaria> no :'( [01:34:01.0000] <takkaria> it is pretty bad when people making serious arguments start reminding me of Colbert, though [01:36:00.0000] <MikeSmith> takkaria: well, arguments of partisans from both/all sides remind me of Colbert sometimes [01:36:01.0000] <MikeSmith> or Beckett [01:36:02.0000] <MikeSmith> and everybody gets their turn to abuse the scare quotes [01:36:03.0000] <MikeSmith> that's the great fun of them [01:37:00.0000] <MikeSmith> scare quotes discriminate against no one [01:37:01.0000] <MikeSmith> they welcome all of us to use the freely and liberally when they suit or purposes or agenda [01:38:00.0000] <Philip`> As long as nobody dares use unbalanced quotes, I'm happy [01:40:00.0000] <Philip`> Hmph, Google retiring Page Creator will break some of my canvas tests :-( [01:40:01.0000] <MikeSmith> Philip`: that makes me think, I dispense with using scare quotes at all, and just say at the end of each message, (scare quotes implied) [01:40:02.0000] <MikeSmith> that leaves it up to reader to insert them where necessary [01:40:03.0000] <MikeSmith> adds some interactivity [01:41:00.0000] <Philip`> Adds some "interactivity", eh? [01:42:00.0000] <Philip`> Wow, you're right [01:42:01.0000] <MikeSmith> you got it, man [01:42:02.0000] <MikeSmith> we're starting a new trend [01:42:03.0000] <MikeSmith> #whatwg innovating, as usual (scare quotes implied) [01:43:00.0000] <takkaria> /me grins [01:43:01.0000] <takkaria> /me grins (sscare quotes implied) [01:43:02.0000] <takkaria> yeah, I prefer the latter [01:43:03.0000] <MikeSmith> heh [01:43:04.0000] <MikeSmith> on that note, I got to drop off [01:43:05.0000] <MikeSmith> back later [01:43:06.0000] <Dashiva> "takkaria" grins [02:29:00.0000] <takkaria> Dashiva: how Zen [02:51:00.0000] <jgraham> BTW I had a look at the mozilla layout table detection code and it is really really simple but hard to implement outside a browser because it requires access to computed style [02:51:01.0000] <jgraham> I supsect it is not that effective but it would be interesting to test [02:51:02.0000] <takkaria> heh, Shelley is brilliant [02:53:00.0000] <Philip`> jgraham: That also makes it hard to execute on a large sample of pages, because you need to have downloaded all the stylesheets too, so you can only really run it live and not on pre-downloaded sets of pages [02:54:00.0000] <jgraham> Philip`: Yes, I know [02:57:00.0000] <Philip`> takkaria: Because of some specific quote, or just in general? [02:58:00.0000] <takkaria> "However, perhaps if there were more people involved in both the effort and the decision process, you wouldn't have to resort to such trickery. Because this form of trickery will not succeed.", amongst others [02:58:01.0000] <Philip`> s/quote/email or blog post or blog comment or Twittering/ [02:58:02.0000] <takkaria> and the brilliant accusations of Hixie having a bias against experts [02:59:00.0000] <Hixie> i couldn't work out what she meant by trickery, that came out of left field for me [02:59:01.0000] <takkaria> I would love to know what Shelley's mental caricature of Hixie is [03:00:00.0000] <Philip`> takkaria: Ah, okay - I didn't bother reading any emails in that thread that were longer than my screen [03:00:01.0000] <takkaria> Hixie: I can't figure it out either, maybe she's accusing you of tricking people into thinking you're unbiased or something? [03:01:00.0000] <takkaria> but it won't succeed, so you shouldn't bother trying [03:03:00.0000] <Dashiva> Sometimes I get curious about what would happen if Hixie said "Okay, whatever" and stopped editing the W3C version. It's a shame we can't spawn a parallel universe to see the outcome without having to suffer it ourselves. [03:06:00.0000] <takkaria> Nick Levinson is also an eternal source of amusement [03:07:00.0000] <takkaria> I think he must just not quite grok how the web works [03:08:00.0000] <Dashiva> Does he realize that downloading a web page to view it is already copying the contents? We're pirates just by using browsers. [03:12:00.0000] <takkaria> there's a FOX News clip I saw where there's a congressman talking about his effort to make 2010 the "Year of the Bible". there's a bit in it where he talks about the government "stealing our children's futures" by the way we're dealing with the recession, and I couldn't help thinking "it's alright, they'll find new ones on piratebay" [03:22:00.0000] <Philip`> Fortunately human population numbers are increasing, so if we steal from future generations then we'll each get a significant benefit and the costs will be much more diluted by the time future generations have to pay for them [03:23:00.0000] <jgraham> Philip`: Aren't populations only increasing in poor areas of the world? [03:23:01.0000] <jgraham> s/only/mainly/ I guess [03:23:02.0000] <Philip`> Also, standards of living increase over time, so we're stealing from the rich (i.e. future people) and giving to the poor (i.e. us) [03:24:00.0000] <takkaria> http://www.youtube.com/watch?v=F25hYD5vs_4 it's quite a brilliant little clip actually [03:24:01.0000] <Philip`> jgraham: That's a bonus, since those people will end up paying a larger share of the costs instead of our children [03:24:02.0000] <takkaria> he tries to sell making 2010 year of the bible as "it's all about freedom, really" [03:25:00.0000] <Dashiva> Freedom to follow rules [03:27:00.0000] <takkaria> he even lets himself go on a mini-rant about the state taking all of our money and private property at teh end, proper republican stuff [03:27:01.0000] <takkaria> and presumably our children's futures, too [03:28:00.0000] <Philip`> Does everyone on American TV always shout, or is it just bad audio balancing in that clip? [03:28:01.0000] <takkaria> on Fox, they generally shout [03:29:00.0000] <Dashiva> Is anything Nick wrote in this last post on-topic? [03:30:00.0000] <remysharp> hi there, regarding the implementation status popup on the spec page - are the implementation statuses maintained, i.e. updated regularly? [03:30:01.0000] <remysharp> I'm specifically looking at DnD - my tests have it working in Firefox 3.5 only - and not webkit as the docs suggest [03:32:00.0000] <Philip`> remysharp: http://developer.apple.com/documentation/appleapplications/conceptual/safarijsprogtopics/Tasks/DragAndDrop.html makes it sound like it's implemented in WebKit [03:35:00.0000] <remysharp> Looks aren't everything ;-) [03:35:01.0000] <remysharp> s/look/sound/ [03:35:02.0000] <remysharp> darnit - anyway. [03:36:00.0000] <remysharp> no - it doesn't work for love nor money when I've tried it. [03:36:01.0000] <Rik`> remysharp: webkit implements DnD pretty much like IE [03:36:02.0000] <remysharp> and I believe I've tried taking fairly tiny steps [03:37:00.0000] <Rik`> remysharp: http://www.whatwg.org/demos/2008-sept/dnd/dnd.html [03:40:00.0000] <remysharp> right, that's perfect - thanks. I need to compare against what and where I got my code from... [03:40:01.0000] <remysharp> thanks! [03:55:00.0000] <remysharp> sorry Rik` one last thing - I've just been testing now, and finding that I can't drag a div (which is the source of all my problems!) is that outlined in the specs, and I've just missed it? [03:56:00.0000] <Rik`> in webkit you'll have to use -webkit-user-drag: element [03:57:00.0000] <remysharp> good lord. okay, thanks - perfect - that cracked it [03:58:00.0000] <remysharp> native DnD is suppose to ease the burden of having pure JS solutions, but having to jump though random hoops like that (via webkit) just make it harder :-( [03:58:01.0000] <Rik`> this is a recent API [03:59:00.0000] <remysharp> yeah, I guess that's fair enough. [03:59:01.0000] <remysharp> one more question if you don't mind? [04:00:00.0000] <remysharp> I couldn't get the DnD to start in Firefox unless I specifically used dataTransfer.setData - is it a requirement that /something/ should be including in the drag [04:00:01.0000] <remysharp> ? [04:01:00.0000] <Rik`> i don't know [04:01:01.0000] <remysharp> I meant more from the spec pov - or same answer? :-) [04:01:02.0000] <Rik`> the mozilla documentation is pretty complete normally [04:01:03.0000] <Rik`> I just played with that on friday which is why I have some answers :) [05:02:00.0000] <Hixie> ok, content models should be easier to navigate now [05:02:01.0000] <Hixie> because the category definitions have the list of elements that aply [05:02:02.0000] <Hixie> apply [05:07:00.0000] <Lachy> Hixie, it would be easier to have a table of all elements that indicated which categories each is in [05:07:01.0000] <Hixie> we'll have that later too [05:07:02.0000] <Lachy> I had planned to make a table for that [05:07:03.0000] <Hixie> http://www.whatwg.org/specs/web-apps/current-work/#index [05:07:04.0000] <Lachy> ok [05:29:00.0000] <takkaria> hmm, abarth's sniffing draft doesn't define behaviour for application/octet-stream [05:30:00.0000] <gavin_> is that ever sniffed? [05:30:01.0000] <gavin_> I thought that was consistently ignored by all sniffers [05:31:00.0000] <takkaria> not by Opera, it seems [05:31:01.0000] <takkaria> and IE's docs say that it tries to determine whether it's text or binary and does stuff based on that [05:33:00.0000] <takkaria> it certainly seems that it's worth sniffing some application/octet-stream stuff, like video or audio, just so the UA knows what app to hand it off to [05:33:01.0000] <Lachy> http://html5doctor.com/ [06:18:00.0000] <Lachy> http://stackoverflow.com/questions/1055214/is-there-a-way-to-make-html5-video-fullscreen [06:19:00.0000] <Lachy> it's a shame we don't have an API to make full screen videos [06:19:01.0000] <Lachy> IMHO, the arguments against doing so are bogus, since Flash can already do it [06:21:00.0000] <Philip`> Do the arguments include the argument that you could take a screenshot of a browser then put it in <video> then make it fullscreen and spoof the browser chrome? [06:21:01.0000] <Lachy> although, it's entirely possible for browsers to provide their own full screen UI [06:22:00.0000] <Philip`> At least Flash has a popup message saying to press Escape to leave fullscreen mode, or whatever it says [06:22:01.0000] <Philip`> (though irritatingly my keyboard's volume keys also cause it to leave fullscreen mode) [06:22:02.0000] <Lachy> sure, and there's no reason a browser couldn't do something similar [06:23:00.0000] <Lachy> a browser would likeliy have to show video controls when the mouse moves anyway, which makes it even more difficult for such an attack to be successful [06:23:01.0000] <roc> there are clever ways to neutralize that popup message [06:23:02.0000] <Lachy> roc, how? [06:24:00.0000] <roc> I can't find the link [06:25:00.0000] <Lachy> if it is possible, it seems like it would have to be a bug in Flash [06:25:01.0000] <roc> basically you display a lot of messages on the screen that look just like Flash's one [06:25:02.0000] <roc> the user is totally confused and doesn't know which one to read [06:25:03.0000] <roc> then it all goes away and you carry on [06:26:00.0000] <gavin_> it's a bug in humans ;) [06:26:01.0000] <roc> I do think we should support fullscreen video [06:27:00.0000] <roc> but I think it shouldn't be under script control [06:27:01.0000] <Philip`> You could show some text saying "Close your eyes for five seconds, then open them again for a surprise!", then switch to fullscreen while the user's eyes are closed [06:27:02.0000] <Lachy> ok, file a bug with The Human Genome Project to get that fixed. [06:28:00.0000] <roc> alright, sleep time [06:28:01.0000] <Lachy> Philip`, browsers would have to apply techniques similar to popup blockers, so that a script can't enter full screen without the user clicking [06:28:02.0000] <Lachy> (or some other user interaction) [06:29:00.0000] <Philip`> You could show some text saying "Close your eyes, click your mouse, wait five seconds, then open your eyes again for a surprise!", then switch to fullscreen while the user's eyes are closed [06:29:01.0000] <Lachy> browsers should of course provide their own native full screen UI, such as double clicking the video or selecting full screen from the context menu [06:30:00.0000] <Lachy> but then the user would probably move their mouse, see the video controls and realise that it's a video. [06:31:00.0000] <Lachy> Also, the attacker would have a hard time mimicking a user's individual desktop to to fool the user that easily [06:33:00.0000] <Lachy> basically, until someone can deomonstrate a practical attack vector that isn't simply hypothesising about ludicrous scenarious, I won't believe it's a serious issue [06:37:00.0000] <Lachy> also, the attack has to have some benefit. Simply being able to show something on a user's screen isn't really useful, unless it could somehow be used to, e.g., steal passwords or gain