| 00:06 | <Hixie> | new record for sending me feedback in awkward ways: e-mailing me a 500KB PDF with the feedback |
| 00:06 | <Hixie> | (only two pages, could have been done in about 30 lines in an e-mail) |
| 00:06 | <TabAtkins> | Uh, what? About what? |
| 00:06 | <Hixie> | some typos |
| 00:07 | <TabAtkins> | I wonder what sort of toolchain they employed to produce the PDF. |
| 00:35 | <Philip`> | Hixie: I think I'll use it too infrequently to remember the character or the key combination needed to enter it, so I'll probably stick with the boring and sane approach of using a bookmark keyword instead |
| 00:36 | <Hixie> | boooriiing |
| 00:44 | <Philip`> | It's good to be boring in this case, because it leaves more time to think about other cases that are more interesting |
| 00:45 | <Philip`> | like trying to find a domain name that is meaningful in both its Unicode and Punycode forms |
| 00:49 | <TabAtkins> | Just use non-ascii vowels, so you have txt spk in the unencoded part. |
| 00:49 | <TabAtkins> | And then do some computation to find set of vowels that'll produce something cool in the encoded part. |
| 01:20 | <TabAtkins> | Hixie: Your argument re:<select required> seems inconsistent. Sure, you can't not submit a value with <select>. You can't not submit a value with <input type=text> either, though. If you're being consistent, you either have to say that @required shouldn't apply to <input type=text>, or that submitting the empty string as the value of an <input type=text> doesn't count as "submitting a value". |
| 01:29 | <Hixie> | you can't submit <select> without a value. You can submit <input type=text> without a value. |
| 01:29 | <Hixie> | I don't understand the inconsistency. |
| 01:29 | <Hixie> | that wasn't the main argument, though, the main argument is "baby steps" |
| 01:29 | <TabAtkins> | No you can't. <form><input name=foo><input type=submit></form> will produce "?foo=". |
| 01:30 | <TabAtkins> | That's a value - the empty string. |
| 01:30 | <Hixie> | i suppose you could look at that as having a value |
| 01:30 | <Hixie> | i don't think most users would :-) |
| 01:31 | <TabAtkins> | Take it from a different tack - why do you think that that scenario is materially different from <form><select name=foo><option value="">Hahaha</option></select><input type=submit></form> |
| 01:31 | <Hixie> | because there the user has selected "Hahaha" |
| 01:31 | <TabAtkins> | No they haven't. They've loaded a page and hit submit. |
| 01:31 | <TabAtkins> | Precisely the same as the previous form. |
| 01:32 | <TabAtkins> | And it produces an identical query string - "?foo=". |
| 01:32 | <Hixie> | i don't think the query string result is relevant here |
| 01:32 | <Hixie> | the point i was making was that from the user's point of view, the <select> has a value (Hahaha) whereas the <input> does not (it's empty). |
| 01:32 | <TabAtkins> | You have some inscrutable notion of what it means to "submit a value", then, which appears to an outside observer as "I'm making random choices for no good reason". |
| 01:33 | <TabAtkins> | In my book, identical behavior producing identical outcomes points pretty strongly to them being similar. |
| 01:34 | <Hixie> | i think if you ask someone who knows nothing about html or forms whether an empty text field has a value, and whether a select box with a "Hahaha" item selected has a value, i am confident they would say the former did not, and the latter did. |
| 01:34 | <Hixie> | whether i'm right or wrong here is academic though since as i said, i'm not using this as an argument for anything -- i agree we should add required="" to <select>. |
| 01:34 | <TabAtkins> | Assume that my example was semi-serious, and the default choice was "Please choose one...". |
| 01:34 | <Hixie> | "Please choose one..." should imho be non-conforming as an <option> value. |
| 01:35 | <TabAtkins> | (Since someone who knew nothing about HTML or forms wouldn't understand that the visible text and the submitted value can be different.) |
| 01:35 | <Hixie> | (i don't see how to make it non-conforming though) |
| 01:35 | <TabAtkins> | Hixie: That opinion makes it impossible to use a dropdown for which there isn't a good default value. |
| 01:35 | <Hixie> | sure. just pick the first one. |
| 01:35 | <TabAtkins> | ...like I just said, that makes it impossible to use a dropdown for which there isn't a good default value. |
| 01:36 | <Hixie> | it's what "real" UIs do (as opposed to the UI disaster that the Web has turned into) |
| 01:36 | <TabAtkins> | "The first one" is the default value by necessity due to how <select> works, not necessarily the actual correct default. |
| 01:37 | <Hixie> | show me an Apple-created UI where there's a <select>-equivalent that doesn't default to something valid. |
| 01:37 | <TabAtkins> | That's not a real argument, Hixie. If I can't, that could just mean that Apple is only using <select> for things with a decent default value. |
| 01:38 | <Hixie> | yes, because doing otherwise is bad UI :-) |
| 01:38 | <Hixie> | however, again, my opinion here is irrelevant, as noted above -- I agree we should add these features to HTML in due course |
