| 00:13 | <Hixie> | nessy: yt? |
| 00:13 | <nessy> | starting to wake up... |
| 00:14 | <Hixie> | you wouldn't happen to be in the bay area in the near future by any chance would you? |
| 00:14 | <nessy> | so, yes :) |
| 00:14 | <nessy> | nothing planned, no |
| 00:14 | <nessy> | why? |
| 00:14 | <Hixie> | k |
| 00:14 | <Hixie> | i'm starting to look at this subtitles stuff |
| 00:14 | <Hixie> | if you were around i would have suggested lunch so i could get a brain dump |
| 00:14 | <Hixie> | but no worries |
| 00:15 | <Hixie> | i can read also :-) |
| 00:15 | <nessy> | I've got to reply to your request and to henry's proposal etc |
| 00:15 | <nessy> | got lots to do - was just swamped with other stuff this week |
| 00:15 | <nessy> | (last week, really) |
| 00:15 | <nessy> | but shoot any questions you have |
| 00:15 | <Hixie> | i've 192 e-mails to read through on the subject (mostly to whatwg; most of the public-html-a11y ones i read as they passed and didn't save) |
| 00:16 | <Hixie> | so i'm sure i'll have questions :-) |
| 00:16 | <nessy> | best intro is probably http://blog.gingertech.net/2010/04/11/introducing-media-accessibilit-into-html5-media/ |
| 00:16 | <nessy> | gives a background |
| 00:16 | <nessy> | most discussions on whatwg were about the caption format - ppl don't like DFXP/TTML |
| 00:17 | <nessy> | so mostly about developing a new format (microformats story all over?) |
| 00:17 | <Hixie> | cool, will read |
| 00:17 | <nessy> | won't go well with the a11y ppl, but I don't really care what format we use - I think it will be many over time |
| 00:18 | <Hixie> | yeah if we can avoid inventing yet another format that would certainly be good |
| 00:18 | <nessy> | are you starting to dig into it now? |
| 00:18 | <Hixie> | yeah |
| 00:18 | <nessy> | bummer - I wanted to give you a list of samples and use cases |
| 00:18 | <nessy> | will try and get to that today then |
| 00:18 | <Hixie> | oh it'll take me a while i'm sure |
| 00:18 | <Hixie> | no rush |
| 00:19 | <Hixie> | at this point i'm mostly just trying to ground my sense of what use cases are real and what are hypothetical |
| 00:19 | <Hixie> | by looking at what people actually do with subtitles |
| 00:19 | <Hixie> | for example i thought that there was a need to handle individual words being emphasised, but i've yet to find a single example of that in the real world |
| 00:19 | <Hixie> | though i have found two examples of ruby! |
| 00:20 | <Hixie> | i can't say i was ultra-pleased by that, given how ruby is like 10 times harder to deal with than italics :-P |
| 00:21 | <nessy> | most styled captions are commercial stuff |
| 00:21 | <nessy> | I'll see if I can find you some examples |
| 00:21 | <Hixie> | that would be awesome |
| 00:34 | <nessy> | with the use cases - are you ok if I take some screenshots of commercial stuff that may not be generally accessible, but a good example? |
| 00:35 | <Hixie> | yeah totally |
| 01:01 | <daedb> | Hixie: Dragon Age: Origins has individual words emphasized (in bold), if video games count as examples :) |
| 01:08 | <Hixie> | daedb: sure, take screenshots and upload them to the wiki page :-) |
| 01:10 | <daedb> | I'll try to remember that next time I see it (or see if Google can find a relevant screenshot)... it has italics too, btw, for character actions (like sighing) :p |
| 01:59 | <MikeSmith> | about http://www.w3.org/Bugs/Public/show_bug.cgi?id=7688 |
| 01:59 | <MikeSmith> | what's POLA? |
| 02:00 | <Hixie> | something of least something? |
| 02:00 | <Hixie> | i've been playing Bad Company 2 too much. I just got spam about some sort of digital camera and the subject line mentioned 5x optical zoom and my brain instantly junked the e-mail on the basis that anything lower than 12x wasn't worth looking at |
| 02:01 | <doublec> | principle of least astonishment |
| 02:02 | <nessy> | Hixie: can I also add examples of other time-aligned text to that wiki page? |
| 02:02 | <Hixie> | sure |
| 02:02 | <nessy> | e.g. chapter markers, linguistic annotation |
| 02:02 | <nessy> | ok, cool |
| 02:02 | <nessy> | and: yes, textual audio descriptions are just text with time markers |
| 02:03 | <MikeSmith> | doublec: thanks |
| 02:03 | <MikeSmith> | and an honorable mention to Hixie for "something of least something" |
| 02:35 | <nessy> | Hixie: are you sure the whatwg wiki lets me upload images? |
| 02:36 | <nessy> | I can't find a means and get the feeling that it's disabled in the mediawiki setup |
| 02:36 | <nessy> | also, http://wiki.whatwg.org/wiki/Special:ListFiles is empty |
| 02:41 | <Hixie> | nessy: oh you'll have to upload them elsewhere first |
| 02:41 | <nessy> | ah I see ... |
| 02:41 | <Hixie> | nessy: if you need somewhere to upload them use http://software.hixie.ch/utilities/cgi/uploader/uploader |
| 02:42 | <Hixie> | password is w643kJ6Gv43q3 |
| 02:42 | <nessy> | ah, cool, thanks! |
| 02:55 | <nessy> | Hixie: good reading - http://main.wgbh.org/wgbh/pages/mag/services/captioning/faq/sugg-styles-conv-faq.html ;-) |
| 03:14 | <nessy> | I think I'm finding it hard to provide an example of musical notes in captions, since all online systems failed to render musical notes |
| 03:15 | <nessy> | I keep coming across ppl complaining about it ;-) |
| 03:15 | <nessy> | it's kinda ironic! |
| 03:23 | <nessy> | yay, found one! |
| 04:19 | <nessy> | Hixie: you'd enjoy watching http://svg-wow.org/audio/animated-lyrics.html |
| 04:23 | <MikeSmith> | annevk: is Daniel around? |
| 04:24 | <annevk> | sitting next to me |
| 04:24 | <annevk> | he's also on skype |
| 04:24 | <annevk> | I just told him about it, seems he's ok |
| 04:24 | <MikeSmith> | great |
| 04:24 | <annevk> | could discuss it more over lunch |
| 04:24 | <MikeSmith> | yep |
| 05:51 | <boblet> | Hixie: yt? |
| 06:53 | <annevk> | so TTML does not have ruby support it seems |
| 06:53 | <annevk> | already falls short of catering to in the wild use of captioning |
| 06:53 | <annevk> | while being fricking complex |
| 06:54 | <annevk> | genius |
| 06:56 | <hsivonen> | is ruby actually used for captioning in the wild? |
| 06:57 | <annevk> | http://wiki.whatwg.org/wiki/Timed_tracks#Ruby |
| 06:57 | <hsivonen> | which side of 80/20 inside locales that use ruby for anything at all? |
| 06:58 | <annevk> | it seems that in Japanese it would be required as otherwise meaning would be ambigious |
| 06:58 | <othermaciej> | that is some crazy exmples |
| 06:58 | <othermaciej> | *examples |
| 06:59 | <annevk> | the page is fascinating |
| 07:01 | <hsivonen> | well, at least the most complex cases aren't for accessibility |
| 07:01 | <hsivonen> | more like gratuitous "just because we can" for translation or karaoke |
| 07:01 | <hsivonen> | (I don't mean ruby here. I mean stuff on the page in general) |
| 07:04 | <hsivonen> | a lot of the plain text translations look very ugly cmpared to what we get here on tv. and I thought what we have wasn't pretty |
| 07:12 | <MikeSmith> | ruby annotations are used quite commonly in Japanese subtitles for films and TV programs and such (not just for karaoke or whatever other edge cases) |
| 07:14 | <hsivonen> | ok. seems like a FAIL for TTML then |
| 07:14 | <MikeSmith> | yeah |
| 07:16 | <MikeSmith> | a lot of the other Japanese caption examples there are not things I at least typically see in normal use |
| 07:16 | <hsivonen> | what we get is pretty readable comapred to many of the Latin examples: http://hsivonen.iki.fi/screen/subtitles.jpg |
| 07:17 | <MikeSmith> | interesting that that uses stroking |
| 07:21 | <annevk> | is CNN not an example of somewhat complex inline formatting btw? |
| 07:21 | <nessy> | what is required to support ruby? |
| 07:21 | <annevk> | ruby markup |
| 07:22 | nessy | goes google |
| 07:22 | <annevk> | and some CSS styling if you want to place it elsewhere prolly |
| 07:22 | <nessy> | i18n is so fascinating! |
| 07:22 | <annevk> | kind of weird the i18n group hasn't catched that with TTML |
| 07:23 | <annevk> | maybe they don't consider it a requirement |
| 07:23 | <nessy> | I'm not sure TTML is even properly finalised yet |
| 07:24 | <annevk> | it's in CR |
| 07:24 | <annevk> | which means it passed Last Call review |
| 07:24 | <nessy> | so there is a <ruby> tag in HTML? |
| 07:24 | <annevk> | yeah |
| 07:25 | <nessy> | cool - I didn't follow those discussions |
