| 02:59 | <MikeSmithX> | Hixie: about shortnames on for the WDs to publish tomorrow, annevk mentioned "I believe we decided that md would be microdata", but I can't recall a decision about that. Can you? |
| 03:07 | <othermaciej> | MikeSmithX: hmm, the TransReq asked for approval of a new shortname but did not cite the shortname |
| 03:07 | <othermaciej> | I do recall discussing "microdata" as the shortname |
| 03:08 | <MikeSmithX> | othermaciej: well, I think it would be an improvement for clarity to use it as the shortname instead |
| 03:08 | <othermaciej> | I think so too |
| 03:08 | <MikeSmithX> | also, another issue is the URI "http://www.w3.org/TR/markup/"; |
| 03:09 | <othermaciej> | our approval does not seem to have record of a shortname, nor was Sam specific about what shortname he was requesting |
| 03:09 | <othermaciej> | yeah H:TML probably should not use that as a shortname |
| 03:09 | <MikeSmith> | yeah, the lack of explicit mention of the shortname requests was my fault for not making it clear to Sam that they needed to be included |
| 03:10 | <MikeSmith> | about H:TML, I suggest we request http://www.w3.org/TR/html-markup/ instead |
| 03:10 | <MikeSmith> | about microdata, I suggest we consider qualifying that short name further as well |
| 03:11 | <MikeSmith> | I we want consistency with the existing RDFa shortname, it would be http://www.w3.org/TR/microdata-in-html/ |
| 03:12 | <MikeSmith> | if we don't care too much about consistency, http://www.w3.org/TR/html-microdata/ |
| 03:13 | <MikeSmith> | and perhaps http://www.w3.org/TR/canvas-2dcontext/ |
| 03:13 | <MikeSmith> | or even http://www.w3.org/TR/canvas-2d/ |
| 03:14 | <othermaciej> | html-microdata is getting kinda long for a short name |
| 03:14 | <othermaciej> | I would say just microdata - if there are other specs to disambiguate later, they can get a longer name |
| 03:14 | <MikeSmith> | OK |
| 03:14 | <MikeSmith> | how about "2dcontext"? |
| 03:14 | <othermaciej> | canvas-2d or canvas2d sounds better to me than 2dcontext, if not qualified otherwise |
| 03:14 | <MikeSmith> | OK |
| 03:15 | <MikeSmith> | I can send mail to the webmaster (reply to Sam's pub request) shortly |
| 03:15 | <MikeSmith> | othermaciej: are you OK with http://www.w3.org/TR/html-markup/ ? |
| 03:16 | <othermaciej> | MikeSmith: what shortname does the diffs document get published with? |
| 03:16 | <othermaciej> | just for comparison |
| 03:16 | <MikeSmith> | http://www.w3.org/TR/html5-diff/ |
| 03:20 | <MikeSmith> | <twitter!squaredeye> .@w3c Are their any specs in future releases of CSS for blending modes with colors? |
| 03:20 | <MikeSmith> | ...not sure what the even means |
| 03:20 | <MikeSmith> | TabAtkins: ↑ |
| 03:21 | <othermaciej> | html-markup sounds ok to me |
| 03:21 | <othermaciej> | MikeSmith: maybe he means ability to do Porter-Duff compositing operations other than SrcOver? |
| 03:21 | <othermaciej> | or, like, photoshop blend modes, like dodge and burn |
| 03:21 | <MikeSmith> | ah |
| 03:24 | <TabAtkins> | Mikesmith: Not sure what the dude means, but othermaciej seems like he has a decent handle on it. |
| 03:24 | <MikeSmith> | OK |
| 03:24 | <TabAtkins> | Assuming that's approximately right, then no. |
| 03:25 | <TabAtkins> | At least, nothing in the pipeline about how colors blend. |
| 03:25 | <othermaciej> | blend modes are not really a color issue so he might just be confused |
| 03:25 | <TabAtkins> | If you can reply and ask for more details, I can give a better answer. |
| 03:26 | <MikeSmith> | maybe he meant more like "blending modes, with colors" .. though not really clear, either |
| 03:26 | <MikeSmith> | TabAtkins: yeah, I will |
| 03:26 | <TabAtkins> | Back to Mass Effect and beer. Basically the best night possible. |
| 03:27 | <MikeSmith> | ah, I now notice he also tweeted, ".@webkit are their any experiments with blending modes (ie: multiply) with colors in the browser?" |
| 03:27 | <MikeSmith> | TabAtkins: what's Mass Effect? |
| 03:27 | <othermaciej> | I assume he means blend modes then |
| 03:27 | <MikeSmith> | yeah |
| 03:27 | <MikeSmith> | http://twitter.com/squaredeye |
| 03:28 | <othermaciej> | and we do have -webkit-background-composite |
| 03:28 | <othermaciej> | to set the compositing mode |
| 03:28 | <othermaciej> | in addition to having different compositing modes in canvas |
| 03:35 | <paul_irish> | othermaciej: earlier on the WOFF topic.. i just wanted to clarify.. you said it'd be easier to just make woff fonts installable on mac os, right? |
| 03:36 | <othermaciej> | paul_irish: it is simpler for us to treat them like any other font than to try to make the font system only handle them for WebKit |
| 03:37 | <paul_irish> | roger that. i'm just kinda wtf'ing because i thought the intention of the woff format was to avoid a trivial-to-install scenario |
| 03:37 | <othermaciej> | I'm sure that is some people's intent, I don't think there is any such requirement in the spec though |
| 03:38 | <paul_irish> | hehe. this is true. :) |
| 03:39 | <paul_irish> | othermaciej: is it okay if i mention this to some webfont friends of mine? |
| 03:48 | <othermaciej> | paul_irish: then it might sounds like I am making a commitment about future Apple products, which I definitely am not |
