| 00:00 | <Philip`> | At least it looks much better in Opera than when I last looked - it's very broken, but not so broken that you can't see how broken it is :-) |
| 00:01 | <Philip`> | (I think the rendering in IE6 counts as 'too broken' - it just looks like a really boring and slightly buggy page, and doesn't tempt you with the promise of interestingness if only a few small browser bugs were fixed) |
| 00:02 | <Hixie> | i'm not really targetting IE |
| 00:02 | <Hixie> | they have bigger fish to fry, imho |
| 00:02 | <Hixie> | i'd be very interested in suggestions for more tests |
| 00:02 | <Hixie> | i've run out of the things i wanted to test :-) |
| 00:04 | <Philip`> | (Konqueror gives an Opera-like level of very-brokenness) |
| 00:14 | <Philip`> | (Ooh, KDE4's Konqueror does surprisingly well - it appears to get the whole layout correct, and it just fails the 1st and 3rd buckets and says "0%" at the end) |
| 00:15 | <zcorpan_> | time to hunt for khtml bugs then? :) |
| 00:15 | <Hixie> | click the "Acid3" text to get a report of what failed |
| 00:17 | <Philip`> | 34, 35, 37, it says |
| 00:18 | <Philip`> | but half the time it asks me if I want to Save As / Open / Cancel the empty.txt file |
| 00:19 | <Philip`> | (The other half the time, it fails doesn't ask but it fails 48 and 49) |
| 00:19 | <othermaciej> | WebKit gets to 95%, positions the boxes slightly wrong, gets two box colors wrong, and gets the color of the text "Acid 3" wrong |
| 00:19 | <Philip`> | s/fails// |
| 00:20 | <othermaciej> | fails 34, 35, 37, 48, 49 |
| 00:20 | <othermaciej> | I feel like it would be cheating to start fixing before the test is done |
| 00:21 | <nlogax> | where can i find the test? (sorry, new in here) |
| 00:23 | <othermaciej> | I bet 34 is because comments are not preserved in the DOM |
| 00:23 | <othermaciej> | http://hixie.ch/tests/evil/acid/003/ |
| 00:32 | <zcorpan_> | othermaciej: presuming that the test uses sgml parsing rules as the spec -- what does sgml say about comments? (may they be dropped as in xml?) |
| 00:33 | <othermaciej> | zcorpan_: I think the DOM spec says that implementations may include comment nodes in the DOM or not |
| 00:35 | <othermaciej> | SGML does not say anything about parsing into a DOM afaik |
| 00:35 | <othermaciej> | HTML5 may require preserving comment nodes; not sure if this is a compat issue |
| 00:37 | <zcorpan_> | ok... surprised that the dom spec would say anything about it |
| 00:38 | <Dashiva> | Is the reference rendering updated? |
| 00:44 | <Hixie> | hm, good point, the test shouldn't rely on comments |
| 00:44 | <Hixie> | feel free to fix the bugs, btw |
| 00:44 | <Hixie> | for the love of kittens don't _not_ fix bugs because acid3's early versions trigger them :-P |
| 00:45 | Dashiva | promotes test as an example of 'corner cases with no relevance to the real web, which can safely be ignored' |
| 00:45 | <Hixie> | oh? |
| 00:46 | <othermaciej> | zcorpan_: SGML (and I think XML) predate the existence of the DOM spec, so I don't think they had the opportunity to talk about parsing into a DOM |
| 00:46 | <Dashiva> | Just to be the first to do so ;) |
| 00:47 | <zcorpan> | othermaciej: yeah, but the xml spec says that comments may be dropped |
| 00:47 | <Hixie> | removed the comment thing from the test |
| 00:48 | <zcorpan> | Hixie: will there be 100 tests when it's done? |
| 00:48 | <othermaciej> | zcorpan: yeah, I think it is talking about processing in a more general way |
| 00:48 | <othermaciej> | WebKit is borking the layout much worse now, though still failing the same number of tests |
| 00:48 | <Hixie> | really? |
| 00:48 | <Hixie> | that's weird |
| 00:49 | <Hixie> | i only removed the comment |
| 00:49 | <Hixie> | zcorpan: if y'all give me enough things to test :-) |
| 00:49 | <zcorpan> | Hixie: ok :) |
| 00:49 | <Lachy> | good morning |
| 00:49 | <zcorpan> | Lachy: morning |
| 00:49 | <othermaciej> | Hixie: it's pretty weird to me too - I have a saved copy of the older test though, so pretty sure it's a real phenomenon |
| 00:49 | <Hixie> | odd |
| 00:49 | <Hixie> | wonder what i did |
| 00:50 | <Hixie> | reload? |
| 00:50 | <Hixie> | i put the comment back |
| 00:50 | <Hixie> | does it render right now? |
| 00:50 | <othermaciej> | Hixie: yes |
| 00:50 | Lachy | wonders how a comment could be used in a test |
| 00:50 | <othermaciej> | I wonder if this is a parsing bug of some sort |
| 00:50 | <Hixie> | othermaciej: looks like your html parser fails to imply <head> or something |
| 00:51 | <othermaciej> | Hixie: that's possible |
| 00:51 | <Hixie> | brb, getting food |
| 00:51 | <othermaciej> | I bet looking for browser-specific quirks in Dojo, Prototype, MochiKit and similar would be a good source of tests |
| 01:10 | <Lachy> | Hixie, your definition of the placeholder attribute conflicts with the <input placeholder> attribute implemented by Safari |
