| 00:36 | <Dashiva> | DOCTYPE Decoration: When web designers add a proper DOCTYPE declaration at the beginning of an HTML document, but then don’t bother to write valid markup for the rest of it. |
| 06:05 | <hsivonen> | AryehGregor: umm, right. I got my thinking about the + selector mixed up. |
| 07:14 | <JoePeck> | "style sheet" or "stylesheet" or is each appropriate for different situations? |
| 07:18 | <annevk> | yup |
| 07:23 | <MikeSmith> | JoePeck: as far as prose in specs, the CSS specs seem to use "style sheet" consistently |
| 07:24 | <MikeSmith> | ..while the XSL specs use "stylesheet" |
| 07:24 | <JoePeck> | MikeSmith: yes, that was the first placed I looked and I noticed that |
| 07:24 | <JoePeck> | interesting. Do you find yourself leaning to one side or the other? |
| 07:27 | <MikeSmith> | JoePeck: I suppose being consistent with the CSS specs is best |
| 07:28 | <JoePeck> | MikeSmith: thanks for your thoughts! |
| 07:28 | <MikeSmith> | the HTML5 spec also seems to use "style sheet" in prose, except for one exception (which I guess could be considered an editorial bug) |
| 07:32 | <MikeSmith> | hsivonen: I'm looking at http://bugzilla.validator.nu/show_bug.cgi?id=438 (about style@scoped needing to be allowed in flow content) |
| 07:32 | <MikeSmith> | and I noticed your comment in common.rnc: |
| 07:32 | <MikeSmith> | common.inner.flow = |
| 07:32 | <MikeSmith> | ( text & common.elem.flow* ) # REVISIT <style scoped> |
| 07:35 | <MikeSmith> | so... trying to figure if it's possible to express in the schema or not |
| 07:36 | <annevk> | good times http://www.w3.org/2010/05/07-hcg-minutes.html (W3C Member-only) |
| 07:37 | <MikeSmith> | hsivonen: or if it instead needs to be handled with a new Checker class or something |
| 07:51 | <hsivonen> | MikeSmith: It seems to me this should be expressible in a schema |
| 07:52 | <hsivonen> | the error messages will suck, though |
| 07:52 | <MikeSmith> | ok |
| 07:52 | <MikeSmith> | I'll mess around with it and see what I can come up with |
| 07:52 | hsivonen | wonders if anyone actually has plans for implementing <style scoped> in a browser |
| 07:53 | MikeSmith | would really like to know the answer to that too |
| 07:54 | <hsivonen> | hmm. JF's email to public-html makes it look like consulting captioning experts were sufficient. |
| 07:54 | gsnedders | wants two things: a git interface in PHP, and a ARM toolchain including X. |
| 07:55 | <hsivonen> | I find it unbelievable that any consultation with browser developers could have lead to using XSL-FO as the layout basis |
| 07:55 | <hsivonen> | seems like a failure to consult with a notable stake holder group! |
| 08:07 | <gsnedders> | hsivonen: HTML5 back on by default in Minefield? |
| 08:08 | <nessy> | it almost seems like that first 10 or so years of W3C work in this century was done mostly without consulting with browser developers… |
| 16:00 | <annevk> | foolip, aah good point |
| 16:00 | <annevk> | foolip, I think language-per-cue can be done, e.g. <narrator lang=en> or some such |
| 16:01 | <annevk> | good points* |
| 16:01 | <annevk> | I meant more the general picture is pretty good, not necessarily WebSRT specifics |
| 16:01 | <annevk> | (though those are pretty good too, imo) |
| 16:07 | <zcorpan_> | if it's per-cue, then maybe it should be a "cue setting" |
| 16:10 | <annevk> | oh right |
| 16:10 | <annevk> | we could even have a generic <span> |
| 16:11 | <annevk> | though someone might call bloat |
| 16:11 | <zcorpan_> | i think multilanguage subtitles are on the maybe wrong side of the 80% rule |
| 16:13 | <foolip> | annevk: do we want attributes in WebSRT elements? |
| 16:14 | <foolip> | so far it's only "voices", the element name that is parsed, right? |
| 16:15 | <annevk> | I think that BNF is no longer entirely accurate and that Hixie found some reason to make that part generic |