| 03:48 | <othermaciej> | we haven't even started on implementing WOFF yet |
| 03:49 | <paul_irish> | gotcha. |
| 03:49 | <paul_irish> | my expectation is that people wont be very pleased. :) |
| 03:49 | <othermaciej> | I can't even commit to implementing it at all ever |
| 03:49 | <paul_irish> | k. thanks much. |
| 03:50 | <othermaciej> | if we do, then I am not sure we have huge motive to make such fonts hard to install |
| 03:50 | <othermaciej> | if we end up participating in the WG then we may be able to discuss our plans |
| 03:50 | <othermaciej> | in more specifics |
| 03:50 | <othermaciej> | if we have plans by then |
| 03:50 | <paul_irish> | aye |
| 03:50 | <othermaciej> | I expect font people are most worried about piracy on Windows |
| 03:50 | <othermaciej> | Windows is the main platform where piracy happens, from what I hear |
| 03:51 | <othermaciej> | at least for other media |
| 03:51 | <paul_irish> | i pirate on mac all the time. :) |
| 03:51 | <miketaylr> | shhhh |
| 03:51 | <paul_irish> | oh fu- |
| 03:59 | <jcranmer> | most people I know don't like macs |
| 04:02 | <paul_irish> | especially not after today. sheesh |
| 04:08 | <othermaciej> | why, did a Mac go rogue and kill its trainer today? |
| 05:18 | <TabAtkins> | MikeSmith: http://en.wikipedia.org/wiki/Mass_Effect |
| 05:20 | <MikeSmith> | TabAtkins: I see.. glad it's not this: http://en.wikipedia.org/wiki/Mass_effect_(medicine) |
| 05:23 | <TabAtkins> | Heh, indeed. |
| 05:31 | <MikeSmith> | hmm, I don't think Jonas's assertion that "Validation only checks for a certain, very limited, types of errors." is entirely accurate |
| 05:32 | <othermaciej> | it only checks for syntax errors |
| 05:32 | <othermaciej> | not other kinds of errors |
| 05:34 | <MikeSmith> | I would not classify syntax errors as "very limited types of errors" |
| 05:34 | <othermaciej> | Jonas is probably speaking from the programmer's perspective |
| 05:34 | <othermaciej> | the errors we spend most of our time dealing with are not syntax errors |
| 05:35 | <MikeSmith> | right, I understand that |
| 05:35 | <MikeSmith> | but a markup validator is essentially a lint checker |
| 05:35 | <othermaciej> | right |
| 05:35 | <MikeSmith> | a static checker |
| 05:36 | <othermaciej> | I think static checkers, lint-like tools, and validators are all awesome things |
| 05:36 | <othermaciej> | yet I agree with Jonas that the errors they can find are but a subset of the possible interesting errors |
| 05:36 | <othermaciej> | though, for document-like content it's probably a higher fraction of errors you may care about |
| 05:37 | <othermaciej> | than for app-type content |
| 05:37 | <othermaciej> | we compile WebKit at a crazy high warning level |
| 05:37 | <MikeSmith> | yeah, agreed |
| 05:39 | <MikeSmith> | checking/validating the initial markup of some particular a Web app that changes state is of limited usefulness |
| 05:53 | <othermaciej> | our warning flags: -Werror -Wmissing-prototypes -Wnon-virtual-dtor -Wnewline-eof -Wall -Wextra -Wcast-align -Wcast-qual -Wchar-subscripts -Wextra-tokens -Wformat=2 -Winit-self -Wmissing-format-attribute -Wmissing-noreturn -Wpacked -Wpointer-arith -Wredundant-decls -Wundef -Wwrite-strings -Wno-trigraphs |
| 05:53 | <othermaciej> | (-Werror makes the build break if someone violates one of these) |
| 05:57 | <JonathanNeal> | Hey all |
| 06:31 | <MikeSmith> | othermaciej: I realize that the TAG specifically suggested a RelaxNG schema and that it should be an appendix to the H:TML document |
| 06:31 | <MikeSmith> | and that may be what they think they want |
| 06:31 | <othermaciej> | MikeSmith: I just filed the bug based on what they said - if something else would make them happy, I have no complaint |
| 06:31 | <MikeSmith> | but I don't think that's what they really want |
| 06:32 | <othermaciej> | actually, I'll go farther - I don't even care if we do make them happy, as long as they get the same due process right to be heard as anyone else |
| 06:33 | <MikeSmith> | OK, well, if we keep that bug summary as specifically adding and appendix with and RelaxNG schema to the H:TML doc, then what I am inclined to do is change it to resolved/wontfix with a comment that I don't plan to add it |
| 06:33 | <MikeSmith> | I dunno, maybe that's the best way to do it |
| 06:34 | <othermaciej> | or you could make it "partially accepted" with a resolution of doing something different than what was asked |
| 06:34 | <othermaciej> | or you could get the TAG's agreement on repurposing the bug |
| 06:34 | <othermaciej> | many options |
| 06:34 | <MikeSmith> | OK, yeah, I suppose I should follow up with them about it directly on the TAG list |
| 06:35 | <MikeSmith> | I will suggest to them that we repurpose that bug |
| 06:38 | <othermaciej> | do you have a concrete alternate approach in mind? |
| 06:42 | <MikeSmith> | othermaciej: not really |
| 06:42 | <othermaciej> | hmm |
| 06:42 | <MikeSmith> | well, not short of just publishing the set of formalisms |
| 06:42 | <othermaciej> | I probably shouldn't be the first vote on these accessibility surveys |
| 06:43 | <othermaciej> | is there a way to delete my answer? |
| 06:43 | <MikeSmith> | othermaciej: yeah |
