| 03:32 | <variable> | Has datagrid been removed from the WHATWG version of HTML5? |
| 03:36 | <variable> | or in general from HTML? |
| 03:43 | <miketaylr> | removed in general, yes |
| 04:14 | <MikeSmith> | was going to tell variable that for rationale on why that was dropped, he should look for message with implementor feedback from David Hyatt |
| 04:15 | <MikeSmith> | and maybe messages from last July or August from Adrian Bateman |
| 04:15 | <MikeSmith> | and Hixie did post a message at the time that was dropped, explaining the rationale for not including it at this time |
| 04:22 | <llrcombs> | any chanse that HTML5 <video> will get playlist support? |
| 04:22 | <roc> | playlist? |
| 04:24 | <llrcombs> | as in... |
| 04:24 | <llrcombs> | <video> |
| 04:24 | <llrcombs> | <playlist> |
| 04:24 | <llrcombs> | <item src="video1.mov"> |
| 04:24 | <llrcombs> | <item src="video2.mov"> |
| 04:24 | <llrcombs> | </playlist> |
| 04:24 | <llrcombs> | </video> |
| 04:24 | <llrcombs> | play those 2 <item>s sequentially |
| 04:25 | <doublec> | llrcombs, comment out in the WHATWG spec is the "proposed v2 (actually v3!) features:" |
| 04:25 | <llrcombs> | link? |
| 04:26 | <doublec> | llrcombs, and playlists is part of that: "playlists, with the ability to get metadata out of them (e.g. xspf)" |
| 04:26 | <llrcombs> | please? |
| 04:26 | <doublec> | llrcombs, http://www.whatwg.org/specs/web-apps/current-work/multipage/video.html |
| 04:26 | <llrcombs> | thx |
| 04:26 | <doublec> | llrcombs, nothing specced though |
| 04:26 | <doublec> | just a mention of it |
| 04:26 | llrcombs | can't find said mention |
| 04:27 | <doublec> | llrcombs, view source on that page |
| 04:27 | <doublec> | llrcombs, and search for 'playlist' |
| 04:27 | <llrcombs> | ahh, commented |
| 04:28 | <llrcombs> | ooh, other question |
| 04:28 | <llrcombs> | nvm, it's on that page :P |
| 04:28 | <doublec> | hehe |
| 04:29 | <llrcombs> | actually, yeah |
| 04:29 | <llrcombs> | what filetype should a <track type="captions"> be? |
| 04:32 | <doublec> | llrcombs, text/srt I think |
| 04:33 | <doublec> | llrcombs, see here http://www.whatwg.org/specs/web-apps/current-work/multipage/video.html#syntax-0 |
| 04:33 | <llrcombs> | is it supported by any browsers? |
| 04:33 | <doublec> | no |
| 04:33 | <llrcombs> | lol "zero or more" |
| 04:33 | <llrcombs> | meh |
| 04:33 | <doublec> | It's still under discussion |
| 04:33 | <llrcombs> | what about subtitles? |
| 04:33 | <doublec> | same |
| 04:33 | <llrcombs> | meh |
| 04:34 | <llrcombs> | no ETA, I'd expect? |
| 04:34 | <doublec> | most usage of subtitles and captions in html5 video uses javascript |
| 04:34 | <llrcombs> | as of NOW! |
| 04:34 | <llrcombs> | e.g. YouTube |
| 04:35 | <doublec> | llrcombs, see here for example: http://www.bluishcoder.co.nz/2008/12/srt-subtitles-with-html5-video.html |
| 04:35 | <llrcombs> | well, once it's supported in WebKit, I'll write a JS parser to convert YouTube captions to WebSRT's, and add a <track> automagically to my <video> with a data: URL to the autocreated SRT |
| 04:36 | <llrcombs> | (see http://combsconnections.tk/extensions/ ) |
| 04:36 | llrcombs | is impressed by his own download count |
| 08:33 | <kennyluck> | MikeSmithX? |
| 08:34 | <MikeSmithX> | yeah |
| 08:35 | <MikeSmith> | hey, it's bobchao ! |
| 08:35 | <MikeSmith> | hey Bob |
| 08:35 | <MikeSmith> | nice to meet you |
| 08:35 | <bobchao> | MikeSmith: Hi Mike! |
| 08:35 | <bobchao> | nice to meet you too :) |