| 01:11 | <zcorpan> | <input placeholder> is something that should be part of html5 imho, the lack of that feature is worked around all over the place (mostly abusing value="") |
| 01:12 | <othermaciej> | we plan to submit <input type="search"> and related stuff for consideration in HTML5 |
| 01:12 | <othermaciej> | I asked someone on my team to write it up |
| 01:13 | <Lachy> | class="search" was predefined for that, but type=search has a cool rendering in safari |
| 01:14 | <othermaciej> | class="search" is for search forms, not search fields, afaik |
| 01:14 | <Lachy> | zcorpan, placeholder was discussed back in 2004 or 2005, and I think Hixie said it would be considered for Web Forms 3 |
| 01:14 | <othermaciej> | "It must only be used on the following elements: aside, body, form, p, section, span" |
| 01:15 | <Lachy> | oh, I thought it could be used on input too |
| 01:15 | <othermaciej> | <input type="search"> and placeholder on non-search fields (which we originally intended only for search but accidentally enabled for text fields too) are both in some actual use on the web |
| 01:15 | <othermaciej> | mostly Apple-oriented sites but not exclusively |
| 01:16 | <Lachy> | I would love to use placeholder, but I've been forced to fake it with js in the past |
| 01:16 | <Hixie> | Lachy: d'oh! |
| 01:18 | <Lachy> | hixie, maybe try relevant="" |
| 01:19 | <Hixie> | placeholder="" had the great property that it would sound silly to put it on a hidden tab panel |
| 01:19 | <Hixie> | relevant="" doesn't really hav ethat |
| 01:23 | <zcorpan> | "Well-formed document requirement—HTML5 defines a well-formed document according to the requirements specified in section 4 Appendix C of the XHTML 1.0 specification." -- http://www.devx.com/webdev/Article/34389/0/page/2 |
| 01:29 | <othermaciej> | I think hyatt is just emotionally attached to a doctype that looks the way doctypes do today |
| 01:30 | <othermaciej> | because his post on the matter makes a hash of the relevant arguments |
| 01:39 | <Philip`> | Does SGML/XML/anything care about the actual content of the "-//W3C//DTD ..." string in the doctype, and become unhappy if it was e.g. "5.0"? (If not, has anybody suggested any reasons why there's any point in using that string, given that it's impossible to remember and is confusing since there's no DTD?) |
| 01:44 | <othermaciej> | Philip`: I don't know if it is needed or not - some other people commented on it, but I think there are strong reasons to prefer a version attribute over a doctype anyway |
| 06:34 | <othermaciej> | hmm, I still can't figure out why the missing comment in Acid3 breaks our display |
| 06:34 | <annevk> | you do reparsing? |
| 06:35 | <annevk> | if that doesn't make much sense in this context, feel free to ignore |
| 06:36 | <annevk> | Maybe the innerHTML parsing mode should be renamed to HTML fragment with context parsing |
| 06:37 | <annevk> | so people don't get the feeling it has to do with scripting... |
| 06:40 | <othermaciej> | that could be good yeah |
| 06:46 | <othermaciej> | hmm, looks like WebKit does in fact keep comments in the DOM (though strictly speaking it is optional) |
| 07:09 | <othermaciej> | annevk: it is a parsing issue, but not due to reparsing - <script> in an implicit head ends up between head and body |
| 07:19 | <Hixie> | othermaciej: what does your dom look like below <html? |
| 07:20 | <othermaciej> | Hixie: I made a reduction which illustrates the issue |
| 07:20 | <othermaciej> | just a sec |
| 07:21 | <othermaciej> | Hixie: http://paste.lisp.org/display/40272 |
| 07:21 | <othermaciej> | weird bug |
| 07:21 | <othermaciej> | I'm not sure why that hoses the rendering though |
| 07:22 | Hixie | looks |
| 07:22 | <Hixie> | there's a style rule |
| 07:22 | <Hixie> | that matches head + body > div > p |
| 07:22 | <Hixie> | so you'll start/stop matching it |
| 07:22 | <annevk> | whoa, that's some quirky parsing right there :) |
| 07:23 | <othermaciej> | annevk: I think Opera puts both of those scripts between head and body |
| 07:23 | <othermaciej> | firefox puts both in head |
| 07:23 | <Hixie> | opera doesn't have a head |
| 07:23 | <Hixie> | in that document |
| 07:23 | <annevk> | right |
| 07:23 | <Hixie> | it doesn't imply <head> tags at all |
| 07:23 | <othermaciej> | ok, that, then |
| 07:23 | <annevk> | who needs <head> anyway |
| 07:23 | <Hixie> | (causes all kinds of subtle problems, that's why acid3 has an implied <head>) |
| 07:23 | <othermaciej> | my test is not very discriminating, just tells me if it went in <head> or <html> |
| 07:24 | <Hixie> | the live dom viewer is probably what you want to be using |
| 07:24 | <Hixie> | http://software.hixie.ch/utilities/js/live-dom-viewer/ |
| 07:25 | <othermaciej> | mainly I'm interested in the WebKit bug which I can figure out w/ the built-in Web Inspector, but that is indeed useful for comparisons |
| 07:25 | <annevk> | Hixie, placeholder clashes with <input placeholder> from WebKit, do you need an e-mail on that? |