unauthorised access [06:38:00.0000] <rocSleep> I sympathize with that point of view. However, there have been several cases of people saying that, and then going ahead and standardizing behaviour which everyone ends up depending on, and then those ludicrous scenarios become reality [06:38:01.0000] <rocSleep> :visited-based history sniffing for example [06:39:00.0000] <Philip`> I'd guess a useful fullscreen mode would need to allow pages to supply custom controls, for the same reasons that non-fullscreen <video> allows custom controls [06:39:01.0000] <Lachy> rocSleep, not doing something because of some unproven hypothetical attack vectors that don't accually allow an attacker to do anything is silly [06:39:02.0000] <Philip`> in which case you could use the custom controls to interact with the fooled user [06:39:03.0000] <Lachy> Philip`, no custom controls on full screen video [06:40:00.0000] <Philip`> Lachy: That makes it much less useful [06:40:01.0000] <Lachy> that would be a silly and unnecessary thing to allow [06:40:02.0000] <gavin_> I don't think fullscreen video would need to allow custom controls to be useful [06:40:03.0000] <Lachy> no it doesn't [06:40:04.0000] <Philip`> Why would it be more silly and unnecessary than on non-fullscreen videos? [06:40:05.0000] <Lachy> when a user goes full screen, the browser's native controls should be sufficient [06:40:06.0000] <gavin_> and I don't think a custom-control-less fullscreen would be "must less useful" [06:41:00.0000] <gavin_> just like I don't think custom-control-less normal <video> isn't "must less useful" [06:41:01.0000] <Lachy> because custom controls generally need to integrate with the page design. But in full screen mode, the rest of the page isn't visible, so that [06:41:02.0000] <Lachy> isn't as necessary [06:42:00.0000] <Lachy> it would also be much more a much more difficult thing to do, since you'd have to somehow allow the user to overlay HTML content on top of the full screen video, which is outside of the normal browser viewport [06:43:00.0000] <Lachy> s/the user/the author/ [06:44:00.0000] <gavin_> bah, how did I manage to type "must" intead of "much" twice? [06:45:00.0000] <Philip`> So if e.g. the BBC iPlayer wants to have a button to enable display of subtitles, and wants to switch from Flash to <video>, it will no longer be able to do that? [06:45:01.0000] <Philip`> (...in fullscreen mode) [06:45:02.0000] <Lachy> hmm, I didn't notice. I read "much" in both cases anyway :-) [06:46:00.0000] <gavin_> perhaps not [06:46:01.0000] <gavin_> but that doesn't mean that the feature is "much less useful" [06:46:02.0000] <Lachy> the browser would have to provide UI for enabling subtitles [06:46:03.0000] <gavin_> only that it doesn't address that specific use case [06:46:04.0000] <Philip`> Well, "less useful" then :-) [06:46:05.0000] <Lachy> Philip`, can you send me a link to a iplayer video with subtitles so that I can see how it works? [06:46:06.0000] <Philip`> Lachy: Not unless you have a UK IP address [06:47:00.0000] <Lachy> /me connects to a UK based VPN to bypass the regional restrictions [06:47:01.0000] <Philip`> Ah [06:47:02.0000] <Philip`> Lachy: Try e.g. http://www.bbc.co.uk/iplayer/episode/b00lkmth/Top_Gear_Series_13_Episode_2_(new_series)/ and then play and wait a few seconds until the actual programme starts [06:48:00.0000] <Philip`> and then look for the "S" button [07:40:00.0000] <jgraham> I wonder if it is OK to make new years resolutions in June? [07:41:00.0000] <jmb> sure. just define a year as the period of time running from one June to the next [07:41:01.0000] <Philip`> The start of the year is entirely arbitrary, so it's just as good to pick any other time [07:42:00.0000] <Philip`> The length of the year isn't so arbitrary, given the planet we live on - you might be considered weird if you intend your resolution to last for a Martian year instead [07:45:00.0000] <jgraham> The length of the year seems irrelevant since new year's resolutions are defined to start at a time and theoretically last forever but in practice generaly last a few days [07:48:00.0000] <Philip`> Oh, okay [07:48:01.0000] <Philip`> But Wikipedia says "While the traditional time to make New Year's resolutions is New Year's Day, it is perfectly acceptable to wait until late January, or even early February to make these goals." so presumably June is unacceptable [07:50:00.0000] <jgraham> Oh. well that's no use [07:50:01.0000] <Phae> you could lump it in with general "spring cleaning" type jobs. [07:52:00.0000] <gsnedders> /me hopes he has a working computer now [07:53:00.0000] <Philip`> gsnedders: Your presence here suggests that you do [07:56:00.0000] <Philip`> "Thank you for your past interest in VMware. As part of our routine scheduled maintenance, we will be removing email addresses and associated subscription information from our marketing database for contacts who have not updated their profile and/or subscription preferences within the last 6 months." [07:56:01.0000] <Philip`> That's probably the first time I've ever seen a company intentionally reduce the reach of their email marketing [08:06:00.0000] <Lachy> jgraham, the financial year ends at the end of June in some countries, so you could do a New Financial Year's Resolution [08:18:00.0000] <gsnedders> What's a sane PDF reader for Ubuntu? [08:19:00.0000] <hsivonen> the bundled one [08:20:00.0000] <Philip`> cat and the PDF file format specification [08:23:00.0000] <Lachy> LOL [08:25:00.0000] <jgraham> It seems like the PDF specification is only avaliable as a PDF, which presents a bootstrapping problem [08:25:01.0000] <Philip`> That's not a problem - you can use Acrobat Reader to print the PDF specification [08:26:00.0000] <Philip`> then use the printed spec and cat as a more sane PDF reader [08:27:00.0000] <gsnedders> jgraham: That's no different to HTML, but I think PDF claims to be logical. [08:28:00.0000] <Lachy> gsnedders, at least with HTML specifications, you can read HTML 2.0, implement that, which then allows you to read HTML 3.2 and 4.01, and after implementing those, you can implement HTML5, throw out your previous work on HTML 2.0, 3.2 and 4.01, and you're all done [08:29:00.0000] <gsnedders> Lachy: Or you could just download the PDF version of HTML 5 to start with [08:30:00.0000] <Lachy> but you first need to find the link to the PDF version, which is linked