| 01:39 | <TabAtkins> | sicking pointed out a good reason to address the issue now, rather than later. |
| 01:39 | <Hixie> | sicking works on a browser that hasn't even implemented what we have so far |
| 01:39 | <Hixie> | (what was the reason?) |
| 01:40 | <Hixie> | (oh, new e-mail) |
| 01:40 | Hixie | reads |
| 01:40 | <sicking> | Hixie: the person asking for @required is the person implementing validation in firefox |
| 01:40 | <sicking> | Hixie: i.e. mounir |
| 01:41 | <Hixie> | well then he's uniquely positioned to convince me that it's time to add it :-) |
| 01:41 | <Hixie> | he need but implement the spec :-) |
| 01:41 | <sicking> | Hixie: it's in patches, waiting for toolkit stuff to land |
| 01:41 | <sicking> | Hixie: we have the whole thing implemented (i.e. validation) |
| 01:42 | <sicking> | Hixie: displaying error messages is the hard part, UI often is |
| 01:42 | <sicking> | Hixie: another reason to do this now is that we have browser vendors focus on this stuff. If we wait it's likely to take much longer to get focus readjusted |
| 01:43 | <sicking> | Hixie: it's like 30 extra minutes to implement @required on <select> if you're already in the code doing all the other validation stuff |
| 01:43 | <Hixie> | and placeholder="" |
| 01:43 | <Hixie> | and the other stuff people asked for |
| 01:43 | TabAtkins | is surprised it would even take that long. |
| 01:43 | <Hixie> | there's tons of stuff in this list of stuff people want in forms |
| 01:43 | <sicking> | Hixie: it's weeks worth of work to refocus, implement, add addtional tests, and push out the door |
| 01:43 | <Hixie> | it's not like required="" is the only thing we'd be adding in v2 |
| 01:44 | <sicking> | *shurg* |
| 01:44 | <sicking> | there's currently a feature that is 90% finished in the spec. no reason not to add the last 10% |
| 01:45 | <Hixie> | i gave the reason, so there's not no reason |
| 01:45 | <Hixie> | there might be good reasons to add it now though, such as the compat issue in the e-mail you sent |
| 01:45 | <Hixie> | your e-mail is in my list of things to respond to again |
| 01:48 | <Hixie> | (man, i don't normally get browser vendors telling me I shouldn't be saying no to feature requests :-P) |
| 01:57 | <sicking> | Hixie: that's not really true, is it. Many of the features originate from browser vendors, no? |
| 01:59 | <Hixie> | Many of the ones we add to the spec, yes |
| 01:59 | <Hixie> | I reject almost all feature requests. |
| 02:00 | <Hixie> | if I didn't, you'd have long stopped paying any attention to me |
| 02:00 | <TabAtkins> | Hixie: jschuh is saying it would be *way* easier to implement @srcdoc if the URL of the browsing context was a data: url producing the same page. |
| 02:01 | <TabAtkins> | (rather than "about:srcdoc".) |
| 02:01 | <Hixie> | tell him to send an e-mail saying why |
| 02:01 | <TabAtkins> | kk, he will do. |
| 02:02 | <jamesr> | why is it about:srcdoc? |
| 02:02 | <Hixie> | wasn't clear what else it should be |
| 02:02 | <Hixie> | using data: seems like it would have implications I'm not sure we want |
| 02:03 | <Hixie> | but off-hand I can't remember what those are |
| 02:05 | <Hixie> | oh, i remember |
| 02:05 | <Hixie> | it would lead to all kinds of weirdness with the origin. |
| 02:05 | <Hixie> | we might end up having to make sure we defined the origin of data:-from-srcdoc differently from data:-from-elsewhere |
| 02:05 | <Hixie> | I wanted to avoid ever having to end up there. |
| 02:07 | <TabAtkins> | Not necessarily. We might have to define some reasonably sane behavior for data urls used in an <iframe sandbox=allow-same-origin>, though. |
| 02:07 | <TabAtkins> | (Right now, I think that <iframe src=data:foo sandbox=allow-same-origin></iframe> is still unique-origin.) |
| 02:07 | <Hixie> | not necessarily, no, but it was a danger i figured we should avoid if at all possible |
| 02:13 | <Twisol> | If I may ask, what exactly is this channel used for, compared to the mailing list? |
| 02:13 | <TabAtkins> | Similar things, just in IRC-sized bites. |
| 02:13 | <TabAtkins> | Also: trolling |
| 02:14 | <Twisol> | Makes sense. |
| 02:15 | <TabAtkins> | Though, serious ideas here often just get told to send an email to the list. |
| 02:15 | <Hixie> | Twisol: this is the water cooler basically |
| 02:15 | <Twisol> | Makes sense, sure. |
| 02:15 | <TabAtkins> | (So more attention is paid to them, more eyes are on them, and they enter Hixie's response queue.) |
| 02:15 | <Hixie> | or the peanut gallery |
| 02:15 | <Twisol> | Oh, perfect explanation. |
| 02:15 | <Twisol> | Water cooler does it for me |
| 02:16 | <Twisol> | And is this just for core HTML5, or does it include other specs like WebSockets? |
| 02:16 | <TabAtkins> | other specs |
| 02:16 | <Hixie> | there's not really any one topic |
| 02:16 | <TabAtkins> | Mostly anything web-related. |
| 02:16 | <Twisol> | Okay, thanks. |
| 02:16 | <Hixie> | yeah |
| 02:16 | <TabAtkins> | Unless the Opera guys are up, in which case it's also anything viking-related. |
| 02:16 | <Hixie> | i also like to talk about model railways but everyone else just ignores me :-( |