| 07:25 | <annevk> | http://www.whatwg.org/html#the-ruby-element |
| 07:31 | <othermaciej> | MikeSmith: btw did you change the bugzilla mail setup for public-html? |
| 07:32 | <hsivonen> | MikeSmith: subtitles ha used white text with black outline for decades over here. |
| 07:32 | <Hixie> | nessy: i'm removing the cnn example because it doesn't seem to be synchronised with the video -- let me know if it is in fact synchronised |
| 07:33 | <hsivonen> | s/ha/have/ |
| 07:33 | <nessy> | let me check |
| 07:33 | <Hixie> | do you have any more info on the other ticker example? |
| 07:33 | <Hixie> | is it just reproducing the subtitle below the video? |
| 07:33 | <Hixie> | or is it something else? |
| 07:34 | <nessy> | oh, that was the one under ticker text? |
| 07:34 | <hsivonen> | isn't ticker by definition something else? |
| 07:34 | <nessy> | it is |
| 07:35 | <nessy> | the TV example (CNN) I wanted to use as an example how it shows on TV |
| 07:35 | <annevk> | but it comes with the video and should supposedly be exposed in some way... |
| 07:35 | <nessy> | but if you only want current web examples, that's fine |
| 07:35 | <MikeSmith> | othermaciej: yeah, I set the user prefs on the public-html⊙wo bugzilla account to disable all e-mail for the account |
| 07:35 | <nessy> | so the one that is currently under "ticker text" has both, captions AND ticker text |
| 07:35 | <hsivonen> | annevk: <marquee> |
| 07:35 | <Hixie> | i'm just looking for timed tracks, doesn't have to be web-only -- in fact, non-web is more interesting |
| 07:36 | <Hixie> | but the cnn tickers seem to just be a continuous unsynchronised stream of information that happens to be overlaid, you wouldn't want to use the same mechanism for those as for subtitles |
| 07:36 | <annevk> | hsivonen, that means you need to deploy different video streams |
| 07:36 | <nessy> | you will notice underneath the video a line of "NOW: Roubini: new regulations...." |
| 07:36 | <othermaciej> | MikeSmith: thanks! |
| 07:36 | <nessy> | that is the ticker text |
| 07:36 | <Hixie> | nessy: right but that's the same as the subtitle |
| 07:36 | <nessy> | just by chance |
| 07:36 | <Hixie> | nessy: is there somewhere i can experiment with that example in the wild? |
| 07:36 | <Hixie> | i don't really understand what it is showing |
| 07:36 | <nessy> | also, I am not sure if that is actually a subtitle or "burnt-in" text |
| 07:37 | <nessy> | there are multiple javascript libraries for doing ticker text |
| 07:37 | <hsivonen> | I wouldn't count newscast title of the story as a subtitle |
| 07:37 | <nessy> | let me see if I can track that one down |
| 07:37 | <hsivonen> | just like I wouldn't count mivie titles |
| 07:37 | <nessy> | hsivonen I don't think that's what that is - it's just intermediate |
| 07:38 | <othermaciej> | hmm |
| 07:38 | <othermaciej> | it looks like someone finished the mass move |
| 07:38 | <othermaciej> | that or bugzilla lied to me about the 235 bugs left |
| 07:38 | <Hixie> | the chapter stuff is interesting... wonder what to conclude from it |
| 07:39 | <Hixie> | there's such a wide variety of ways it is exposed that it might make more sense to just support it in the sense of exposing the data in the API but not doing anything with it otherwise |
| 07:40 | <hsivonen> | nessy: ok. I wouldn't count the cable news summary of the last couple of minutes, either |
| 07:40 | <nessy> | yeah, I was wondering about the chapters |
| 07:40 | <nessy> | if they are given inside the file, you will only get them successively normally |
| 07:40 | <nessy> | but for external tracks it is easy to grab them all in one go and display e.g. as a bar at the bottom to click directly to |
| 07:41 | <Hixie> | well certainly we shouldn't disallow the UA from showing UI for them, yeah |
| 07:42 | <Hixie> | annotations also seem like something we should probably leave to sites to figure out, giving them just an API so they don't have to do the timing themselves |
| 07:43 | <Hixie> | basically let them give us a file with the annotation data then have them hook into notifications for when to render each one |
| 07:44 | <annevk> | wouldn't they handle the file as well then? |
| 07:45 | <annevk> | otherwise you might as well add time data to the file format... |
| 07:46 | <Hixie> | ? |
| 07:46 | <hsivonen> | fwiw, the reason why I wouldn't count titles and summary of the past couple of minutes is that they are customarily considered to be under the rule of the art department of the show |
| 07:46 | <nessy> | Hixie, for ticker text I am thinking of something like at http://www.scriptocean.com/dticker.html with time markers synchronised to the video |
| 07:47 | <Hixie> | does anyone do that? |
| 07:47 | <hsivonen> | so even if we had an elaborate mechanism for them, chances are that the art dept would still want to have full control |
| 07:47 | <hsivonen> | and treat them as part of the image |
| 07:47 | <Hixie> | i mean, as hsivonen says, that kind of thing would just be a title in most cases and so burnt in |
| 07:47 | <zcorpan> | why is http://dev.w3.org/html5/core/ not a 410 Gone? |
| 07:47 | <hsivonen> | and they are data on their own right as opposed to being a synchronized alternative of other data |
| 07:47 | <nessy> | Hixie: are you referring to me? |
| 07:47 | <Hixie> | nessy: you and hsivonen yeah |
| 07:48 | <zcorpan> | http://www.theopensourcery.com/keepopen/?p=2205 points to it |
| 07:48 | <nessy> | no, the text is not normally burnt in - only on TV |
| 07:48 | <hsivonen> | Hixie: to me, that kind of thing seems like something you'd put on the surrounding HTML page |
| 07:48 | <nessy> | it is an example of text that comes in at a certain time and leaves at a different |
| 07:48 | <hsivonen> | Hixie: not logically part of the video |
| 07:49 | <annevk> | zcorpan, you could commit a Redirect 410 /html5/core/ .htaccess line I suppose |
| 07:49 | <nessy> | possibly - but then there are definitely times when ticker text is synchronised to the video |
| 07:49 | <Hixie> | hsivonen: it's something i'd expect to see if i uploaded the video to my ipod or played it to my tv, so i'd expect it as part of the video data. |
| 07:49 | <annevk> | zcorpan, not entirely sure if Redirect was part of the syntax, but I thought it was |
| 07:49 | <zcorpan> | annevk: oh right i have cvs access |
| 07:50 | <Hixie> | nessy: would you agree that movie credits are text that comes in at a certain time and leaves at a different time but should be burnt in? |
| 07:50 | <hsivonen> | Hixie: in that case, I'd treat it as nice to have but not as a core requirement for the Web |
| 07:50 | <nessy> | it has very similar properties to captions and subtitles, which is why I think we can have a generic interface to deal with all of them |
| 07:50 | <hsivonen> | Hixie: uploading to iPod is not part of the Web mindset |
| 07:50 | <zcorpan> | annevk: actually i don't have access to dev.w3.org |
| 07:50 | <nessy> | movie credits are part of the video production process, therefore burnt in |
| 07:50 | <hsivonen> | Hixie: you are supposed to navigate to the original page using Mobile Safari on iPod Touch |
| 07:50 | <Hixie> | hsivonen: it's open content, i should be able to do whatever i like with it |
| 07:51 | <Hixie> | hsivonen: screw that :-P |
| 07:51 | <Hixie> | hsivonen: it's video, i want to treat it as video. |
| 07:51 | <hsivonen> | Hixie: depends on what kind of rathole of requirements come with your wishes |
| 07:51 | <Hixie> | nessy: i would expect the kind of thing i see at |
| 07:51 | <nessy> | well, all external but linked text is either closer or less closely related to the video |
| 07:51 | <Hixie> | nessy: ... http://www.scriptocean.com/dticker.html to be part of the video production process also |
| 07:52 | <nessy> | but they are all synchronised to it, so they can all be treated the same way |
| 07:52 | <nessy> | no, ticker text is not part of the video production process - it is created by the publisher normally, not by the producer |
| 07:53 | <Hixie> | well the only time i've ever seen such ticker text is in news shows, where it's definitely part of the production |
| 07:53 | <Hixie> | but if there are other examples i'm certainly eager to have my eyes opened :-) |
| 07:53 | <nessy> | just because the producer and publisher are the same ;) |