| 06:43 | <othermaciej> | ah, there is |
| 06:43 | <MikeSmith> | that brings to mind one other thing I wanted to ask about |
| 06:44 | <othermaciej> | I will register my opinion later if it seems appropriate |
| 06:44 | <othermaciej> | I do like both proposals |
| 06:44 | <Hixie> | MikeSmith: there are short names assumed in the various drafts, look at the "current version" urls |
| 06:44 | <MikeSmith> | othermaciej: I think it's likely we will get agreement to proceed with those |
| 06:45 | <MikeSmith> | Hixie: looking now |
| 06:45 | <othermaciej> | yeah, I am just looking to be supportive |
| 06:45 | <othermaciej> | but I do not want to inadvertently influence people for or against too much |
| 06:45 | <othermaciej> | I gotta go |
| 06:46 | <MikeSmith> | Hixie: ah, OK |
| 06:48 | <MikeSmith> | "microdata" seems fine, but I already sent mail suggesting "canvas-2d", so I guess I need to wait unless/until anybody replies to that |
| 06:49 | <MikeSmith> | if PLH or nobody else thinks it needs to be changed, then we can just leave it as-is |
| 06:50 | <Hixie> | canvas-2d is a bit misleading since there's no canvas element in the spec :-) |
| 06:53 | <MikeSmith> | Hixie: I don't think the shortname necessarily implies it's a spec for the canvas element |
| 06:54 | <MikeSmith> | and the abstract does read, "This specification defines the 2D Context for the HTML canvas element. |
| 06:54 | <MikeSmith> | and the abstract does read, "This specification defines the 2D Context for the HTML canvas element" |
| 06:54 | <MikeSmith> | oops |
| 06:54 | <MikeSmith> | anyway |
| 06:54 | <MikeSmith> | I don't feel strongly about it |
| 06:55 | <MikeSmith> | it's just that I let the genie out of the bottle already, as far as having proposed it, so I need to wait to see what the responses are |
| 06:55 | <MikeSmith> | I guess |
| 06:56 | <MikeSmith> | Hixie: anyway, I wanted to say something related to Maciej's comment that he probably should not be the first vote on the accessibility surveys |
| 06:59 | <MikeSmith> | Hixie: which is, I think if you genuinely want to help reduce the amount of unproductive discussion on the public-html list, one thing you can consider doing more of is waiting to post replies on threads that you know are going to be contentious |
| 07:01 | <nessy> | othermaciej: I made the mess larger by replying to your questions on the a11y list :-) |
| 07:18 | <Hixie> | MikeSmith: i almost always do |
| 07:20 | <nessy> | foolip: do you think the media track API is not ready yet? |
| 07:22 | <MikeSmith> | Hixie: OK. I think even it non-contentious it helps others to allow time for some people form their own opinions |
| 07:22 | <MikeSmith> | *non-contentious discussions |
| 07:46 | <Hixie> | MikeSmith: in general i only respond to threads weeks after they've finished, because i just stuff all incoming e-mail into my folders |
| 07:46 | <Hixie> | MikeSmith: actually for public-html mail now i tend to just not reply at all, and just wait for a bug to be filed |
| 07:46 | <Hixie> | occasionally there are exceptions, of course, like when someone asks me a question directly |
| 07:48 | <MikeSmith> | Hixie: that sounds reasonable to me |
| 07:49 | <MikeSmith> | we need to get the overall list traffic down on public-html |
| 07:50 | <MikeSmith> | I think moving more discussions to bugzilla is one good way to do that |
| 07:51 | <MikeSmith> | and encouraging people to comment in bugzilla instead |
| 12:09 | <asmodai> | mmm |
| 12:09 | <asmodai> | http://www.theregister.co.uk/2010/03/02/virgin_america_html_flash/ |
| 12:09 | <asmodai> | Guess that made the rounds too? |
| 12:12 | <annevk> | not on this channel |
| 12:13 | <asmodai> | Hadn't seen it in my scrollback buffer, but hey, could've missed it. :) |
| 12:16 | <boblet> | anyone have an opinion on using <dl> for forms? (<dt> for labels, <dd> for <input> etc) |
| 12:18 | hsivonen | wishes Lufthansa dropped Flash, too |
| 12:18 | <Philip`> | boblet: Why not use <label> for labels and <input> for <input>? |
| 12:19 | <boblet> | Philip`: sorry, I meant putting <label> in <td> etc |
| 12:19 | <boblet> | sheesh |
| 12:19 | <Philip`> | Seems like anything else is just a presentation concern |
| 12:19 | <boblet> | in <dt> |
| 12:20 | <boblet> | Philip`: agreed, but replying to the latest HTML5Doctor Q&A, and wondering why <dl> feels wrong for me (I’d say a list, or maybe a table for a very complex form, eg multi-col) |
| 12:21 | <Philip`> | "Virgin anticipates moving this new site to HTML 5, once it's cleared standards ratification at the World Wide Web Consortium (W3C)" - in 2022? |
| 12:22 | <boblet> | I was impressed they were using it as a buzzword |
| 12:22 | <Philip`> | "the vast majority of web content [...] are built using [Flash]" - huh? |
| 12:22 | <boblet> | HTML5 = Web 2.1? or Web 3.0? ;-) |
| 12:25 | <jgraham> | Philip`: Yeah, the grammar in that quote sounds all wrong |
| 12:25 | <MikeSmith> | hey lazychannel, what's the hg command-line syntax for pulling a specific revision in mercurial? |
| 12:25 | <hsivonen> | hg pull -r 431432421 -u |
| 12:25 | <MikeSmith> | hsivonen: thanks |
| 13:03 | <zcorpan> | http://my.opera.com/core/blog/2010/03/03/everything-you-need-to-know-about-html5-video-and-audio-2 |