| 08:35 | <kennyluck> | Yeah, I want to introduce you guys. |
| 08:35 | <kennyluck> | bobchao don't get onto IRC often, though. |
| 08:35 | <MikeSmith> | bobchao: I'm looking forward very much to visiting Taiwan next month |
| 08:35 | <bobchao> | MikeSmith: I'm really exciting that you can come to COSCUP :) |
| 08:36 | <MikeSmith> | kennyluck: IRC is the greatest thing since Google Wave! everybody should use it |
| 08:36 | <kennyluck> | bobchao works on random stuff including promoting Firefox |
| 08:36 | <MikeSmith> | ah great |
| 08:36 | <kennyluck> | He has a very nice website, but I am afraid it's in Chinese |
| 08:36 | <MikeSmith> | bobchao: most of the attendees at this event will be people familiar with open-source software? |
| 08:37 | <kennyluck> | http://gfx.tw/ |
| 08:37 | <MikeSmith> | kennyluck: I can run it through Google Translate or whatever... |
| 08:37 | <MikeSmith> | wow |
| 08:37 | <MikeSmith> | great domain name |
| 08:37 | <bobchao> | MikeSmith: most of them are familiar with FLOSS, if not all :) |
| 08:37 | <MikeSmith> | good good |
| 08:37 | <bobchao> | (But surely we still have some beginner in this area) |
| 08:37 | <kennyluck> | What is FLOSS? |
| 08:37 | <MikeSmith> | OK |
| 08:37 | <kennyluck> | (sorry) |
| 08:38 | <bobchao> | kennyluck: http://en.wikipedia.org/wiki/Free_and_open_source_software |
| 08:38 | <MikeSmith> | Free/Libre Open Source Software |
| 08:38 | <bobchao> | free/libre/open source software |
| 08:39 | <MikeSmith> | bobchao: nice site design |
| 08:39 | <kennyluck> | It allows you do drag-and-drop three of your favorate Firefox features. |
| 08:40 | <bobchao> | MikeSmith: thanks, we are still working on improve it :) |
| 08:40 | <kennyluck> | And produces a sticker that you can paste on your blog or whatever. |
| 08:40 | <kennyluck> | I hope it could be translated in English, but it seems that it has some layout issues. |
| 08:40 | <MikeSmith> | bobchao: do you know Gen Kanai? |
| 08:41 | <bobchao> | MikeSmith: yes, he's working for Mozilla Co. |
| 08:41 | <MikeSmith> | yep |
| 08:41 | <MikeSmith> | I know Gen, as well |
| 08:41 | <MikeSmith> | and a number of other folks at Mozilla Japan |
| 08:41 | <kennyluck> | IIRC, bobchao was once invited to Tokyo and gave a talk at a Moziila event. |
| 08:41 | <bobchao> | I know a few people in Mozilla Japan, too. :P |
| 08:42 | <bobchao> | yeah, last year |
| 08:42 | <MikeSmith> | bobchao: when did you visit? |
| 08:42 | <MikeSmith> | oh yeah |
| 08:42 | <MikeSmith> | I think I remember Gen blogging about that |
| 08:42 | <MikeSmith> | and photos |
| 08:42 | <bobchao> | to share what happened in Taiwan community |
| 08:42 | <MikeSmith> | excellent |
| 08:43 | <kennyluck> | Yeah, MikeSmith, he is also with me when I was talking to III. |
| 08:43 | <kennyluck> | I hope he can help us more. :) |
| 08:43 | <MikeSmith> | me too |
| 08:44 | <bobchao> | Guys I have to go, need to meeting with peers in office. |
| 08:44 | <MikeSmith> | I hope we can help bobchao promote the Open Web Platform in Taiwan |
| 08:44 | <MikeSmith> | bobchao: cheers |
| 08:44 | <MikeSmith> | thanks for dropping by |
| 08:44 | <bobchao> | chat to u guys later |
| 08:44 | <kennyluck> | Sure. See you then. |
| 08:44 | <bobchao> | MikeSmith: sure I would like to, love Open. see ya :) |
| 08:46 | <kennyluck> | Hmm... I have to say, it's very hard to drag people into the IRC world. |
| 08:46 | <kennyluck> | I never like any other IRC client beside Colloquy. |
| 08:49 | <MikeSmith> | No IRC, No Life. |