| 02:17 | <Twisol> | Haha :) |
| 02:18 | <Hixie> | hah, http://gavinsharp.com/irc/whatwg.html has the most used words as 1. would, 3. should, and 6. could |
| 02:18 | <Twisol|away> | Ha! |
| 05:46 | <wirepair> | here's an interesting one... |
| 05:46 | <wirepair> | what would cause http://localhost to cause a file to render one way |
| 05:46 | <wirepair> | yet http://127.0.0.1 to render differently, in firefox? |
| 05:47 | wirepair | boggles |
| 05:59 | <Twisol> | How does it differ? O_o |
| 06:01 | <wirepair> | i'm using extjs and the examples are missing borders |
| 06:01 | <wirepair> | or the top pixel of widgets is gone |
| 06:01 | <wirepair> | it's the same exact code, but if i type in http://127.0.0.1 it renders just fine |
| 06:29 | <Hixie> | wirepair: are you referring to anything by hostname, which could trigger cross-origin checks? |
| 06:32 | <Twisol> | localhost is just a local hostname alias (it's in /etc/hosts on my computer), so that's definitely possible... |
| 07:11 | <annevk5> | gah |
| 07:11 | <annevk5> | http auth is terrible |
| 07:41 | <annevk5> | Hixie, you just introduced "Theh" |
| 07:42 | <Hixie> | bummer. |
| 07:42 | <Hixie> | thanks |
| 07:45 | <annevk5> | can someone run http://software.hixie.ch/utilities/js/live-dom-viewer/saved/584 in Internet Explorer? |
| 07:45 | <annevk5> | bonus points for IE9 |
| 08:22 | <annevk5> | got results from Peter` |
| 08:22 | <annevk5> | http://www.w3.org/Bugs/Public/show_bug.cgi?id=10326 |
| 08:59 | <Ms2ger> | annevk5, does WebKit support document.width? |
| 08:59 | <annevk5> | I think so |
| 09:04 | <Peter`> | Ms2ger: yes |
| 09:04 | <Ms2ger> | Thanks |
| 09:05 | <annevk5> | :/ |
| 09:05 | <Peter`> | annevk: you weren't really conclusive.. |
| 09:06 | <annevk5> | just annoying that they copied it |
| 09:06 | <annevk5> | it's not in IE / Opera |
| 09:43 | <annevk5> | ah crap |
| 09:43 | <annevk5> | HTML5 does not actually define in detail how to serialize XML documents |
| 09:45 | <Ms2ger> | Sounds like something for XML5 |
| 09:48 | <jgraham> | Ms2ger: nice play :) |
| 09:49 | <annevk5> | i see what you did there ;p |
| 09:49 | <Ms2ger> | Seriously though, I think that would be a better place to spec it |
| 09:50 | <annevk5> | if anyone was serious about putting XML5 in a browser I would |
| 09:50 | <gsnedders> | Maybe you sohuld just spec XML5 to be what people are willing to implement? |
| 09:50 | <gsnedders> | I hear that works well. |
| 09:52 | <jgraham> | Well XML serialization could be a different spec |
| 09:52 | <annevk5> | could be in a web dom core |
| 09:53 | <annevk5> | there's not even basic support in browsers for createDocument |
| 09:54 | <annevk5> | document.implementation.createDocument(null,null,null) does not give an empty document :/ |
| 09:54 | <gsnedders> | Is that not an illegal hierarchy? |
| 09:55 | <jgraham> | annevk5: Web DOM Core? hahah |
| 09:55 | <jgraham> | a |
| 09:56 | <annevk5> | gsnedders, rtfs |
| 09:56 | <annevk5> | jgraham, it's gonna happen |
| 09:56 | <gsnedders> | annevk5: No, it's easier to ask here. |
| 09:57 | <gsnedders> | annevk5: And who is going to write it? |
| 09:57 | <annevk5> | me maybe, xhr is more or less done now |
| 09:57 | <annevk5> | would be nice to work on something non-networked |
| 10:00 | <annevk5> | but really if you can't even rely on createDocument to work how am I supposed to test all this :/ |
| 10:01 | <annevk5> | maybe do a request first to get a document via responseXML |
| 10:05 | <annevk5> | would also make it easier to test the crappy encoding requirement i guess |
| 15:57 | <annevk5> | Ms2ger++ |
| 17:52 | <Philip`> | jgraham: "There is the question of what to do for e.g. the <canvas> tests that use a different harness." - change them to use the same harness, perhaps? |
| 18:05 | <TabAtkins> | Sigh. Getting nerd-sniped by Leif. |
| 18:05 | <TabAtkins> | Under what conceivable definition of "legacy" could "is compatible with ASCII" be construed as making something more legacy? |
| 18:06 | <Workshiva> | In a world where everyone uses UTF-16 |
| 18:10 | <TabAtkins> | Granted. |
| 18:10 | <TabAtkins> | Luckily, we don't live in such a world. |
| 18:11 | TabAtkins | kinda wishes we lived in a utf-32 world, which is just like the ascii world only 4 times the size. |
| 18:13 | <TabAtkins> | The spec doesn't define any special behavior for <iframe src=data:foo sandbox=allow-same-origin/>, does it? The iframe is still cross-origin, since data urls are unique-origin by default? |
| 18:15 | <erlehmann> | remysharp, can you help me with something IE-related? it seems your IE fix for new HTML5 element styling creates some kind of null namespace, which makes innerHTML then return something like <:figure></:figure> |
| 18:16 | <erlehmann> | any way to fix that short of simple and ugly string manipulation? |
| 18:16 | <remysharp> | I probably can't help - only because I'm about to head out - but do you have a live example? i.e. in something like jsbin.com? |
| 18:17 | <erlehmann> | remysharp, that plugin i created for GSoC is active at this page and generates the <figure> http://gsoc2010.dieweltistgarnichtso.net/?p=63 |
