| 00:46 | <Hixie> | ok I'm going to create a Google Calendar, and it'll list sections for review each week |
| 00:46 | <Hixie> | anyone interested in helping set this up? |
| 01:04 | <Hixie> | hm, this doesn't seem like a good way of doing this |
| 01:05 | Hixie | ponders |
| 01:08 | <MikeSmith> | Hixie: I might otherwise offer to help but I'm the middle of (re)learning Java so that I can try to write useful patches for v.nu conformance-checking backend without making my lack of coding skills obvious |
| 01:22 | <Hixie> | MikeSmith: heh |
| 01:24 | <Hixie> | maybe the better solution is to just get people to see if the spec satisfies their pet peeve rather than trying to get people to review the spec section by section |
| 01:24 | <Hixie> | since the latter really isn't something most people have the skillset to do sanely |
| 01:28 | <Philip`> | There are still some people who would be willing and able to review many sections, and those people's interests would overlap to some extent, and so it would be good to prioritise them to maximise coverage |
| 01:28 | <Philip`> | s/prioritise/organise/ |
| 01:30 | <othermaciej> | I can probably do a lot of useful review but I have no idea how to prioritize it in a secion-based way |
| 01:31 | <MikeSmith> | hmm, Glazman resigning from CSS WG |
| 01:31 | <MikeSmith> | http://www.glazman.org/weblog/dotclear/index.php?post/2009/04/01/Moving-on |
| 01:32 | <MikeSmith> | oh |
| 01:32 | <MikeSmith> | April Fools |
| 01:32 | <MikeSmith> | ? |
| 01:32 | <MikeSmith> | maybe I should actually read the posting |
| 01:33 | <MikeSmith> | ah, got me |
| 01:37 | <MikeSmith> | Hixie: maybe Sierk Bornemann's comments in "Disregard of RFC 4329 and IANA MIME Media Types" are an April Fools joke |
| 01:44 | <Hixie> | MikeSmith: it's an old bug, i doubt it |
| 01:44 | <Hixie> | Philip`: any suggestions on how to do that? |
| 01:46 | <MikeSmith> | Hixie: I think it planned it well in advance. He seems like a really clever guy. I like especially the part where he says, "the browser teams of Mozilla, KDE Konqueror, Safari and Opera... did not have such problems like you in accepting and implementing these ... They simply did it. Without questioning the decision of IANA and IETF." |
| 01:47 | <MikeSmith> | s/think it planned/think he planned/ |
| 01:47 | <Hixie> | i wish i could think you were right |
| 02:22 | <Philip`> | Hixie: No |
| 02:31 | <Philip`> | Hixie: "the MIME type used to refer to JavaScript in this specification is text/javascript, since that is the most commonly used type." - most commonly used where? http://canvex.lazyilluminati.com/misc/stats/scripttypes2.html says the HTTP Content-Type 1.5 years ago was almost always "application/x-javascript" instead |
| 02:31 | <Philip`> | Hixie: Also: s/JavsScript/JavaScript/ |
| 02:37 | <Hixie> | hm, most commonly used isn't what i meant |
| 02:37 | <Hixie> | most recognisable, maybe? |
| 02:38 | <Hixie> | it's most commonly used in <script type="">, i think |
| 02:38 | <Hixie> | but i haven't checked recently |
| 02:39 | <Philip`> | http://canvex.lazyilluminati.com/misc/stats/scripts2.html - it was the most common <script type> by far |
| 02:43 | <Hixie> | ok |
| 02:43 | <Hixie> | then i'll claim that's what i meant :-P |
| 02:43 | <Hixie> | after all, i don't think i talk about the HTTP type anywhere |
| 03:22 | <Hixie> | Philip`: yt? |
| 03:31 | <Hixie> | if anyone has any comments on http://wiki.whatwg.org/wiki/Reviewing_HTML5 let me know, otherwise i'll post it to the list and the blog tomorrow |
| 04:21 | Hixie | sends aria last call comments |
| 04:36 | <MikeSmith> | Hixie: so is it not your plan for the HTML5 spec itself to define which ARIA attributes are conformant on particular elements? |
| 05:33 | <Hixie> | MikeSmith: as far as i can tell, the ARIA spec disallows that. I sent them e-mail asking if they could change that. |
| 05:34 | <MikeSmith> | Hixie: I see. I wasn't aware that the ARIA spec explicitly disallowed it. |
| 05:35 | <MikeSmith> | Hixie: I'm wondering how useful it will be in the long run to put ARIA-attribute checking into conformance checkers |
| 05:35 | <MikeSmith> | I mean, isn't the real use-case for ARIA that the attributes get added to the DOM dynamically? |
| 05:36 | <MikeSmith> | instead of in markup |
| 05:44 | <Hixie> | yeah |
| 05:45 | <Hixie> | i'm hoping we will get dynamic conformance checkers on the long run |
| 05:45 | <Hixie> | not holding my breath though |
| 06:33 | <sayrer> | hmm, these what blog posts have become increasingly snide |
| 06:34 | <sayrer> | whatwg blog posts |
| 06:34 | <sayrer> | I would like to publish a whatwg blog post about that |
| 06:36 | <Hixie> | go ahead :-) |
| 06:36 | <Hixie> | the blog is open to anyone |
| 06:36 | <sayrer> | that's what I thought |
| 06:36 | <sayrer> | how do I do it? |
| 06:37 | <Hixie> | if you register for an account and log in, you can write the post |
| 06:37 | <Hixie> | then ask lachy to submit it |
| 06:37 | <sayrer> | alright, let's see |
| 06:38 | <Hixie> | keep things nice and happy though :-) |
| 06:38 | Hixie | is off to play his game console |
| 06:38 | <Hixie> | bbl |
| 06:38 | <sayrer> | I think markp is failing at nice and happy, fwiw |
| 07:17 | hsivonen | notes that markp's copy and paste from IRC logs lower cases content |
| 07:20 | <hsivonen> | Hixie: as benh says, we'll probably be stuck with <nav role=navigation> indefinitely. so that probably needs to be made conforming even for the long term accompanied by IRC, blog and twitter whining how it sucks that the W3C has a separate WG patching HTML only from an accessibility point of view |
| 07:21 | <hsivonen> | Hixie: (and whining how implementors opted to implement role=navigation first instead of exposing <nav> right away) |
| 07:24 | <hsivonen> | annevk42: I think in general, ARIA should just stay away from XSD. (nmtoken or otherwise) |
| 07:26 | <hsivonen> | annevk42: regarding ARIA working around MS but being supported in IE8 nonetheless: do you believe that all top four would have some support for ARIA if ARIA weren't designed in such a way that no vendor can block it? |
| 08:12 | <Hixie> | sayrer: how so? |
| 08:12 | <sayrer> | Hixie, just snide stuff like |
| 08:13 | <sayrer> | regarding the meta charset thing |
| 08:13 | <sayrer> | "This is mainly to address the desire of a few overly vocal authors to be able to serve the same markup in both text/html and application/xhtml+xml modes." |
| 08:13 | <sayrer> | or |
| 08:13 | <Hixie> | is that snide? isn't it precisely accurate? |
| 08:14 | <sayrer> | do you add things because people are overly vocal? |
| 08:14 | <sayrer> | I thought not |
| 08:14 | <Hixie> | in that particular case, yes, it was added because sam complained loudly |
| 08:14 | <Hixie> | i probably wouldn't have added it if he hadn't asked for it |
| 08:15 | <Hixie> | because i think serving polyglot documents is silly |
| 08:15 | <sayrer> | did he make a bad argument? |
| 08:15 | <Hixie> | but he had some use cases for it |
| 08:15 | <Hixie> | so in it goes |
| 08:15 | <Hixie> | no, what makes you ask that? |
| 08:15 | <sayrer> | so, it seems snide to require people to make their case, and then call them "overly vocal" for doing so |
| 08:16 | <sayrer> | and then there's |
| 08:16 | <sayrer> | "Those who complained that HTML 5 was 'too bloated' will have a little less to complain about now that several parts of it have been published as separate documents. On the other hand, critics who complained about these things as a cover for other agendas will have to continue complaining a little while longer." |
| 08:16 | <Hixie> | again, that just seems accurate |
| 08:16 | <Hixie> | how is that snide? |
| 08:17 | <Hixie> | i think you're reading far more into mark's humorous comments than is intended |
| 08:17 | <sayrer> | who has a hidden agenda? |
| 08:17 | <Hixie> | the only way that last comment could be offensive is if someone had a hidden agenda, which no-one does as far as i know |
| 08:17 | <sayrer> | but the comment seems to claim they exist |