| 13:03 | <zcorpan> | hmm, i managed to screw up the url. oh well |
| 13:03 | <gsnedders> | n00b |
| 13:08 | <Philip`> | Cool URIs don't change, but screwed up URIs should change so that they become cool |
| 13:09 | <mut> | hello |
| 13:09 | <boblet> | I dunno, I think you could add some more keywords in there yet |
| 13:09 | <boblet> | hey mut |
| 13:10 | <mut> | im trying to make a site with canvas |
| 13:10 | <mut> | well its all working pretty well for me so far |
| 13:10 | <boblet> | zcorpan: what about adding -ogg-vorbis-h264 at the end? |
| 13:10 | <mut> | but im using .translate(), i just wondered how do you reset the origin to 0,0 and clear the canvas? |
| 13:14 | <mut> | like im drawing something based on user interaction based on a form, but when they select something, i wanna clear the canvas, and start again |
| 13:16 | <mut> | ah if i change/reset the width/height of the canvas it "resets" the canvas... :) |
| 13:18 | <Philip`> | mut: Usually you should use save()/restore() to reset the origin |
| 13:18 | <Philip`> | and clearRect to clear the bitmap |
| 13:18 | <Philip`> | though if you really want to reset the entire canvas then changing size will work too |
| 13:21 | <mut> | oh |
| 13:21 | <mut> | save() restore() |
| 13:21 | <mut> | hmm ill take a look |
| 13:22 | <mut> | so i save() when the canvas is clear, and then do restore()? |
| 13:24 | <Philip`> | They don't affect the canvas bitmap at all |
| 13:24 | <Philip`> | save() just remembers the current origin transformations, fill styles, etc |
| 13:24 | <Philip`> | and restore() resets them back to the last saved values |
| 13:24 | <Philip`> | Oh, wait |
| 13:24 | <Philip`> | That's what you said |
| 13:24 | <Dashiva> | Wow, Jobs actually called animating an object "original technology"? |
| 13:25 | <mut> | ok |
| 13:25 | <Philip`> | so, yes :-) |
| 13:25 | <mut> | errm but when i try that, the canvas dosent clear |
| 13:25 | <mut> | oh i nead to clearrect toto |
| 13:25 | <mut> | too |
| 13:25 | <mut> | heh |
| 13:25 | <Dashiva> | Is clearrect cheaper than .width = .width? |
| 13:26 | <Philip`> | Dashiva: Yes, in theory, because it has to do less work |
| 13:26 | <Philip`> | (although implementations might be suboptimal and do more work in the common case, in order to handle more general cases) |
| 13:27 | <mut> | philip, i have used transform to move things around, but i want to reset the origin to the canvas 0,0 |
| 13:27 | <mut> | err can i give you a link? |
| 13:27 | <Philip`> | Sure |
| 13:28 | <mut> | cheers |
| 13:28 | <mut> | it just keeps stacking images |
| 13:28 | <mut> | if that makes sense |
| 13:31 | <Philip`> | It's probably best to not call ctx.save in init(), and instead make the drawStairs function do "ctx.clearRect(...); ctx.save(); ...do all the drawing...; ctx.restore()" |
| 13:31 | <mut> | ok |
| 13:33 | <Philip`> | That way the canvas would always be in a 'normal' state, except temporarily while it's inside the drawing function (where it'll mess around with the state but it'll restore everything at the end) |
| 13:34 | <mut> | ok thats almost there :) cheers |
| 13:35 | <mut> | heh wierd, its just messing up on one small thing now |
| 13:35 | <mut> | thanks for that philip, *back to work |
| 13:35 | <mut> | :) |
| 13:47 | <Creap> | What would you suggest to use to accomplish something like <nobr>, for instance to keep the unit on the same row when writing "10 GB" |
| 13:47 | <Creap> | any particular element or just a span with some custom class |
| 13:48 | <Creap> | or 10 GB? |
| 13:48 | <Philip`> | "£10" :-) |
| 13:49 | <Philip`> | Oops |
| 13:49 | <Philip`> | I thought you said GBP |
| 13:49 | Philip` | isn't good at reading today |
| 13:50 | <Creap> | hehe |
| 13:50 | <Philip`> | nbsp seems to me the semantically appropriate thing when you really want a space that doesn't break |
| 13:51 | <Philip`> | though maybe it renders narrower than you'd want |
| 13:51 | <Creap> | true, I didn't even think of that, makes the eyes bleed because of the misuse |
| 13:51 | <Creap> | or well, I thought of that, but not until I've already asked ;) |
| 13:52 | <Philip`> | I wouldn't really recommend using for layout purposes, but it sounds like a good fit here :-) |
| 13:54 | <jgraham> | You can always insert a literal U+00A0 character. That would be less ugly, but possibly also harder to maintain |
| 14:00 | <asmodai> | zcorpan: Mmm, looks like my.opera.com has some issues now. |
| 14:01 | <zcorpan> | yeah, i noticed. maybe my post was too long |
| 14:01 | <asmodai> | zcorpan: Oh noes! You broke it |
| 14:02 | <asmodai> | Ah, there it is again |
| 14:02 | <karlcow> | http://www.theregister.co.uk/2010/03/02/virgin_america_html_flash/print.html |
| 14:02 | <mut> | hey philip any chance of 1 more quick piece of advice? |
| 14:03 | <mut> | same again, but it all works |
| 14:03 | <mut> | except the bottom part of the drawing dosent reset |
| 14:03 | <mut> | its bizzare |
| 14:03 | <asmodai> | zcorpan: Guess you guys moved away from Python/Genshi for it nowadays? |
| 14:04 | <zcorpan> | asmodai: i have no idea what's used for my.opera |
| 14:04 | <asmodai> | Shoot |