| 07:53 | <nessy> | I've seen movies on TV with ticker text that is definitely not created by the production of the movie |
| 07:54 | <Hixie> | but it's synchronised with the movie? |
| 07:54 | <nessy> | e.g. when 9-11 happened, there was plenty of ticker text running over the bottom of a movie |
| 07:54 | <Hixie> | sure but that isn't timed text |
| 07:54 | <Hixie> | it's just a ticker overlay |
| 07:54 | <Hixie> | has nothing to do with the video |
| 07:54 | <nessy> | roughly synchronised in that case only - synchronised between specific larger sections |
| 07:55 | <annevk> | zcorpan, done |
| 07:55 | <nessy> | as I said: text that runs in parallel with video or audio can be more or less tightly related to the video content |
| 07:55 | <nessy> | that changes nothing in the fact that it is running in parallel to the video |
| 07:55 | <annevk> | zcorpan, lets see if we get complaints o_O |
| 07:56 | <zcorpan> | nice |
| 07:57 | <nessy> | I do see your point about the difference between a newscast and move examples though, with the newscast the semantic link is stronger |
| 07:57 | <Hixie> | nessy: semantically, it is the difference between something that runs as part of a <video> element, and something that runs as part of wholy unrelated javascript with wholy unrelated <div>s or <canvas> or ARIA live regions or whatnot, which just happens, because of some CSS, to be rendered over the video |
| 07:57 | <Hixie> | nessy: that's a huge difference |
| 07:57 | <Hixie> | nessy: since we're talking about what <video> should provide |
| 07:57 | <othermaciej> | is "synchronized with the video but over the top of it" distinct from "synchronized with the video but not on top of it?" |
| 07:58 | <Hixie> | othermaciej: not especially, no |
| 07:58 | <Hixie> | othermaciej: i mean, the former might be handled natively while the latter would likely need JS support, but that's all |
| 07:58 | <othermaciej> | is time-based relationship to what the video is currently showing the key semantic relation? |
| 07:58 | <nessy> | I guess the question is: will it be going along its own timeline or along the timeline of the video |
| 07:59 | <Hixie> | othermaciej: they both would need to be expressed in some timed track format |
| 07:59 | <Hixie> | nessy: right |
| 07:59 | <nessy> | even if it goes along its own timeline for most of it, there is definitely a link at some point |
| 07:59 | <othermaciej> | I'm thinking about the ever-popular "video of presentation with synchronized HTML slideshow below" example |
| 07:59 | <nessy> | you won't have a ticker text on a video that starts ticking along without the video playing |
| 08:00 | <Hixie> | nessy: not for the 9/11 subtitles. they would have shown those subtitles irrespective of what movie was showing. |
| 08:00 | <nessy> | not if the video was paused |
| 08:00 | <Hixie> | well there's no way to pause it |
| 08:00 | <Hixie> | since it's all burnt in in that case |
| 08:00 | <nessy> | I'm talking about having a video like a youtube video on a site |
| 08:00 | <nessy> | TV is different |
| 08:01 | <Hixie> | if youtube had breaking-news tickers, they wouldn't pause when you pause the video. |
| 08:01 | <nessy> | if it's burnt-in it's none of our business anyway |
| 08:01 | <Hixie> | burnt-in subtitles tell us what people need on the web |
| 08:01 | <Hixie> | it's entirely our business |
| 08:01 | <Hixie> | it's what i'm using to evaluate the proposals :-) |
| 08:01 | <nessy> | if I go to a youtube page and the video sits there paused, I don't think I would expect the ticker text to be running |
| 08:02 | <nessy> | yes, TV gets you some way there, but we have a lot more flexibility on the Web ;) |
| 08:02 | <othermaciej> | YouTube has time-synchronized popover ads |
| 08:02 | <othermaciej> | but those don't change or go away, they just appear at a certain time |
| 08:02 | <boblet> | Hixie: I found out something interesting about ruby in Korean |
| 08:03 | <nessy> | I think we need to offer all that TV can do, but more :) |
| 08:03 | <boblet> | They’re kind of phasing kanji/hanzi out, and current usage is to have the hangul first, followed inline by the kanji in parentheses, so basically the opposite of the default display… |
| 08:04 | <Hixie> | nessy: i don't, fwiw :-) i think we should offer a small set of functionality in v1, and see what people feel they are truly lacking. baby steps and all. |
| 08:04 | <nessy> | that's fair enough |
| 08:04 | <boblet> | ruby associates rt with preceding non-rt ruby base text, and CSS3 Ruby doesn’t have this ruby-position, so currently impossible |
| 08:04 | <nessy> | we don't have to do ticker text in the first version |
| 08:04 | <Hixie> | i'm trying to find a web page that has tickers and video |
| 08:04 | <nessy> | I can, however, see a need |
| 08:04 | <Hixie> | but my network is acting poorly |
| 08:05 | <annevk> | boblet, CSS3 Ruby is somewhat obsolete as it is though and will be updated with HTML5 in mind |
| 08:05 | <nessy> | I found it difficult to find a video on the Web with ticker text actually, which is why that example came up |
| 08:05 | <nessy> | but let me look a bit more |
| 08:05 | <Hixie> | boblet: fwiw, "kanji/hanzi", "hangul", and "kanji" are all greek to me. :-) |
| 08:05 | <Hixie> | boblet: what does it mean in terms of html? |
| 08:06 | <boblet> | annevk: yeah, I read some of www-style on it. Interesting that the Korean usage is outside of CSS3 Ruby tho |
| 08:06 | <nessy> | for the craziest use of text with video ever got to http://www.masternewmedia.org/video_internet_television/video-annotation/online-video-annotation-takes-a-giant-leap-20070405.htm |
| 08:06 | <boblet> | Hixie: :) |
| 08:06 | <nessy> | I can find an example of almost anything you can think of on that page |
| 08:06 | <annevk> | boblet, if it's inline you don't need ruby markup though |
| 08:07 | <annevk> | boblet, you can just use normal typography :) |
| 08:07 | <annevk> | "normal" to be i18n-correct |
| 08:07 | <boblet> | Hixie: normally <ruby>base text <rt>ruby text</ rt></ruby> to give ruby text above in supporting browsers or after in non-supporting browsers (default) |
| 08:07 | <boblet> | annevk: but that wouldn’t be *semantic* :P hehe |
| 08:07 | <othermaciej> | kanji and hanzi are the japanese and chinese names respecitvely for chinese ideographs |
| 08:08 | <othermaciej> | hangul is the korean alphabet |
| 08:08 | <annevk> | boblet, semantics are overrated |
| 08:08 | <boblet> | Hixie: current Korean usage is ruby text inline *before* base text, but old CSS3 ruby-position has every side but before |
| 08:08 | <boblet> | annevk: not to pedants |
| 08:09 | <annevk> | boblet, sure, but they're wrong |
| 08:10 | <annevk> | boblet, most of HTML is focused on pragmatism, not pure semantics |
| 08:10 | <boblet> | Hixie: also because no <rb> anymore source order is fixed it’s not possible to rearrange in source |
| 08:10 | <nessy> | ok, I found the Roubini example - I think what I thought was ticker text is actually an unrelated title div |
| 08:11 | <boblet> | (note that I’m not suggesting this is important enough to warrant <rb> as Korean is moving away from using anything that needs ruby) |
| 08:11 | <nessy> | Hixie: I don't mind about dropping ticker text from time-aligned text types and waiting till the market catches up |
| 08:11 | <boblet> | annevk: heh |
| 08:11 | <Hixie> | boblet: wait so koreans write <ruby><rt>ruby text</rt>base text</ruby> and it works in IE? |
| 08:11 | <Hixie> | nessy: k |
| 08:12 | <Hixie> | nessy: i'll remove that section for now |
| 08:12 | <nessy> | Hixie: after all, I waited for 10 years for the Web to catch up with my ideas of how hyperlinked video and captions should be done ;) |
| 08:12 | <nessy> | ok |
| 08:12 | <zcorpan> | http://dvcs.w3.org/hg/html/raw-file/tip/tests/submission/Microsoft/interface/anchor_href.htm - the framework seems to need unnecessary boilerplate and has the arguments backwards in assertEquals (compared to how i'd order them) |
| 08:12 | <boblet> | Hixie: I have yet to see a code example or make a test case |
| 08:13 | nessy | thinks there is a business idea in video ticker text offerings… ;) |
| 08:13 | <Hixie> | boblet: then i don't understand what you mean by "ruby text inline *before* base text" |
| 08:13 | <boblet> | Hixie: based on display of an image I was sent demonstrating current usage |