| 08:18 | <sayrer> | not very productive |
| 08:18 | <sayrer> | I wouldn't call it "precisely accurate" at all |
| 08:18 | <Hixie> | given how productive mark is and how useful his posts are, i think it's reasonable to let him inject some of his humour in there |
| 08:18 | <Hixie> | in what way is it not accurate? |
| 08:19 | <sayrer> | I think it's reasonable to let him write whatever he wants |
| 08:20 | <sayrer> | you just said no one has a hidden agenda, but Mark wrote |
| 08:20 | <sayrer> | "critics who complained about these things as a cover for other agendas" |
| 08:21 | <sayrer> | I would call that "haha only serious" |
| 08:21 | <Hixie> | why, are there people with hidden agendas?! |
| 08:21 | <sayrer> | oh probably |
| 08:21 | <Hixie> | o_O |
| 08:21 | <Hixie> | who? |
| 08:22 | <sayrer> | I dunno, the problem with the comment is that it places some people above reproach |
| 08:22 | <Hixie> | how ever so? |
| 08:23 | <Hixie> | are you seriously suggesting that mark was suggesting that there really are people with hidden agendas?? |
| 08:23 | <sayrer> | yes |
| 08:23 | <Hixie> | ok in that case allow me to disillusion you |
| 08:24 | <Hixie> | mark may have a sense of humour that some would call "british" |
| 08:24 | <sayrer> | oh, I understand that kind of humour |
| 08:24 | <Hixie> | but i assure you that he's not actually suggesting there are people acting in bad faith |
| 08:24 | <sayrer> | Agree to disagree |
| 08:24 | <Hixie> | he likes needling people (just see his blog) |
| 08:24 | <Hixie> | apparently he needled you this time :-) |
| 08:26 | <sayrer> | no, I just think he's smug, and I don't mind saying so |
| 08:26 | <sayrer> | it's not happy and nice |
| 08:26 | <Hixie> | mark is one of the most self-effacing guys out there |
| 08:27 | <sayrer> | ok |
| 08:33 | <MikeSmith> | +1 to more jackassery on the whatwg blog |
| 08:33 | <sayrer> | http://twitter.com/diveintomark/status/1344958151 |
| 08:34 | <MikeSmith> | the whatwg blog should ideally reflect the diversity of opinion on #whatwg, and the color/character of #whatwg discussions |
| 08:34 | <sayrer> | what a self-effacing guy |
| 08:35 | <Hixie> | are you sure you know what "smug" and "self-effacing" mean? that twitter had nothing at all to do with mark, how could it possibly be either? |
| 08:35 | <Hixie> | i'm very confused about your attitude here |
| 08:35 | <sayrer> | yes, I know what both of those mean. do you? |
| 08:36 | <Hixie> | yes, and both require one to say something about oneself |
| 08:36 | <Hixie> | that twitter said nothing about mark |
| 08:36 | <sayrer> | that isn't true |
| 08:37 | <Hixie> | smug: Exhibiting or feeling great or offensive satisfaction with *oneself* |
| 08:37 | <Hixie> | self-effacing: Not drawing attention to *oneself* |
| 08:37 | <sayrer> | what do you need to do to pass judgement on others, in public? |
| 08:38 | <Hixie> | who is passing judgment on something? |
| 08:38 | <sayrer> | especially regarding their moral fiber or motivations |
| 08:38 | <Hixie> | sam himself has acknowledged that his history document missed numerous key points of the history of html, how is mark's statement again anything but factual? |
| 08:39 | <sayrer> | would you say it is "unencumbered as it was by facts, context, or perspective of any kind" ? |
| 08:39 | <Hixie> | yes, the whole document was woefully inaccurate and wildly devoid of context |
| 08:39 | <sayrer> | so, no facts, context, or perspective of any kind? |
| 08:39 | <Hixie> | (literally starting with the first sentence) |
| 08:40 | <Hixie> | it certainly had an interesting perspective |
| 08:40 | <sayrer> | no facts? |
| 08:40 | <Hixie> | so possibly the statement was inaccurate in that regard |
| 08:40 | <Hixie> | i'll certainly grant you that mark, like myself, is prone to exaggeration |
| 08:40 | <sayrer> | no context? |
| 08:40 | <sayrer> | of any kind? |
| 08:41 | <sayrer> | which part of the statement was true? |
| 08:42 | <Hixie> | mark's twitter was an exaggerated indication of inaccuracy in sam's document, and sam's document was indeed wildly inaccurate |
| 08:43 | <Hixie> | sure, it was exaggerated for humorous value, just like mark's blog posts have humorous comments which you apparently take seriously |
| 08:43 | <sayrer> | by exaggerated, you mean wrong? |
| 08:43 | <Hixie> | no, i mean exaggerated. |
| 08:43 | <Hixie> | why are you offended at an obvious joke (people with hidden agendas? really? you thought that was serious?), and why do you now try to insult someone who has spent so much of his time helping us bring html to a greater audience? |
| 08:44 | <sayrer> | he gets paid to that, afaik |
| 08:44 | <Hixie> | so...? |
| 08:45 | <sayrer> | anyway, you think "HTML is being developed outside of the W3C by a number of browser implementers, excluding Microsoft. " |
| 08:45 | <sayrer> | is wrong? |
| 08:45 | <Hixie> | yes |
| 08:45 | <sayrer> | why? |
| 08:46 | <Hixie> | HTML is being developed inside the W3C by 350 people and outside the W3C by over 950 people, and Microsoft contributes roughly equal amounts to html both inside and outside the w3c |
| 08:46 | <Hixie> | the statement is wildly misleading from start to finish |
| 08:46 | <sayrer> | you mean exaggerated? |
| 08:46 | <Hixie> | no, i mean misleading |
| 08:46 | <Hixie> | there's no exaggeration there |
| 08:47 | <Hixie> | if anything it's the opposite |
| 08:47 | <Hixie> | misleading through omission |
| 08:47 | <sayrer> | I don't know. you mean the ommission of development inside the W3C? |
| 08:48 | <sayrer> | too many mms. anyway, there is some W3C mailing list traffic |
| 08:48 | <Hixie> | and the omission of people other than browser vendors, who form the bulk of the contributors, and the omission of the way microsoft's involvement is not any substantially more productive in the w3c than out |
| 08:48 | <Hixie> | but i'm really not interested in discussing sam's document, i don't think it is particularly important |
| 08:48 | <sayrer> | OK |
| 08:48 | <Hixie> | (i and others have already mentioned the numerous mistakes with it) |
| 08:48 | <Hixie> | (to him) |
| 08:49 | <sayrer> | it's mostly quotes |
| 08:49 | <sayrer> | I guess there's a bit of commentary alongside |
| 08:50 | <Hixie> | if you're interested in a significantly more balanced and accurate history of html, btw, MikeSmith wrote one on the esw wiki |
| 08:50 | <MikeSmith> | heh |
| 08:50 | <MikeSmith> | well, there are some that think that is not more balanced or more accurate at all |
| 08:50 | <MikeSmith> | which is why my name is at the top of it now |
| 08:51 | <MikeSmith> | I was asked to put it there to make it clear that, well, it was written by an actual person |
| 08:51 | <Hixie> | i said "more balanced", not that it was perfect :-P |
| 08:51 | <MikeSmith> | a person with opinions |
| 08:51 | <sayrer> | what is the link? |
| 08:51 | <MikeSmith> | http://esw.w3.org/topic/HTML/history |
| 08:51 | <sayrer> | I wonder if it spends a lot of time discussing XHTML |
| 08:52 | <MikeSmith> | there are subpages like http://esw.w3.org/topic/HTML/DraconianErrorHandling |
| 08:52 | <MikeSmith> | sayrer: yeah, both flavors |
| 08:52 | <MikeSmith> | e.g., http://esw.w3.org/topic/HTML/XHTML2/objections |
| 08:53 | <sayrer> | frankly, this one just looks uncontroversial |
| 08:53 | <sayrer> | so I guess there is less to argue about |
| 08:53 | jgraham | wasn't aware that controversy was a goal |
| 08:54 | <sayrer> | it is orthogonal to accuracy |
| 08:54 | <MikeSmith> | sayrer: anyway, you will notice that Mark Pilgrim's name shows up in that history a lot |
| 08:54 | <sayrer> | yours? |
| 08:54 | <MikeSmith> | yeah |
| 08:54 | <sayrer> | where? |
| 08:54 | <MikeSmith> | everywhere |
| 08:54 | <MikeSmith> | he has been like a kind of Greek chorus throughout this whole tragic comedy |
| 08:55 | <sayrer> | in the linked documents? |
| 08:55 | <MikeSmith> | I probably could have dropped a whole more more Pilgrim in there |
| 08:55 | <sayrer> | boy oh boy |
| 08:55 | <MikeSmith> | but I tried to use it judiciously |