| 08:50 | <annevk> | hmm, someone from Google pointed out a bug in CORS when using it to prevent <canvas> tainting |
| 08:50 | <kennyluck> | I keep asking myslelf, If I were a Windows would I get onto IRC. My answer is no |
| 08:50 | <annevk> | for that to work <img> does the credentialed request thingie |
| 08:50 | <annevk> | so per CORS the server needs to reply with the allow-credentials and allow-origin header |
| 08:50 | <annevk> | again per CORS that means that allow-origin cannot be * |
| 08:51 | <annevk> | but per other rules <img> fetch requests do not include a suitable origin header value |
| 08:51 | <annevk> | I guess maybe that should change, but I thought we had a reason not to, or something |
| 09:03 | <MikeSmith> | annevk: this is maybe a lesson why it's a good idea to record the rationale for changes |
| 09:03 | <MikeSmith> | even it it's just through commit messages |
| 09:03 | <MikeSmith> | or comments in the HTML source for the spec |
| 09:04 | <annevk> | nothing really changed; either I forgot something or I just did not think it through fully |
| 09:04 | <annevk> | well, I think what I expected is that the way <img> fetching works today will change |
| 09:04 | <MikeSmith> | ? |
| 09:04 | <annevk> | but it is no longer clear to me it will |
| 09:04 | <MikeSmith> | why would it change? |
| 09:05 | <annevk> | to support CORS |
| 09:05 | <MikeSmith> | oh |
| 09:05 | <annevk> | I guess that is still a reasonable expectation |
| 09:06 | <annevk> | only if we do not want to do that we have to figure out something else |
| 09:06 | <MikeSmith> | no suggestions from whomever it as a Google who brought to bug up? |
| 09:07 | <annevk> | he just asked some questions |
| 09:08 | <MikeSmith> | I see |
| 09:10 | <MikeSmith> | so maybe a question to take to the public-webapps list |
| 09:11 | <annevk> | I suggested that, yes |
| 09:14 | <MikeSmith> | can you remind me where we are at with plans for moving CORS to LC? |
| 09:14 | <MikeSmith> | wasn't there some discussion of a date? |
| 09:14 | <MikeSmith> | target deadline? |
| 09:14 | <annevk> | we had a deadline for renaming headers |
| 09:15 | <annevk> | that passed, so unless earth gets a second moon that won't change anymore |
| 09:15 | <MikeSmith> | good |
| 09:15 | <MikeSmith> | yeah, I remember that now |
| 09:15 | <annevk> | someone volunteered for the security section, but he appears somewhat busy |
| 09:16 | <annevk> | and I'm in no rush |
| 09:16 | <MikeSmith> | so are that (security) and this current bug the only thing blocking us from starting an LC round? |
| 09:16 | <MikeSmith> | why no rush? |
| 09:17 | <annevk> | yeah, not even sure this is a bug |
| 09:17 | <MikeSmith> | ok |
| 09:17 | <MikeSmith> | would be nice to get this moved ahead to a point where you can be done with it for now and focus on other stuff |
| 09:17 | <MikeSmith> | *focus more on other stuff |
| 09:18 | <annevk> | yeah, I suppose |
| 09:18 | <annevk> | there's quite a few projects on my "wishlist" |
| 09:18 | <boblet> | anyone know of a microdata vocabulary curation site like http://prefix.cc/ for RDFa? |
| 09:18 | <annevk> | testing, DOM Core, encodings |
| 09:19 | <MikeSmith> | annevk: we can go to LC without the security section, mark lack of it as a Known Issue |
| 09:19 | <annevk> | I have some ideas on CSS too, but it's just too much work |
| 09:19 | <MikeSmith> | annevk: which could help to get somebody motivated to contribute content for the Security section |