| 18:18 | <erlehmann> | Click on the embed button, receive markup. |
| 18:19 | <remysharp> | okay, I'm starting up windows. That said, I'm pretty sure I'm going to say: It's because innerHTML is fucked in IE when it comes to unknown elements... |
| 18:22 | <erlehmann> | well, you may notice that the <video> element gets copied without that kind of null namespace |
| 18:22 | <erlehmann> | hmmm, is there a way to temporarily reverse the IE styling fix? |
| 18:24 | <erlehmann> | remysharp, the video has an embed button too now. |
| 18:25 | <remysharp> | a regex :) |
| 18:25 | <remysharp> | so which IE? IE9? /hopefully... |
| 18:27 | <Hixie> | TabAtkins: data: URLs aren't cross-origin by default |
| 18:27 | <TabAtkins> | They're not? I was sure they were unique-origin, and were special-cased to be same-origin in some places. Is it the other way around? |
| 18:28 | <Hixie> | depends on implementation at the moment. See the WHATWG spec for details. |
| 18:28 | <TabAtkins> | urgh, which part? |
| 18:30 | <Hixie> | i forget |
| 18:30 | <Hixie> | try origin |
| 18:30 | <Hixie> | or search for "data:" |
| 18:32 | <TabAtkins> | "For Document objects and images... if a Document or image was obtained in some other manner (e.g. a data: URL typed in by the user...) The origin is a globally unique identifier." |
| 18:33 | <TabAtkins> | The other references to data: are to a redirect to a data:, and a data: found in another Document. |
| 18:37 | <Ms3ger> | TabAtkins, surely that would be elem.dataset.jwBitrate |
| 18:38 | <TabAtkins> | Ms3ger: (a) yeah, I'm dumb. dataset, not data. (b) ah, I didn't realize it converted dashes to camelcase. |
| 18:39 | <Hixie> | ooh, an Ms3ger! |
| 18:39 | <Ms2ger> | Where? |
| 18:39 | <Hixie> | d'oh |
| 18:40 | <remysharp> | erlehmann: sorry dude, my time is up - gotta leave - fusion was running slow. |
| 18:42 | <TabAtkins> | Hixie: Ah, no, here's the relevant line: "If url does not use a server-based naming authority, or if parsing url failed, or if url is not an absolute URL, then return a new globally unique identifier." |
| 18:42 | <TabAtkins> | Doesn't that mean that data: urls are unique-origin by default? |
| 18:42 | <Hixie> | i thought there was a line about when you navigated to a data: URL |
| 18:42 | Hixie | looks |
| 18:43 | <Hixie> | If a Document or image was generated from a data: URL found in another Document or in a script |
| 18:43 | <Hixie> | The origin is the origin of the Document or script that initiated the navigation to that URL. |
| 18:43 | <Hixie> | (which by the way means the origin of the Document isn't the same as the origin of its URL) |
| 18:44 | <TabAtkins> | Oh, I was *totally* misreading that line. |
| 18:46 | <TabAtkins> | I don't understand what you mean by "isn't the same", though, Hixie, unless you're referring to some of the stranger special-cases that can affect the origin of the original document. |
| 18:46 | <TabAtkins> | Sigh. Never mind. I got it. |
| 18:47 | <Hixie> | i mean the URL itself (the string that starts data:...) has an origin, and that origin isn't the same as the origin of the Document object (whose URL is data:...) |
| 18:47 | <TabAtkins> | The url itself is unique-origin. The Document created by following the url or using it in an iframe is the origin of the including document. |
| 18:47 | <Hixie> | the origin of the url is generally only used to see if something is going to be cross-origin, after that it's rarely used if at all |
| 19:47 | <Hixie> | you have _got_ to love the w3c's spirit of cooperation: http://lists.w3.org/Archives/Public/public-html/2010Aug/0099.html |
| 19:49 | <AryehGregor> | Top-posting, too. Terrible. |
| 19:50 | <AryehGregor> | (Can you come up with a spec the W3C publishes where it's already not the change controller?) |
| 19:53 | <jgraham> | Philip`: Are you volunteering? :) |
| 19:54 | <gsnedders> | Anyone have any suggestion of where to go in Scotland for a few days? |
| 19:55 | <cheeser> | i loved Glasgow |
| 19:56 | <gsnedders> | I was thinking more of the in-the-middle-of-nowhere places |
| 19:56 | <gsnedders> | (Also: I'm living in Glasgow for at least the next four years) |
| 19:56 | <cheeser> | ha! fair enough. |
| 19:57 | <gsnedders> | (And don't suggest St Andrews either, I've lived there for 17) |
| 19:58 | <jgraham> | gsnedders: The caledonian canal is nice |
| 19:58 | <jgraham> | Or go up to skye or somewhere |
| 19:59 | <gsnedders> | I was thinking of Skye, somewhere around the Great Glen, or maybe following the Great Highland Way from Glen Coe up to Fort William |
| 20:02 | <gsnedders> | (I'd be tempted to do the entirity of the Great Highland Way, but that takes longer than I want, really) |
| 20:06 | <gsnedders> | … or maybe Mull… meh. I do'nt know. |
| 20:07 | <jgraham> | Whisky tasting in islay? |
| 20:07 | <gsnedders> | Heh. My sister was trying to convince me of that. |
| 20:08 | <gsnedders> | But really, does that seem my sort of thing? |
| 20:08 | <jgraham> | Umm |
| 20:22 | gsnedders | wonders whether to make another (6th today) cup of tea… |