| 08:55 | <sayrer> | like... SVG Tiny 1.2 Candidate Recommendation published. |
| 08:55 | <MikeSmith> | Mark's like a fine spice |
| 08:55 | <sayrer> | what is that about? |
| 08:55 | <MikeSmith> | filler |
| 08:56 | <sayrer> | haha |
| 08:56 | <MikeSmith> | to make the peanut gallery happy |
| 08:56 | <sayrer> | oh, XML On The Web Has Failed |
| 08:58 | <sayrer> | it has an entire section titled "Enter RFC 3023" |
| 08:58 | <MikeSmith> | sayrer: feel free to add anything to that history page, if you find gaps |
| 08:58 | <sayrer> | no thanks |
| 08:58 | <MikeSmith> | well, you can even say mean stuff there that you know will really piss people off |
| 08:59 | <MikeSmith> | but just put a smiley after it |
| 08:59 | <MikeSmith> | and then everything will be OK |
| 08:59 | <sayrer> | oh right |
| 08:59 | <MikeSmith> | that's the magic of the smiley |
| 08:59 | <sayrer> | british humour |
| 08:59 | <MikeSmith> | I want to make a smiley that's a little icon of me smiling |
| 08:59 | <MikeSmith> | but flipping the bird at the same time |
| 08:59 | <MikeSmith> | I would use that one a lot if I had it |
| 09:00 | <sayrer> | I'm surprised this hasn't gotten old yet |
| 09:00 | <sayrer> | "you're a cunt :)" |
| 09:00 | <sayrer> | so many forms of that on the internet |
| 09:01 | <jgraham> | MikeSmith: Just make a really popular typeface and thn use some PUA glyphs for extra smilie characters |
| 09:02 | <MikeSmith> | sayrer: you know what playing the dozens is? |
| 09:02 | <sayrer> | yes |
| 09:02 | <MikeSmith> | well, there's only one way to lose that game |
| 09:03 | <sayrer> | I thought there were a couple |
| 09:03 | <sayrer> | but which did you have in mind? |
| 09:04 | <MikeSmith> | the one where you get genuinely mad or hurt by what the other dude says |
| 09:06 | <sayrer> | you can also lose by being an actual asshole |
| 09:06 | <MikeSmith> | nope |
| 09:06 | <MikeSmith> | not in that game |
| 09:07 | <MikeSmith> | it's irrelevant |
| 09:07 | <MikeSmith> | I'm not defending it as a great game, by the way |
| 09:07 | <sayrer> | there are lots of seemingly rude things that are allowed, sure |
| 09:07 | <zcorpan__> | "The fact that different user agents deal with protocol violations in different ways is a good thing." -- http://lists.w3.org/Archives/Public/ietf-http-wg/2009AprJun/0030.html |
| 09:08 | <takkaria> | uh-oh |
| 09:10 | <MikeSmith> | zcorpan__: that is a gem |
| 09:10 | <MikeSmith> | [[ |
| 09:10 | <MikeSmith> | It cannot be standardized in a way that would be safe for safety- |
| 09:10 | <MikeSmith> | critical |
| 09:10 | <MikeSmith> | environments such as health care, where failure to display the errors |
| 09:10 | <MikeSmith> | could very well result in serious injury or death. |
| 09:10 | <MikeSmith> | ]] |
| 09:10 | <MikeSmith> | solid gold, that |
| 09:10 | <MikeSmith> | the "serious injury or death" card |
| 09:10 | <sayrer> | yeah, close to godwin's law |
| 09:11 | <sayrer> | on the other hand, they aren't protocol violations if everyone handles them the same way |
| 09:11 | <Hixie> | aah, mikesmith is ready roy's latest offering |
| 09:11 | <jgraham> | "The correct way to interoperate with broken Web |
| 09:11 | <jgraham> | content is to display a very large error message that explains why |
| 09:11 | <jgraham> | it is broken" |
| 09:12 | <jgraham> | I see Roy is pitching for that UI led job he's always coveted |
| 09:12 | <MikeSmith> | sayrer: they are because the Bible and/or the Founding Fathers said they should be violations |
| 09:12 | <jgraham> | *lead |
| 09:13 | <takkaria> | I do wonder what to do with a problem like Roy |
| 09:13 | <MikeSmith> | jgraham: very well aligned with market realities |
| 09:13 | <hsivonen> | It would be "fun" to generate a version of the HTML5 parser for Gecko that halts on the first parse error |
| 09:13 | <Hixie> | sayrer: i think there are things that are clearly errors even if all UAs interoperate on the behaviour |
| 09:13 | <Hixie> | sayrer: e.g. <span></em> is clearly an error |
| 09:13 | <jgraham> | takkaria: I guess sending him off to work as a children's nanny won't work? |
| 09:14 | <sayrer> | <b><i>hmm</b></i> ? |
| 09:14 | <Hixie> | that's another example |
| 09:14 | <Hixie> | the markup doesn't match the resulting DOM at all |
| 09:14 | <sayrer> | where is the error? |
| 09:15 | <sayrer> | we should have a good reason to call it an error |
| 09:15 | <sayrer> | it's certainly ugly |
| 09:15 | <takkaria> | jgraham: I have a feeling his approach to children when they do soething wrong is display a very large erorr message that explains what they've done wrong, probably not too lenient on the error-correcting side of things... |
| 09:15 | <annevk42> | if you put a character after </b> you cannot tell what DOM you get |
| 09:15 | <annevk42> | unless you know how an HTML parser works |
| 09:15 | <annevk42> | therefore it makes sense to flag the simple case as well |
| 09:15 | <sayrer> | we're assuming a conformant HTML parser |
| 09:16 | <sayrer> | what else would we talk about? |
| 09:16 | <annevk42> | I am too |
| 09:16 | <MikeSmith> | takkaria: we don't want to do anything. he's actually a secret double agent of the whatwg cabal |
| 09:16 | <annevk42> | I'm just saying that getting nodes duplicated is not what you expect to happen |
| 09:16 | <sayrer> | but you can check it in your browser and see if it is what you want |
| 09:16 | <MikeSmith> | he appears in times of trouble to act as a sort of common foil to unite the cabal |
| 09:16 | <takkaria> | how about we call things parse errors where the DOM and the text/html don't intuitively match up? |
| 09:18 | <annevk42> | sayrer, that doesn't really lead to understanding how it works or why it works in unexpected ways |
| 09:18 | <sayrer> | so parse errors have a pedagogical goal? |
| 09:18 | <jgraham> | sayrer: What if it isn't what you want? How do you tell where you made a mistake? |
| 09:19 | <sayrer> | a fair point! |
| 09:19 | <sayrer> | the difference between a C compiler and a lint tool, right? |
| 09:20 | <jgraham> | Hmm. Not sure that's a good analogy. biab |
| 09:20 | <takkaria> | I don't know what other goal they would have |
| 09:21 | <takkaria> | HTML prior to HTML5 never had parse errors |
| 09:21 | <roc> | it's important to have a common enemy to unite against |
| 09:22 | <roc> | hsivonen: you need to implement that strict parser, ship it to Roy and force him to use it exclusively |
| 09:23 | <sayrer> | "The error handling for parse errors is well-defined" |
| 09:23 | <sayrer> | pretty much says it all |
| 09:24 | <annevk42> | I doubt it says the same to you and me |
| 09:24 | <sayrer> | well, what does it say to you? |
| 09:25 | <zcorpan__> | it's just a statement of fact, might not actually be true |
| 09:26 | <zcorpan__> | just like this one http://html5.org/tools/web-apps-tracker?from=2951&to=2952 |
| 09:27 | <Hixie> | zcorpan__: btw, please send comments on typos and things by e-mail, i can't track them when you just mention them on IRC |
| 09:27 | <annevk42> | sayrer, that requirements for authors are different from implementors |
| 09:27 | <sayrer> | when I see <b><i>hmm</b></i> |
| 09:28 | <sayrer> | I definitely agree there is an aspect "this may not mean what you think it means" |
| 09:28 | MikeSmith | likes roc idea of inviting people to try browsing the Web with a strict-HTML-parser browser for one day or whatever. or even 1 hour. or ten minutes |
| 09:28 | <annevk42> | actually, the simple case does what you think it means |
| 09:28 | <sayrer> | beside the point |
| 09:28 | <annevk42> | problems starts arising when you enter a few more characters between the tags |
| 09:28 | <sayrer> | but also invalid? |
| 09:28 | <sayrer> | or a parse error? |
| 09:28 | <annevk42> | sure |
| 09:28 | <sayrer> | why, it does what I think it means |
| 09:29 | <Hixie> | <b>a<i>b</b>c</i> is different from <x>a<y>b</x>c</y>, too |
| 09:29 | <annevk42> | it's more likely you made a typo |
| 09:29 | <zcorpan__> | Hixie: k |