| 14:04 | <asmodai> | I know it was Python with Genshi at one point, since I had some contact with the dev guys on your side |
| 14:05 | <zcorpan> | ah |
| 14:05 | <gsnedders> | asmodai: my.opera uses Perl |
| 14:06 | zcorpan | gotta go catch a train |
| 14:12 | <asmodai> | ah, sorry |
| 14:12 | <asmodai> | was widgets.opera.com |
| 14:12 | <asmodai> | gsnedders: http://my.opera.com/WebApplications/blog/2008/02/07/widgets-opera-com-2-0 |
| 14:12 | <gsnedders> | asmodai: Yeah, that uses Python |
| 14:13 | <asmodai> | Still does? |
| 14:13 | <gsnedders> | Yeah |
| 14:16 | <Philip`> | mut: Hmm, that might be because you're calling clearRect after having called scale, so the cleared area will be scaled too, so it might not be covering the whole canvas |
| 14:19 | <mut> | hmmm |
| 14:19 | <mut> | ok cheers |
| 14:19 | <mut> | ill take a look |
| 14:21 | <mut> | philip` I dont think its that, because, it seems to leave behind the last drawing of the bottom tread |
| 14:21 | <mut> | if that makes sense |
| 14:21 | <mut> | so whereever it was last, gets left behing |
| 14:21 | <mut> | sometimes the new one is further down than that |
| 14:22 | <mut> | (although im sure that the scale thing isnt helping :) |
| 14:28 | <Philip`> | mut: Not sure if this is the problem, but I think you need to call beginPath more |
| 14:29 | <Philip`> | Typically you should always do "beginPath(); moveTo and lineTo etc to set up the path; fill() or stroke();" |
| 14:29 | <mut> | oh ok |
| 14:29 | <Philip`> | because fill/stroke don't reset the current path, so it will continue to exist for your next path-drawing operation |
| 14:29 | <Philip`> | and so you'll keep building up a longer and longer path, and keep refilling/restroking the early parts lots of times |
| 14:30 | <Philip`> | e.g. in the drawStairRun loop |
| 14:30 | <Philip`> | and it looks like that'll also be drawing the bits of path left over from the end of the previous time you redrew the screen |
| 14:30 | <mut> | Ok, ill go and have a look (only found canvas this morning! so Im on the bottom of the learning curve) ill go and sort out the beginpath stuff out first and see if that helps |
| 14:30 | <Philip`> | so beginPath should help :-) |
| 14:30 | <mut> | ok ta |
| 14:32 | <mut> | Im liking this canvas stuff, all the functions like .arc etc are very similar to CNC coding, so its coming quite quickly to me |
| 14:33 | <jgraham> | hsivonen: http://blogs.msdn.com/ie/archive/2010/03/02/how-ie8-determines-document-mode.aspx |
| 14:34 | <paul_irish> | i was kinda curious why they were replicating henri's flowchart two years later... |
| 14:35 | <paul_irish> | like they forgot to hit the publish button on this one? |
| 14:36 | <Rik|work> | I think they're trying to prove that "a picture is worth a thousand words" is a wrong adage |
| 14:36 | <hsivonen> | jgraham: interesting. I need to verify that my flowchart matches |
| 14:37 | <mut> | Philip` sorted, thanks |
| 14:40 | <jgraham> | hsivonen: I think they have a few things you don't and you have a few things that they don't |
| 14:41 | <Philip`> | mut: Good :-) |
| 14:48 | <hsivonen> | paul_irish: maybe they didn't forget to publish it earlier but it took this long to map the thing out :-) |
| 14:49 | <paul_irish> | hahaha |
| 14:52 | <Philip`> | Word and Excel really ought to detect when you're pressing printscreen and disable the spelling/grammar checker underlines |
| 14:53 | <lazni> | I think a lower level thingy catches the pressing |
| 14:57 | <mut> | i always hit print preview before i printscreen cus of that |
| 15:21 | <karlcow> | 77% of French people do not understand the ballot screen for browsers http://www.silicon.fr/fr/news/2010/03/02/une_majorite_d_internautes_ne_connait_pas_le_ballot_screen_de_microsoft |
| 15:23 | <Philip`> | 23% understand the concept of web browser choice? That's much higher than I'd expect |
| 15:26 | <mut> | :P |
| 15:26 | <jgraham> | Philip`: That's roughly the same percentaqge that already use something other than IE I expect |
| 15:29 | <Philip`> | jgraham: I thought most of those only used not-IE because a family member installed Firefox for them and set it as the default |
| 17:59 | <TabAtkins> | So, hey, does anyone know of any examples of video on the web that are *not* done through <video> or plugins? |
| 17:59 | <TabAtkins> | I don't know if this actually exists or not. |
| 18:00 | <TabAtkins> | And more specifically, if it *does* exist, does it act like <video> with respect to letter-boxing when the container isn't the right size? Or does it try and fully scale itself like an image would? |
| 18:00 | <Philip`> | You mean native browser support for videos? |
| 18:00 | <TabAtkins> | I guess. Not sure. Hakon says there is some. I doubt this. |
| 18:01 | <TabAtkins> | It's the point upon which our argument was hinging on today in the conf call. |
| 18:01 | <Philip`> | There's animated GIFs |
| 18:01 | <Philip`> | Maybe also http://msdn.microsoft.com/en-us/library/ms533606(VS.85).aspx |
| 18:02 | <TabAtkins> | They resize like images, right? |
| 18:02 | TabAtkins | has several animated gifs on hand, and can check. |
| 18:02 | <Philip`> | I don't know what else exists that isn't <embed>/<object>, though |
| 18:02 | <Philip`> | TabAtkins: Yes, they're just the same as any image |