| 08:14 | <othermaciej> | zcorpan: I like the WebKit framework for script tests a *lot* better |
| 08:14 | <Hixie> | boblet: can you upload it so i can see? |
| 08:14 | <othermaciej> | both the output and the code there are icky |
| 08:14 | <boblet> | wip ;-) |
| 08:14 | <annevk> | I think boblet is saying they write "blah (blah)" |
| 08:14 | <annevk> | and therefore don't need ruby at all, but since it's called ruby pedants want to use ruby markup? |
| 08:14 | <boblet> | but the (blah) is what the blah would be in Japanese |
| 08:15 | <nessy> | haha, I just found a better example for ticker text, but also from Yahoo, see http://media.rgemonitor.com/images/blogs/10082008tt.jpg |
| 08:15 | <boblet> | annevk: there are some benefits in having reading semantically marked up you know. surely you’ve seen a few kanji by now? :P |
| 08:16 | <nessy> | and this: http://www.freewaregenius.com/wp-content/uploads/2009/11/Free-Stocks-Ticker-Screenshot.jpg |
| 08:16 | <nessy> | anyway - can all be done with javascript of course, so I don't really mind |
| 08:16 | <boblet> | Hixie: http://oli-studio.com/temp/ruby-korean.jpg |
| 08:16 | <Hixie> | nessy: the second one doesn't even have a video! |
| 08:17 | <nessy> | ah, you're right! |
| 08:17 | <Hixie> | boblet: ah well i'm happy to say that you can do that in text/plain |
| 08:17 | <Hixie> | nessy: i think that pretty much guarantees that i'm right that it's not linked to the video :-P |
| 08:17 | <boblet> | top line shows Japanese-style ruby, with ruby text above base text |
| 08:17 | <nessy> | oh, there are plenty of tickers that go without a video, that's for sure |
| 08:18 | <Philip`> | othermaciej: I think I got something slightly over 400 bug emails - was that the total number of bugs that you were changing (including the 235)? |
| 08:18 | <othermaciej> | Philip`: 493 |
| 08:18 | <nessy> | but if no callbacks are required and all that is required is to start the thing on video playback and to stop it on video pause/end, then indeed the time-alignment is minimal |
| 08:19 | <boblet> | Hixie: true, but as mentioned there are linguistic benefits of indicating the phonetics of something. also if all we’re worried about with ruby is placement above it’s far more a presentational element than I thought :) |
| 08:19 | <Hixie> | nessy: i'm not even convinced there's that -- i |
| 08:19 | <Hixie> | nessy: ...i've yet to see a video with a ticker that isn't burnt in but that pauses when you pause the video |
| 08:19 | <nessy> | yeah, I saw :) |
| 08:19 | <annevk> | boblet, ruby markup makes sense for the top line in http://oli-studio.com/temp/ruby-korean.jpg but is not needed at all for the other lines |
| 08:19 | <nessy> | on the Web the tickers aren't bunt in, but they also may not stop, as you are saying, so same effect |
| 08:20 | <annevk> | boblet, I fully support ruby markup for examples similar to the top line (as used in Japan) |
| 08:20 | <MikeSmith> | boblet: that Korean markup case you mentioned is interesting (the logogram in parens after the phonetic reading, instead of the other way around) |
| 08:20 | <boblet> | hey Mike |
| 08:20 | <Hixie> | boblet: if all you want to do is render <ruby>bbb<rt>aaa</rt></ruby> as aaabbb then the answer is to use CSS (not ruby CSS, just regular CSS) |
| 08:21 | <Hixie> | boblet: though CSS doesn't yet support rendering things out of order like that, it's something the wg intends to provide eventually |
| 08:21 | <Hixie> | nessy: well if they don't stop with the video, they're definitely not timed tracks |
| 08:21 | <Hixie> | anyway |
| 08:21 | <nessy> | yeah, fair enough |
| 08:23 | <boblet> | Hixie: based on CSS3 Ruby module it seemed like that presentation slipped through the cracks. I’ll try to email about it tho, so hopefully it can be included in CSS3 Ruby mark 2 |
| 08:24 | <Hixie> | boblet: it shouldn't be in the CSS ruby module |
| 08:24 | <Hixie> | boblet: it's not ruby presentation |
| 08:24 | <Hixie> | boblet: it's just regular inline rendering but with out of order flow |
| 08:25 | <boblet> | Hixie: then ruby-position: after also shouldn’t be there huh. ok that makes sense |
| 08:25 | <boblet> | which module would that end up in? |
| 08:25 | <Hixie> | what does 'ruby-position: after' do? |
| 08:25 | <Hixie> | no idea what module it should be in... TabAtkins or annevk might know |
| 08:25 | <boblet> | Hixie: according to that spec same rendering as non-supporting browser |
| 08:26 | <Hixie> | that seems like a waste of time then, they could have just used display:inline :-) |
| 08:26 | <boblet> | Hixie: so you read my article after all ;) |
| 08:26 | <boblet> | j/k, although I mentioned it in there |
| 08:43 | <boblet> | MikeSmith: thanks for the feedback |
| 09:05 | <jgraham> | Random, probably rather obvious, point; for sufficiently simple captions it's not clear to me that I want them overlayed on the video in the same way that they are on the TV. For fullscreen it makes sense but if I am watching something that only takes up 50% of my screen height I might want the subtitles to sit under the video |
| 09:06 | <jgraham> | zcorpan: It is not clear to me that there is a "natural" order for the arguments of assertEquals |
| 09:08 | <Philip`> | assertEquals(1+1, 2); |
| 09:08 | <Philip`> | assertEquals(2, 1+1); |
| 09:08 | <Philip`> | The first sounds more natural to me |
| 09:08 | <zcorpan> | jgraham: i often omit message, so from that point of view at least message should be last |
| 09:09 | <jgraham> | zcorpan: I tend to agree but I guess the idea here is that the message is mandatory |
| 09:09 | <othermaciej> | Philip`: agree, (test, expectation) seems like the natural order |
| 09:09 | <jgraham> | I'm not sure I agree |
| 09:10 | <jgraham> | If you wrote it like assert(1+1 == 2) I would agree |
| 09:11 | <jgraham> | But assert equals 2, 1+1 also makes sense |
| 09:11 | <jgraham> | Not typical english word order but logical |
| 09:12 | <gsnedders> | Balloons exploading are kinda loud. |
| 09:12 | <zcorpan> | gsnedders: happy birthday |
| 09:12 | <gsnedders> | zcorpan: thanks. |
| 09:14 | <jgraham> | zcorpan: In any case it's not clear to me that having lots of debates about the precise best format for testsuites is a good use of time. If microsoft are happy to write tests in this format, that is fine |
| 09:14 | <Philip`> | They should make vacuum-filled balloons |
| 09:14 | <Philip`> | Vacuum is much lighter than helium or even hydrogen, and the balloons could never explode |
| 09:15 | <jgraham> | Although I can't say that I would be inclined to use it myself |
| 09:15 | <jgraham> | (the only other option seems to be adopting some preferred framework in which case it is worthwhile to debate all the specifics) |
| 09:16 | <jgraham> | Which I guess is a point worth rasing on the list |
| 09:16 | <zcorpan> | jgraham: sure |
| 09:19 | <othermaciej> | jgraham: imagine if you also had assertFalse(), assertTrue(), assertThrowsException() |
| 09:19 | <othermaciej> | in each case you would want the test expression as the first parameter I think |
| 09:19 | <othermaciej> | so for consistency, assertEquals should also have it first, and the expectation second |
| 09:19 | <othermaciej> | my 2c anyway |
| 09:21 | <jgraham> | othermaciej: assertThrows should take some object to indicate the expected type and properties of the exception |
| 09:21 | <jgraham> | So that isn't an obvious counter-example |
| 09:30 | <othermaciej> | jgraham: it also needs to take a string to eval instead of an expression to do anything useful |
| 09:32 | <Philip`> | Isn't it better to pass a function(){} to call than a string to eval? |
| 09:32 | <othermaciej> | this is what WebKit's script tests look like: http://trac.webkit.org/export/57879/trunk/LayoutTests/fast/dom/HTMLAnchorElement/set-href-attribute-hash.html |
| 09:32 | <jgraham> | othermaciej: I would pass a function |
| 09:33 | <othermaciej> | and view the linked script-tests/set-href-attribute-hash.js to see what the actual test scripts look like |
| 09:33 | jgraham | has written an assertThrows with the signature assertThrows(expectedException, callable, message) |
| 09:33 | <othermaciej> | I think both the output and the source are more readable than Microsoft's format |
| 09:34 | <othermaciej> | (or whoever came up with that) |
| 09:34 | <othermaciej> | we use strings for both the test expression and the expectation |