| 09:29 | <sayrer> | I agree that is possible |
| 09:29 | <sayrer> | but it's not an interoperability problem |
| 09:30 | <Hixie> | sayrer: do you think we should allow <blink>, too? |
| 09:30 | <sayrer> | yes |
| 09:30 | <Hixie> | ok |
| 09:30 | <takkaria> | if you mentally do s/parse error/warning/, does that solve your problems? |
| 09:30 | <Hixie> | i am confident that your opinion is far from the consensus |
| 09:30 | <sayrer> | not really |
| 09:30 | <sayrer> | no big deal |
| 09:30 | <sayrer> | it's not like I'm going to propose blink |
| 09:31 | <sayrer> | since, well, who cares |
| 09:31 | <annevk42> | if you don't believe in having differences for authors and impl at all I can see why we differ from opinion :) |
| 09:31 | <sayrer> | I think there is value in helping authors understand what they have written |
| 09:31 | <zcorpan__> | sayrer: it's easy to make everything allowed - just don't read the spec and don't use a validator |
| 09:32 | <sayrer> | that is true for authors |
| 09:32 | <zcorpan__> | yes i meant as an author |
| 09:32 | <zcorpan__> | if you're not an author, why do you care about conformance requirements on authors? :) |
| 09:32 | <sayrer> | I am obviously an author |
| 09:33 | <sayrer> | but I don't care about them at all, true |
| 09:33 | <sayrer> | I have noticed that most of the controversy in the spec centers around author requirements |
| 09:34 | <Hixie> | removing teh author requirements won't remove the controversies |
| 09:34 | <zcorpan__> | removing the author requirements doesn't mean that the controversy is removed, too |
| 09:34 | <Hixie> | it would add a lot more :-) |
| 09:34 | <sayrer> | how so? |
| 09:34 | <Hixie> | well for a start i would start bitching :-) |
| 09:34 | <sayrer> | but why? |
| 09:34 | <sayrer> | is more the question |
| 09:34 | <zcorpan__> | people would go "omg html5 allows <font> again! and tag soup! what a mess!" |
| 09:35 | <Hixie> | because i think have conformance rules for authors is an extremely valuable tool for quality assurance and for learning the language |
| 09:35 | <sayrer> | sounds like what the IETF would call a BCP document |
| 09:35 | <sayrer> | and that is something that would be valuable, I agree |
| 09:36 | <sayrer> | I am not sure I would call them "conformance rules" |
| 09:36 | <sayrer> | still thinking of a better way to put it |
| 09:36 | <annevk42> | if you do not want authoring conformance at all it might be easier to start with that next time rather than why <i><b>test</i></i> is non-conforming |
| 09:37 | <annevk42> | would save me some time |
| 09:37 | <Hixie> | call them what you like, so long as not following them is considered an error |
| 09:37 | <Hixie> | and is not allowed |
| 09:37 | <sayrer> | it is obviously allowed |
| 09:37 | <sayrer> | as HTML5 is written currently |
| 09:38 | <Hixie> | given that html5 explicitly disallows the use of incorrect syntax and incorrect semantics, "obvious" is not the word i would use |
| 09:38 | <Hixie> | "not" maybe would be the word |
| 09:39 | <sayrer> | hmm, there don't seem to be a lot of teeth there |
| 09:40 | <sayrer> | like, Apple's HTML5 homepage had incorrect semantics |
| 09:40 | <Hixie> | teeth? |
| 09:40 | <sayrer> | well, if you use English incorrectly, people might think of less of you, or maybe not even understand you |
| 09:40 | <Hixie> | right, and it's incorrect |
| 09:41 | <Hixie> | just like using incorrect html |
| 09:41 | <Hixie> | it's not allowed by english syntax and grammar |
| 09:41 | <Hixie> | it's not allowed by html syntax and grammar |
| 09:41 | <sayrer> | but with HTML5, you can do that and people will read your web page just fine |
| 09:41 | <Hixie> | same concept |
| 09:41 | <sayrer> | and not notice at all |
| 09:41 | <Hixie> | so? |
| 09:41 | <Hixie> | why would we want them to notice |
| 09:41 | <sayrer> | we don't |
| 09:42 | <zcorpan__> | you can make them notice by having a draconian html parser which is allowed in the spec |
| 09:42 | <sayrer> | so why wag a finger over grammar |
| 09:42 | <zcorpan__> | the v.nu html parser is draconian for certain cases in streaming mode |
| 09:42 | <roc> | the only real solution is for Google to dish out search points based on HTML5 conformance |
| 09:43 | <sayrer> | adsense rates are down, might not work ;) |
| 09:43 | <MikeSmith> | sayrer: authors love to be told exactly what to do, even they don't understand the reasons |
| 09:43 | <sayrer> | yep |
| 09:43 | <annevk42> | Hixie, proposed blog post seems fine |
| 09:44 | <annevk42> | Hixie, though I'd indicate they can e-mail public-html-comments as well |
| 09:44 | <annevk42> | Hixie, that's easier to post to than whatwg⊙wo |
| 09:44 | <Hixie> | annevk42: good idea |
| 09:44 | <sayrer> | so the main reason for these conformance rules is so that people won't say bad things? |
| 09:45 | <sayrer> | like "ooooh, tag soup! <blink>!" |
| 09:46 | <annevk42> | it's recommending how to author HTML |
| 09:46 | <annevk42> | it doesn't prevent anyone from doing something else |
| 09:46 | <sayrer> | actually, it's requiring |
| 09:46 | <Hixie> | there are dozens of reasons for having authoring requirements, many of which have been discussed tonight |
| 09:46 | <Hixie> | - helping people learn how to write effective markup |
| 09:46 | <annevk42> | sayrer, only if you follow the spec :) |
| 09:46 | <Hixie> | - helping people track down typos and other errors |
| 09:46 | <MikeSmith> | sayrer: authors also love to have conformance rules so that they can beat other authors over the head with them, when those other authors make mistakes |
| 09:46 | <Hixie> | - helping people write markup that reflects what they intend |
| 09:47 | <Hixie> | - encouraging best practices |
| 09:47 | <sayrer> | that last one seems bogus |
| 09:47 | <MikeSmith> | we need to support that for-beating-others-over-the-head-with use case |
| 09:47 | <Hixie> | - satisfying the needs of authors who feel the need for rules to guide them |
| 09:47 | <sayrer> | or at least redundant |
| 09:47 | <Hixie> | you may agree with some or none of these |
| 09:48 | <sayrer> | I actually find the list quite valuable |
| 09:48 | <sayrer> | are there more? |
| 09:48 | <zcorpan__> | - make people aware of accessiblity issues? at least that seems to be a point pf wants to make about alt |
| 09:48 | <Hixie> | there's lots, yes |
| 09:48 | <sayrer> | do they all start with "help people..." |
| 09:48 | <Hixie> | zcorpan__: yeah, that's similar to the best practices thing |
| 09:48 | <sayrer> | ? |
| 09:48 | <Hixie> | sayrer: no idea |
| 09:49 | <sayrer> | well, I guess your current list is enough for me to work with |
| 09:50 | <sayrer> | but I would like to know whether any of the author conformance requirements actually have an effect on conformant HTML parsers |
| 09:50 | <sayrer> | it doesn't seem that they do |
| 09:50 | <sayrer> | which I think is a fine design, I should add |
| 09:51 | <Hixie> | author conformance requirements have nothing to do with parsers |
| 09:51 | <annevk42> | http://www.w3.org/mid/e9dffd640904011708l3dd36d28xc6e797791777f4df⊙mgc maybe we need WTP, like HTTP, but for Web browsers :p |
| 09:52 | <zcorpan__> | Web HTTP |
| 09:53 | <sayrer> | Hixie, ok, it's almost like they could be in a different document |
| 09:55 | <Hixie> | i think that would be a maintenance nightmare |
| 09:55 | <Hixie> | but as it stands the html5 spec can be filtered to show only the author stuff |
| 09:56 | <Hixie> | which is a middle ground that seems to address the same needs |
| 09:56 | <sayrer> | can it do the opposite? |
| 09:56 | <Hixie> | only show implementation stuff? no, but it can highlight implementation stuff, which is about as good |
| 09:56 | <Hixie> | (in practice implementations need to know what author rules are because some implementations are conformance checkers) |
| 09:57 | <Hixie> | (and annotation the document in more detail would be far too much work) |
| 09:57 | <sayrer> | is the author stuff marked up with a specific class? |
| 09:57 | <Hixie> | (for me, at least) |