| 18:02 | <Philip`> | (at least when you use them in <img>) |
| 18:03 | <TabAtkins> | Yeah. |
| 18:03 | <TabAtkins> | Ah, nobody cares about HTML+TIME. |
| 18:03 | <Philip`> | Seems strange to hinge an argument on the behaviour of clearly obscure technologies where you don't even know if they exist :-) |
| 18:04 | <TabAtkins> | Well, if they *do* exist, we have to add a new value that means "do whatever is appropriate for the media". If they don't, then the existing ones are fine. |
| 18:05 | <Philip`> | Why do we have to do what they do? |
| 18:05 | <Philip`> | <video> will fail to support a load of things that people do already |
| 18:06 | <Philip`> | (When you say "plugins", do you just mean Flash, or do you also mean to exclude things like WMP/QuickTime/RealPlayer plugins that were used before Flash video came along?) |
| 18:07 | <TabAtkins> | Anything that would invoke the plugin architecture. |
| 18:07 | <TabAtkins> | Also, sorry, was referring to a CSS discussion, not HTML. |
| 18:08 | <TabAtkins> | Specifically, the image-fit property, which determines how replaced content scales in its content box. |
| 18:08 | <TabAtkins> | So, those other things you mentioned count as plugins. |
| 18:08 | <Philip`> | Oh |
| 18:13 | <TabAtkins> | Hm. I need to eat a distinctly Texan lunch today, since it will be one of my last few before I move. Wonder where I should go? |
| 18:13 | <Philip`> | I suggest going to Texas |
| 18:13 | <TabAtkins> | Step 1: Done. |
| 18:14 | <Philip`> | Sounds like you're well on the way to having a genuine Texan lunch, then |
| 18:15 | <TabAtkins> | I also need to finally go buy myself a cowboy hat tonight, since it is a conspicuous lack in my life. |
| 18:15 | <TabAtkins> | Gotta play up the "y'all" and "ma'am" when I move. |
| 20:08 | <AryehGregor> | MikeSmith, what exactly is the relationship between validator.nu and validator.w3.org? Could my problem be fixed in validator.nu and it would also work on the W3C validator? The latter is the one that most people use. |
| 20:13 | <MikeSmith> | AryehGregor: the current W3C validator is essentially a Perl script that wraps around James Clark's old C-based "sp" validation engine, which is an SGML tool that does DTD-based validation. if it either encounters a <!doctype html> doctype or is manually told to check a doc as HTM5, what that perl script does is, it uses the validator.nu REST API to validate the document, then processes the results of that and reports them |
| 20:14 | <AryehGregor> | Oh, so it's literally requesting results from validator.nu, not just running the same software? Interesting. |
| 20:15 | <MikeSmith> | AryehGregor: no, actually, to be precise, it's usually a locally-installed headless instance of the validator.nu backend |
| 20:15 | <AryehGregor> | Ah. |
| 20:15 | <AryehGregor> | It seems like I'd want 1) validator.nu to say it's valid HTML5 if it is, even with a confusing doctype, and 2) the Perl script to try passing things to validator.nu if it has an obsolete but conforming doctype but doesn't validate according to that doctype. |
| 20:15 | <AryehGregor> | Does that sound right? |
| 20:16 | <MikeSmith> | AryehGregor: yeah |
| 20:16 | <AryehGregor> | Okay, thanks. |
| 20:17 | <MikeSmith> | but I would not recommend calling those doctypes "obsolete but conforming" in discussions on the qa-dev and www-validator lists |
| 20:18 | <MikeSmith> | you would likely just get people taking issue with that wording rather than listening to what you're proposing |
| 20:19 | <AryehGregor> | Er, okay. |
| 20:19 | <AryehGregor> | That's the term the HTML5 spec uses, right? |
| 20:19 | <MikeSmith> | right, but even the intent of that is the specific context of HTML4 |
| 20:19 | <MikeSmith> | *HTML5 |
| 20:20 | <MikeSmith> | there are lot of people who don't care about validating their documents as HTML5 at this point and who may never want to change or update their content to get it validate against HTML5, if it's already valid HTML4 or XHTML1 |
| 20:20 | <MikeSmith> | which is fine |
| 20:21 | <AryehGregor> | Right. I'd say only kick it to validator.nu if it fails validation under the given doctype. |
| 20:22 | <MikeSmith> | even in that case, I don't think you will get a warm reception if you propose that it does that automatically, or especially if it does it without reporting their errors against whatever doctype the document actually uses |
| 20:23 | <MikeSmith> | for one thing, it would mean validation of every invalid document would take longer, because it would get automatically checked twice |
| 20:25 | <AryehGregor> | Maybe we could use a secret code, like adding <!-- html5 --> right after the doctype to trigger HTML5 validation mode? |
| 20:25 | <AryehGregor> | Hacky, but it would be a lot simpler, I imagine. |
| 20:25 | <MikeSmith> | icky indeed |
| 20:26 | <MikeSmith> | I don't know what the target users would be who would take the time to do that |
| 20:26 | <AryehGregor> | Or could the HTML5 obsolete-but-conforming doctypes be modified so that they specify the same DTD/entities, but are not actually conforming HTML4 or whatever? |
| 20:26 | <AryehGregor> | Take the time to do what? |