| 09:34 | <othermaciej> | that way shouldBe (our version of assertEquals) can catch exceptions |
| 09:34 | <othermaciej> | so it can report unexpected throw instead of making the rest of the test bomb out |
| 09:35 | <othermaciej> | otherwise you have to either surround every test assertion with try/catch, or pass everything as closures |
| 09:35 | <jgraham> | I don't like using strings; making everything rely on eval can have surprising results |
| 09:36 | <othermaciej> | in practice it hasn't caused a problem |
| 09:36 | <othermaciej> | other than people getting confused about the need to double-quote actual strings |
| 09:37 | <othermaciej> | anyway, strings are not the main point |
| 09:37 | <othermaciej> | my point was (a) readable output; (b) readable test script |
| 09:37 | <jgraham> | I think quoting hell plus the lack of syntax highlighting for the test code is enough to put me off |
| 09:37 | <othermaciej> | could easily convert the system to use all closures |
| 09:39 | <othermaciej> | here is one with examples of more of the types of assertions: http://trac.webkit.org/export/57879/trunk/LayoutTests/fast/js/regexp-compile.html |
| 09:39 | <jgraham> | We have a whole bunch of javascript tests that use assertEquals(expected, actual, [message]) type syntax, use try/catch on the backend to handle failures and uses closures for cases where the test is expected to throw |
| 09:39 | <jgraham> | It has worked pretty well |
| 09:39 | <othermaciej> | and the corresponding JS source: http://trac.webkit.org/export/57879/trunk/LayoutTests/fast/js/script-tests/regexp-compile.js |
| 09:40 | <jgraham> | (nothing that I can link to sadly) |
| 09:41 | <othermaciej> | anyway, say what you will about the details of the assert function parameters, I think our output is the most useful I have seen for script tests |
| 09:41 | <othermaciej> | the color makes it easy to spot failures at a glance, it copies to plain text well, it doesn't have pointless tables, etc |
| 09:41 | <jgraham> | However, I think the main difficulty with DOM tests (rather than pure ECMAScript tests) is dealing with async results\ |
| 09:42 | <jgraham> | e.g. if you want to test that some event is fired when it ought to be |
| 09:42 | <othermaciej> | another reason to use strings is to include the test expression in the test output |
| 09:42 | <othermaciej> | so you don't have to say what ou tested a second time |
| 09:42 | <othermaciej> | that's definitely a problem with *some* DOM tests |
| 09:43 | <othermaciej> | not with the attributes of the <a> element though |
| 09:43 | <jgraham> | Well no |
| 09:43 | <jgraham> | But in general, if you want a framework that deals well with testing DOM, it has to test async stuff well |
| 09:43 | <jgraham> | That is the problem that I don't feel I have ever found a good solution to |
| 09:44 | <othermaciej> | we use ad hoc methods |
| 09:44 | <jgraham> | (A good general solution rather than just making it up for each specific case) |
| 09:44 | <othermaciej> | the most annoying is if you want to test that some async event fires eventually, and one possible bug is that it never fires |
| 09:44 | <othermaciej> | then you kinda have to add an arbitrary timeout, unless you know of a later event that will likely fire even in the bug case |
| 09:44 | <jgraham> | Yeah, but that's a pretty typical case |
| 09:45 | <gsnedders> | othermaciej: But the problem with outputting the JS source is that doing it as the WebKit tests do relying upon eval seems bad |
| 09:45 | <othermaciej> | we end up just relying on the fact that our test framework times out eventually |
| 09:45 | <jgraham> | So it sucks if you have to reinvent the wheel every single time |
| 09:46 | <othermaciej> | gsnedders: I hate eval as much as the next guy, but the string thing works really well in this case - too bad JS doesn't have macros |
| 09:47 | <jgraham> | othermaciej: If you write tests with one test per function, you can toString the function to get the text of the test |
| 09:47 | <zcorpan> | i now have a default timeout of 2s (many tests depend on network so need quite a long timeout) and require an explicit end() when the test is done in my framework |
| 09:47 | <jgraham> | it's not quite as simple |
| 09:47 | <zcorpan> | it works pretty well for async tests |
| 09:48 | <gsnedders> | othermaciej: Until you come across cases where you have bugs and eval has bugs the normal codepath doesn't (or vice-versa) |
| 09:48 | Philip` | does assertEqual(1+1, 2, "1+1", "2"); in order to be able to output the JS source without relying on eval |
| 09:48 | <othermaciej> | jgraham: true, and then you are just trading off the oddities of eval and the awkwardness of quoting for the oddities of code in a function and the awkwardness of saying function() { } all the time |
| 09:48 | <othermaciej> | Philip`: I think repeating yourself is a worse programming sin than eval |
| 09:48 | <Philip`> | ...with an ugly custom preprocessor language that converts "@assert 1+1 == 2;" into that syntax automatically |
| 09:48 | <Philip`> | so that I don't have to repeat anything :-) |
| 09:49 | <othermaciej> | using a preprocessor is cheating :-) |
| 09:49 | <gsnedders> | othermaciej: But functions work better in the case of JITing compilers that compile whole functions at once, as then you can just loop the function |
| 09:49 | <jgraham> | It's true that tests-in-a-function can have problems if you want to test something that specifically depends on global scope |
| 09:49 | zcorpan | does assertEquals(1+1, 2, "1+1") or just assertEquals(1+1, 2) |
| 09:49 | othermaciej | still wonders what y'all think of the output format, as opposed to the use of eval |
| 09:50 | <jgraham> | These problems are perhaps more pronounced if you are trying to test the ECMAScript engine itself rather than the DOM |
| 09:50 | <gsnedders> | The output format is I don't think an issue, and is really just a bikeshed :) |
| 09:51 | <othermaciej> | http://dvcs.w3.org/hg/html/raw-file/tip/tests/submission/Microsoft/interface/anchor_href.htm is really ugly |
| 09:51 | <jgraham> | othermaciej: It's pretty nice. Not quite to die for though ;) |
| 09:51 | <othermaciej> | and looks like it would be hard to spot failures |
| 09:51 | <othermaciej> | tables with double borders on everything, srsly |
| 09:51 | <jgraham> | Yes, I would turn off the table borders and use some colour there |
| 09:51 | <othermaciej> | so 1998 |
| 09:52 | <othermaciej> | I don't get why it's a table at all |
| 09:52 | <othermaciej> | I also can't tell what it's actually testing from reading |
| 09:52 | zcorpan | 's output is a <pre> which says "PASS" if it passed, or "FAIL" followed by the failing messages if it failed |
| 09:52 | <othermaciej> | since the descriptions are arbitrary strings and don't show the code it ran |
| 09:53 | <othermaciej> | readable test output makes it much easier for implementors to look at the test and fix their implementations accordingly |
| 09:53 | <othermaciej> | which is why I think it's important, more so than the details of how test assertions are coded |
| 10:02 | <jgraham> | othermaciej: I need to read the source code in either case to be confident I know what is going on |
| 10:03 | <othermaciej> | which is why it helps for the source to be readable too, which is why I dislike the giant try/catch blocks, and the fact that it's hard to map the table output to the test assertions in the source |
| 10:03 | <jgraham> | Giant try/catch blocks are unnecessary |
| 10:04 | <jgraham> | You should be able to do function testAddition() {assertEquals(2, 1+1)} |
| 10:05 | <jgraham> | Then you only need to dive in deeper if you broke the harness itself not the test |
| 10:05 | <hsivonen> | Does IE9 mode have a default border on linked images? |
| 10:05 | <jgraham> | (and you should make the harness simple to avoid that) |
| 10:06 | <jgraham> | and you can use e.g. the <details> element in the output to show the full test source |
| 10:06 | <zcorpan> | jgraham: hmm, <details> seems to mean the framework is getting too complex |
| 10:07 | <jgraham> | zcorpan: In the output? |
| 10:07 | <jgraham> | I don't see why |
| 10:07 | <jgraham> | Not all tests can be a single line of code |
| 10:08 | <othermaciej> | test assertions should generally be one line, though I could imagine multiple additional lines of setup code |