| 09:57 | <sayrer> | I haven't looked since you did your great big edit |
| 09:57 | <Hixie> | the implementation requirements have class="impl" |
| 09:58 | <sayrer> | I can't believe that HTTP WG thread |
| 09:58 | <annevk42> | (but in a way that you cannot extract all class="impl" and have it make sense) |
| 09:58 | <Hixie> | right, imeplementation requirements share definitions with author stuff |
| 09:58 | <Hixie> | anyway |
| 09:58 | <Hixie> | i should sleep |
| 09:58 | <Hixie> | nn |
| 09:59 | <MikeSmith> | cool to see the http mailing list providing some real entertainment value again |
| 09:59 | <MikeSmith> | imaginary numbers or no less important than real numbers |
| 09:59 | <takkaria> | annevk42: I wonder how large the non-html/js/jpeg portion of the web is |
| 10:00 | <annevk42> | large enough for the HTTP WG to exist and not care about it |
| 10:00 | <annevk42> | not care about the html/js/jpeg portion, that is |
| 10:01 | <zcorpan> | so they care about the ms word and pdf portion? |
| 10:02 | <annevk42> | harhar |
| 10:02 | <MikeSmith> | why make handling of existing Web content such a huge priority when we have these whole undiscovered galaxies where clients and servers exchange stuff other than things like HTML, Javascript, and JPGs? |
| 10:02 | <MikeSmith> | the logic is perfectly clear |
| 10:02 | <sayrer> | I suggested publishing Barth's document as informational |
| 10:02 | <sayrer> | everyone shook their head at me |
| 10:03 | <MikeSmith> | sayrer: I think they shook their beards |
| 10:03 | <annevk42> | really? I'd think that's what the HTTP WG guys want |
| 10:03 | <sayrer> | they didn't want to publish it without a specific section in the HTTP standard saying "something might happen here" |
| 10:03 | <MikeSmith> | the council of wise men who don't suffer fools lightly |
| 10:04 | <sayrer> | I don't really see why |
| 10:04 | <annevk42> | it might kill people apparently |
| 10:04 | <sayrer> | since the value of the information will be the same, no matter what the HTTP spec says |
| 10:04 | <MikeSmith> | I hope Adam has learned his lesson feels sufficiently chastised |
| 10:04 | <MikeSmith> | all that nonsense about doing real research and collecting data. who cares about that? |
| 10:04 | <annevk42> | I think it should be more than informational personally |
| 10:04 | <sayrer> | why? |
| 10:04 | <MikeSmith> | it will be |
| 10:05 | <sayrer> | what does it matter? |
| 10:05 | <MikeSmith> | sayrer: because it will get implemented |
| 10:05 | <MikeSmith> | or something like it |
| 10:05 | <sayrer> | W3C standards are only recommendations |
| 10:05 | <annevk42> | sayrer, by that metric we might as well not bother at all and just make it publicly available |
| 10:05 | <annevk42> | e.g. by sending a copy to www-archive |
| 10:05 | <sayrer> | JSON is only an informational RFC |
| 10:05 | <MikeSmith> | sayrer: the ones that aren't just recommendations are called "implementations" |
| 10:06 | <MikeSmith> | or "implemented" |
| 10:06 | <annevk42> | sayrer, I don't quite get that either, though I understand JSON has some interop issues |
| 10:08 | <sayrer> | DNS is an informational RFC too |
| 10:08 | <sayrer> | the status doesn't matter |
| 10:08 | <MikeSmith> | sayrer: it's moot point anyway |
| 10:08 | <annevk42> | sayrer, mkay |
| 10:08 | <sayrer> | what do you mean? |
| 10:09 | <MikeSmith> | they clearly aren't going to publish it regardless |
| 10:09 | <sayrer> | you think? |
| 10:09 | <sayrer> | why? |
| 10:09 | <zcorpan> | hmm. here's an idea -- have the same tree builder as in html but use a different namespace and do namespace fixups on attributes when in foreign |
| 10:09 | <zcorpan> | then support /> on "unknown" elements everywhere and do element case fixup everywhere |
| 10:10 | <MikeSmith> | sayrer: because it's too clearly aligned with the real world |
| 10:10 | <sayrer> | I think it is too conservative actually |
| 10:10 | <sayrer> | maybe they would go for it! |
| 10:10 | <annevk42> | zcorpan, idea for what? |
| 10:11 | <zcorpan> | annevk42: handling svg and mathml in html in an html-y way |
| 10:11 | <zcorpan> | allowing proper speculative tokenization |
| 10:12 | <zcorpan> | even past <svg> ... <select><style> ... |
| 10:13 | <zcorpan> | hsivonen: ^ |
| 10:13 | <zcorpan> | sicking: ^ |
| 10:14 | <annevk42> | I'm not really following, but ok |
| 10:14 | <zcorpan> | instead of having insertion modes, have a flag for foreign content and use the insertion modes you would use for html (except you insert elements with a different namespace) |
| 10:16 | <Philip`> | Hixie: No |
| 10:16 | <Philip`> | but I am now |
| 10:16 | Philip` | is having fun upgrading his kernel on Gentoo and switching to a new wireless driver architecture and wondering why kdm no longer displays anything on the screen |
| 10:17 | <annevk42> | zcorpan, and then have special handling for some elements? |
| 10:18 | <annevk42> | zcorpan, e.g. <image> |
| 10:18 | <annevk42> | zcorpan, but why can't you speculate past <svg> currently? |
| 10:19 | <hsivonen> | annevk42: currently, you can do speculative tree building past <svg> but you can't do cheap speculative tokenization without tree building |
| 10:20 | <zcorpan> | annevk42: forgot about <image> |
| 10:20 | <hsivonen> | I want to instrument the HTML5 parser in Gecko and get some data about the feasibility of doing speculative tree building only before I express more opinions on this topic |
| 10:20 | <zcorpan> | but yeah, you could check the flag for <image> |
| 10:21 | <hsivonen> | zcorpan: case fixup everywhere might be a good idea regardless of speculative parsing if we end up with a magic mapping for selectors and getElementsByTagName |
| 10:22 | <annevk42> | bz said that making selectors and gEBTN work like HTML5 currently requirescould work |
| 10:22 | annevk42 | much rather has that |
| 10:24 | <hsivonen> | annevk42: it would bloat all interned node ns/local/prefix interned struct by another pointer and would require changes to code is unexpected places |
| 10:24 | <hsivonen> | annevk42: but clearly, storing pre-lowercased local names for each node would work |
| 10:25 | <annevk42> | you could also do 2 pointer checks at matching time |
| 10:25 | <hsivonen> | annevk42: what's the other pointer check? |
| 10:25 | <annevk42> | isHTMLElement && equallowercasetoken || equaltoken |
| 10:26 | <annevk42> | might need parenthesis there |
| 10:26 | <hsivonen> | if there were pre-lowercased atoms for each node, it would make sense to have those on all nodes anyway to avoid an extra check |
| 10:27 | <zcorpan> | if the parser camelcases everywhere, should createElement similarly camelcase? |
| 10:27 | <hsivonen> | anyway, doing it "right" adds some RAM footprint and requires a review of a lot of code only to support non-conforming cases that currently Gecko, WebKit and IE don't support and only Opera currently supports |
| 10:28 | <hsivonen> | zcorpan: maybe, but you'd still get the wrong namespace |
| 10:28 | <annevk42> | hsivonen, that wouldn't allow for gEBTN("textarea") to not match textArea but gEBTN("textArea") to match textArea |
| 10:28 | <zcorpan> | hsivonen: yep |
| 10:28 | <hsivonen> | annevk42: that can be solved by not supporting textArea |
| 10:29 | <annevk42> | same example for any other camelCase name |
| 10:29 | <zcorpan> | there are no other conflicting camelcase names |
| 10:29 | <zcorpan> | or are there? |
| 10:29 | <hsivonen> | AFAIK, textArea is the only conflict |
| 10:29 | <annevk42> | I don't think so, but you might still not want them to match case-insensitively |
| 10:30 | <hsivonen> | and textArea is a bad idea in the first place since it duplicates CSS formatter functionality |
| 10:30 | <annevk42> | and it also doesn't help with createElementNS() nodes |
| 10:30 | <annevk42> | and I think it would be nice if it did work with createElementNS nodes |
| 10:30 | <hsivonen> | annevk42: it helps with conforming createElementNS() nodes |
| 10:30 | <annevk42> | only if you change createElementNS |
| 10:31 | <zcorpan> | if you want to select elements with a certain case, use getEBTNNS |