| 20:26 | <AryehGregor> | I'd do it in MediaWiki, that's my concern here. The current setup has MediaWiki not validating even if it's valid HTML5, if in well-formed XML mode. |
| 20:27 | <MikeSmith> | I meant, take the time to add a special-purpose comment like that |
| 20:27 | <MikeSmith> | but I think you'd get a warmer reception if the suggestion were for an opt-in mechanism that prompts users to choose whether they want to re-validate the doc at HTML5 or not |
| 20:27 | <AryehGregor> | I meant, I'll add the comment to MediaWiki if it's the only way to validate. |
| 20:27 | <MikeSmith> | I see |
| 20:27 | <AryehGregor> | I'm concerned about people sticking the page in the validator and getting a "This document is invalid!" |
| 20:28 | <AryehGregor> | Nobody's going to notice a revalidate button unless it's unreasonably prominent. |
| 20:28 | <AryehGregor> | (since most such documents will not be valid HTML5 either) |
| 20:28 | <AryehGregor> | I don't know much about doctypes -- is there some way to make the doctype invalid HTML4/XHTML1 while still keeping the same behavior in browsers and other XML processors? |
| 20:29 | <AryehGregor> | Or is every part that's specified essential? |
| 20:29 | <MikeSmith> | I don't think there's any way to do that, no |
| 20:29 | AryehGregor | thinks it all looks essential, although he doesn't understand all of it. |
| 20:30 | <MikeSmith> | only thing I can think of is that SGML does not require the system ID |
| 20:31 | <MikeSmith> | that is, the URI part in the doctype |
| 20:31 | <MikeSmith> | the second quoted string |
| 20:31 | <MikeSmith> | it only needs to first |
| 20:31 | <AryehGregor> | That's necessary for XML validators to figure out the entities, though. |
| 20:31 | <MikeSmith> | leaving off the system ID makes it invalid XML |
| 20:31 | <AryehGregor> | Hmm. |
| 20:31 | <MikeSmith> | actually, non-well-formed XML, I gues |
| 20:31 | <AryehGregor> | Well, a comment would be hacky, but I had actually thought of it earlier anyway, because it will also serve to inform human readers that the document is actually HTML5. |
| 20:32 | <AryehGregor> | Lest they get confused. |
| 20:32 | <AryehGregor> | Do you think that would stand a chance of getting approved? |
| 20:32 | <MikeSmith> | I very much doubt it |
| 20:32 | <AryehGregor> | Bleh. |
| 20:32 | <MikeSmith> | I think most people consider specialized comments of that kind to be an anti-pattern |
| 20:33 | <MikeSmith> | that's the kind of thing that SGML/XML processing instructions are meant for |
| 20:33 | <MikeSmith> | but such processing instructions are not valid in HTML5, for good reasons |
| 20:33 | <AryehGregor> | Can you think of anything else that might possibly get approved? Maybe I should just explain the problem and outline the pros and cons I see with each possible solution and hope someone finds something acceptable beyond "call it invalid"? |
| 20:34 | <MikeSmith> | AryehGregor: explaining the problem first would be good, yeah |
| 20:34 | <AryehGregor> | Well, I was planning on doing that anyway. :) |
| 20:34 | <AryehGregor> | I'll give that a shot when I have the time. |
| 20:35 | <AryehGregor> | Thanks. |
| 20:35 | <MikeSmith> | OK. proposing specific solutions is fine, too, but the risk is always that people then get hung up on arguing about the merits or faults of the proposed solution rather than thinking about the problem |
| 20:36 | <AryehGregor> | Right. |
| 20:36 | <AryehGregor> | Got to go now, I'll post a message to one of those lists soon. |
| 20:51 | <MikeSmith> | <twitter!andymantell> @w3c Still don't get why there is no previous sibling combinator in CSS3, it'd be so so useful - any chance this could make it in one day? |
| 20:55 | <Dashiva> | That would lead to possible cycles |
| 20:57 | <TabAtkins> | MikeSmith: I can answer that. Gimme a sec; busy right now, then I'll give you a details response. |
| 20:57 | <MikeSmith> | TabAtkins: thanks |
| 20:58 | <MikeSmith> | or feel free to reply directly to him on twitter if you want |
| 20:58 | <MikeSmith> | Dashiva: performance issue? |
| 21:03 | <Dashiva> | I'm not sure if it's only performance or if it could lead to loops in the layout as well |
| 21:04 | <TabAtkins> | No, just within selectors you can't get cycles. |
| 21:04 | <TabAtkins> | It's performance reasons, though previous-sibling and parent combinators are okay enough that implementors are all right with doing them. |
| 21:04 | <Dashiva> | Oh, my bad then |
| 21:04 | <TabAtkins> | We just haven't started writing Selectors 4 yet. |
| 21:05 | <TabAtkins> | They're pretty much certain to be in there. |
| 21:06 | <TabAtkins> | Basically browsers have a somewhat unintuitive performance profile that makes all the reverse-versions of the current 4 combinators pretty inefficient, even though they can be done fairly easily in selectors engines like Sizzle. |
| 21:07 | <TabAtkins> | It's because nearly all selectors matching in browsers happens on an incomplete DOM as a page is being loaded, while js-based selectors engines operate on complete doms. |
| 21:09 | <TabAtkins> | And the existing combinators happen to work great with that, as they only need information that is guaranteed to already be in the DOM. |