| 07:25 | <Hixie> | yeah lachy said |
| 07:25 | <Hixie> | what i need is a better name :-) |
| 07:27 | <othermaciej> | what is it supposed to do? |
| 07:28 | <annevk> | irrelevant, not-relevant-now, hidden |
| 07:29 | <othermaciej> | so placeholder is basically a presentational attribute for display: none |
| 07:30 | <annevk> | othermaciej, http://html5.org/tools/web-apps-tracker?from=776&to=777 |
| 07:30 | <annevk> | othermaciej, made "semantic" and only for specific use cases, yes |
| 07:33 | <othermaciej> | it seems like "placeholder" is the wrong meaning for the given example |
| 07:33 | <othermaciej> | the input is a placeholder for the video |
| 07:33 | <othermaciej> | the video isn't a placeholder for anything while hidden |
| 07:33 | <othermaciej> | at least as I understand placeholder (temporary substitute for something else) |
| 07:34 | <othermaciej> | it seems like "hidden" would be the best word, but then you have to wonder why not just use the CSSOM, or a user-defined class that styles to display: none |
| 07:36 | <othermaciej> | placeholder does seem like it captures a common use case, but it's pretty limited, since it is harder to hook in author-controlled animation of the new content appearing/disappearing, which might be desirable |
| 07:36 | <othermaciej> | it's also hard to tell the cases where it should be used from ones where it shouldn't so it seems easy to abuse |
| 07:37 | <othermaciej> | I assume it would be considered wrong to use this for hover menus, for instance |
| 07:37 | <annevk> | it does mention that at the end |
| 07:39 | <annevk> | animations are a good point |
| 07:45 | <othermaciej> | so yeah, it doesn't sound like a compelling feature to me just yet |
| 07:50 | <Hixie> | othermaciej: we need a way to do what the examples show without any css (i.e. non-presentational hiding of content that isn't usable/relevant) |
| 07:51 | <Hixie> | othermaciej: i'm all in favour of having ways to hook animation to it |
| 07:52 | <othermaciej> | Hixie: can you clarify why being able to do it without CSS is a requirement? (I don't necessarily disagree, I am just unaware of the reasoning behind it) |
| 07:52 | <Hixie> | because css is a presentation-layer thing, not semantic |
| 07:52 | <Hixie> | but this is a semantic |
| 07:53 | <Hixie> | or more pragmatically, because css can be disabled |
| 07:53 | <Hixie> | because lync has no css |
| 07:53 | <Hixie> | lynx |
| 07:53 | <Hixie> | etc |
| 07:53 | <othermaciej> | does lynx have JavaScript? |
| 07:53 | <Hixie> | it could do |
| 07:54 | <Hixie> | the point is css is the wrong way to do this |
| 07:54 | <Hixie> | it's a semantic thing |
| 07:54 | <Hixie> | i need to say here's a section that isn't relevant, hide it |
| 07:54 | <Hixie> | it's not a stylistic thing |
| 07:55 | <annevk> | non-stylistic-hidden |
| 07:55 | <othermaciej> | I guess I'm not sure how to tell what reasons to make something hidden are semantic and which aren't |
| 07:57 | <Hixie> | if you can think of a way to apply an alternate style sheet that would show that section, then it's stylistic |
| 07:57 | <Hixie> | if you would never want to show it, it's semantic |
| 07:58 | annevk | wonders if these type of explanations should be in the spec |
| 07:59 | <annevk> | maybe a section that explains markup semantics... |
| 07:59 | <othermaciej> | seems to me like something that's hidden in the markup could just be in a comment, if it is truly not part of the semantics of the document, and if your UA has JS it probably has CSS too |
| 08:00 | <Hixie> | the problem is that a comment isn't parsed |
| 08:00 | <Hixie> | you want something that's in the DOM |
| 08:00 | <Hixie> | e.g. so it can be conformance checked |
| 08:02 | <othermaciej> | maybe the attribute should name should be "irrelevant" or "notrelevant" |
| 08:03 | <othermaciej> | it still doesn't sound terribly compelling to me, but I can see why you want it |
| 08:04 | <Hixie> | i've switched it to irrelevant for now |
| 08:05 | <Hixie> | i would love to make it more useful |
| 08:05 | <annevk> | so with animations you would set this after the animation has run I suppose |
| 08:05 | <othermaciej> | I'm trying to think how you'd make it tie in well with animation approaches, which are generally CSS-based |
| 08:05 | <Hixie> | we basically need a Core Animation of Web technologies |
| 08:06 | <Hixie> | i suppose SMIL is intended to be it, but i'm not impressed by SMIL |
| 08:06 | <othermaciej> | I sure hope someone works on such a completely hypothetical thing |
| 08:06 | <Hixie> | indeed |
| 08:07 | <annevk> | othermaciej, was that sarcastic? |
| 08:07 | <othermaciej> | it was a fancy way of saying "no comment" |
| 08:07 | <annevk> | if it was, I don't get it |
| 08:08 | <annevk> | Hixie, the problem is that SVG inflicts SMIL upon the world |
| 08:09 | <annevk> | And SVG animation is being implemented (supported)... |