| 10:08 | <jgraham> | Yes, that's what I mean |
| 10:09 | <zcorpan> | well, would you want the <details> to work in existing browsers? |
| 10:09 | <jgraham> | Since the setup code is just as important as the assertion it makes sense to display it |
| 10:09 | <zcorpan> | you can view source |
| 10:10 | <jgraham> | zcorpan: Of course you can view source. othermaciej seems to be arguing that there is value in not having to view source for each fail |
| 10:10 | <jgraham> | Making details work somewhat in existing browsers is not the world's greatest challenge |
| 10:12 | <othermaciej> | jgraham: I don't mind if view source is required sometimes, what I find nice is if the fail message itself (a) often gives useful info in itself about what went wrong and (b) is easy to match up with the test assertion code in the source |
| 10:12 | <othermaciej> | I guess having those messages satisfies (b) but not (a), assuming the descriptions are unique |
| 10:55 | <Lachy_> | Does anyone in here know what should happen in the following window.stop() test? There's not much interop in this case between browsers |
| 10:55 | <Lachy_> | http://html5.lachy.id.au/output?data=%3C%21DOCTYPE+html%3E%0D%0A%3Cscript%3E%0D%0Adocument.write%28%22A%22%29%3B%0D%0Awindow.stop%28%29%3B%0D%0Adocument.write%28%22B%22%29%0D%0A%3C%2Fscript%3E&type=text%2Fhtml%3B+charset%3DUTF-8 |
| 11:11 | <MikeSmith> | boblet: about your ruby article, one suggestion to maybe consider is adding some images of ruby annotations in non-Web contexts |
| 11:12 | <MikeSmith> | I was pointing out some to anne this morning on the train and I think he took some pictures |
| 11:12 | <MikeSmith> | but anyway, I think for people outside of Japan or Asia, it might help to show them some examples of real-world use of ruby |
| 11:15 | <gsnedders> | Apparently I'm meant to be responsible now. That seems bad. |
| 11:17 | <Lachy> | gsnedders, you're allowed to be completely irresponsible on your birthday. |
| 11:17 | <Lachy> | as long as there is alcohol involved |
| 11:19 | <hsivonen> | gsnedders: did you reach the legal drinking age in Sweden? |
| 11:19 | <jgraham> | This could be messy... |
| 11:20 | <Lachy> | if Sweden is anything like Norway, then I think at 18, you're allowed to drink low percentage alcohol, like beer |
| 11:20 | <Lachy> | and I guess wine too |
| 11:20 | <jgraham> | I don't think Sweden is like Norway |
| 11:21 | <jgraham> | He can't *buy* strong alcohol |
| 11:21 | <Lachy> | oh, then that's good. He'll be allowed to drink any alcohol. Anyone got a bottle of Austrian Rum for him? |
| 11:21 | <gsnedders> | I think, from what people were saying at lunch, I'm allowed to drink alcohol up to 3.5% |
| 11:22 | <gsnedders> | Lachy: If they got it with the intention of giving it to me, they've broken the law. |
| 11:22 | <jgraham> | gsnedders: Pointer? Wikipedia at least doesn't obviously agree with that |
| 11:23 | <MikeSmith> | gsnedders: the good part is that if we engage in felony crimes and get caught, you can go to the same prison with me .. instead of you being charged as a minor and going to junior-criminal jail |
| 11:23 | <MikeSmith> | you get to go to the real prison |
| 11:23 | <gsnedders> | MikeSmith: I could from 15 |
| 11:24 | <MikeSmith> | well, yippee-aye-kai-yai-yay |
| 11:24 | <gsnedders> | Oh, so I can drink alcohol legally. I can get the light beer sold in supermarkets. |
| 11:24 | <Philip`> | gsnedders: If you hypnotise somebody into buying it and giving it to you, does that count as intent? |
| 11:25 | <Lachy> | gsnedders, has there been a party planned for you that you know about yet? |
| 11:25 | <gsnedders> | But I can't go to the state-owned shop and buy anything. |
| 11:25 | <gsnedders> | Philip`: IANAL |
| 11:25 | <jgraham> | gsnedders: Yes, that is my understanding |
| 11:25 | <gsnedders> | Lachy: yes |
| 11:25 | <MikeSmith> | gsnedders: if we you were already fully prepared for serious criminal work at 15, I don't know what we've been waiting for, then. Let's get out and do some damage |
| 11:25 | <gsnedders> | MikeSmith: fuck yeah. |
| 11:27 | <jgraham> | gsnedders: Also I would like to reconfirm that Fleetwood Mac are indeed dire |
| 11:27 | <MikeSmith> | gsnedders: you got the right "can do" spirit |
| 11:29 | <gsnedders> | jgraham: You're wrong. |
| 11:32 | <jgraham> | On the other hand, the new Jónsi album is, 1.5 tracks in, pure awesome |
| 11:32 | <boblet> | MikeSmith: good idea — will add a couple |
| 11:32 | <MikeSmith> | friends, can anybody tell me where the latest Notifications API draft spec is? |
| 11:33 | <MikeSmith> | this? -> http://www.chromium.org/developers/design-documents/desktop-notifications/api-specification |
| 11:33 | <MikeSmith> | and is Mozilla also implementing it? or is it just Chromium project that plans to at this point? |
| 11:34 | <MikeSmith> | boblet: you might be able to even find some examples of add with vertical text and ruby |
| 11:35 | <boblet> | I’d like to do a haiku like that, but need to get IE working so I can actually get a screenshot |
| 11:35 | <zcorpan> | iirc, it's legal to give alcohol to children, but not to sell |
| 11:35 | <boblet> | broke the VM image :/ |
| 11:35 | <MikeSmith> | I think you can find plenty of books and magazine or newspaper articles, but I think an ad with color and different font sizes and such would be more example-worthy |
| 11:35 | <zcorpan> | you're allowed to buy strong drinks in pubs but not in systembolaget |
| 11:35 | <zcorpan> | when you're 18 |
| 11:35 | <boblet> | MikeSmith: aah, you mean images? yeah good idea |
| 11:35 | <boblet> | bbl |
| 11:36 | <MikeSmith> | boblet, yeah, actually images from train ads or whatever |
| 11:36 | <MikeSmith> | zcorpan: what is systembolaget? like vinmonopolet? |
| 11:37 | <Philip`> | boblet: Could use something like browsershots.org when you want IE screenshots |
| 11:37 | <zcorpan> | MikeSmith: yeah |
| 11:38 | <jgraham> | MikeSmith: Yeah, it's the government run "if-we-don't=sell-all-the-alcohol-and-not-on-a-Sunday-or-after-3-on-a-Saturday-then-you-will-all-become-alchoholics-obviously store" |
| 11:38 | <MikeSmith> | heh |
| 11:39 | <zcorpan> | a few years ago it was all of saturday also |
| 11:40 | <MikeSmith> | liquor stores are surprisingly harder to break into than pharmacies |
| 11:41 | <jgraham> | Until a few months ago there was also a government monopoly on selling pharmacuticals |
| 11:41 | <jgraham> | Then someone noticed that the only other countries with the same restiction were Cuba and North Korea |
| 11:41 | <jgraham> | At that point I guess it became a bit embarassing |
| 11:42 | <MikeSmith> | the Axis of Evil: Cuba, North Korea, and Sweden |
| 11:43 | <hsivonen> | gsnedders: well, congratulations on aging to a point where statutory conditions flip |
| 11:53 | <MikeSmith> | me finds http://dev.w3.org/2006/webapi/WebNotifications/publish/ |
| 11:53 | <MikeSmith> | hsivonen: do you recall anybody from Mozilla posting about intentions to implement that? |
| 11:54 | zcorpan | does s/loadedfirstframe/loadeddata/ in https://developer.mozilla.org/En/Using_audio_and_video_in_Firefox but other things in there are probably also outdated |
| 11:55 | <zcorpan> | e.g. i don't get any canshowcurrentframe event |
| 11:55 | <MikeSmith> | I seem to remember recent discussion of the NOtifications api but can't recall where.. maybe the whatwg list? |
| 12:00 | <gsnedders> | hsivonen: thanks |
| 12:05 | <MikeSmith> | maybe I'm thinking about the notifications feature that gavin has been working on... |
| 12:05 | <MikeSmith> | which is unrelated, I think |
| 12:06 | <othermaciej> | MikeSmith: there has been discussion on webapps |
| 12:06 | <othermaciej> | dunno if it is of the same or a related API |
| 12:06 | <othermaciej> | I think it is related |
| 12:06 | <MikeSmith> | ok |
| 12:06 | <MikeSmith> | I'll go back and look there again |
| 12:07 | <MikeSmith> | ah, OK, I see now - http://lists.w3.org/Archives/Public/public-webapps/2010JanMar/thread.html#msg893 |
| 12:07 | <MikeSmith> | last month, not this month |
| 12:07 | <MikeSmith> | having so much fun I lose track of the time |
| 12:08 | <MikeSmith> | oh |
| 12:10 | othermaciej | takes a peek at http://damowmow.com/playground/htmlwg/chart.html and decides to cry himself to sleep |