| 16:21 | <foolip> | I also need to add a use case for multiple voices. In karaoke duets singer A, B and A+B can be 3 different colors |
| 16:29 | <annevk> | foolip, how is that not covered by <a> <b> <ab> at the start of a line? |
| 16:43 | <foolip> | annevk: oh, maybe it is that simple |
| 16:44 | <foolip> | I'm quite keen to see how the HTML-like stuff is going to turn out |
| 16:46 | <annevk> | I'm guessing that instead of a DOM they'll just generate CSS boxes directly |
| 17:00 | <hsivonen> | annevk: I'm not convinced that a special-purpose tokenizer and a special-purpose css frame constuctor are nice to have |
| 17:03 | <othermaciej> | special-purpose css frame constructor? |
| 17:04 | <mhausenblas> | hey lazyIRC, which browsers support CORS, ATM? |
| 17:04 | <hsivonen> | othermaciej: for rendering timed text using the css formatter without a DOM |
| 17:06 | <othermaciej> | oh |
| 17:06 | <mhausenblas> | any overview/implementation report on CORS available? |
| 17:06 | <othermaciej> | seems easier to make a fake DOM, if you already have an existing css layout engine |
| 17:06 | <othermaciej> | mhausenblas: Safari, Chrome, Firefox, partial support in IE via XDomainRequest |
| 17:06 | <mhausenblas> | thanks, othermaciej |
| 17:26 | <variable> | http://wiki.whatwg.org/wiki/Rationale#B.2C_I.2C_EM.2C_and_STRONG --> any comments |
| 17:26 | <variable> | Lachy: around? |
| 17:28 | <Lachy> | variable, yo |
| 17:29 | <annevk> | hsivonen, you already have the latter for <canvas> |
| 17:29 | <annevk> | hsivonen, and we have lots of tokenizers too and this one would be a lot simpler than the HTML one, which would be a benefit for non-HTML impls |
| 17:29 | <variable> | Lachy: I sent a blog post for review --> could you look at it and let me know if it is good? |
| 17:30 | <Lachy> | variable, will do shortly |
| 17:30 | <variable> | kk |
| 17:35 | <jgraham> | http://glazman.org/JSCSSP/ hmm if that were MIT licensed I could port it to python and use it to implement proper CSS sanitization in html5lib |
| 17:35 | <JonathanNeal> | Mornin' |
| 17:36 | <jgraham> | Although I guess it was ported from Gecko iun the first place so the license is rather fixed |
| 17:36 | <variable> | Lachy == Lachlan Hunt -- right? |
| 17:37 | <variable> | jgraham: isn't gecko tri-licensed |
| 17:37 | <jgraham> | variable: Yeah |
| 17:37 | <variable> | IIRC it is mozilla+gpl+???? |
| 17:37 | <jgraham> | + LGPL |
| 17:37 | <variable> | ? |
| 17:37 | <variable> | jgraham: isn't gecko tri-licensed ? |
| 17:37 | <variable> | sorry bad copy/paste |
| 17:37 | <variable> | can't you use it under one of the other licenses? |
| 17:38 | <jgraham> | I could _use_ it under LGPL, sure |
| 17:38 | <jgraham> | But derivatives would have to be LGPL also, I guess |
| 17:38 | <variable> | true |
| 17:39 | <jgraham> | so I couldn't incoroprate the whole thing into an MIT licensed project and keep the MIT license |
| 17:40 | <variable> | jgraham: ah I see - damn gpl |
| 17:42 | <jgraham> | Well I don't really have anything aginst the LGPL. It just doesn't work for me in this case |
| 17:42 | <variable> | jgraham: in general I don't like the gpl for exactly this reason ;) |
| 17:42 | <variable> | I thought for a moment that one of the licenses was MIT/BSD - thats all |
| 17:43 | <Lachy> | variable, yes |
| 17:43 | <variable> | all: http://wiki.whatwg.org/wiki/Rationale#B.2C_I.2C_EM.2C_and_STRONG --> any comments |
| 17:46 | <foolip> | meh, no image upload on the wiki? |
| 17:53 | <foolip> | bayimg to the rescue |
| 18:09 | <Aleoss> | What is the compatibility of application/rss+xml VS application/atom+xml VS text/xml for a feed? |
| 18:10 | <Lachy> | variable, published |
| 18:12 | <Lachy> | variable, you've been upgraded to Author. You can now publish without moderation |
