| 00:11 | <zcorpan> | Hixie: have you considered if it makes sense for microdata to use img src if there's also srcset? |
| 00:15 | <tantek> | zcorpan, what's the use case? |
| 00:15 | <tantek> | and is anyone publishing microdata or microformats with <img srcset> ? |
| 00:16 | <tantek> | any URLs to real world examples? |
| 00:16 | <zcorpan> | no, just asked if it was a case that has been considered |
| 00:18 | <zcorpan> | considering if two features make sense together can be a useful thing to do without starting with a use case |
| 00:18 | <tantek> | I'm still waiting for more real world experience reports re: use of just srcset |
| 00:18 | <tantek> | nevermind with formats |
| 00:58 | <MikeSmith> | zcorpan: will try to implement new script restrictions in the validator soon but may not get to it til end of the year holidays |
| 01:05 | <Hixie> | zcorpan: never considered them together |
| 01:05 | <Hixie> | http://www.w3.org/mid/1385505093.3251.101.camel@chacal is pretty awesome given the w3c's fork of whatwg specs |
| 10:24 | <Ms2ger> | "sometimes I feel we're building more and more complex structures on top of foundations we barely understand" |
| 10:57 | <jgraham> | barely? |
| 10:57 | <jgraham> | That seems unreasonably optimistic |
| 11:04 | <smaug____> | very much so |
| 11:04 | <smaug____> | Ms2ger: where is that from ? |
| 11:06 | <Ms2ger> | https://twitter.com/johnallsopp/status/405615160586280960 apparently |
| 11:31 | <gsnedders> | hsivonen, annevk: what's the status with getting rid of locale specific fallback encodings? is all the discussion just the odd chat in here? (there seems to be no emails I can see, looking quickly?) |
| 11:32 | <hsivonen> | gsnedders: no status for getting rid of it for IP number URLs, .com, .org or .net |
| 11:32 | <hsivonen> | gsnedders: for other TLDs, I have a patch |
| 11:32 | <hsivonen> | gsnedders: but there's a handful of TLDs that need more research |
| 11:32 | <gsnedders> | hsivonen: Do we have data for those TLDs as to how accurate it is? |
| 11:32 | <hsivonen> | gsnedders: and I lack a good landing plan |
| 11:32 | <hsivonen> | gsnedders: no |
| 11:33 | <hsivonen> | gsnedders: most are no-brainers |
| 11:33 | <hsivonen> | gsnedders: but there are some non-obvious ones |
| 11:35 | gsnedders | is tempted to try and get actual data on this, and see whether bi-grams over the first 512 bytes (that's what we use for <meta detection, right?) yields anything useful, esp. for "generic" TLDs |
| 11:36 | <gsnedders> | In other news, I probably should do uni work instead of procrastinate doing stuff like this :) |
| 11:36 | <hsivonen> | gsnedders: For landing plan, I see 3 options: 1) on by default, 2) off by default and 3) randomly on or off with telemetry for A/B testing |
| 11:37 | <hsivonen> | gsnedders: I'd appreciate you researching stuff like this |
| 11:38 | <gsnedders> | hsivonen: While I'm tempted to try and do something along these lines for MSci thesis, I think that's probably too far off to be much use (that'd be a year and a quarter or so away from being done); may add this to list of things to play around with next summer. Probably won't have time over Christmas to do anything like this, given I have enough uni to catch up on. |
| 11:39 | <gsnedders> | So, yeah, summer is probably the best outcome :( |
| 11:40 | <gsnedders> | Anyhow, to lecture. ttyl! |
| 11:40 | <hsivonen> | gsnedders: I see |
| 11:40 | <hsivonen> | gsnedders: I think I'll either do research or make guesses before summer |
| 11:55 | <annevk> | gsnedders: I think it's 1024 bytes |
| 12:26 | <annevk> | "The fact that you (or even I)" lol |
| 12:27 | <annevk> | http://lists.w3.org/Archives/Public/www-tag/2013Nov/0190.html some meme material right there |
| 12:29 | <annevk> | Non-Unicode locales?! |
| 12:29 | <annevk> | That thread went stupid fast |
| 13:11 | <gsnedders> | hsivonen: Still, might find some improvement on guesses. |
| 14:35 | <zewt> | cool, to remove email search notifications on ebay i'm required to agree to a new contract |
| 14:35 | <zewt> | don't agree? permanent spam! |
| 15:53 | <gsnedders> | I'm going to a talk on congestion control in WebRTC in 15 minutes. What I ought I be questioning? |
| 15:54 | <gsnedders> | s/I ought/ought/ |
| 15:55 | <JonathanNeal> | I haven't followed the living spec in a while, has anything replaced <hgroup>? |