from the HTML version [08:30:01.0000] <gsnedders> Lachy: HTML can be read well enough as a text format though [08:30:02.0000] <Philip`> Someone should print out a copy of the HTML5 spec, then drop it in a trench on a building site, so that a future generation's equivalent of Time Team will dig it up even if there's been an intervening collapse of civilisation and loss of digital information [08:31:00.0000] <gsnedders> Under the rubble, preferably. [08:31:01.0000] <Lachy> you're assuming the person attempting to read the specification is familiar with some form of HTML, and won't be frightened away by all that markup [08:34:00.0000] <Dashiva> Also, if they don't have a browser, they probably don't have internet access to the PDF version either [08:36:00.0000] <Lachy> Dashiva, wget [08:36:01.0000] <Philip`> Dashiva: gopher [08:36:02.0000] <Lachy> Philip`, is the spec even available through the gopher protocol? [08:37:00.0000] <Dashiva> Lachy: If they can wget the spec, they can wget a browser [08:37:01.0000] <Philip`> Lachy: Don't know; if not, someone should set up a server for it [08:38:00.0000] <hsivonen> wonders if Workers address the threading use cases of Google Native Client [08:38:01.0000] <Lachy> Philip`, it depends if someone can create some market demand for such a service [08:38:02.0000] <hsivonen> /me wonders... [08:39:00.0000] <Philip`> Lachy: It doesn't need any market demand, it just needs someone sufficiently bored and strange to set it up [08:40:00.0000] <Lachy> Philip`, someone like you, perhaps? [08:40:01.0000] <Philip`> Dashiva: Maybe they're using an IP-over-SMS protocol and will be bankrupt before they manage to download an entire browser, but would just about be able to download the HTML5 spec [08:40:02.0000] <Philip`> Lachy: I'm not sufficiently bored at the moment [08:42:00.0000] <Lachy> I wonder if anyone has implemented an IP-over-SMS service yet? [08:44:00.0000] <Lachy> I wonder if anyone would be bored enough to manually type in the bytes on their phone to construct an IP packet that is itself divided into 160 byte sequences? [08:45:00.0000] <Philip`> You could build a device out of Lego to press the buttons for you automatically [08:45:01.0000] <Dashiva> Surely you'd want to use base64 instead :) [08:47:00.0000] <Philip`> Depends on whether you want to optimise bandwidth usage or data entry time [08:48:00.0000] <Lachy> Philip`, given the limitations of character entry on common mobile devices, encoding the entire packet as base64 might be the only viable option [08:48:01.0000] <Philip`> e.g. it'd probably be much quicker to turn off predictive typing and use base10, but you'd have to send more messages [08:49:00.0000] <Lachy> well, not the only viable option, since you could conceivably create your own encoding scheme [08:49:01.0000] <Philip`> I don't think my phone's dictionary contains SGVsbG8gd29ybGQ= so it would complain a lot when I tried to type it in [08:49:02.0000] <Lachy> but that would be a silly waste of time (unlike the rest of this project) [08:50:00.0000] <Dashiva> How about just sending one SMS per byte, having the message contain the byte value as a number 0-255? [08:50:01.0000] <Dashiva> Then you could have jumbo SMSes where you type more than one number in each SMS, separated by non-digits [08:51:00.0000] <Philip`> SMSs are a bit unreliable so you'd probably have to add sequence numbers and ACKs [08:51:01.0000] <Philip`> else the higher layers aren't going to cope very well with all the packet corruption [08:53:00.0000] <Lachy> Dashiva, once you have IP, you can then use TCP for that [08:53:01.0000] <Philip`> You need each SMS to contain the octet, the sequence number, and a SHA512 HMAC to ensure integrity, and then it'll be fine [08:54:00.0000] <Dashiva> And the ACK? [08:55:00.0000] <Lachy> you could get away with just the sequence number. Adding all the rest would take up too much of each message's 160 byte sequence [08:56:00.0000] <Philip`> Lachy: SMS probably loses more than one message out of every ~1500, and TCP won't recover from 100% packet loss, so you need some layer 2 error-correction mechanism [08:57:00.0000] <Lachy> if you get a 100% packet loss, then the connection would simply timeout and you could assume transmission failure [08:57:01.0000] <Philip`> Also, TCP won't help at all with UDP applications like real-time games and Skype [08:57:02.0000] <Lachy> SMS-over-IP isn't designed for real time applications [08:58:00.0000] <Philip`> Lachy: If transmission failure is pretty much guaranteed, it's a bit of a rubbish protocol [08:58:01.0000] <Dashiva> You could do a regular call and use voice software to read out the bytes [08:58:02.0000] <Lachy> oops, IP over SMS [08:58:03.0000] <Philip`> so we ought to design it to work around those problems [08:59:00.0000] <Philip`> Dashiva: Good thinking - the voice channel should be much more reliable than SMS [08:59:01.0000] <Lachy> I'd love it if someone wrote an April Fools RFC for this protocol :-) [09:00:00.0000] <takkaria> mm, Skype-over-IP-over-SMS would be beautiful [09:01:00.0000] <Lachy> hey, you could use Twitter for multicasting [09:01:01.0000] <gavin_> I imagine most SMS traffic ends up being sent over an IP network at some point [09:01:02.0000] <Lachy> but then you'd be restricted to 140 character sequences [09:01:03.0000] <gavin_> so really it's IP-over-SMS-over-IP [09:03:00.0000] <Philip`> Mmm, IP-over-Twitter [09:04:00.0000] <gsnedders> takkaria: You're upstairs, right? [09:04:01.0000] <takkaria> gsnedders: yeah [09:05:00.0000] <hsivonen> We should move TPAC to Sweden [09:05:01.0000] <Lachy> IP-over-twitter would be the same as IP-over-SMS, except that the IP-over-twitter protocol would need to specify a way to detect and strip the username from the beginning of each message [09:06:00.0000] <Philip`> Presumably IP-over-SMS and IP-over-Twitter would be able to interoperate, using Twitter's SMS features [09:06:01.0000] <hsivonen> /me notes that twitter applies the length limit by UTF-16 (or characters?) while SMS applies it by bytes [09:07:00.0000] <hsivonen> so twitter gives you 140 characters even if you use stuff that's not in the original GSM encoding and turns SMS into an 80-code unit UTF-16 message [09:07:01.0000] <hsivonen> /me wonders how twitter's gateway deals with that [09:08:00.0000] <Dashiva> Splits into two