| 08:09 | <Hixie> | there's room for more than one solution |
| 08:09 | <Hixie> | especially when one is as bad as SVG |
| 08:14 | <annevk> | I suppose |
| 08:14 | <annevk> | I think it would have to be significantly better though to persuade implementors, but maybe that's not hard |
| 08:17 | <Hixie> | how do you mean? |
| 08:18 | <annevk> | Well, having two sets of technologies to do the same thing doesn't seem very desirable |
| 08:19 | <annevk> | IE already has SMIL working for HTML (HTML+Time proposal to the W3C) |
| 08:19 | <annevk> | otoh, they had VML too... |
| 08:19 | <annevk> | s/had/have/ |
| 08:24 | <othermaciej> | they also have WPF/E now |
| 08:25 | <annevk> | Yeah, this Flash competitor |
| 08:25 | <annevk> | Silverlight iirc |
| 08:35 | <hsivonen> | Is Flash still bundled with Vista? |
| 08:36 | hsivonen | is very surprised about hyatt's opinions on versioning |
| 08:36 | <othermaciej> | Flash is not in fact bundled with Vista |
| 08:37 | <hsivonen> | othermaciej: ooh. radical |
| 08:37 | <hsivonen> | othermaciej: that means MS is ready to break the Web! |
| 08:38 | <hsivonen> | Silverlight in Windows-only and IE/Firefox-only, right? |
| 08:38 | <annevk> | I think Safari too |
| 08:38 | <hsivonen> | does the Firefox version work with Opera? |
| 08:38 | <hsivonen> | annevk: oh. where do I get a Mac version? |
| 08:38 | <annevk> | Maybe I was misinformed |
| 08:39 | <annevk> | I only read about it in w3c-svg-wg archives |
| 08:39 | <othermaciej> | I think they intend to have it work on Safari, dunno if that has been released yet |
| 08:40 | <hsivonen> | it seems that there is a "Silverlight CTP for Macintosh" |
| 08:42 | <hsivonen> | the Silverlight FAQ uses user-unfriendly styling/scripting. have to use Lynx to read |
| 08:44 | <hsivonen> | apparently, "/E" in the MS world view means Windows and Mac |
| 08:47 | <annevk> | The '<' doesn't seem to be specialcased at all in IE's tokenizer |
| 08:48 | <annevk> | For element names, attribute names and unquoted attribute values that is |
| 08:58 | <othermaciej> | hsivonen: it means that at least as long as it takes them to kill Flash |
| 09:02 | <annevk> | This basically means that the SHORTTAG TAGC OMISSION feature of SGML is obsolete... |
| 09:03 | <othermaciej> | what's the SHORTTAG TAGC OMISSION feature? |
| 09:03 | <annevk> | <test</test> being parsed as <test></test> |
| 09:03 | <annevk> | <p><img src=""</p> like <p><img src=""></p> etc. |
| 09:04 | <annevk> | it basically means special casing < |
| 09:04 | <virtuelv> | heh, I now have a new showcase of how to make shiny accessible JS+DOM-based UI, http://www.satama.nl/ </offtopic> |
| 09:06 | <othermaciej> | ah |
| 09:06 | <othermaciej> | does any browser do that? |
| 09:07 | <annevk> | Firefox |
| 09:07 | <annevk> | maybe Safari? |
| 09:08 | <annevk> | It also means that <p</p>test will give you a 'p<' element with a 'p' attribute with 'test' as textContent |
| 09:31 | <annevk> | othermaciej, what's the DOM view in Safari for http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C%21DOCTYPE%20html%3E%3Ctest%3C/test%3E |
| 09:31 | <othermaciej> | annevk: in webkit trunk, it's HTML --> BODY --> TEST |
| 09:32 | <othermaciej> | with no text content |
| 09:34 | <annevk> | IE has a "TEST<" element which has a "test" attribute |
| 09:35 | <annevk> | "TEST<" also occurs before "HTML" but we shouldn't follow that I think |
| 09:35 | <annevk> | This is pretty trivial to fix in the html5lib tokenizer though it breaks a bunch of tests |
| 09:36 | <othermaciej> | it |
| 09:36 | <annevk> | maybe http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C%21DOCTYPE%20html%3E%3Cbody%3E%3Cp%20%3C/p%3Etest is a better test |
| 09:36 | <othermaciej> | it's really hard to tell how much of IE's crazier parsing rules should be followed |
| 09:37 | <othermaciej> | annevk: the text is sibling to the P and the P has no funny attributes |
| 09:38 | <annevk> | Well, cwilso is right with suggesting there's an IE web and an "other browser" web I think. Some application frameworks are relying on the IE namespaced DOM for instance and have some patched version for other browsers. |
| 09:38 | <annevk> | othermaciej, I think WebKit supports the SGML feature in that case, at least partially |
| 09:39 | <annevk> | In IE "test" is a child text node of <p> which has <="" and p="" attributes |
| 09:41 | <othermaciej> | it is sometimes hard to withhold my aesthetic judgment on these parsing oddities |
| 09:42 | <annevk> | I would personally prefer IEs behavior |
| 09:42 | <annevk> | Less work for Opera and simplifies the code of html5lib |
| 09:43 | <annevk> | As the affect of having less characters that do something special |
| 09:44 | <othermaciej> | the only part that seems nasty is having an attribute named "<" |
| 09:44 | <othermaciej> | "<" is clearly already a special character, question is only whether this is also true while parsing a tag or only when parsing text |