| 20:41 | <Workshiva> | You're such a brit |
| 21:20 | <annevk5> | haha ms3er :p |
| 21:21 | <annevk5> | euh ms3ger |
| 21:21 | <Hixie> | AryehGregor: what i find especially galling is that what i originally proposed is that both organisations be joint change controllers |
| 21:21 | <Hixie> | AryehGregor: but w3c is like "no, i think we just want ultimate control" |
| 21:21 | <Hixie> | because, you know, the w3c has shown how incredibly good they are at maintaining HTML |
| 21:21 | <Hixie> | so it makes lots of sense that we'd have the W3C as sole change controller |
| 21:21 | <AryehGregor> | In practice, does it matter who the change controller is? |
| 21:22 | <annevk5> | this is such a non debate |
| 21:22 | <annevk5> | finding a way to get rid of IANA would be more productive |
| 21:22 | <Hixie> | AryehGregor: apparently not since as far as i can tell it's impossible to actually register MIME types |
| 21:23 | <annevk5> | also, didn't we recently learn that "change controller" has no meaning at all? |
| 21:23 | <AryehGregor> | Then why not just drop the issue and give them their token victory? |
| 21:23 | <Hixie> | annevk5: http://wiki.whatwg.org/wiki/Registries |
| 21:23 | <Hixie> | annevk5: also true |
| 21:23 | <AryehGregor> | No one will know or care who the change controller is. |
| 21:23 | <Hixie> | AryehGregor: it |
| 21:23 | <Hixie> | er |
| 21:23 | <Hixie> | AryehGregor: it's the principle of the thing :-) |
| 21:23 | <annevk5> | Hixie, one step ahead of me :) |
| 21:23 | <AryehGregor> | Better to save energy and political capital to push for more WHATWG influence in things that actually matter, like stuff authors might see (related to the specs themselves, say). |
| 21:24 | <AryehGregor> | Being stubborn because of the principle of the thing is not a wise strategy unless you don't want any concessions from the other party at all, which is not the case here. |
| 21:27 | <Hixie> | being stubborn makes sense when you're not getting any concessions though |
| 21:27 | <Hixie> | which we're not as far as i can tell |
| 21:27 | <AryehGregor> | No, you should be generous so that you can stockpile a list of times when you were generous when you actually want to ask for concessions. |
| 21:28 | <AryehGregor> | We're getting plenty of concessions from the W3C's point of view, given that their point of view is we shouldn't exist. |
| 21:28 | <Hixie> | name one |
| 21:28 | <AryehGregor> | (speaking in generalities here) |
| 21:28 | <AryehGregor> | Having any mention of the WHATWG in the W3C version of the spec? |
| 21:28 | <Hixie> | name one thing that has actually been done counter to what the w3c want, which the w3c could have done differently |
| 21:28 | <Hixie> | they couldn't not have a mention of the WHATWG in the W3C version, I'd just stop editing. |
| 21:29 | <AryehGregor> | Thus you've forced that concession from them. |
| 21:29 | <AryehGregor> | Still a concession. |
| 21:29 | <gsnedders> | Are such ultimatums benefical to the spec? |
| 21:29 | gsnedders | is not convinced they are |
| 21:29 | <Hixie> | at this point i don't see much advantage to having the w3c spec at all |
| 21:30 | <jgraham> | Every time someone says "change controller" I think of the fat controller |
| 21:30 | <Hixie> | hah |
| 21:30 | <AryehGregor> | Well, if your goal is to increase friction between the W3C and WHATWG until the status quo is no longer tenable, you aren't doing a terrible job of it, but it would be nice if we could get that over with more quickly instead of having it drag on for years. |
| 21:31 | <jgraham> | I really don't think this issue is worth getting upset over |
| 21:31 | <jgraham> | MIME types are mostly useless and ignored on the internet at the best of times |
| 21:32 | <Hixie> | i'm not really getting upset over it, i just think the w3c is being assinine |
| 21:32 | <Hixie> | w3c staff, that is |
| 21:32 | <jgraham> | What it says in some field that almost no one knows exists on the registration form for a type is really quite irrelevant |
| 21:33 | <annevk5> | fwiw, I doubt W3C Members would disagree with the W3C Team here |
| 21:33 | <Hixie> | jgraham: unless it ever gets actually registered, in which case if the w3c once again dropped the ball on html, we'd be stuck unable to update the registration |
| 21:34 | <AryehGregor> | Does it matter? |
| 21:34 | <Hixie> | at least in theory, though as anne pointed out, (a) apparently they ignore that field when they feel like it and (b) apparently it's impossible to register types anyway in practice |
| 21:34 | <AryehGregor> | I mean, can't you just let the registration be wrong? |
| 21:34 | <Hixie> | AryehGregor: then people like Julian will argue with you about what software will do, and e.g. servers like Apache won't update their configurations to match reality. |
| 21:34 | <Steve^> | Question: If I use <address> outside of an article, then my document must have a body element? |
| 21:34 | <nessy> | or supercede it |
| 21:35 | <Hixie> | Steve^: all documents must have <body> elements (though the <Body> tag itself can be omitted, it's implied) |