| 21:09 | <TabAtkins> | If you are trying to match div > span, when a span appears you are guaranteed that, if it has a div parent, you'll have already seen it and know about it because of the way HTML's structure works. |
| 21:10 | <TabAtkins> | While "span < div" doesn't offer the same promise - when you see a div, you then have to wait until all its children have been loaded before you can be sure if it has a span child. |
| 21:13 | <TabAtkins> | MikeSmith: So there you go. Try condensing that down to 140 characters. ^_^ |
| 21:14 | <TabAtkins> | I want :has-child() *so bad*, but I understand why it hasn't appeared so far. It will appear in the future, though. |
| 21:14 | TabAtkins | mainly wants it so he can do "label:has-child(:checked)". |
| 21:18 | <MikeSmith> | TabAtkins: thanks |
| 21:23 | <MikeSmith> | TabAtkins: you mentioned that implementors are OK with doing previous-sibling and parent combinators |
| 21:30 | <MikeSmith> | Hixie: I need some help in getting the Microdata draft through pubrules |
| 21:31 | <MikeSmith> | TabAtkins: so you are saying that implementors are planning to add support for previous-sibling? |
| 22:38 | <Hixie> | MikeSmith: here |
| 22:40 | <MikeSmith> | Hixie: I am told we can't have URLs for images and other things in examples that don't resolve |
| 22:40 | <MikeSmith> | particularly to third-party sites |
| 22:40 | <MikeSmith> | e.g., the oreilly.com URLs in the examples |
| 22:41 | <MikeSmith> | so I think we should change those to example.com URLs |
| 22:43 | <Philip`> | example.com is a third-party site, and most URLs on it won't resolve |
| 22:43 | <Philip`> | so that seems an inconsistent policy |
| 22:43 | <MikeSmith> | yeah, but we know they are not meant to reolve |
| 22:45 | <MikeSmith> | I believe in part the logic of this is that we don't want somebody contacting O'Reilly to complain about a broken URL that we created |
| 22:46 | <MikeSmith> | Hixie: the bogus dev.w3.org house.jpeg and mailbox.jpeg URLs as well |
| 22:46 | <MikeSmith> | Philip`: example.com is not actually a third-party site |
| 22:46 | <MikeSmith> | it is non-site |
| 22:47 | <Philip`> | It looks like a site to me |
| 22:47 | <Philip`> | Not a very exciting one, admittedly |
| 22:47 | <Philip`> | It's got a web server and everything |
| 22:48 | <MikeSmith> | well, it's a site that tells you to go read an RFC to find out why there's actually no content at the site |
| 22:48 | <TabAtkins> | MikeSmith: Yeah, that's the indication so far. Previous adjacent sibling and parent aren't *too* bad, and they're useful enough to justify the cost. General previous sibling (reverse of ~) and ancestor, though, are much worse and can't justify themselves as well. |
| 22:49 | <MikeSmith> | TabAtkins: OK |
| 22:49 | <Hixie> | MikeSmith: looking... |
| 22:50 | <MikeSmith> | Philip`: and it's site that we can create arbitrary fully-qualified domain names for without somebody nagging me to fix them |
| 22:50 | <MikeSmith> | s/create/invent|make up/ |
| 22:51 | <TabAtkins> | Hehe, I love people who don't know how mailing lists work, or how they got on them. |
| 22:51 | <TabAtkins> | Specifically, I love the line "I'M SO FATUP WITH THIS!!!". |
| 22:51 | <Hixie> | MikeSmith: the oreilly URLs were provided by an O'Reilly employee, so I presume that they are valid RDF URLs (which don't resolve by design) |
| 22:52 | <Hixie> | MikeSmith: i couldn't find any dev.w3.org house.jpeg URLs |
| 22:53 | <MikeSmith> | Hixie: the example in http://dev.w3.org/html5/md/#associating-names-with-items |
| 22:54 | <Hixie> | wtf |
| 22:54 | <Hixie> | that's not in the source |
| 22:54 | <Hixie> | what on earth |
| 22:55 | <Hixie> | c.f. http://www.whatwg.org/specs/web-apps/current-work/complete.html#associating-names-with-items |
| 22:56 | <MikeSmith> | weird |
| 22:56 | <Hixie> | very weird |
| 22:56 | <MikeSmith> | I hope I'm not to blame for this |
| 22:56 | <Hixie> | it's in the version i check in |
| 22:56 | <Hixie> | what on earth is going on here |
| 22:57 | Hixie | walks through this pipeline |
| 22:58 | <MikeSmith> | Hixie: about the O'Reilly URLs, if you can point me to an e-mail message or IRC log or whatever where those came from, I can pass it on |
| 22:59 | <Hixie> | http://gavin.carothers.name/2009/08/13/trying-to-understand-microdata-rdfa/ |
| 23:00 | <MikeSmith> | thanks |
| 23:02 | <Hixie> | (this is his google profile http://www.google.com/profiles/gcarothers ) |
| 23:04 | <Hixie> | apparently the URL gets screwed up by my filter-for-w3c script |
| 23:04 | <Hixie> | aha, i see the problem |
| 23:06 | <MikeSmith> | Hixie: it's still not clear whether those are real or just examples he's using for illustration purposes |
| 23:07 | <MikeSmith> | I would really rather we just replace them with example.com URLs instead of me needing to argue with the publication team to convince them we're not violating the policy |
| 23:08 | <Hixie> | k |
| 23:08 | <Dashiva> | <dd><span itemprop="http://purl.org/dc/terms/creator">Wil Wheaton</span></dd> |
| 23:08 | <Dashiva> | Is the span really necessary there? |
| 23:09 | <TabAtkins> | No. |
| 23:14 | MikeSmith | steps away for a bit |
| 23:48 | <Hixie> | MikeSmith: ok, checked in changes to fix those two problems |