| 18:16 | <Aleoss> | Should it be text/xml or application/xml? They are both valid MIME types under RFC3023 but which is the technically correct method? |
| 18:19 | <Philip`> | Aleoss: text/xml apparently has weird charset requirements, so application/xml is likely to be less confusing and/or less wrong |
| 18:22 | <Aleoss> | Thanks. |
| 18:26 | <Philip`> | (...though if you're using it for e.g. XHTML then you should probably use application/xhtml+xml instead because that's more conventional) |
| 18:42 | <oal> | Is this the html5 channel? |
| 18:42 | <Philip`> | oal: Yes |
| 18:43 | <oal> | I'm trying to make use of the Canvas element for massive amounts of text |
| 18:43 | <oal> | And I want to color certain words as I type |
| 18:43 | <oal> | Do I have to redraw the whole canvas to avoid getting double letters? |
| 18:44 | <oal> | And are there any good, in depth docs on the canvas element yet? |
| 18:49 | <Philip`> | If you want to dynamically update a small region of the canvas, you could do a clearRect and then set a clipping rectangle and then draw the next text into that area |
| 18:50 | <Philip`> | (though it's hard to know precisely what rectangle to update since you don't know the precise font metrics) |
| 18:50 | <oal> | Hmm, if I use say a monospaced font in a given size at all times? |
| 18:50 | <Philip`> | (but it should be possible to make a reasonable overestimate) |
| 18:51 | <oal> | Philip`, if I do use clearRect etc, will that area be the only space redrawn, leaving the rest untouched? |
| 18:51 | Dashiva | reads about using two iPads to multitask. Loses faith in humanity. |
| 18:52 | jgraham | assumes that you really need canvas and that you wouln't be better off doing the text with some retained mode feature like er, well HTML, or SVG |
| 18:52 | <jgraham> | Dashiva: Now that is a business model |
| 18:52 | <oal> | I'd like to use canvas to learn something new, other than that, I don't have to ;) |
| 18:53 | <Philip`> | oal: If you do clearRect(x,y,w,h); save(); rect(x,y,w,h); clip(); /* draw stuff */; restore(); then it will only draw into the given rectangle and won't affect anything outside there |
| 18:53 | <jgraham> | Maybe the next version of the OS will allow you to wirelessly copy + paste between multiple iPads :) |
| 18:53 | <oal> | Philip`, which will then be a lot faster? :) |
| 18:53 | <oal> | with massive amounts of text |
| 18:54 | <Philip`> | jgraham: Someone should make an app that displays the clipboard as one of those funny pixel grid things, so an app on the other iPad can use a camera to pick up the data and move it into its own clipboard |
| 18:55 | <jgraham> | That would be awesome |
| 18:55 | <Philip`> | oal: It probably wouldn't be much faster if you're redrawing all the text, but you can write some high-level code to only redraw the lines that you want |
| 18:55 | <Philip`> | so you're not trying to draw massive amounts of text each time |
| 18:56 | <oal> | But wouldn't your example above only redraw the region specified? |
| 18:58 | <Philip`> | If you draw some text which is outside the clipping region, the browser would still have to process all the characters and turn them into glyph shapes and see if they're inside the clipping region |
| 18:58 | <Philip`> | so you only save the cost of converting the glyph shapes into pixels |
| 18:58 | <Philip`> | (probably) |
| 18:58 | <Philip`> | so it's faster if you don't attempt to draw the text at all |
| 18:58 | <oal> | Ok, I see |
| 18:59 | <oal> | Say I wanted to erase some text again, could I use clearRect then, where the characters I want to remove are? |
| 19:00 | <Philip`> | Yes, if you know where they are |
| 19:00 | <Philip`> | You have to be careful about stuff like other characters extending into the space used by the characters you want to delete |
| 19:00 | <oal> | Thanks, Philip`. I'll go get my hands dirty with some coding ;) |