| 21:35 | <jgraham> | Hixie: So the problem scenario here is a) IANA gets its shit together enough that it is registered as W3C, b) W3C stops working on HTML and c) IANA (or someone) then refuses to do (something) with future non W3C revisions of HTML even if they are the ones that have implementations |
| 21:35 | <Steve^> | ah of course, thanks |
| 21:37 | <Hixie> | jgraham: right, something which has already happened (except (c) happened by default because they just stopped having their shit together) |
| 21:37 | <hober> | the way I see it, the change controller issue is simply a proxy issue for "who is ultimately in charge here" |
| 21:38 | <hober> | becuase, as has been mentioned, the iana process doesn't really matter |
| 21:39 | <hober> | so by pushing resolution on this issue, which isn't actually harming anyone by leaving it the way it is, people push toward an ultimate settlement of which organization is ultimately responsible for html |
| 21:39 | <Hixie> | yes |
| 21:39 | <Hixie> | and clearly the w3c shouldn't be the only such organisation |
| 21:41 | <hober> | if you have any interest in a (***WIKITRAP***) historical analogy, I think this is a pretty good fit: http://edward.oconnor.cx/2007/12/the-third-dail |
| 21:41 | <Steve^> | I see that address must be inside flow content, that's certainly not what I expected |
| 21:42 | <annevk5> | joint change controller would make sense to me too, maybe plh hadn't considered that |
| 21:42 | <annevk5> | could be that he missed it |
| 21:43 | <annevk5> | but if the W3C wants sole ownership that doesn't seem really bad either; it wasn't a problem in 2004 and there's no reason it will be in the future |
| 21:44 | <Hixie> | how was it not a problem in 2004? |
| 21:45 | <hober> | it wasn't a problem pre-2001 or so |
| 21:45 | <Hixie> | pre-2001 the w3c hadn't given up on html yet |
| 21:46 | <hober> | indeed |
| 21:49 | <annevk5> | Hixie, it didn't stop (de facto) updates to the media type |
| 21:50 | <Hixie> | yes it did |
| 21:50 | <jgraham> | Hixie: From the W3C's point of view it is an understandable position to take. It is not clear that the W3C clearly understands where it adds value so it tries to protect all the obvious written evidence of it being important and in control |
| 21:50 | <AryehGregor> | Should someone point out the possibility of co-sponsorship to plh on the list? |
| 21:50 | <annevk5> | I guess you care more about IANA correctness than I do |
| 21:50 | <jgraham> | (I think the W3C does add value to be clear) |
| 21:50 | <Hixie> | annevk5: people are _still_ arguing that you can send XHTML 1.1 as text/html |
| 21:51 | <Hixie> | jgraham: understandable, yes. reasonable, no. |
| 21:51 | <Hixie> | jgraham: i never said i didn't understand it :-) |
| 21:51 | <annevk5> | that seems kind of irrespective of who is the change controller though |
| 21:51 | <Hixie> | just that it was inane |
| 21:52 | <annevk5> | that is more a case of people-like-arguing |
| 21:52 | <Hixie> | annevk5: maybe. hard to tell since they haven't been accepting any registrations (E.g. SVG still isn't registered, ten years later) |
| 21:52 | <annevk5> | yeah, I lost faith in IANA a long time ago |
| 21:53 | <jgraham> | Hixie: You could try addressing the underlying issues then. It is not clear to me that W3C have any reason to back down from their position |
| 21:53 | <annevk5> | and the IETF process that mostly guards it |
| 21:53 | <Hixie> | jgraham: underlying issue to what? |
| 21:53 | <jgraham> | Hixie: Why they want to assert ownership when the question arises |
| 21:53 | <jgraham> | assuming they are not happy with joint ownership |
| 21:53 | <Hixie> | jgraham: what do you think the underlying issue is? |
| 21:54 | <jgraham> | Hixie: I said above |
| 21:54 | <Hixie> | that they only add the patent policy as value? |
| 21:54 | <Hixie> | isn't that up to them to fix? |
| 21:54 | <Hixie> | i don't see why it'd be our role to shore up the w3c's relevance |
| 21:55 | <Hixie> | i don't really see any need for standards organisations to exist, frankly |
| 21:55 | <jgraham> | Hixie: That they percieve that having joint ownership or non-ownership would represent a substantive value-loss to their members |
| 21:56 | <Hixie> | we just need some patent reform and we could replace the whole edifice (w3c, ietf, whatwg, etc) with just some public mailing lists and version control repos for each spec |
| 21:56 | <Hixie> | jgraham: i have no idea how we would address that usefully |
| 21:57 | <jgraham> | Hixie: Nor do I, but I would not expect them to back down unless you can |
| 21:57 | <Hixie> | i don't expect them to back down either |
| 21:58 | <jgraham> | So if you can't, given that the relevance of the issue rests on the basis of a bunch of hypotheticals that can be effectively controlled, it doesn't seem worth spending any time on |
| 21:58 | <jgraham> | Just concede and move on |
| 21:58 | <Hixie> | who's spending any time on it? |
| 21:59 | <Hixie> | i haven't done anything since rejecting the bug, as far as i can tell |
| 21:59 | <jgraham> | In particular I don't expect it to be so easy for W3C to drop the ball on HTML again. They did that experiment and it failed |