messages? :) [09:10:00.0000] <Philip`> Truncates? Crashes? [09:11:00.0000] <Lachy> someone should test it. I don't have Twitter sending me SMSs, so i can't [09:12:00.0000] <Philip`> Apparently SMS does "160 7-bit characters, 140 8-bit characters, or 70 16-bit characters" [09:13:00.0000] <Philip`> (since it's 140 octets) [09:13:01.0000] <Philip`> (according to http://en.wikipedia.org/wiki/SMS#Message_size) [09:14:00.0000] <gsnedders> hsivonen: Twitter applies NFC, then limits by number of codepoints used. [09:14:01.0000] <gsnedders> (or at least did a few months ago when I got bored and tried to work it out) [09:16:00.0000] <Lachy> well, if we're encoding the message as Base64, then normalising to NFC won't actually affect the message anyway [09:17:00.0000] <Philip`> The SMS spec says it's "UCS-2" and references ISO/IEC 10646 [09:18:00.0000] <hsivonen> Philip`: what's the 8-bit encoding? [09:18:01.0000] <takkaria> can someone with IE8 go to http://takkaria.org/tests/octet-stream/ and see if hitting 'Start' does anything? [09:18:02.0000] <Philip`> and doesn't give any details on e.g. how surrogates should be handled [09:18:03.0000] <hsivonen> the 160 char thing works with accented Western European chars [09:18:04.0000] <Philip`> hsivonen: "8 bit data is user defined" [09:19:00.0000] <hsivonen> ah. [09:19:01.0000] <Philip`> hsivonen: "Character table: User Specific" [09:19:02.0000] <hsivonen> Russian goes the 16 bit path [09:19:03.0000] <takkaria> yeah, it kinda relies on different phones using the same charsets [09:19:04.0000] <hsivonen> which is why it pays off to use ad hoc Romanizations of Russian [09:19:05.0000] <Lachy> I would assume the defacto standard would probably be Windows-1252 [09:19:06.0000] <Lachy> at least in Western markets [09:19:07.0000] <hsivonen> Lachy: I wouldn't assume so [09:20:00.0000] <Philip`> The table in http://en.wikipedia.org/wiki/SMS#GSM shows the 7-bit characters [09:20:01.0000] <hsivonen> GSM had a crazy legacy of its own [09:20:02.0000] <Philip`> 8-bit sounds like it's for data (e.g. sending photos), not text [09:21:00.0000] <Philip`> They've got a whole two bits for identifying the character encoding, so they could expand it with an extra choice in the future [09:21:01.0000] <Lachy> wow, I always thought SMS just used ASCII [09:22:00.0000] <hsivonen> Lachy: it was invented by people who had accented use cases [09:23:00.0000] <Philip`> The 7-bit thing also has escape codes to shift into a different character set, with three currently defined (Turkish, Spanish, Portugese) [09:27:00.0000] <Philip`> If you use the single shift mechanism, then sending 0x1B 0x?? will look up the second byte in the character table for the message's language (defined in a header field) [09:28:00.0000] <Philip`> or there's another mode which uses a language-specific table by default, and I think uses 0x1B to escape to the standard 7-bit table [09:29:00.0000] <Philip`> This system does sound somewhat evolved, i.e. an accumulation of hacks [09:34:00.0000] <Philip`> /me was looking at http://www.3gpp.org/ftp/Specs/archive/23_series/23.038/23038-820.zip via ftp://ftp.3gpp.org/Specs/html-info/23038.htm via http://en.wikipedia.org/wiki/SMS#cite_note-3GPP_23.038-25 [09:38:00.0000] <hsivonen> hmm. I guess if Google Native Client takes off, it'll further cement the position of x86 instructions as the universal byte code that future CPUs need to implement [09:39:00.0000] <hsivonen> /me wonders how GCC & Native Client compares to javac & HotSpot [09:41:00.0000] <Philip`> It's particularly bad if people use NaCl to do computation-intensive work (which is its main purpose) using SSE/SSE2/etc [09:41:01.0000] <Philip`> because they're probably not much fun to emulate efficiently on a different CPU architecture [09:43:00.0000] <hsivonen> are there autovectorizers for any non-vector languages and do they work? [09:44:00.0000] <Philip`> The Intel C++ Compiler can autovectorise loops [09:44:01.0000] <Philip`> and sometimes it works [09:45:00.0000] <Philip`> http://software.intel.com/en-us/articles/vectorization-with-the-intel-compilers-part-i/ [09:48:00.0000] <Philip`> (As far as I'm aware, anyone who actually cares about vectorisation in C++ will use explicit SIMD intrinsics (or write assembly code) because compilers are rarely magic enough) [09:50:00.0000] <hsivonen> ok [09:52:00.0000] <Philip`> Hmm, apparently GCC has -ftree-vectorize [09:53:00.0000] <Philip`> which apparently works for C and C++ and Fortran and Java etc [09:55:00.0000] <Philip`> (to a limited, buggy extent) [09:58:00.0000] <takkaria> gsnedders: I'm no longer upstairs [12:34:00.0000] <gavin_> Lachy: http://www.bunnyhero.org/2008/05/10/scaring-people-with-fullscreen/ fwiw [12:47:00.0000] <Lachy> gavin_, as I expected, a totally ineffective and useless attack [12:48:00.0000] <Lachy> since the "Press ESC..." message appears right in the middle of the screen, it's the most visible [12:48:01.0000] <gavin_> I couldn't find it the first two times I tried [12:48:02.0000] <gavin_> (was looking at the center) [12:48:03.0000] <Lachy> granted, the demo isn't the best since it uses jibberish for the other text, making them easy to ignore [12:49:00.0000] <Lachy> I don't know how it's possible to miss, it's so obvious [12:50:00.0000] <Lachy> also, the way I suggested implementing full screen mode in browsers would make that attack virtually ineffective, since pressing any keys would reveal the browser's native video controls [13:15:00.0000] <Midler> gavin_: i couldnt find it either [13:15:01.0000] <Midler> however [13:16:00.0000] <Midler> making a random white & black dots (lika ants war on tv) [13:16:01.0000] <Midler> might make it harder to find [13:19:00.0000] <Midler> does html5 support webforms 2.0? [13:20:00.0000] <Lachy> wow, Midler, I really don't get how it's even remotely difficult to see. It might be harder if the text boxes contained readable information [13:20:01.0000] <Midler> not me asking [13:20:02.0000] <Midler> I dont even know what it is [13:20:03.0000] <Lachy> Midler, webforms 2 stuff was integrated into HTML5 already [13:20:04.0000] <Midler> ok [13:21:00.0000] <Lachy> basically, it's all the form controls and attributes, like <input>, <select>, <textarea>, and their associated APIs [13:21:01.0000] <Midler> aha [13:21:02.0000] <Midler> well, iam a kind of a newbie =) [13:22:00.0000] <Midler> therefore i hang around in here, hope to learn something [13:22:01.0000] <Midler> =) [13:36:00.0000] <Midler> Lachy: was it webforms that was shown in the google tech talk video? [13:43:00.0000] <Lachy> Midler, if it was about new form controls in HTML5, yes [13:43:01.0000] <Midler> ok [14:48:00.0000] <takkaria> so, Firefox doesn't sniff application/octet-stream, Safari 4 apparently only sniffs application/octet-stream files which look like HTML, and Opera sniffs lots of thigs sent as application/octet-stream depending on file extension [14:58:00.0000] <takkaria> I'm pretty sure sniffing application/octet-stream as text/html is a potential security hole, actually [14:59:00.0000] <Lachy> what does IE do? [14:59:01.0000] <takkaria> no idea, I've not got any way to run it ATM [14:59:02.0000] <Lachy> ok [14:59:03.0000] <takkaria> you could run http://takkaria.org/tests/octet-stream/ if you wanted [15:00:00.0000] <takkaria> I assume it will work on IE, I don't think I use anything it can't handle 2009-06-30 [19:38:00.0000] <heycam> did the Location interface have stringification requirements that were removed? or am i misremembering? [19:39:00.0000] <heycam> i thought i remembered the interface having [Stringifies=href] on it [21:00:00.0000] <shepazu> does the case-insensitivity extend to attribute values, or just element and attribute names? [21:01:00.0000] <heycam> shepazu, which case-insensitivity? [21:03:00.0000] <heycam> i don't think in general the case of attribute values is changed: http://www.whatwg.org/specs/web-apps/current-work/multipage/syntax.html#attribute-value-%28double-quoted%29-state [21:04:00.0000] <heycam> when parsing [21:11:00.0000] <shepazu> ok, thanks, heycam [22:54:00.0000] <othermaciej> Hixie is a brave man [23:24:00.0000] <hober> indeed. I can't imagine going near the codec discussion [23:38:00.0000] <shepazu> Hixie: is heycam correct that attribute values are not parsed in a case-insensitive manner? [23:40:00.0000] <shepazu> (to unwind that a bit: my understanding is that attribute values, unlike element and attribute names, are left as the case they are encountered, and not matched to different values in the DOM) [23:52:00.0000] <Hixie> shepazu: correct -- iirc the only processing the parser does on attribute values (other than finding their delimiters) is &entity; expansion [23:57:00.0000] <shepazu> Hixie: ok, good... I was worried for a moment about path commands (relative and absolute) [23:59:00.0000] <shepazu> now, what implications does that have on things like <text text-anchor="miDDle"/> ? will that be "error-corrected", somehow permitted as is, or fail? [00:13:00.0000] <heycam> shepazu, afaict, that's purely within our (svg's) purview [00:17:00.0000] <shepazu> heycam: then that opens up a question we should be addressing, no? [00:17:01.0000] <heycam> yeah i think it would be good to talk about that [00:18:00.0000] <Hixie> shepazu: same as in XML. [00:18:01.0000] <heycam> especially for consistency with style="text-anchor: miDDle", which is allowed i think [00:18:02.0000] <shepazu> heycam: the one place it really matters, IIRC, is path@d [00:19:00.0000] <shepazu> oh, and anyplace where an @id is referenced [00:19:01.0000] <Hixie> personally I would encourage you to keep as much as possible case-sensitive as you can [00:19:02.0000] <Hixie> but that if you do have anything case-insensitive, you make it work the same in XML as in text/html [00:20:00.0000] <shepazu> Hixie: we would certainly do the latter [00:20:01.0000] <Hixie> but that's just me :-) [00:20:02.0000] <shepazu> you've argued for case-insensitivity for element and attribute names... this seems a bit inconsistent [03:29:00.0000] <takkaria> gsnedders: have you written content-type sniffing tests for abarth's draft? someone said you might have done [03:31:00.0000] <gsnedders> takkaria: I started playing around with ome but never got anywhere really [03:31:01.0000] <gsnedders> takkaria: gimme a min [03:32:00.0000] <gsnedders> takkaria: Ah, http://gsnedders.html5.org/tests/content-type.php is what I have [03:33:00.0000] <gsnedders> heh. Opera 10b thinks i'm on a slow network. [03:35:00.0000] <takkaria> gsnedders: hm useful, is the source for that around anywhere? [03:35:01.0000] <gsnedders> no [03:35:02.0000] <takkaria> ok :) [03:35:03.0000] <gsnedders> /me can grab it if you want [03:36:00.0000] <takkaria> no, don't worry, it was more out of curiosity than anything [03:37:00.0000] <gsnedders> takkaria: http://gsnedders.html5.org/tests/ct.tgz [03:38:00.0000] <takkaria> gsnedders: ta [04:27:00.0000] <virtuelv> Hixie: do you have anything written on Channel messaging yet? or are there requirements available anywhere? [04:34:00.0000] <Hixie> virtuelv: Channel messaging? [04:47:00.0000] <gsnedders> takkaria: yt? [04:47:01.0000] <takkaria> gsnedders: yup [04:48:00.0000] <gsnedders> takkaria: Had lunch? [04:48:01.0000] <takkaria> nope :) [04:48:02.0000] <gsnedders> takkaria: Want to go and get lunch? :) [04:48:03.0000] <gsnedders> /me uh, forgot to bring lunch [04:49:00.0000] <takkaria> ah [04:49:01.0000] <takkaria> I have curry to reheat [04:49:02.0000] <gsnedders> (and everyone around here has had lunch) [05:56:00.0000] <virtuelv> Hixie: yes [05:56:01.0000] <virtuelv> 8.1 Event definitions [05:56:02.0000] <virtuelv> Messages in server-sent events, Web sockets, cross-document messaging, and channel messaging use the message event. [05:57:00.0000] <virtuelv> the link to channel messaging points to a description not yet written [06:02:00.0000] <virtuelv> I'm having a hard time understanding where this fits into anything [06:07:00.0000] <virtuelv> and is there any particular reason step six in the description uses Window instead of WindowProxy? [06:07:01.0000] <virtuelv> «Queue a task to dispatch the event created in the previous step at the Window object on which the method was invoked. The task source for this task is the posted message task source.» [06:08:00.0000] <virtuelv> ah, nevermind that last one [06:14:00.0000] <gsnedders> What's a good hex editor? [06:14:01.0000] <jgraham> gsnedders: I don't know what a good one is but ghexedit or something is a hex editor [06:15:00.0000] <jgraham> Oh it's called ghex [06:17:00.0000] <Philip`> I used to like Biew [06:30:00.0000] <jgraham> /me notes that not having ready access to windows makes testing jas harder [06:30:01.0000] <jgraham> JAWS [06:31:00.0000] <Philip`> /me wonders how hard it can be to get access