| 12:10 | gsnedders | hugs othermaciej |
| 12:10 | <gsnedders> | You'll get over this tough time… |
| 12:11 | <othermaciej> | look at that trendline! |
| 12:14 | <jgraham> | othermaciej: I would be more worried about the fact that we are clearly due for a change of chair |
| 12:14 | <Lachy> | othermaciej, it's showing that the number of closed issues is steadily increasing, which is a good thing |
| 12:14 | <othermaciej> | jgraham: don't tempt me! |
| 12:14 | <jgraham> | That graph is predicting that you have a 1 in 3 chance of a nasty accident! |
| 12:15 | <othermaciej> | Lachy: that is a good thing - just need to get the number of closed issues decreasing |
| 12:15 | <othermaciej> | er |
| 12:15 | <othermaciej> | get the number of open issues decreasing |
| 12:16 | <othermaciej> | I think the closed issues have a much steeper slope under the current set of chairs than ever before, but it's kinda hard to see on this graph |
| 12:16 | <daedb> | So, what's the best place to get a free website (without ads) these days? Wordpress? |
| 12:17 | <Lachy> | at 2 out of the 3 chair changes so far, the graph did see a sudden rapid decline in open issues, so maybe that could be a good thing. |
| 12:38 | <nessy> | hsivonen: finally got time to read your proposal properly |
| 12:38 | <nessy> | I mostly like it ;) |
| 12:38 | <hsivonen> | nessy: nice |
| 12:38 | <hsivonen> | nessy: what's "mostly" about? |
| 12:38 | <nessy> | you'll see my reply |
| 12:39 | <hsivonen> | ok |
| 12:39 | <nessy> | I think we might need to deal with time-overlapping segments |
| 12:39 | <nessy> | but there may be a way to fit that onto the proposal |
| 12:39 | <nessy> | I'm not too worried about it for now |
| 12:40 | <nessy> | your proposal generally feels right |
| 12:40 | <nessy> | and I like the idea of calling events |
| 12:40 | <nessy> | anyway - gotta go - sorry |
| 12:41 | <Lachy> | which proposal are you referring to? |
| 12:43 | <hsivonen> | Lachy: http://lists.w3.org/Archives/Public/public-html/2010Apr/0365.html |
| 12:44 | <jgraham> | hsivonen: Your proposal explicitly prevents having the subtitles below the video (rather than overlaid) |
| 12:46 | <jgraham> | I know this is not common in non-web situations, but it unclear to me whether this is by design or just a limitattion of the typical medium |
| 12:47 | <MikeSmith> | http://www.w3.org/Submission/2010/SUBM-WOFF-20100408/ WOFF spec has been published as a W3C Member Submission |
| 12:48 | <hsivonen> | jgraham: to have them outside the video's CSS box, I think it's necessary to spec how to target the timed text to a non-anonymous iframe |
| 12:48 | <hsivonen> | jgraham: because establishing the rendering space for the captions outsize the <video> box using merely pseudos would suck |
| 12:49 | <hsivonen> | jgraham: so in that case, you'd need a concrete element to establish the box sanely |
| 12:49 | <jgraham> | Works for me as a suggestion |
| 13:11 | <nessy> | hsivonen: are you still there? |
| 13:14 | <nessy> | you mentioned SVG 1.2 Tiny textArea in the email to Sean - what is the status of that in HTML5? |
| 13:35 | <hsivonen> | nessy: http://lists.w3.org/Archives/Public/public-html/2009Sep/0005.html |
| 13:36 | <nessy> | thanks |
| 13:37 | <nessy> | Hixie: why did you remove all the other examples of time-aligned text from http://wiki.whatwg.org/wiki/Timed_tracks ? |
| 13:41 | <nessy> | ah I just found http://wiki.whatwg.org/wiki/Use_cases_for_API-level_access_to_timed_tracks |
| 13:42 | <nessy> | but I think the lyrics part should go back into the other page |
| 13:47 | <hsivonen> | I guess I should find an example of italics in otherwise plaintext subtitling |
| 13:48 | <hsivonen> | too bad the method for finding such an example is watching a lot of tv |
| 14:00 | <nessy> | hsivonen: yeah, italics, bold and colors |
| 14:02 | <hsivonen> | I don't remember seeing bold or colors on tv |
| 14:03 | <jcranmer> | karoake subbing for foreign theme songs ? |
| 14:03 | <jcranmer> | that uses multiple colors... |
| 14:04 | <zcorpan> | i see colors occasionally (e.g. indicating different speakers) |
| 14:04 | <zcorpan> | but that's a whole line, not just a word |
| 14:06 | <jcranmer> | hmm, umineko used multple colors for some lines because the color of text made a difference |
| 14:09 | <Philip`> | zcorpan: http://philip.html5.org/misc/subtitles-3.xml has differently coloured speakers on a single line |
| 14:09 | <Philip`> | <p begin="00:15:38.48" id="p205" end="00:15:41.76"><span tts:color="yellow">That's not very good looking. </span>You're<br />right. <span tts:color="cyan">It isn't, is it?</span></p> |
| 14:09 | <Philip`> | (Default text colour is white) |
| 16:31 | <TabAtkins> | Who's nessy? Silvia? |
| 16:42 | <zcorpan> | nessy said she would reply to hsivonen's email, and i see a reply to his email from Silvia |
| 16:42 | <TabAtkins> | That's what I'm basing my guess on. |
| 16:43 | TabAtkins | doesn't like it when people have email and irc names that are significantly different. |
| 16:43 | zcorpan | takes note but won't change his nick :) |
| 16:44 | <TabAtkins> | That's fine. Just change your email instead. |
| 16:44 | <zcorpan> | i had zcorpan in my email address before |
| 16:44 | <zcorpan> | but now i have to use simonp⊙oc |
| 16:44 | <TabAtkins> | You can adjust your display name to show zcorpan. |
| 16:45 | <zcorpan> | naw |
| 16:45 | <zcorpan> | whois or google will reveal my name |
| 16:46 | <TabAtkins> | Now you're just being difficult. |
| 16:54 | <gsnedders> | TabAtkins: pingy? |
| 16:55 | <TabAtkins> | gsnedders: No ping. Going to breakfast. Will reping when I'm back. |
| 16:55 | <gsnedders> | TabAtkins: I'm going out to supper so no ;P |
| 16:55 | <TabAtkins> | Dammit, ok. |
| 16:55 | <TabAtkins> | What's up. |
| 16:56 | <gsnedders> | No time now ;P |
| 16:56 | <TabAtkins> | Nuuuuu! |
| 16:57 | <gsnedders> | No, ni! |
| 16:57 | <zcorpan> | jgraham: i haven't had much need for deep equality checking |
| 16:57 | <TabAtkins> | http://www.harkavagrant.com/index.php?id=253 <-- Most adorable squid ever. |
| 16:58 | <jgraham> | zcorpan: I have needed it a few times |
| 16:58 | <jgraham> | One possible implementation strategy (that I think jQuery/qUnit uses) is just to implement assertObjectEquals |
| 16:59 | <jgraham> | that checks deeply, and use it for lots of different types of object |
| 16:59 | <jgraham> | (instead of also having e.g. assertArrayEquals, etc.) |
| 16:59 | <TabAtkins> | So, gsnedders, for real should I go to breakfast or talk? |
| 16:59 | <gsnedders> | But you basically can't do anything with unknown properties or check things that aren't enumerable |
| 16:59 | <gsnedders> | TabAtkins: Breakfast |
| 17:00 | <jgraham> | TabAtkins: Talk |
| 17:00 | <Philip`> | function deeplyEqual() { return uneval(a) === uneval(b); } |
| 17:00 | <jgraham> | function isEvil(user) {return user.irc_nick === "Philip`"} |
| 17:01 | <gsnedders> | or if it is Philip on other servers |
| 17:01 | <gsnedders> | Philip`: And what is the defined behaviour of uneval? |
| 17:01 | <Philip`> | gsnedders: I define it to be what happens when I do it in Firefox |
| 17:02 | <gsnedders> | Philip`: What version of Firefox, what platform? |
| 17:02 | <Philip`> | (I do actually use uneval for equality checking tests, but only for non-browser-based SpiderMonkey-using code) |
| 17:02 | <zcorpan> | window unevals to {} |
| 17:02 | <Philip`> | gsnedders: 1.5, BeOS |
| 17:03 | <zcorpan> | sorry |
| 17:03 | <zcorpan> | ({}) |
| 17:03 | <jgraham> | I don't think Firefox 1.5 ever ran on BeOS |
| 17:03 | <jgraham> | did it? |
| 17:03 | <Philip`> | http://www.bebits.com/appver/3143 |
| 17:04 | <Philip`> | Looks like it did |
| 17:04 | <jgraham> | Yeah, I was just looking at that |
| 17:05 | jgraham | would like to play with Haiku sometime |
| 17:22 | <AryehGregor> | Hixie, is there some particular reason that uploads are disabled on the wiki, or should I go ahead and enable them? |
| 17:22 | <AryehGregor> | (We should really support uploads-in-database, if only so they can be enabled by default . . .) |
| 17:24 | <AryehGregor> | I hypothesize that video games have more markup in subtitles because they're really targeted at everyone, not just the hard-of-hearing. In many games, not all lines are even voice-acted, so subtitles have to be on. In other cases the speaker might be far away or drowned out by loud noises, since dialog isn't always fully scripted. |