| 15:56 | <gsnedders> | JonathanNeal: It's still in the WHATWG spec, just not in the W3C one. |
| 15:57 | <gsnedders> | JonathanNeal: To my knowledge, it has no replacement in the W3C one. |
| 16:01 | <JonathanNeal> | gsnedders: good to know, thanks. is there a W3C channel for taking this up? |
| 16:03 | <gsnedders> | JonathanNeal: There'#html-wg on irc.w3.org, but you're just as likely to get an answer here by waiting |
| 16:04 | <JonathanNeal> | Waiting for who or what? |
| 16:04 | <gsnedders> | Other people to be around :) |
| 16:26 | <JonathanNeal> | gsnedders: are there other notable differences between what the specs say? |
| 16:37 | <Domenic_> | can of worms... |
| 16:39 | <Domenic_> | basically every time SteveF writes an article about some exciting HTML5 feature, e.g. on html5doctor.com you can guarantee it's something he made up for the W3C fork. |
| 16:48 | <annevk> | Hmm. Spotify took Eve away and now they put it back. DRM *shakes fist* |
| 16:54 | <JonathanNeal> | Domenic_: speaking of Steve, are there thoughts around http://rawgithub.com/w3c/subline/master/index.html ? |
| 16:54 | <JonathanNeal> | That comes up every year or so and then sits there and then life just goes on. |
| 16:54 | <Domenic_> | JonathanNeal: I am not the person to ask, sorry, just a bystander on the sidelines of the great W3C forking effort :) |
| 17:21 | <Ms2ger> | jgraham, I think we agreed that https://critic.hoppipolla.co.uk/showcomment?chain=807 was fine; want to resolve? :) |
| 17:23 | <jgraham> | Ms2ger: I think we agreed that you were going to convert the whole file :p |
| 17:23 | <Ms2ger> | jgraham, we did, but surely not in this PR :) |
| 17:24 | <jgraham> | Well |
| 17:25 | <jgraham> | Since you can have multiple commits in a single PR, I wouldn't actually object |
| 17:25 | <Ms2ger> | Dammit |
| 18:07 | <gsnedders> | hsivonen: I might try and start to play around with stuff while I'm back at my parents over xmas, not sure how far I'll get though |
| 18:53 | <JonathanNeal> | If I want to markup a subheading (or subtitle, alternative title, tagline), should I use <hgroup>? What is the W3C alternative since they have removed <hgroup>? |
| 18:58 | <Ms2ger> | Use hgroup |
| 19:37 | <Domenic_> | did ... did someone just reference the HTML4 spec? |
| 19:40 | <JonathanNeal> | Steve's response to my question above was a link to http://blog.paciellogroup.com/2013/10/html5-document-outline/ |
| 19:47 | <SteveF> | jonathonNeal: hgroup is as dead as <center>, <font> etc |
| 19:50 | <JonathanNeal> | With much time to reflect since the last update, what are your thoughts on <subhead>, SteveF? |
| 19:51 | <SteveF> | JonathonNeal: difference is Google like <font> and <center> more https://docs.google.com/presentation/d/1gnQg-3jQSytv60ozANjM7BK-OhiP7elf13P5oXHT6CM/present#slide=id.g11be9b862_114 |
| 19:52 | <JonathanNeal> | I don't see such markup. WayBackMachine link? |
| 19:54 | <SteveF> | JonathanNeal: there does not appear to be any great enthusiasm from dev's for a a feature to identify subheadings full stop |
| 19:55 | <SteveF> | JonathanNeal: i just looked at google.co.uk in Firefox 25/windows and font/center are present |
| 19:56 | <JonathanNeal> | re: second statement, that's bizarre. |
| 19:57 | <JonathanNeal> | re: first statement, except from me every few months full stop? |
| 19:57 | <SteveF> | JonathanNeal: that slide was from a month or so ago on https://www.google.fr/ just looked again and 2 <font> 1 <center> |
| 19:58 | <JonathanNeal> | SteveF: yes, I saw it too, just now. |
| 20:08 | <Domenic_> | while we're asking semantics questions... for progress from trade date to maturity date, I am thinking <progress max="number of days between them" value="days between now and trade date" />. Thoughts? I was originally thinking <meter> but this is fairly progresion-ey. |
| 20:29 | <gsnedders> | hsivonen: "http://s.validator.nu/html5/html5full-rdfa.rnc":1:0: error: Syntax error. happens when running build/build.py all given that server appears to not resolve |
| 20:35 | <Jasper> | Is there any guarantee that modifying an element will cause a reflow to happen synchronously? |
| 20:35 | <Jasper> | e.g. elem.classList.add('foo'); elem.getBoundingRect().width; |
| 20:35 | <Domenic_> | Oh interesting, there's a WHATWG validator page too? |