| 21:59 | <jgraham> | People won't forget that |
| 21:59 | <jgraham> | Hixie: (you are, now) |
| 21:59 | <Hixie> | ranting on irc doesn't count as spending time on something :-) |
| 21:59 | <Hixie> | i'm busy in other windows doing real work :-) |
| 22:02 | <annevk5> | "Messaging API" and "Web Messaging" |
| 22:02 | <annevk5> | gonna be confusing |
| 22:02 | <annevk5> | http://dev.w3.org/2009/dap/messaging/ vs http://dev.w3.org/html5/postmsg/ |
| 22:04 | <Hixie> | annevk5: i assume dap/messaging isn't intended for the web |
| 22:04 | <Hixie> | since giving web pages access to the sms backend on a phone is pretty crazy |
| 22:06 | <abarth> | all of dap is pretty crazy |
| 22:06 | <annevk5> | yeah dunno; never get really concrete answers when I ask such things |
| 22:08 | <annevk5> | abarth, you might wanna add the "user:password" (userinfo) production to things to look at for URL handling (apparently IE still does not handle them for HTTP, but other browsers do, and HTTPbis obsoletes it...) |
| 22:08 | <abarth> | there's two questions: |
| 22:08 | <abarth> | 1) how to parse |
| 22:08 | <abarth> | 2) what to do with the parsed URL |
| 22:09 | <abarth> | do you mean (1) or (2)? |
| 22:11 | llrcombs | drops his sense of logic |
| 22:11 | <llrcombs> | so what's cooking? |
| 22:11 | <abarth> | IANA registries, apparently |
| 22:16 | <llrcombs> | I meant in HTML5 |
| 22:16 | <llrcombs> | but that works too |
| 22:16 | <annevk5> | abarth, I think this would affect 1, at least as far as IE is concerned and where we may or may not want to align |
| 22:17 | <llrcombs> | lolIE |
| 22:17 | <llrcombs> | sorry] |
| 22:17 | <llrcombs> | had to say that |
| 22:22 | <abarth> | annevk5: ok, i have test cases for parsing userinfo. i check to see what IE does. my impression is that they parse it fine by then canonicalize it away or hide it in the UI |
| 22:22 | <annevk5> | abarth, oh, so they simply ignore it |
| 22:23 | <Hixie> | AryehGregor: re your mail -- in practice, the only positions that matter are implementors |
| 22:23 | <annevk5> | abarth, see also http://www.w3.org/Bugs/Public/show_bug.cgi?id=10326 (XHR bug) |
| 22:23 | <abarth> | annevk5: that's my recollection, but i haven't dug back into it yet |
| 22:24 | <annevk5> | abarth, there's no rush really btw; though since you're here, how is the URL coming along? |
| 22:24 | <abarth> | i have a couple test suites and a lot of data |
| 22:24 | <abarth> | i need to crunch the data though :) |
| 22:25 | <AryehGregor> | Hixie, you make the final decision, even if you have no real choice in some cases. |
| 22:25 | <abarth> | generally, IE likes to canonicalize more than FF |
| 22:25 | <abarth> | parsing seems largely consistent |
| 22:26 | <Hixie> | AryehGregor: does one make a decision when one has no choice? |
| 22:26 | <AryehGregor> | Well, you have a choice, just other people's opinions might in some cases make all but one choice clearly stupid. You still have to decide not to be stupid. :) |
| 22:27 | <AryehGregor> | Also, you do make the decision in a lot of cases, anytime implementers don't express a unified enough opinion. |
| 22:28 | <Hixie> | sure, but my decisions are powerless unless the implementors then agree that they were the right decisions |
| 22:29 | <AryehGregor> | Likewise, the decisions of the President of the United States are powerless unless various government officials decide to help enforce them. Power that's dependent on other people's cooperation is still power, just not absolute power. |
| 22:29 | <Hixie> | anyway my point was just that it wasn't so much that i have "decided" to include srcdoc="" in the WHATWG HTML spec regardless of what the W3C HTML WG decides, so much as that is the only choice that isn't, as you put it, stupid |
| 22:30 | <Hixie> | so insofar as that is a decision it's rather a hollow one |
| 22:30 | <AryehGregor> | I was really only saying that I have no say, more than that you do have a say. |
| 22:30 | <Steve^> | actually, the President has certain executive privileges |
| 22:31 | <Hixie> | AryehGregor: i will grant you that in practice i could exert influence more easily than you, but given the way i work in practice we probably both have about as much influence on the spec |
| 22:31 | <Hixie> | AryehGregor: but granted, that's because i've decided to work in a way that makes that the case |
| 22:31 | <Hixie> | and one could imagine that i could change that decision, though i doubt i'd remain editor for long if i did :-) |
| 22:31 | <AryehGregor> | Hixie, that's not even slightly true. You have complete power up until the point that an implementer is dissatisfied enough to say they won't implement what you say. That's a *lot* of influence. |
| 22:32 | <AryehGregor> | (and you have even more power when implementers are irrelevant, like authoring conformance requirements) |
| 22:32 | <AryehGregor> | Mostly, implementers will implement whatever you say unless they think it's a bad idea for some specific reason. You therefore have total freedom to choose from all ideas that implementers don't think are particularly bad. |