| 10:04 | <annevk> | Hixie, the spec points to http://lists.whatwg.org/admin.cgi/commit-watchers-whatwg.org |
| 10:04 | <annevk> | Hixie, it should point to http://lists.whatwg.org/listinfo.cgi/commit-watchers-whatwg.org |
| 10:08 | <MikeSmith> | annevk - just tried to subscribe to the list but got a non-delivery message: |
| 10:08 | <MikeSmith> | [[ |
| 10:09 | <MikeSmith> | <commit-watchers-request⊙lwo>: host |
| 10:09 | <MikeSmith> | lists.whatwg.org[66.33.216.179] said: 550 |
| 10:09 | <MikeSmith> | <commit-watchers-request⊙lwo>: Recipient address rejected: |
| 10:09 | <MikeSmith> | User unknown in virtual alias table (in reply to RCPT TO command) |
| 10:09 | <MikeSmith> | ]] |
| 10:09 | <annevk> | Through the interface? |
| 10:10 | <MikeSmith> | anne - yeah. sent me a confirm message, I replied, got the non-delivery notification |
| 10:10 | <annevk> | Interesting. I'd suggesting waiting a couple of hours (like nine, ten) for Hixie so he can fix it |
| 10:12 | <MikeSmith> | k |
| 12:35 | <zcorpan> | anyone know spanish? |
| 12:37 | <nlogax> | a little |
| 12:37 | <citoyen> | Only enough to order beer and dinner |
| 12:37 | <zcorpan> | can you translate this to spanish? :) "This is not correct. Well-formedness is a concept defined in XML 1.0, and namespace-well-formedness is a concept defined in Namespaces in XML 1.0. The concepts do not apply to HTML, and HTML5 does not introduce such a concept for text/html." |
| 12:37 | <annevk> | cerveza |
| 12:38 | <zcorpan> | (which is a response to "HTML5 define como documento bien formado según las especificaciones de la section 4 Appendix C of the XHTML 1.0 specification." -- http://mistervertigo.es/blogs/introduccion-html-5-i-2/ ) |
| 12:38 | citoyen | agrees with annevk's translation |
| 12:38 | <citoyen> | it's all you'll ever need |
| 12:38 | <annevk> | I believe he got that from somewhere else (in English) |
| 12:38 | <zcorpan> | annevk: yes |
| 12:39 | <annevk> | in which case replying in English should cause no problems |
| 12:39 | zcorpan | wanted to reply in spanish mostly for fun :P |
| 12:40 | <zcorpan> | but ok |
| 12:42 | <zcorpan> | commented in english |
| 12:50 | hsivonen | just talked to a reporter about the conformance checker project; wonders what will end up in print |
| 12:56 | <annevk> | A promising student has been doing his master thesis on a conformance checker for XHTML2, a technology frequently used on the web. |
| 13:03 | <Dashiva> | Oh yes |
| 13:04 | <Dashiva> | hsivonen is a role="model" for web-interested people everywhere |
| 14:54 | <krijnh> | annevk, Philip`: I think it's fixed now |
| 14:56 | <Lachy> | Hixie should rename placeholder="" to role="placeholder"! |
| 14:56 | Lachy | ducsk |
| 14:56 | <Lachy> | *ducks |
| 14:58 | <Philip`> | krijnh: Looks to me like it's working - thanks! |
| 14:58 | <krijnh> | Np |
| 14:58 | <krijnh> | Can't test too much now, so if it breaks somewhere, let me know :) |
| 14:58 | <Lachy> | oh, there's an XHTML2 WG telcon going on! This should be fun to watch :-) |
| 14:59 | <Lachy> | in #xhtml on w3 IRC server |
| 14:59 | <krijnh> | http://krijnhoetmer.nl/irc-logs/xhtml/20070425 ;) |
| 15:00 | <krijnh> | Oki, bye :) |
| 15:00 | <Philip`> | You should add extra AJAX for live streaming IRC logs ;-) |
| 15:00 | <Philip`> | (I suppose anybody who wants live streaming IRC could just join the channel, but that's no fun) |
| 15:01 | <Lachy> | just reload every ~30 seconds |
| 15:02 | <Lachy> | I'll let you know when something interesting happens |
| 15:02 | <Philip`> | "when" rather than "if"? |
| 15:03 | <Lachy> | yeah, sorry. I meant "if" |
| 15:04 | <Lachy> | they're not typing the minutes :-( |
| 15:04 | <Lachy> | oh, now they are |
| 15:19 | <virtuelv> | we should stream the logs using server-sent events |
| 15:19 | <Lachy> | that'd be awesome |
| 15:22 | <Philip`> | "[16:24] * Zakim Steven, you typed too many words without commas" - Zakim seems like quite a peculiar bot |
| 15:23 | <Lachy> | I didn't understand that message either. I had no idea zakim was a grammar checker |
| 15:24 | <Lachy> | wow, IE will apparently be implementing role="" |
| 15:25 | <virtuelv> | I could point out that words themselves don't have commas |
| 15:40 | <annevk> | Zakim is the telcon bot |
| 15:41 | <MikeSmith> | that message from Zakim prolly just means it was expecting the word "to" |
| 15:42 | <MikeSmith> | q+ to talk about blah |
| 15:42 | <MikeSmith> | instead of |
| 15:42 | <MikeSmith> | q+ talk about blah |
| 15:43 | <MikeSmith> | (q+ is command to tell Zakim you want to added to the speaker queue during a telcon) |
| 15:43 | <MikeSmith> | s/to added/to be added/ |
| 15:50 | <annevk> | This RDF nonsense makes role= confusing |
| 16:30 | <Philip`> | Hmm, how odd - Safari (at least in old versions) seems to fail if I define a "function native(...)", while no other browser minds :-( |
| 16:31 | <annevk> | -> #webkit |