| 10:31 | <hsivonen> | annevk42: the only issue is if a WG somewhere creates a camelCaseML for XML and then realizes they need to ship an Ajax library to hack support into text/html |
| 10:31 | <annevk42> | Is (isHTMLElement && equallowercasetoken) || equaltoken really that bad? |
| 10:32 | <hsivonen> | annevk42: it's worse than simply (caseFoldAtom == queryAtom) |
| 10:34 | <hsivonen> | I'd be surprised if Opera implemented what it currently does in a way other than doing what Maciej said on the list |
| 10:35 | <hsivonen> | as far as I can tell, adding a pointer to all Gecko node infos and WebKit qualified names would be the Right Way to implmement real ASCII-case-insensitivity |
| 10:40 | <annevk42> | hsivonen, we match createElementNS("x","X") with gEBTN("x") |
| 10:40 | <annevk42> | (not sure that's what I want either) |
| 10:41 | <hsivonen> | annevk42: that neither proves nor disproves my hypothesis of Opera internals :-) |
| 10:42 | <hsivonen> | my hypothesis isn't blackbox testable except by observing memory footprint and timing execution |
| 10:43 | <annevk42> | fair enough |
| 10:55 | <annevk42> | sweet |
| 10:55 | <annevk42> | the Forms WG published a whole bunch of Member confidential information |
| 10:55 | <annevk42> | http://lists.w3.org/Archives/Public/public-forms/2009Apr/att-0000/2009-04-01.html |
| 10:55 | <annevk42> | regarding the AC meeting |
| 11:03 | <annevk42> | not too interesting though |
| 11:04 | <MikeSmith> | annevk42: I should remove that |
| 11:04 | <MikeSmith> | but I don't think I actually have the ability to remove attachments from teh archive |
| 11:04 | <annevk42> | I hope someone makes a copy |
| 11:14 | <hsivonen> | already gone :-/ |
| 11:15 | <MikeSmith> | temporarily gone |
| 11:15 | <annevk42> | I do have a copy |
| 11:18 | Philip` | wonders what genius thought the new wireless drivers should make the LED blink at about 4Hz every time a packet is sent or received |
| 11:24 | <jgraham> | Philip`: Presumably one that doesn't suffer from photosensitive epilepsy |
| 11:26 | <annevk42> | Hixie, your ARIA comments went to the wrong list |
| 11:26 | <annevk42> | Hixie, they need to go to public-pfwg-comments⊙wo |
| 11:35 | <hsivonen> | Hixie: why is there "XXX quirks" on 'hr' start tag? as far as I can tell, none of Gecko, WebKit, Opera and IE8 have that quirk |
| 11:39 | <hsivonen> | hmm. Safari and Opera split the text node on <body>foo</h1>bar |
| 11:41 | <annevk42> | I guess I should stop reviewing non-normative materal in WAI-ARIA |
| 11:45 | <hsivonen> | MikeSmith: I redeployed V.nu |
| 11:46 | <MikeSmith> | hsivonen: cool. hoping I didn't introduce any regresssions |
| 11:49 | <annevk42> | you should probably add tests for code you touch :) |
| 11:50 | <zcorpan> | hsivonen: ie nests the hr in p if there was no text in the p |
| 11:52 | <hsivonen> | zcorpan: ah. I tested with text |
| 11:52 | <hsivonen> | zcorpan: any indication of Web actually needing that? |
| 11:52 | <zcorpan> | seems not since other browsers don't do it |
| 11:53 | <zcorpan> | actually opera nests hr in p if there's another element in between |
| 11:53 | <zcorpan> | but i haven't tested whether styling agrees |
| 11:54 | <zcorpan> | nothing surprising with the styling |
| 11:55 | <hsivonen> | so it seemss the only quirk left in the spec XXX that everyone implements is <p><table> |
| 11:55 | <zcorpan> | gecko nests hr for <p><span><hr>, too |
| 11:56 | <hsivonen> | but it's the <span>, not <hr> that is special, right? |
| 12:07 | <zcorpan> | in gecko yes |
| 12:08 | <zcorpan> | in opera span is just as special as em and i, but unknown elements don't cause nesting |
| 12:09 | <annevk42> | doesn't IE nest <hr> inside <p>? |
| 12:09 | <zcorpan> | annevk42: sometimes |
| 12:10 | <zcorpan> | hsivonen: actually the styling doesn't agree with the dom in opera for <p><span><hr>x |
| 12:10 | <zcorpan> | we style it more like as if the p was closed |
| 12:12 | hsivonen | wonders what the Opera CSS frame constructor is like |
| 12:12 | <zcorpan> | i could tell you but then we'd have to hire you ;) |
| 12:17 | <zcorpan> | hsivonen: ie inserts an empty element for most blockish end tags |
| 12:17 | <zcorpan> | hsivonen: like e.g. </plaintext> |
| 12:18 | <zcorpan> | hsivonen: only </p> is needed for compat :) |
| 12:19 | Philip` | supposes </plaintext> is pretty rare in practice |
| 12:19 | <Philip`> | seeing as even people legitimately using the obsolete element won't use it |
| 12:20 | <Philip`> | but it wouldn't surprise me to see a non-zero number of occurrences of it anyway :-( |
| 12:20 | <hsivonen> | zcorpan: Interesting. that suggests that the </h1> stuff isn't important |
| 12:20 | <zcorpan> | hsivonen: indeed |
| 12:21 | <annevk42> | OMG: http://www.w3.org/TR/wai-aria/#textequivalentcomputation |
| 12:21 | <annevk42> | make that OMG: http://www.w3.org/TR/2009/WD-wai-aria-20090224/#textequivalentcomputation |
| 12:21 | <annevk42> | 'cause hopefully they fix this mess |
| 12:23 | <annevk42> | a lot of this ARIA stuff is just written down way more fricking complex than necessary |
| 12:23 | <annevk42> | you don't need RDF or OWL |
| 12:23 | <annevk42> | not for conformance checkers anyway, which is what the initial claim in the draft is |
| 12:24 | <annevk42> | it's a bit sad that all this is so over engineered, I imagine they spent a lot of time on these things for zero benefit |
| 12:25 | <annevk42> | I'm now actually making comments, though not on most of the editorial mayhem: http://lists.w3.org/Archives/Public/public-pfwg-comments/2009AprJun/ |
| 12:26 | <annevk42> | OMG2: http://www.w3.org/TR/2009/WD-wai-aria-20090224/rdf_model.png |
| 12:32 | <Philip`> | Oh, my LEDs stopped blinking |
| 12:32 | <Philip`> | and I'm not sure why, since I tried to run some script I found on the internet to disable the blinking but it had no effect for at least an hour |
| 12:33 | hsivonen | wonders if there's a generic UML<->RDF mapping |
| 12:33 | <MikeSmith> | annevk42: point taken (about tests for v.nu patches) |
| 12:46 | <MikeSmith> | speaking of playing the dozens, I really miss that mookid conneg dude |
| 12:50 | <hsivonen> | Argh. I broke XSLT. |
| 12:50 | <hsivonen> | hmm. maybe not |
| 12:54 | <hsivonen> | does anyone happen to know if Web compat requires http://mxr.mozilla.org/mozilla-central/source/content/html/content/test/347174transform.xsl to 'work' without neither the XHTML namespace nor the html output method? |
| 13:17 | <hsivonen> | Is ARIA Best Practices in LC or just ARIA itself? |
| 13:21 | <annevk42> | just ARIA |
| 13:28 | <annevk42> | someone argued to me about live-region yesterday |
| 13:28 | <annevk42> | live-regions are apparently dropped |
| 13:28 | <jgraham> | ?! |
| 13:28 | <jgraham> | I thought live regions were supposed to be the most useful part |
| 13:28 | <annevk42> | or is that just a comment? |
| 13:29 | <annevk42> | ah, never mind |
| 13:29 | <annevk42> | it was simplified or something: http://lists.w3.org/Archives/Public/public-pfwg-comments/2009JanMar/0035.html |
| 13:29 | annevk42 | read it too quickly initially |
| 13:29 | <annevk42> | I'm getting sick of all these auto-responders |
| 13:30 | <MikeSmith> | "make tests, not war" |
| 13:32 | <annevk42> | is that a DanC quote? |
| 13:32 | <annevk42> | :) |
| 13:40 | <annevk42> | http://www.w3.org/mid/201536C5-96EB-4029-B4A7-D3FB0EAF25FA⊙ao "I strongly believe that specialized mime-types will be useful in the future!" (MathML 3.0 will have 3 apparently.) |
| 13:42 | <hsivonen> | sigh. does it still have xml:id, too? |
| 13:43 | <annevk42> | I guess some people missed the news about media types, XML, etc. |
| 13:48 | <hsivonen> | has anyone filed a formal objection about xml:id in MathML 3, yet? |
| 13:48 | <annevk42> | i filed a comment |
| 13:49 | <hsivonen> | annevk42: has it been handled yet? |
| 13:49 | <annevk42> | no |
| 13:49 | <hsivonen> | ok |
| 13:50 | <annevk42> | but I filed my comment in April and they have published something since that date |