| 22:32 | <Hixie> | if i were to ignore feedback, sure |
| 22:32 | <Hixie> | but i don't |
| 22:33 | <Hixie> | all the decisions i make are based on what data and arguments people (including myself) can find and put forward |
| 22:33 | <AryehGregor> | But no feedback happens if you could have done things some other way but didn't, unless someone else happens to think the other way is particularly superior. |
| 22:33 | <AryehGregor> | Which is only going to happen in a small minority of cases. |
| 22:34 | <AryehGregor> | There are zillions of things you could have done differently without raising disagreement. Including a lot of fairly significant things, probably. |
| 22:34 | <Hixie> | so if everyone else would have made the same decision i have the power to make the decision unimpeded? that's even more hollow! :-) |
| 22:34 | <Hixie> | but sure |
| 22:34 | <Twisol> | Unrelated question: In your opinion, is it worthwhile to try to support both WebSockets -76 and -75, or would it be more reasonable to wait until a later revision? |
| 22:34 | <AryehGregor> | No, other people would have made a different decision. They just don't care enough about the difference to mention it. |
| 22:34 | <Hixie> | Twisol: both -75 and -76 are obsolete |
| 22:35 | <AryehGregor> | (or they don't bother thinking about it, since they don't have to make the decision) |
| 22:35 | <Hixie> | AryehGregor: fair enough |
| 22:35 | <Twisol> | Oh, really? Where's the most recent draft? |
| 22:35 | <Hixie> | AryehGregor: i think you undersell your ability to effect change on the spec though |
| 22:35 | <Steve^> | If you give people an answer they are more likely to go with it, if you got everyone to think separately beforehand, won't the best answer appear from lack of bias? |
| 22:35 | <annevk5> | Twisol, whatwg.org/C#network or some such |
| 22:35 | <AryehGregor> | I didn't say I have no influence, just that I have no more influence than anyone else. At least, beyond my ability to make more convincing arguments. |
| 22:35 | <Hixie> | AryehGregor: http://www.whatwg.org/specs/web-socket-protocol/ or http://www.whatwg.org/specs/web-apps/current-work/complete/network.html#websocket-protocol if you prefer an HTML version |
| 22:36 | <AryehGregor> | Wrong person. :) |
| 22:36 | <Twisol> | ;) |
| 22:36 | <Twisol> | Thanks, much appreciated. |
| 22:37 | <Hixie> | er, yeah, Twisol|away, what I said to AryehGregor :-) |
| 22:52 | <Hixie> | anyone got IE? what does it alert on https://bug-19967-attachments.webkit.org/attachment.cgi?id=60491 ? |
| 22:53 | abarth | boots a vm |
| 22:54 | <TabAtkins> | string,string,string |
| 22:54 | <TabAtkins> | in IE8 on W& |
| 22:54 | <AryehGregor> | IE6 in ies4linux says string,string,string. |
| 22:54 | <TabAtkins> | 7 |
| 22:54 | <llrcombs> | same with WK |
| 22:54 | <llrcombs> | string,string,string |
| 22:54 | <Hixie> | thanks!! |
| 22:55 | <llrcombs> | anyone know if HTML5 <video> will eventually do playlists? |
| 22:55 | <llrcombs> | or will that always be a task for JS? |
| 22:55 | <TabAtkins> | I think that's noted as a v2 feature? |
| 22:55 | <AryehGregor> | Really? |
| 22:55 | <Hixie> | llrcombs: no plans to add that support so far, but it's occasionally requested |
| 22:55 | <AryehGregor> | That seems like it should be left up to JS, no? |
| 22:55 | TabAtkins | is wrong, apparently. |
| 22:55 | <Hixie> | "v2" in the spec just means people requested it |
| 22:56 | <Hixie> | not that we'll definitely add it |
| 22:56 | <llrcombs> | so I'll pencil you in for a "maybe" |
| 22:59 | <annevk5> | what is a fun test |
| 22:59 | <annevk5> | s/what/that/ doh |
| 23:11 | AryehGregor | is very annoyed at Opera's OOM error on http://aryeh.name/tests/reflection.html, since it means he only has two browsers to look for spec/test bugs in, not three. |
| 23:38 | <gsnedders> | AryehGregor: Uh… Okay. That's weird. |
| 23:38 | <AryehGregor> | Yes, it is. |
| 23:38 | <AryehGregor> | Also annoying. |
| 23:38 | <AryehGregor> | I'd be okay with a workaround if you can come up with one. |
| 23:38 | <AryehGregor> | It first happened a couple of days ago, when I did some refactoring. |
| 23:39 | <gsnedders> | I guess I know what I'm doing tomorrow… |
| 23:39 | <AryehGregor> | Dunno what caused the problem, but if you want to investigate, you can look in the git history to find which one first triggered it. |
| 23:39 | <gsnedders> | s/tomorrow/later today/ |
| 23:39 | <gsnedders> | (I mean, I have a bunch of interns helping with the other thing I'm doing, so…) |
| 23:41 | <TabAtkins> | Urgh, damn you untrusted certificates. |
| 23:41 | TabAtkins | doesn't want to try and resolve this. |
| 23:54 | <TabAtkins> | Ah, never mind, turned out easy. |
| 23:57 | <gsnedders> | Hixie: Have you seen anything about window.parent being readonly causing site-compat problems? |
| 23:57 | <gsnedders> | (like, in any feedback sent on the spec) |
| 23:58 | <Hixie> | sounds vaguely familiar |
| 23:58 | <Hixie> | but i don't recall specifics, if any |
| 23:58 | gsnedders | shrugs |
| 23:58 | <gsnedders> | I'll try and remmeber to send an email tomorrow |