| 16:34 | <hasather> | Philip`: they probably don't allow future reserved words either then |
| 18:20 | Hixie | looks at the responses to his e-mail on www-html and decides to just leave it alone |
| 18:21 | <Hixie> | i don't understand why people think that something that the spec says is non-conformant is somehow a "feature" |
| 18:23 | <annevk> | Bjoern had a nice response |
| 18:23 | <Hixie> | not really |
| 18:23 | <Hixie> | it was a "look at me, i'm so clever" response that totally missed the point |
| 18:29 | annevk | filed a bug on something Acid3 tests today |
| 18:29 | <Hixie> | heh |
| 18:29 | <Hixie> | which one? |
| 18:29 | <annevk> | exposing DOCTYPE in text/html |
| 18:29 | <Hixie> | ah |
| 18:30 | <annevk> | I think we got bugs on the other failures |
| 18:30 | <Hixie> | probably filed by me :-P |
| 18:31 | <Hixie> | * Presentational elements are kept. This includes the HR element, |
| 18:31 | <Hixie> | SMALL, B, and I. This is 2007; NONE of them should be kept. |
| 18:31 | Hixie | wonders how those are presentational anymore |
| 18:32 | <Hixie> | mind you the same person thought <m> was presentational |
| 18:32 | <Hixie> | so... |
| 18:33 | <Hixie> | maybe they're confused and think "presentational" means "easy" and "semantic" means "hard" |
| 18:33 | <Hixie> | that would explain it |
| 18:33 | <Hixie> | lol the same person wants us to remove indeterminate progress bars |
| 18:34 | <annevk> | I was just reading that e-mail |
| 18:34 | <annevk> | I think I agree with his last point |
| 18:34 | <annevk> | but I don't have a better solution |
| 18:35 | <annevk> | (last point of the list) |
| 18:35 | <annevk> | http://www.w3.org/mid/tkrat.c998060f76b1e71c⊙gn is the e-mail, fwiw |
| 18:36 | <Hixie> | not sure why APIs shouldn't be defined |
| 18:36 | <Hixie> | it's not a markup language |
| 18:36 | <Hixie> | it's "web apps 1.0" |
| 18:37 | <annevk> | The APIs that are markup language neutral |
| 18:38 | <Hixie> | i guess i don't really care about "language neutral". the web is html. |
| 18:39 | <Philip`> | It's not WA1 - it's WA1-in-the-context-of-becoming-HTML5 (at least in that email discussion), so the HTML name gives some justification for wanting it to be a markup language |
| 18:43 | <Hixie> | i guess my point of view on that is that i have bigger fish to fry than trying to get the ideal specification silos sorted out |
| 18:54 | <annevk> | Just got an e-mail that says the following: |
| 18:54 | <annevk> | Would save the trouble of going through and adding alt="" to all the |
| 18:54 | <annevk> | non-content elements when validating. The alt attribute should be |
| 18:54 | <annevk> | automatically applied to <object> and <img> at least. |
| 18:54 | <annevk> | (plus quoting my last message to whatwg⊙wo |
| 19:56 | <annevk> | Hixie, Google was a typo? :) |
| 19:56 | <annevk> | being picky, that comment should probably say icon="y.png" now you changed / fixed the name of the company |
| 19:57 | <annevk> | oh, and it's "Yahoo!" |
| 20:21 | <Philip`> | Hixie: Was "Fix a couple of typos" really meant to change the hex pattern to say "<!1DOCTYPE HTML"? |
| 21:18 | <Hixie> | annevk: it's a comment. deal with it. :-P (i only changed it to help track down a bug with the postprocessor) |
| 21:22 | <Hixie> | i love this sentence from htm5: "Implementations that use ECMAScript to implement the APIs defined in this specification must implement them in a manner consistent with the ECMAScript Bindings for DOM Specifications specification, as this specification uses that specification's terminology. [EBFD]" |
| 21:22 | Hixie | notes EBFD doesn't exist yet |
| 21:24 | <annevk> | House on ice |
| 21:34 | <nickshanks> | what was with that Tina woman? |
| 21:34 | <nickshanks> | on the mailing list |
| 21:36 | <bewest> | opinionated |
| 21:36 | <bewest> | like everyone else :-) |
| 21:39 | <annevk> | hah, encoding error on http://www.greytower.net/about/ |
| 21:40 | <bewest> | she does seem to have strange notions of "pragmatic" and what consitutes a sane solution |
| 21:40 | <bewest> | evidently billions of previously authored pages don't constitute evidence suitable for decision making |
| 21:42 | <nickshanks> | bewest: i think a combination of carrot and stick is needed to get web developers to fix things like raw ampersands in documents. |
| 21:43 | <nickshanks> | of course that depends on your definition of 'broken' and 'fixed' in this context |
| 21:44 | <nickshanks> | people here seem to prefer the view that the web is correct and the specs are broken, not the other way around. |
| 21:45 | <annevk> | I would think that most people here think both are broken |
| 21:45 | <Hixie> | indeed |
| 21:45 | <annevk> | But realize the web is way harder to fix |
| 21:45 | <Hixie> | surely to fix the web we just have to educate the web developers |
| 21:46 | <annevk> | They're not the only people creating content, but yes |