| 13:52 | <hsivonen> | annevk42: happens all the time with HTML5 :-) |
| 13:52 | <annevk42> | true |
| 13:53 | <hsivonen> | considering all the MUST around @alt, ARIA sure says a lot of SHOULD for authoring reqs |
| 13:54 | <annevk42> | hehe |
| 13:57 | <MikeSmith> | I wish we could all avoid using the words "formal objection" |
| 13:57 | <MikeSmith> | it is a blunt instrument |
| 13:57 | <jgraham> | MikeSmith: I would like to make a formal objection to that wish |
| 13:57 | <MikeSmith> | jgraham: good smartass-ery |
| 13:58 | <MikeSmith> | we need more of that |
| 13:58 | <MikeSmith> | need to pad out whatwg blog with more |
| 13:59 | <jgraham> | MikeSmith: Actually I thought it was pretty lame. I almost didn't bother. Then I realised that consstency is the hobgoblin of little minds, so I ought to be consistent in my lameness. |
| 13:59 | <MikeSmith> | heh |
| 13:59 | <MikeSmith> | even better |
| 13:59 | <annevk42> | hobgoblin is the name of my dreamhost server |
| 13:59 | <Philip`> | The W3C should require all formal objections to actually be formal, and hand-written on paper with an official letterhead and wax seal |
| 14:00 | <MikeSmith> | yeah, and delivered in person |
| 14:00 | <MikeSmith> | and you only get one per lifetime |
| 14:00 | <jgraham> | And signed in blood |
| 14:00 | <jgraham> | Your own. |
| 14:00 | <jgraham> | Lots. |
| 14:00 | <MikeSmith> | and if you ever try to file a 2nd one, we have somebody who comes and visits you and kills you |
| 14:00 | <jgraham> | Hence the one per lifetime restriction |
| 14:01 | <MikeSmith> | draconian error handling |
| 14:01 | <hsivonen> | like suggesting new C++ features? |
| 14:06 | <hsivonen> | what kind of sharp instruments are available? |
| 14:07 | <MikeSmith> | hsivonen: assertions of technical merit |
| 14:16 | <jgraham> | MikeSmith: Arguments from technical merit don't seem to be sharp enough to cut through philosophical divides |
| 14:18 | <MikeSmith> | blunt objections don't work any better |
| 14:24 | <Philip`> | Blunt objects work better than blunt objections |
| 14:24 | <jgraham> | MikeSmith: Sure |
| 14:25 | <jgraham> | Philip`: Do sharp objects work better than blunt objects? |
| 14:26 | <MikeSmith> | sharp object slide in between the ribs better |
| 14:28 | <zcorpan> | Hixie: why did you change the forum antispam message from "FIVE"? it's annoying to have to copy-paste every time |
| 14:28 | <Philip`> | MikeSmith: If you're trying to slide a blunt object between people's ribs, you're doing it wrong |
| 14:28 | <Philip`> | I suggest smashing them on the back of the head |
| 14:29 | <MikeSmith> | kneecaps |
| 14:31 | <Philip`> | That's less fatal |
| 14:32 | <Philip`> | (That's just an observation of fact, not a comment on whether I think that's a benefit or a drawback) |
| 15:02 | <MikeSmith> | blockquote is not allowed in footer, right? |
| 15:02 | <MikeSmith> | neither in header nor footer? |
| 15:03 | <MikeSmith> | hmm, or it is allowed in footer? |
| 15:03 | <MikeSmith> | if so, I wonder why |
| 15:04 | <Philip`> | If online purchasing systems redirecting me to https://www.securesuite.co.uk (who I've never heard of and have no reason to trust) and telling me to enter my online banking password weren't bad enough, I've now found one that opens that site in a snazzy drop-shadowed iframe, which means I can't tell if the page inside the iframe even has a valid certificate and isn't https://a-really-dodgy-domain.info/ |
| 15:06 | <zcorpan> | MikeSmith: should be allowed if i'm reading the spec correctly |
| 15:06 | <zcorpan> | MikeSmith: blockquote is sectioning root but sectioning root is not sectioning content |
| 15:06 | <MikeSmith> | hmm |
| 15:07 | <MikeSmith> | so why should it not be allowed in header as well? |
| 15:07 | <zcorpan> | it is allowed in header afaict |
| 15:07 | <zcorpan> | header just bans sectioning content |
| 15:09 | <zcorpan> | MikeSmith: i think blockquote was sectioning content before, which is probably why v.nu disallows it |
| 15:09 | <MikeSmith> | zcorpan: |
| 15:09 | <MikeSmith> | zcorpan: OK |
| 15:10 | <zcorpan> | hmm.. <header><blockquote><h1></h1></blockquote></header> seems to be allowed in the content model |
| 15:11 | <MikeSmith> | yeah, 'cause I can't see any constraint assertions in the current spec to indicate that blockquote is not allowed in header or footer |
| 15:11 | <zcorpan> | but the outline ignores teh blockquote |
| 15:11 | <zcorpan> | i think |
| 15:11 | <MikeSmith> | maybe that's the issue |
| 15:11 | <jgraham> | zcorpan: Yes |
| 15:11 | <jgraham> | Because the <h1> is part of the quote so it is not the heading of the current page |
| 15:12 | <zcorpan> | where's gsnedders' outline tool? |
| 15:12 | <MikeSmith> | hsivonen: blockquote in header and footer? |
| 15:14 | <zcorpan> | gsnedders: please add a textarea to your outliner |
| 15:14 | <MikeSmith> | v.nu assertions.sch still says, "The sectioning element blockquote..." |
| 15:14 | <MikeSmith> | which it's not a sectioning element |
| 15:14 | <MikeSmith> | or not now at least |
| 15:15 | <MikeSmith> | makes me thinks maybe the rationale for that constraint is outdated/not in sync with current spec |
| 15:15 | <zcorpan> | so clearly the content model for header needs to be smarter about h1 and blockquote |
| 15:21 | <zcorpan> | jgraham: gsnedders outliner doesn't ignore the blockquote when it's a descendant of header |
| 15:22 | <jgraham> | zcorpan: I guess that's a bug |
| 15:22 | <zcorpan> | jgraham: in the spec or in the impl? |
| 15:23 | <jgraham> | In the impl. |
| 15:23 | <jgraham> | At least, if my understanding of the rationale behind the current spec is right |
| 15:25 | <jgraham> | zcorpan: http://james.html5.org/outliner.html has the same issue so I guess maybe it is a spec bug? |
| 16:01 | <zcorpan> | the pf people seem to think authors are angels and therefore we can expose them to aria-labelledby and aria-describedby etc and disabled users will benefit by both short and long descriptions |
| 16:01 | <zcorpan> | meanwhile authors generally can't use alt correctly (if at all) |
| 16:02 | <jgraham> | The pf people seem to have the alarming notion that authors generally know what they are doing and therefore should be given as much control as possible |
| 16:02 | <zcorpan> | yeah, that's what i meant with "angels" :) |
| 16:03 | <jgraham> | <cynic>Or that authors have no idea what they are doing and so having complex technology results in lots of consulting jobs</cynic> :) |
| 16:03 | <jgraham> | Note that I am not seriously suggesting the second option |
| 16:04 | <jgraham> | Because that would be ascribing to malice what can be adequately explained by naivity |
| 16:04 | <beowulf> | is that malice? |
| 16:05 | <jgraham> | beowulf: Making accessibility intentionally hard to line one's own pockets would be malice |
| 16:06 | <jgraham> | (also: most a11y advocates I have met do not strike me as malicious) |
| 16:08 | <zcorpan> | MikeSmith: i get b.v.nu email even though i'm not explicitly cc-ed |
| 16:08 | <gsnedders> | zcorpan, jgraham: I pretty much exactly implement the spec, so almost certainly spec |
| 16:08 | <jgraham> | gsnedders: Me too, but neither of us have great test coverage so… |
| 16:09 | <jgraham> | (the fact that we have the same bug does suggest a spec bug) |
| 16:09 | <gsnedders> | I just think <http://hg.gsnedders.com/anolis/file/b6d93515d41e/anolislib/processes/outliner.py>; is so close to the spec that it's unlikely to be an impl bug :) |
| 16:09 | <jgraham> | gsnedders: iirc the spec is so hard to understand that anything is posssible |
| 16:09 | <gsnedders> | jgraham: When did you last read the spec? |
| 16:10 | <gsnedders> | jgraham: Hixie made quite a few edits just before he wrote URLs |
| 16:10 | <gsnedders> | (Which is when I was writing that impl.) |
| 16:10 | <zcorpan> | i tried to read the spec and i couldn't tell whether it should ignore sectioning roots in header or not |
| 16:11 | <jgraham> | gsnedders: Maybe before that. After the algorithm changed from the treewalker-based one though |