| 19:01 | <Philip`> | or vice versa |
| 19:01 | <Philip`> | so you probably need to clear some larger region around the text to delete, and then redraw the text you don't want to delete (clipped to that region) to fill in any bits that got cut off |
| 19:01 | <Philip`> | Using HTML for all of this is much easier, of course ;-) |
| 19:02 | <oal> | Hmm, it probably is. Maybe I should do most of it in traditional html and some of the fancy parts with canvas? |
| 19:03 | <Philip`> | Depends on what you're doing |
| 19:04 | <Philip`> | and whether the benefit of using the most appropriate technologies for each part of your system is greater than the pain of integrating multiple technologies |
| 19:05 | <oal> | I'll do some prototyping with both, and see what's easiest and most efficient |
| 19:06 | <Philip`> | Also don't forget SVG |
| 19:06 | <oal> | Haven't been working with svg other than with InkScape |
| 19:06 | <oal> | What are the advantages of using SVG? |
| 19:06 | <Philip`> | Sounds like an opportunity to learn something new :-) |
| 19:07 | <Philip`> | The advantages depend on what you're doing :-p |
| 19:07 | <oal> | Yes, of course. I'll definitely look into it. :) |
| 19:07 | <oal> | Summer is coming and lots of spare time, so it's good to have some plans |
| 19:07 | <Philip`> | It's more graphical than HTML and less procedural than canvas |
| 19:07 | <Philip`> | so it might be good if that's what you want |
| 19:08 | <oal> | Can you for example give names to boxes and control them later, rather than having to redraw the whole canvas to move one object? |
| 19:08 | <Philip`> | Yes |
| 19:08 | <oal> | That's convenient |
| 19:08 | <Philip`> | Indeed |
| 19:10 | <oal> | Any websites with SVG examples? I found this for canvas http://www.canvasdemos.com/ |
| 19:49 | <JonathanNeal> | So now that HTML5 allows embed again, should we be using it for flash? |
| 19:50 | <JonathanNeal> | Or should flash stay in <object>? |
| 19:51 | <Philip`> | You shouldn't be using Flash |
| 19:55 | <JonathanNeal> | Philip`, I know. |
| 20:11 | <Aleoss> | Philip: That is a niave thing to say. Think of websites like YouTube and NewGrounds and IllWillPress and HomeStarRunner and what not? Those sites people go to knowing that it is flash/shockwave content on. |
| 20:18 | <Aleoss> | Fact: Javascript executes faster than HTML5 Canvas drawing does. |
| 20:19 | <Aleoss> | The odd thing about it: HTML5 Canvas uses Javascript to do it's functions. |
| 20:20 | <Aleoss> | Proof: http://webroo.org/2010/01/17/html-5-canvas-vs-flash/ |
| 20:20 | <Aleoss> | Correction: That is the flash one. |
| 20:20 | <Aleoss> | Flash does execute faster as well. |
| 20:21 | <Aleoss> | But to draw a line via JS is faster than drawing a line on a <canvas> element. |
| 20:29 | <annevk> | You do know YouTube is planning on switching to using <video> right? |
| 20:29 | <annevk> | Anyways, is Hixie around already? |
| 20:29 | annevk | wants to complete the multipage setup |
| 21:01 | <variable> | lachy - I know its a /little/ late but thanks ;( |
| 21:01 | <variable> | ;) |
| 21:10 | <roc> | jgraham: we had some old tests that used a kind of render tree dump, but they are completely obsolete and never run |
| 21:10 | <roc> | reftests are the way |
| 21:11 | <annevk> | nn; maybe tomorrow |
| 21:49 | <variable> | accessibility question: are skip links still needed with a <nav> element ? |
| 22:05 | <jgraham> | roc: Thanks |
| 22:10 | <AryehGregor> | . . . why does variable always ask questions and then quit a few minutes later? |
| 22:11 | <ment> | AryehGregor: ... because he can't take the pressure? |
| 22:13 | <Dashiva> | Maybe he reads the logs for answers |
| 22:15 | <roc> | he can't handle the truth |
| 22:15 | <roc> | speaking of which |
| 22:15 | roc | checks the "implement SVG fonts" bug |
| 22:43 | jwalden | snickers |