| 21:46 | <annevk> | Reducing the need to rely on browser specific features is another important thing |
| 21:46 | <jgraham> | Everything is broken / everyone is broken... |
| 21:47 | <annevk> | Fostering interop and such |
| 21:47 | <Hixie> | (i was being sarcastic) |
| 21:47 | <annevk> | oh, duh :) |
| 21:47 | annevk | was wondering about it |
| 21:48 | <Hixie> | you should get your sarcasm detector checked, i think it's off-calibration ;-) |
| 21:48 | <annevk> | in my defense, I still feel jetlagged |
| 21:48 | annevk | should probably get some more sleep |
| 21:49 | <Hixie> | :-) |
| 21:49 | <jgraham> | So, if I wanted to find out how a HTML5 DOM tree for some content looked in an existing browser what would be the most sane way to do it? |
| 21:49 | jgraham | was thinking generating a script that built the DOM |
| 21:49 | <annevk> | http://software.hixie.ch/utilities/js/live-dom-viewer/ |
| 21:49 | <Hixie> | http://software.hixie.ch/utilities/js/live-dom-viewer/ |
| 21:49 | <Hixie> | hah |
| 21:49 | <annevk> | I'm still fast though |
| 21:49 | <annevk> | :p |
| 21:49 | <Hixie> | i beat you that time |
| 21:50 | <annevk> | oh, not in my client... |
| 21:50 | <annevk> | weird |
| 21:50 | <Hixie> | oh, actually, the logs say you did |
| 21:50 | <Hixie> | nevermind :-) |
| 21:50 | <bewest> | yeah, annevk won over here too |
| 21:50 | <Hixie> | clearly google should upgrade from our modem connection to dsl |
| 21:51 | <bewest> | btw, I read some of annevk's and maciej's work on sunday... it was pretty good |
| 21:51 | <bewest> | Window |
| 21:51 | <bewest> | and |
| 21:51 | <bewest> | uh |
| 21:51 | <jgraham> | Does that actually do what I want? I want to take some content, parse it with the HTML5 algorithm and then see how the browser redering of the resulting tree would compare with the rendering of the same source using the browser's native parser |
| 21:51 | <bewest> | evidenty by "read" I mean scanned |
| 21:51 | <Hixie> | jgraham: oh, i misunderstood |
| 21:51 | <othermaciej> | I need to update Window RSN |
| 21:52 | <jgraham> | I don't know if it's a *useful* think to want to do but it seems like one way to see how much HTML5 parsing breaks |
| 21:52 | <jgraham> | s/think/thing/ |
| 21:52 | <Hixie> | jgraham: you'd want something similar, but which had two iframes, one which just got the given markup rendered straight in, and the other which round-tripped through an html5lib script to convert it to valid markup and sent that back |
| 21:52 | <Hixie> | jgraham: it would be extremely useful to have, yes |
| 21:53 | <Hixie> | othermaciej: fwiw i've collapsed most of the window spec back into the html5 spec |
| 21:53 | <Hixie> | othermaciej: there were many things that i needed to define that were just too closely tied to things Window defined |
| 21:53 | <annevk> | we should have html5lib in ECMAScript... |
| 21:53 | <Hixie> | othermaciej: e.g. session history, navigation, content sniffing, etc |
| 21:53 | <Hixie> | browsing contexts |
| 21:53 | <Hixie> | javascript security model |
| 21:54 | <othermaciej> | Hixie: yeah, that makes me feel less concerned, OTOH I still would like to help SVG and CDF with crack removal |
| 21:54 | <othermaciej> | (then the trick is making sure what they say is compatible) |
| 21:54 | <othermaciej> | if Web Apps 1.0 becomes a W3C spec, I can freely plagiarize |
| 21:54 | <Hixie> | yeah, dunno how to solve it really. maybe they should just require html support ;-) |
| 21:54 | <Hixie> | or they should keep out of that level altogether |
| 21:55 | <annevk> | so the thing is that "HTML5" not just defines a markup language but also the web navigation, security etc. model |
| 21:56 | <Hixie> | it _is_ called Web Applications 1.0 |
| 21:56 | <annevk> | right |
| 21:56 | <othermaciej> | Hixie: btw, have you thought about the tension between history state objects and ability to have a URI for your current state that you can bookmark or send to others? |
| 21:56 | <othermaciej> | it seems like they solve the back problem, but not the bookmark or send problem |
| 21:56 | <annevk> | I suppose this ok since to support the web you have to support all of that, but some people would like those features to be implementable without having to implement HTML5 |
| 21:56 | <Hixie> | othermaciej: yeah, there's a red box in there about it. i think pushState() will need two more arguments, URL and title |
| 21:57 | <Hixie> | othermaciej: the problem is with whether it raises any security or usability issues, since the URI might not represent the state at all |
| 21:57 | <Hixie> | annevk: yeah, i'd love it if they weren't so closely related |
| 21:58 | <Hixie> | annevk: that doesn't mean it's easy to do |
| 21:58 | <othermaciej> | Hixie: not being able to change the protocol, host or port from the real one probably addresses the security issues |
| 21:58 | <Hixie> | probably |
| 21:58 | <othermaciej> | agreed there could be usability issues |