| 16:11 | <gsnedders> | jgraham: A lot of my comments were about spec clarity, as there were things where I wasn't sure what to do because the spec was ambiguous |
| 16:12 | <Philip`> | zcorpan: Maybe they think that a non-zero number of authors are angels and therefore if we expose them to aria-labelledby and aria-describedby etc then a non-zero number of disabled users will benefit, and so the effort will have been worthwhile |
| 16:13 | <gsnedders> | zcorpan: Everything should be ignored in a header, me thinks |
| 16:13 | <zcorpan> | Philip`: not if a greater number of authors use aria-labelledby and aria-describedby incorrectly so that users will have a better experience overall if the attributes are ignored |
| 16:13 | <Philip`> | (It's clearly impossible to make everything better for everyone, so the scope is limited to making some things better for some people) |
| 16:14 | <jgraham> | Philip`: That doesn't sound like a good way of making an optimum solution |
| 16:14 | <jgraham> | For example, some disabled people would be better off if no one had spent any time working on aria but had fixed a few existing websites |
| 16:14 | <zcorpan> | gsnedders: what do you use as the heading text in the outline for <header>? |
| 16:14 | <jgraham> | to be more accessible |
| 16:15 | <gsnedders> | zcorpan: See paragraph two of #the-header-element |
| 16:15 | <gsnedders> | "For the purposes of document summaries, outlines, and the like, the text of header elements is defined to be the text of the highest ranked h1–h6 element descendant of the header element, if there are any such elements, and the first such element if there are multiple elements with that rank. If there are no such elements, then the text of the header element is the empty string." |
| 16:15 | <Philip`> | zcorpan: Depends on whether the attributes have a negative effect on user experience when used incorrectly, and it seems easy to imagine that users just won't press the 'access description of this element' key when they know the site isn't going to be using it usefully |
| 16:16 | <zcorpan> | gsnedders: well that's the thing. that paragraph should ignore h1-h6 in descendants that are sectioning roots |
| 16:16 | <gsnedders> | zcorpan: Well, spec bug |
| 16:17 | gsnedders | is just what Hixie told him to :P |
| 16:18 | <zcorpan> | Philip`: presumably the AT would tell the user that there is a description available, though? |
| 16:18 | <jgraham> | gsnedders: That won't stand up in a court of law when your outliner is used in a life-critical system and someone DIES |
| 16:21 | <jgraham> | Sorry dunno what came over me there, guess I should take a rest |
| 16:22 | <jgraham> | Er, anyway, moving on |
| 16:23 | <jgraham> | Philip`: Your argument still seems like a logical fallacy |
| 16:23 | <jgraham> | Since you are asserting that doing anything that helps a non-zero number of people is good, even if it isn't the optimum solution |
| 16:28 | <Philip`> | jgraham: Now you're just engaging in the fallacy of "argument by logical fallacy" |
| 18:22 | gsnedders | has managed to hang tesseract |
| 18:34 | <gsnedders> | Wow. Fun. It hangs ocrad too |
| 18:34 | <gsnedders> | This is one awesome image. |
| 18:35 | <gsnedders> | Ah, ocrad comes out of hang, but output is totally wrong |
| 18:38 | gsnedders | crops image into segments and throws it at tesseract again, and it works straight away |
| 18:38 | <gsnedders> | Oh well |
| 19:37 | <Hixie> | hsivonen: re XXX quirks things -- the XXX comments in the spec are what i put in when i don't want to think about something at all |
| 19:38 | <Hixie> | hsivonen: e.g. if i happen to notice a mistake, don't have the inclination to fix it straight away, and want to move on to something more important (at the time) |
| 19:38 | <Hixie> | hsivonen: so "XXX quirks" might just mean i happen to notice something fishy while doing some other testing, and i should look more closely |
| 19:38 | <Hixie> | hsivonen: it doesn't mean "there is definitely a quirk to add here" |
| 20:15 | <mpilgrim> | there is a pending (non-spam!) draft on the WHATWG blog |
| 20:15 | <mpilgrim> | by sayrer |
| 20:15 | <mpilgrim> | any clue if he's done with it and wants it published? |
| 20:18 | <mpilgrim> | it seems done but still says "draft", not "pending review" |
| 20:19 | <jgraham> | I can deal with that |
| 20:19 | <gavin_> | he is online on moznet |
| 20:19 | <jgraham> | If he wants it published |
| 20:20 | <mpilgrim> | i'm in there anyway, publishing another "this week" |
| 20:21 | Philip` | thinks it's going by a fairly loose definition of "this" |
| 20:22 | mpilgrim | slinks away in shame |
| 20:22 | <mpilgrim> | it's been a hectic quarter |
| 20:23 | <mpilgrim> | i'm catching up now so i can resume regular weekly posts in Q2 |
| 20:23 | <Philip`> | I'm not complaining, just commenting :-) |
| 20:25 | <jgraham> | Maybe we should rename it "Some past week in HTML5, untill Google finally invents time travel, at which point we'll need several new tenses to describe just what is going on anyway" |
| 20:26 | <Philip`> | "A week in HTML5" |
| 20:26 | <mpilgrim> | pretty sure the past tense is taken ;) |
| 20:48 | Hixie | grumbles |
| 20:48 | <Hixie> | openDatabase() should have been a constructor |
| 20:48 | <Hixie> | i hate watching myself screw up the web |
| 20:48 | <gsnedders> | Well it's hardly as if _I_ wrote the spec that's now interoperably implemented! |
| 20:49 | <Lachy> | Hixie, re http://www.w3.org/Bugs/Public/show_bug.cgi?id=6477 - what made you decide to add support for input.height and .width if browsers don't support it? |
| 20:49 | <Hixie> | Lachy: i don't remember. they were added long before i added that comment. |
| 20:49 | <Hixie> | i probably thought browsers did have them |
| 20:50 | <Lachy> | ok |
| 20:50 | <Lachy> | I suppose they shouldn't be too hard to implement. Not sure if it's really worth the effort though. |
| 20:51 | <gsnedders> | Is it bad that more than half the money (in terms of value!) in my wallet is coppers? |
| 20:51 | <Lachy> | gsnedders, that depends if by "coppers" you mean really low value coins, like 1 and 2 pence peices (or whatever currency you use) |
| 20:52 | <gsnedders> | Yeah |
| 20:52 | <jcranmer> | gsnedders: you have copper pound notes in the UK? |
| 20:52 | <gsnedders> | jcranmer: No |
| 20:52 | <gsnedders> | Coppers as in coins worth 0.01 GBP and 0.02 GBP |
| 20:54 | <Lachy> | jcranmer, originally, 1 GBP was actually measured as 1 pound of silver. Imagine carrying several of those around in your wallet :-) |
| 20:54 | <Hixie> | i thought he meant police men |
| 20:54 | <gsnedders> | heh |
| 20:54 | <jcranmer> | " |
| 21:05 | Philip` | was unaware until recently that Isaac Newton was heavily involved with currency reform |
| 21:06 | <Hixie> | it's a big plot point in the baroque cycle |
| 21:06 | <Philip`> | That's how I became aware of it :-) |
| 21:07 | <Philip`> | (It's probably unwise to take everything in those books as strictly true, but Wikipedia seems to agree that he was involved) |
| 21:07 | <Hixie> | indeed |
| 21:08 | <gsnedders> | http://stuff.gsnedders.com/atom-iri/ — start of IRI testing! |
| 21:08 | <gsnedders> | Not that I got very far, because it took me a while to work out how to produce those files |
| 21:35 | <Philip`> | I love how selecting "run as administrator" on a 1.5GB .exe in Vista results in it freezing for ten minutes while silently copying the entire file into c:\windows\temp, and then it runs out of disk space and doesn't clean up after itself and leaves me with 0 bytes of free space |
| 21:39 | <Lachy> | Philip`, which program has a 1.5GB exe file? |
| 21:39 | <Philip`> | Lachy: An installer for some game |
| 22:52 | <Hixie> | are there any exception objects on the Web other than DOMException? And is it ok to make up new ones? |
| 23:04 | <heycam> | Hixie, there's an SVGException. don't see any reason not to make a new one. |
| 23:06 | <Hixie> | ok |
| 23:57 | <Hixie> | ok, the invitation for people to review the spec is Out There |