| 20:35 | <gsnedders> | Jasper: No, because reflows are an implementation detail |
| 20:36 | <Jasper> | Whatever the correct word for the semantics of that is. |
| 20:36 | <gsnedders> | Jasper: The script will block for layout, yes. |
| 20:36 | <gsnedders> | Jasper: That is guaranteed. |
| 20:36 | <Jasper> | Is there any reference for this? |
| 20:36 | <gsnedders> | Jasper: When it reads the bounding rect |
| 20:37 | Jasper | couldn't find anything in the HTML or DOM specs |
| 20:38 | <gsnedders> | Jasper: It's implicit in CSSOM that the bounding rect must be correctly computed |
| 20:38 | <gsnedders> | Jasper: Whether or not it draws to screen is not defined |
| 20:39 | <Jasper> | Right. |
| 20:39 | <Jasper> | gsnedders, well, is the guarantee only for the bounding rect, or for anything that affects layout? |
| 20:39 | <gsnedders> | Jasper: Anything. |
| 20:40 | <gsnedders> | Jasper: The CSSOM is all synchronous. Of course, the implementation is free to delay actually recomputing style as long as possible. |
| 20:40 | <Jasper> | OK. I just couldn't find it in any spec. |
| 20:41 | <gsnedders> | Jasper: It's in the CSSOM spec, by virtue of the interfaces mutating one thing then reading it back somewhere else, all synchronous actions. |
| 20:45 | <Jasper> | gsnedders, I'm scanning through it now, but I can't see anywhere that it says that it is synchronous. |
| 20:45 | <Jasper> | Unless it's implicit by the serialization step. |
| 20:46 | <gsnedders> | Jasper: Basically it defines the steps the host objects must execute when called, and defines their return points. |
| 20:46 | <gsnedders> | Jasper: If they don't return until the end of the function, they are obviously synchronous |
| 20:46 | <gsnedders> | Also, it's a bit of a horrible interaction between various specs here :) |
| 20:46 | <gsnedders> | (with the classList example) |
| 20:46 | <Jasper> | Yeah, I hope you'll admit it's not the easiest thing to parse in the world. |
| 20:48 | <gsnedders> | Someone else might be able to point better at quite how style mutation is defined for classList, but essentially both operations should syncrhonously read/write the layout state. |
| 20:49 | <Jasper> | Oh, duh. |
| 20:49 | <Jasper> | I was looking at the CSSOM spec, when *clearly* I should have been looking at the CSSOM View Module spec |
| 20:49 | <gsnedders> | Yeah, you probably want CSSOM, CSSOM View, *and* HTML for this. |
| 20:49 | <gsnedders> | Maybe CSS 2.1 too. |
| 21:04 | <JonathanNeal> | So, in short, SteveF, you're saying no one implements <hgroup>, no one implements the outline algorithm, so ... be cautious with your <h1-6> tags. Is that right? |
| 21:13 | <Domenic_> | browsers implement <hgroup> (it is HTMLElement not HTMLUnknownElement), but SteveF is saying that their implementation of it is as/less important than their implementation of <center>. |
| 21:27 | <JonathanNeal> | Hmm. Domenic_: do any non-whatwg-w3c content parsers, search engines, screen readers, or accessibility apps respect <hgroup>? |
| 21:29 | <Domenic_> | I don't know; my impression is SteveF has done some research that says the answer is no. |
| 22:22 | <Hixie> | JonathanNeal: it's very early days yet for any of the contemporary html heading stuff, whether <hgroup> or <section> or anything else |
| 22:24 | <Hixie> | JonathanNeal: in practice, if you use hgroup, the worst you'll get is the same as what you get with the same W3C specs have done it for years (a subsection heading where you'd expect a subheading), which in practice is fine for users |
| 22:24 | <Hixie> | JonathanNeal: (plus you get the advantage of being able to style it better) |
| 22:24 | <Hixie> | JonathanNeal: and on the long run software will improve modestly so as to treat it as a sunheading rather than a subsection heading |
| 22:33 | <Hixie> | ok i'm updating acid3 - removing test 67 attributes.removedNamedItemNS |
| 22:35 | <JonathanNeal> | Hixie: thanks for that. |
| 23:00 | <Hixie> | anyone know if IE does navigator.addProtocolHandler() ? |
| 23:12 | <SteveF> | JonathanNeal: if semantic fiction is your bag - go with the whatwg advice on headings |
| 23:15 | <JonathanNeal> | Hixie: not in IE11. |
| 23:18 | <Hixie> | JonathanNeal: thanks |