| 21:58 | <Hixie> | but i need to make sure i look at it very closely before i spec it |
| 21:59 | <Hixie> | and i havnen't gotten around to it yet |
| 21:59 | <jgraham> | Hixie: just rendering the output of html5lib doesn't work in general e.g. http://wordsandpictures.dyndns.org/cgi-bin/parsetree/parsetree.py?uri=&source=%3C%21Doctype+html%3E%0D%0A%3Chtml%3E%0D%0A%3Chead%3E%0D%0A%3Cstyle%3E%0D%0Afoo+p+%7Bcolor%3Agreen%3B%7D%3E%0D%0A%3C%2Fstyle%3E%0D%0A%3C%2Fhead%3E%0D%0A%3Cbody%3E%0D%0A%3Cfoo%3E%0D%0A%3Cp%3EThis+should+be+green%3C%2Fp%3E%0D%0A%3C%2Ffoo%3E%0D%0A%3C%2Fbody%3E%0D%0A%3C%2Fhtml%3E (click View in B |
| 21:59 | <jgraham> | rowser) |
| 21:59 | <othermaciej> | the current de facto approach to this is using a fragment ID |
| 21:59 | <Hixie> | it's on my radar for this quarter, fwiw |
| 21:59 | <annevk> | you can also have usability issues by just giving someone a URL which does browser sniffing |
| 21:59 | <othermaciej> | Hixie: ok, noted |
| 21:59 | <jgraham> | (In firefox) |
| 21:59 | <Hixie> | (but let me know if it becomes more urgent for you and it'll become on my radar for the week in question) |
| 22:00 | <othermaciej> | it's not immediately urgent, I just thought of it because you mentioned session history |
| 22:00 | <Hixie> | jgraham: good point |
| 22:00 | <Hixie> | jgraham: maybe write a script that recreates the DOM using JS? |
| 22:01 | <jgraham> | Right. I guess that's the way forward :) |
| 22:01 | <Hixie> | let me know if there's anything i can do to help, short of actually writing it, because this is something i'd really love to have |
| 22:01 | <Hixie> | i encourage you to check it into the html5 repo :-) |
| 22:02 | <annevk> | does <video> already cover video-support-disabled, can't-do-video, etc.? |
| 22:02 | <annevk> | it doesn't seem like it, but maybe I missed something |
| 22:02 | <Hixie> | what do you mean by "cover"? |
| 22:02 | <annevk> | tell implementors what to do |
| 22:02 | <annevk> | it seems to just say that video should always be rendered |
| 22:03 | <Hixie> | if the implementor supports <video>, then yes, it has to be rendered (though it might not play until the user starts it). |
| 22:03 | <annevk> | where for Lynx it might make sense to offer a download link or something |
| 22:04 | <Hixie> | if you think the spec isn't clear enough, or requires something bad, send a mail with your suggestion, i'll add it to the list |
| 22:04 | <Hixie> | i think you're right and i haven't yet really taken care of handling alternative <video>/<audio> modes |
| 22:17 | <annevk> | btw, you could also modify Hixie's dom viewer script to output the same serialization as html5lib |
| 22:18 | <annevk> | yet another way would be adding a new flag to html5lib to output the hixie html format |
| 22:19 | <annevk> | that shouldn't be that hard either |
| 22:46 | <fantasai> | "User agents should net render elements" s/net/not/ |
| 23:10 | <Hixie> | hm, cool, tbl said he'd be on the call tomorrow |
| 23:12 | <bewest> | hrm email says 5pm. I wonder which timezone |
| 23:13 | <bewest> | oh I see |
| 23:13 | <bewest> | who's invited? |
| 23:13 | bewest | is guessing invited experts aren't invited in general |
| 23:13 | <Hixie> | didn't you get the mail on the list? |
| 23:13 | <bewest> | I'm looking at list email |
| 23:13 | <Hixie> | http://lists.w3.org/Archives/Public/public-html/2007Apr/1458.html |
| 23:14 | <Hixie> | http://www.w3.org/2002/09/wbs/40318/tel26Apr/ |
| 23:14 | <gavin_> | to more directly answer your question, and WG member is invited |
| 23:14 | <gavin_> | that includes Invited Experts |
| 23:15 | <bewest> | yeah, I'm looking at that email... following the references now |
| 23:15 | <bewest> | oh WG Participants are invited |
| 23:15 | <bewest> | I thought member means something special |
| 23:16 | <Dashiva> | Member with capital m does |
| 23:16 | <bewest> | ok thanks |
| 23:16 | <Dashiva> | W3C Member, as opposed to WG members |
| 23:20 | <bewest> | Hixie: when someone submits some feedback, there are basically two possibilities, correct? 1.) you respond relatively quickly or 2.) it goes on some queue |
| 23:21 | <bewest> | if there is no immediate response, is there some way to know if my feedback has been noticed and that I should simply wait? |
| 23:21 | <Hixie> | it's always noticed |
| 23:21 | <bewest> | ah |
| 23:21 | <Hixie> | but if you really want me to check, i can check :-) |
| 23:22 | <Hixie> | (anything sent to whatwg⊙wo that isn't replied to and isn't inane ends up in my folders) |
| 23:23 | <bewest> | ah |
| 23:23 | <bewest> | I'm specifically talking about http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2007-April/010924.html |
| 23:24 | <Hixie> | yeah, that's e-mail 109 of 119 in INBOX.input-for-whatwg-WF2 |
| 23:24 | <bewest> | ok |
| 23:25 | ajnewbold_ | sends the latest hot pharmaceutical offer to whatwg⊙wo |
| 23:25 | <Hixie> | qv. inane. :-) |
| 23:25 | <ajnewbold_> | :P |