| 17:26 | <AryehGregor> | Also, they're always from scripts, so you can copy italics from the scripts if applicable. In unscripted videos, or videos that follow a loose script, there's no transcript to tell you definitively whether something should be italicized or bolded or such. |
| 17:27 | <AryehGregor> | I don't think 3D video games are good examples to base HTML5 video stuff on, except if it's an actual video cutscene rather than an in-game dialog (and actual video cutscenes are rare these days in games). |
| 17:27 | Philip` | notes that Portal (and presumably HL2 too?) has both subtitles (for everyone) and closed captions (for hard-of-hearing) |
| 17:28 | <Philip`> | (where the latter includes textual alternatives to sound effects) |
| 17:28 | <AryehGregor> | Interesting, I didn't know that. |
| 17:28 | <AryehGregor> | Most games don't bother. |
| 17:28 | <Philip`> | I don't remember seeing that option on any others |
| 17:28 | <Philip`> | though I've never really looked for that option on any others |
| 18:11 | <TabAtkins> | Argh I hate bitly and url shorteners in general! |
| 18:15 | <Philip`> | http://bit.ly/9Cm8kU |
| 18:15 | <TabAtkins> | Thanks for that, Philip`. Big help. |
| 18:16 | <Philip`> | Glad to be of assistance |
| 18:18 | <Philip`> | Aren't they doing the world a service by preventing web developers from having to waste time coming up with concise readable URL structures, given that their URLs are all going to get obscured anyhow? |
| 18:18 | <TabAtkins> | Concise readable url structures are all we have! |
| 18:19 | <TabAtkins> | On an unrelated note, are table cells automatically BFCs? |
| 18:19 | TabAtkins | supposes he could test this himself. |
| 18:20 | <TabAtkins> | Yes, they are. |
| 20:10 | <Hixie> | AryehGregor: uploads were disabled so that we wouldn't have to manage getting rid of abusive uploads |
| 20:32 | <KaOSoFt> | Good morning. |
| 20:32 | <TabAtkins> | Holy crap. http://craftymind.com/factory/html5video/CanvasVideo.html |
| 20:33 | <KaOSoFt> | TabAtkins- :o |
| 20:34 | <KaOSoFt> | That looks totally cool. |
| 20:38 | <TabAtkins> | You can do so many awesome/ridiculous things when you aren't having to muck around in plugin land. |
| 20:44 | <paul_irish> | TabAtkins: in his notes he says ... "One feature I didn’t have time to figure out was adding depth to the explosion, so pieces that are closest to ground zero fly up into the air as they sail outward. With full shadow effects this could look really cool." |
| 20:44 | <paul_irish> | hopefully he adds that as well |
| 20:44 | <TabAtkins> | Yus. |
| 21:25 | <danbri> | browser geeks, may I ask for some help w/ rdfa idioms? |
| 21:25 | <danbri> | in rdfa/xhtml it is ok to put <object><span ...rdfablahblah></span></object> in the <head> of a doc... |
| 21:26 | <danbri> | ...if such a block of markup found its way out of the xhtml universe and into html5 or pre-html5 html doc header, ... does anyone know of ugly side-effects of putting object/span there? |
| 21:27 | <Philip`> | danbri: The <object> would force an implicit </head><body> in most (all?) browsers |
| 21:31 | <Philip`> | (It seems pretty bogus that HTML4 (hence XHTML) allows <object> in <head>, since it doesn't work like that) |
| 21:34 | <danbri> | so if this is done, it should be the last section of <head> at least? |
| 21:35 | <jgraham> | It can't end up in head at all in HTML5/real text/html browsers |
| 21:35 | <danbri> | well this design is for the xhtml flavour, but it would be nice if it was non-destructive if it leaked elsewhere |
| 21:35 | <danbri> | what would the practical consequences be? |
| 21:36 | <danbri> | (hopefully the html/rdfa wg will come up with some microdata/rdfa hybrid in future and i won't have to be asking such things next year...) |
| 21:37 | <jgraham> | danbri: Play around with http://livedom.validator.nu/?%3C!DOCTYPE%20html%3E%0A%3Chead%3E%0A%3Cobject%3E%3C%2Fobject%3E |
| 21:38 | <jgraham> | Basically the practical consequences wouldbe that anything that expected the object to end up in head would break and there would be a high chance of collateral damage from things moving out of head that shouldn't |
| 21:39 | <danbri> | so only rdfa parsers would be looking for it in head... |
| 21:39 | <Philip`> | The practical consequences would be that HTML5 parser developers would curse you for exploiting misfeatures in the HTML4 DTD |
| 21:39 | <danbri> | ...and if the <object> block was the last thing in </head> then nothing else would be bumped into body... |
| 21:40 | <Philip`> | and it would abort in pure streaming HTML5 parsers |
| 21:40 | <Philip`> | when it hits the explicit <body> later |
| 21:42 | <Dashiva> | A non-legacy timed text format that doesn't support ruby... heh |
| 21:44 | <Philip`> | danbri: <head><link etc><object></object><body foo> would (in non-streaming parsers) give the same DOM as <head><link etc></head><body foo><object></object> so that should indeed work |
| 21:48 | <danbri> | and spans inside the object would survive? page display and function generally ok in currently deployed browser? |
| 21:51 | <Philip`> | danbri: As far as I'm aware, yes |
| 21:51 | <Philip`> | although it looks like IE6 doesn't put the <object> element in the DOM at all |
| 21:52 | <Philip`> | and puts the <span> element as a child of the <head> |
| 21:52 | <Philip`> | Oh, and then it crashes |
| 21:52 | <Hixie> | note that putting <object> in <head> is non-conforming in XHTML5 and HTML5 |
| 21:56 | <danbri> | thanks guys |
| 21:56 | <danbri> | so what's this about 'would abort'? I thought html5 parsers were softer, gentler more accepting creatures... |
| 21:57 | <Philip`> | In slightly rare cases, like if you want SAX output without potentially-unlimited internal buffering in the parser, there are some errors you can't correctly recover from |
| 21:58 | <Philip`> | e.g. <body foo><body bar> (which should be equivalent to <body foo bar> but it's too late if you emit the body element after reading the first tag) |
| 21:59 | <Philip`> | which (unless I'm mistaken) is what happens when you put non-head content inside a <head> (so it emits an implicit body element) and then have a body tag later |
| 22:01 | <Hixie> | yeah, if you have a streaming parser (unusual for web browsers, but probably not unusual for rdfa processors) then that would likely cause an abort |
| 22:01 | <Hixie> | unless the streaming processor basically buffers everything |
| 22:01 | <Hixie> | but then it's not really streaming |
| 22:01 | <danbri> | ok well it seems the folk i'm helping are gonna stick with 'meta tags' for now |
| 22:02 | <danbri> | which gives some breathing time at least |
| 22:08 | <erlehmann> | the examples are pure awesome :3 http://wiki.whatwg.org/wiki/Timed_tracks |
| 22:12 | <TabAtkins_> | erlehmann: I just love the one with a nun yelling that someone should get an afro. |
| 22:13 | <erlehmann> | TabAtkins_, i love all of it because it provides a rare insight into what kind of cultural background the people who create the spec have. |
| 22:13 | <erlehmann> | something you cannot easily discern from other material |
| 22:16 | <Philip`> | Hmm, the BBC iPlayer has subtitles using TTML with colours, but the Flash player still only displays them all in white |
| 22:16 | <Philip`> | which is particularly confusing when there's multiple speakers on a single line, differentiated by colour |
| 22:18 | <Hixie> | erlehmann: most of them are the result of me searching for "subtitle" or "soustitre" on google images |
| 22:18 | <Hixie> | erlehmann: with various other keywords |
| 22:18 | <Hixie> | like "anime" |
| 22:20 | <erlehmann> | owww |
| 22:20 | <erlehmann> | and i thought |
| 22:21 | <erlehmann> | but meh, it may give away the origin of subtitling to the overlords using timed text 4357 A.D. |
| 22:53 | <TabAtkins_> | Looks like a few people on www-style have been hit by a virus. Interesting. |
| 23:00 | <daedb> | Hixie: Do you still want an example screenshot of subtitle with individual words emphasized? |
| 23:02 | <Dashiva> | http://dashiva.net/misc/1271498287684.jpg - Like this? :) |
| 23:04 | <daedb> | Yeah, sorta. |
| 23:05 | <daedb> | I took a screenshot from a game though, so not the same style as that one ^^ |
| 23:56 | <Traveler9> | hi |