| 23:19 | <JonathanNeal> | SteveF: you haven't said <hgroup> will cause any breakage in and of itself, and I'm not sure whether you're saying I should be careful with <h1-6> usage. Sorry if I missed you addressing the latter point quite clearly. |
| 23:21 | <JonathanNeal> | I bring up the <h1-6> stuff since <hgroup> is strictly for them. And, related, <subhead> is basically the liberal version of <hgroup>, right? |
| 23:22 | <Hixie> | (what's <subhead>?) |
| 23:23 | <JonathanNeal> | Hixie: http://html5doctor.com/howto-subheadings/ and the June 2013 revised http://rawgithub.com/w3c/subline/master/index.html |
| 23:24 | <Hixie> | man, and some people accuse _me_ of nih syndrome... |
| 23:24 | <SteveF> | JonathanNeal: the post i linked to previously explains why relying on the outline for doc structure is bogus, so need to use h1-h6, <hgroup> is same as <div class="hgroup"> as far as anything other than CSS styling |
| 23:25 | <JonathanNeal> | <subhead> was SteveF's response to the community of devs saying they wanted some semantic indication of subheadings if <hgroup> was going to be pulled. |
| 23:26 | <Hixie> | SteveF is right that if you want today's tools and yesterday's tools to generate good outlines, you should stick to h1-h6 and avoid the newer h1-only approaches |
| 23:26 | <Hixie> | that's somewhat orthogonal to hgroup, though |
| 23:27 | <Hixie> | and it's true not just of the h1-only thing, but of any new feature of html |
| 23:27 | <Hixie> | it takes years for these things to deploy widely |
| 23:27 | <Hixie> | as far as hgroup goes, in such legacy UAs it just falls back to the same thing as the w3c has been doing for years with subheadings (e.g. on all their specs) |
| 23:28 | <Hixie> | which is technically suboptimal, but nobody really notices other than OCD people like me, so it's fine for a fallback :-) |
| 23:28 | <SteveF> | not all specs: http://www.w3.org/html/wg/drafts/html/master/ |
| 23:29 | <Hixie> | yes well the less i say about that document the better |
| 23:29 | <SteveF> | yeah probably true |
| 23:30 | <Hixie> | (sure would be nice if you stopped copying all my work though, if you're just going to mess it up afterwards) |
| 23:32 | <SteveF> | I don't copy just mess |
| 23:33 | <JonathanNeal> | After 6 months, it seemed like it was worth seeing what progress has been made to address subheadings. |
| 23:35 | <Hixie> | JonathanNeal: yeah |
| 23:35 | <JonathanNeal> | It sounds like one approach is to follow a spec that gives subheadings semantic meaning but relies on future compatibility, and the other is to drop any semantic hope for the time being, and just rely on divs to style. |
| 23:35 | <Hixie> | JonathanNeal: the problem with <hgroup> is that the FUD that the w3c spec has added to the mix means that it'll be even longer than normal for this to get implemented widely, unfortunately |
| 23:35 | <Hixie> | JonathanNeal: almost like a self-fulfiling prophecy |
| 23:36 | <SteveF> | JonathanNeal: like i said semantic fiction go for it |
| 23:39 | <SteveF> | ride the unicorn http://blog.paciellogroup.com/wp-content/uploads/2013/10/pinkunicorn.png |
| 23:43 | <JonathanNeal> | SteveF: go for what, or which, rather? I was trying to see both points of view. To be clear, in your mind, SteveF, you don't think there's a foreseeable future (say, the next +6 months) with semantic subheadings, subtitles, taglines and bylines in the W3C spec? |
| 23:50 | <SteveF> | JonathanNeal: if somebody comes along to move it forward then we may see progress, otherwise no, until such times that there is some meaningful way to convey subheading semantics, the requirement in the HTML spec is don't use multiple headings unless you intend it to be a new section or subsection as that is the semantic info that is conveyed: |
| 23:50 | <SteveF> | JonathanNeal: http://www.w3.org/html/wg/drafts/html/master/common-idioms.html#sub-head |
| 23:52 | <JonathanNeal> | That's pretty clear. Thanks. The part I'd like some clarification on is the "somebody" who "comes along to move it forward". Who does this somebody need to be or what do they need to do to move it forward? |
| 23:53 | <JonathanNeal> | SteveF: ^ |
| 23:57 | <SteveF> | JonathanNeal: anybody who is interested see http://www.w3.org/html/wg/wiki/ExtensionHowTo on mechanics, if you want to take the subhead proposal and work on it go ahead https://github.com/w3c/subline pull/push/fork break remake |