to Windows [06:32:00.0000] <takkaria> jgraham: I think there should be some vmware images around on the network somewhere, if that's any use [06:32:01.0000] <Philip`> It's not like you're trying to get access to a SunOS 4 machine or something [06:34:00.0000] <jgraham> takkaria: Well I have windowws 7 in Virtualbox but I don't think it likes my audio. Plus I don't really want to spend my weekends in the office [06:35:00.0000] <virtuelv> gsnedders: you could also have a look at Bless as a hex editor [06:35:01.0000] <virtuelv> (and since we're on such tools, presumably in a Gnome environment - Meld is by far the nicest diff viewer) [06:36:00.0000] <takkaria> you say 'meld', I say 'less' :) [06:40:00.0000] <Philip`> gsnedders: If the file you want to hex edit is actually executable code, then http://www.hex-rays.com/idapro/ is great [06:40:01.0000] <gsnedders> Philip`: It is not [06:40:02.0000] <Philip`> gsnedders: Then you should find some executable code and use IDA on that [06:44:00.0000] <Philip`> /me wishes Google hadn't decided to start looking all ugly with a wide margin around the search results page [09:30:00.0000] <Philip`> http://everything2.com/title/W3C - "There is not a single browser on the market that complys with all W3C standards, not even Microsoft Internet Explorer." [09:30:01.0000] <Philip`> (Thu Nov 02 2000 at 2:39:22) [09:30:02.0000] <Philip`> The first part of that sentence is still true, but I can't imagine anyone nowadays saying the second part [09:31:00.0000] <Philip`> I guess things haven't really changed that much, they've just shuffled around a bit [09:38:00.0000] <tantek> It may not be possible for any single implementation of anything to comply with all W3C standards - have you seen how many there are? http://www.w3.org/TR/ [10:04:00.0000] <takkaria> things which suck: when your internet connection will only show you your broadband company's activation page and it's all in swedish [10:07:00.0000] <Philip`> All you need to do is activate the internet connection and then put the page through Google Translate to see what it said [10:35:00.0000] <Midler> takkaria: U want help with swedish? [10:35:01.0000] <Midler> Or is it way to late? [10:35:02.0000] <takkaria> I've sorted it now [10:36:00.0000] <Midler> In that way, Sweden is just like poland and other countries [10:36:01.0000] <Midler> U have ascii characters that u've never used...almost [10:36:02.0000] <Midler> =) [10:36:03.0000] <takkaria> my swedish is picking up rapidly [10:36:04.0000] <takkaria> but that's what happens when you sit and translate things word-by-word [10:37:00.0000] <Midler> :D [10:37:01.0000] <Midler> why not using google translate or something similar? [10:40:00.0000] <takkaria> I have time on my hands, and a swedish-english dictionary [10:41:00.0000] <takkaria> if I'm in sweden for twelve weeks, I might as well use all the opportunities I can to learn the language [10:41:01.0000] <Midler> oh, thats a long time [10:42:00.0000] <takkaria> I'm going to have a go at reading the paper in a bit [10:43:00.0000] <Midler> I guess comics might be an easy way to start. Seing their expressions makes it easier to understand what they say...i guess [11:00:00.0000] <remysharp> is there anyone here that knows a bit about the offline apps/manifest? [11:01:00.0000] <remysharp> I've got this url in Firefox (from what I can tell, RC3 == final release): http://html5demos.com/offlineapp - allow it to store and refresh I get a content encoding issue - [11:02:00.0000] <remysharp> does this look like a bug to do with FF or is this something I've done wrong? [11:02:01.0000] <remysharp> the manifest is being served under the right mime type (or should be - I did change it) [11:02:02.0000] <remysharp> /me works fine in Safari 4 + iPhone version. [11:26:00.0000] <beowulf> remysharp: i have the same problem [11:26:01.0000] <remysharp> cool - then it's not just me [11:26:02.0000] <remysharp> or - you've got same issue on my url or your own test? [11:27:00.0000] <beowulf> remysharp: when i try a hello world with a manifest I get a Content Encoding Error with ff3.5 [11:27:01.0000] <beowulf> remysharp: http://www.carisenda.com/sandbox/storage/ [11:27:02.0000] <remysharp> aye - cool - then it's a bug [11:28:00.0000] <remysharp> great. damn, that's some badass functionality and it doesn't work in FF :-( [11:28:01.0000] <beowulf> well, it would appear to be worse than doesn't work, it prevents you from using it [11:28:02.0000] <remysharp> Is that you then, beowulf == Stephen ? \o. hi if so! [11:29:00.0000] <beowulf> unless we've both done something stupid :) [11:29:01.0000] <beowulf> remysharp: hi! :) [11:31:00.0000] <remysharp> :) [11:31:01.0000] <beowulf> remysharp: it seems to work ok in ff 3.05 [11:31:02.0000] <beowulf> 3.0.5 [11:32:00.0000] <remysharp> then it's definitely a regression bug. [11:32:01.0000] <remysharp> perfect, that's what I needed. [11:32:02.0000] <remysharp> thanks! [11:33:00.0000] <beowulf> /me checks his logs... [11:34:00.0000] <beowulf> ah, it asked me if it was ok to store stuff but doesn't look at the manifest [11:35:00.0000] <remysharp> my server logs do when I'm checking it [11:35:01.0000] <remysharp> or do you mean FF3.0.5? [11:35:02.0000] <beowulf> yeah, i'm trying 3.05 (which confused me by asking if it was ok to store stuff offline) [11:40:00.0000] <ttepasse> Is it just me or is whatwg.org down? [11:41:00.0000] <ezyang> http://downforeveryoneorjustme.com/whatwg.org [11:41:01.0000] <ttepasse> thanks. [11:43:00.0000] <Midler> ezyang: Now i just need to remember that DNS name... [11:43:01.0000] <Midler> :P [11:43:02.0000] <beowulf> remysharp: does your server use mod_deflate? [11:43:03.0000] <remysharp> nope [11:43:04.0000] <remysharp> but it does use gzip [11:43:05.0000] <ezyang> http://lmgtfy.com/ [11:44:00.0000] <ezyang> ;-) [11:44:01.0000] <Midler> hahaha [11:44:02.0000] <Midler> well, my brain only has place for a byte [11:44:03.0000] <Midler> oops to low [11:45:00.0000] <Midler> 16bytes [11:45:01.0000] <Midler> i dont count .com [12:56:00.0000] <Philip`> How boring, Yahoo fixed their iso-2022-kr XSS vulnerability quickly :-( [12:56:01.0000] <Philip`> /me wonders how many other sites are vulnerable to the same issue [12:56:02.0000] <ezyang> linky? [12:57:00.0000] <Philip`> ezyang: http://search.yahoo.com/search?p=%EC%84%9E%E8%81%BD%E9%9D%91%E5%AD%83%EC%88%8D%E8%A8%8A%E6%98%B1%E7%A9%BF%E5%88%B7%E5%A5%84%E5%89%94%EA%B5%89%EB%82%91%EC%84%9E%E6%98%8C%E4%BE%84%E5%BE%9E%EC%97%86&ei=UTF-8&eo=iso-2022-kr [12:57:01.0000] <Philip`> in Chrome [12:57:02.0000] <Philip`> before today [12:57:03.0000] <ezyang> holy crap long url [12:57:04.0000] <Philip`> Safer than pasting Unicode into IRC :-) [12:58:00.0000] <ezyang> my irc client doesn't actually make it easy for me to open that :-( [12:58:01.0000] <ezyang> (the url marches off the side of my terminal) [12:58:02.0000] <Philip`> http://krijnhoetmer.nl/irc-logs/whatwg/20090630#l-406 [12:58:03.0000] <ezyang> sweet [12:59:00.0000] <ezyang> So, what does the text "do"? [13:00:00.0000] <Philip`> $ echo '섞聽靑孃숍訊昱穿刷奄剔굉낑섞昌侄從없' | iconv -t iso-2022-kr [13:00:01.0000] <Philip`> )C</title><script>alert(123)</script>x [13:00:02.0000] <ezyang> whoa :-P [13:01:00.0000] <ezyang> Ok, so this only affects sites that let you set the page's output encoding [13:01:01.0000] <Philip`> so Yahoo encodes the text into that string of bytes, and Chrome doesn't support iso-2022-kr so it decodes as iso-8859-1 (or utf-8 or something) and executes the script [13:01:02.0000] <Philip`> (All other browsers do support iso-2022-kr, as far as I can tell) [13:01:03.0000] <ezyang> That's... a chrome bug [13:02:00.0000] <ezyang> Also, I still don't understand why search engines let you twiddle the output encoding [13:02:01.0000] <Philip`> It's not a bug for a browser to not support a relatively obscure encoding [13:03:00.0000] <Philip`> Even if it did, you might be able to find aliases that are supported by a server but not by the browser [13:03:01.0000] <Philip`> Twiddling the output encoding does seem like a pretty stupid idea [13:03:02.0000] <Philip`> but Google and Yahoo both allow it, so I guess they've got reasons [13:04:00.0000] <Philip`> Ultraseek allows it too, and didn't even forbid EBCDIC when I last tried it, so you can XSS it that way too [13:04:01.0000] <ezyang> So, the only good reason is accessibility [13:05:00.0000] <ezyang> For those pauvre browsers that don't support UTF-8 [13:05:01.0000] <ezyang> Search engines like Google have a weird notion of what accessible needs to be [13:06:00.0000] <ezyang> I wonder what UTF-8 penetration is these days [13:07:00.0000] <Philip`> It doesn't help those browsers unless the users intentionally follow a link with the query parameter that sets the output encoding [13:07:01.0000] <ezyang> I thought that Google would do tricky user agent detection? [13:12:00.0000] <Philip`> If it was doing that, why would it put the output encoding in the query parameter instead of doing the tricky user agent detection when serving the results page? [13:56:00.0000] <gsnedders> jgraham: Can't get a external HD shipped to SE from amazon.co.uk :P [14:32:00.0000] <eighty4> gsnedders: buy one from dustinhome, I bet they're way better then crapy amazon [14:40:00.0000] <gsnedders> eighty4: And several hundred kronor more expensive [14:41:00.0000] <eighty4> so :) [14:44:00.0000] <eighty4> not several [14:44:01.0000] <gsnedders> several with what I need to buy :) [14:45:00.0000] <eighty4> ok [14:45:01.0000] <eighty4> I'm thinking of buying this: http://www.amazon.co.uk/Samsung-Spin-Point-Internal-Drive/dp/B000XZ5JBW/ref=sr_1_19?ie=UTF8&s=electronics&qid=1246394761&sr=1-19 from dusinhome [14:45:02.0000] <gsnedders> Amazon isn't the place to buy internal HDs :) [14:46:00.0000] <eighty4> costs 865 sek from amazon and 895 from dusin [14:50:00.0000] <Dashiva> gsnedders: Why not buy it, you know, from somewhere in Sweden? [14:50:01.0000] <gsnedders> Dashiva: It costs more? [14:50:02.0000] <gsnedders> Dashiva: Also, the Swedish websites are, uh, in Swedish. [14:53:00.0000] <Dashiva> I'm pretty sure Samsung Spin Point is the same :P [14:54:00.0000] <gsnedders> Dashiva: But actually buying stuff is more challenging :P [14:54:01.0000] <Dashiva> And surely you aren't the person to back away from a good challenge [14:56:00.0000] <Dashiva> Why must people insist on using obviously flawed analogies? [14:58:00.0000] <gsnedders> Dashiva: Where's the fun in coherent analogies? [14:58:01.0000] <Dashiva> I was thinking maybe the goal of an analogy was communication, not fun :) [14:59:00.0000] <gsnedders> Dashiva: You misguided bugger [15:23:00.0000] <Sirisian> I'm confused. whatwg is linked from the W3C site, but what is it? [15:23:01.0000] <Sirisian> Just a group of developers? [15:23:02.0000] <Hixie> http://wiki.whatwg.org/wiki/FAQ#What_is_the_WHATWG.3F [15:23:03.0000] <Sirisian> oh nvm [15:23:04.0000] <Sirisian> yeah just saw it [15:23:05.0000] <Sirisian> it's also on the main site. I didn't scroll down [15:29:00.0000] <Sirisian> Something I noticed. I've been looking through the Editor's Draft for the HTML 5 stuff. It has a section for the context menus and it says "When an element's context menu is requested [...] the UA must fire a simple event called contextmenu that bubbles and is cancelable at the element for which the menu was requested." I was going to post in the forum with a long post, but why isn't this extended to mouse up and mouse down [15:29:01.0000] <Sirisian> to set a standard for preventDefault for like middle click events by the browser. (Making a standard that says it should be "cancelable"). [15:30:00.0000] <Hixie> there is another spec to handle the actual user interaction events [15:30:01.0000] <Hixie> like mousedown and mouseup [15:30:02.0000] <Hixie> separate from html5 [15:30:03.0000] <Hixie> (it applies to things like svg also) [15:30:04.0000] <Hixie> that spec hasn't been written yet though [15:30:05.0000] <Sirisian> You mean for how browsers should react and work with them? [15:30:06.0000] <Hixie> yeah [15:30:07.0000] <Sirisian> oh is there a link? I must have missed it [15:30:08.0000] <Hixie> and how they should be cancelable and stuff [15:30:09.0000] <Hixie> it doesn't exist yet [15:32:00.0000] <Sirisian> oh that's why. Well that's good to here. Right now FF and IE have methods to allow a web developer to fully take over the mouse. I've been looking forward for a while when that idea is passed onto other browsers. Thanks for the response. [15:32:01.0000] <Sirisian> hear [16:18:00.0000] <roc> I find it unfathomable that in 2009 people still argue that HTML5 should have draconian parsing [16:23:00.0000] <othermaciej> wait, what? who? [16:50:00.0000] <roc> http://www.w3.org/Bugs/Public/show_bug.cgi?id=6777#c13