| 09:20 | <annevk> | I'm not sure the lack of it will help us addressing LC comments |
| 09:20 | <annevk> | especially not when it comes to the anti-camp at Google |
| 09:20 | <MikeSmith> | annevk: you start working on CSS much again and you will be sucked in and never re-emerge for another N year |
| 09:20 | <MikeSmith> | LC gives people more incentive to get off their asses |
| 09:21 | <MikeSmith> | and actually take the time to comment |
| 09:21 | <MikeSmith> | we can do multiple LC rounds |
| 09:21 | <MikeSmith> | WD pretty much has come to mean "ignore me" |
| 09:21 | <MikeSmith> | I guess I should not be the one saying that |
| 09:22 | <MikeSmith> | what I meant to say is, Some people would say that WD pretty much has come to mean "ignore me" |
| 09:22 | <MikeSmith> | the IRC heard me wrong |
| 09:23 | <MikeSmith> | some people would say that the space between FPWD and LC is a very quiet, calm, uninteresting place |
| 09:23 | <MikeSmith> | kind of a limbo |
| 09:23 | <MikeSmith> | or purgatory |
| 09:23 | <MikeSmith> | a place where you don't want to be |
| 09:24 | <MikeSmith> | you want to be in Heaven |
| 09:24 | <MikeSmith> | boblet: I know of none such |
| 09:25 | <MikeSmith> | nudge Manu about it maybe |
| 09:25 | <MikeSmith> | to ask |
| 09:25 | <boblet> | MikeSmith: atm the only vocabularies I’ve found are Hixie’s and Google{s |
| 09:25 | <MikeSmith> | https://twitter.com/manusporny |
| 09:25 | <boblet> | k |
| 09:25 | <MikeSmith> | Manu is the Man |
| 09:26 | <MikeSmith> | when it comes to RDFa |
| 09:26 | <annevk> | MikeSmith, dunno, LC of workers and all were also ignored |
| 09:26 | <MikeSmith> | precisely |
| 09:27 | <MikeSmith> | hmm, that's not the word I meant |
| 09:27 | <annevk> | I would think so |
| 09:27 | <annevk> | it contradicts what you said earlier :) |
| 09:28 | <MikeSmith> | I've been awake for a little while |
| 09:28 | <MikeSmith> | I got a little bit of sleep |
| 09:28 | <MikeSmith> | I think |
| 09:28 | <MikeSmith> | I can't remember |
| 09:29 | <annevk> | I slept shorter than I would have liked too |
| 09:29 | <MikeSmith> | anyway, if nobody comments on an LC, then that makes creating the disposition of comments pretty easy to write |
| 09:29 | <MikeSmith> | and it makes it pretty easy to get the Director to transition it to CR |
| 09:30 | <MikeSmith> | we can't exactly lay out the red carpet for people every time we do an LC |
| 09:31 | <MikeSmith> | if they don't care enough to comment, it can't be that important to them |
| 09:31 | <MikeSmith> | so we go with what we got, and take it to CR |
| 09:32 | <MikeSmith> | this is the way we really are *supposed* to be doing it |
| 09:32 | <MikeSmith> | as far as W3C process |
| 09:32 | <MikeSmith> | keep stuff moving through the pipeline |
| 09:33 | <MikeSmith> | it is of course infinitely more important that the specs actually get implemented |
| 09:33 | <MikeSmith> | and there's no point in moving them along if they are not being implemented |
| 09:33 | <MikeSmith> | or if implementors show no interest or movement at all in implementing them |
| 09:34 | <MikeSmith> | but that ain't the case with this stuff |
| 09:34 | <annevk> | I think caution is warranted though if not many comments are received |
| 09:34 | <MikeSmith> | sure it is |
| 09:35 | <annevk> | either it means most comments were addressed pre Last Call, or it is something bad |
| 09:35 | <annevk> | in case of Web Workers and EventSource and all it might actually be that most comments have already been addressed |
| 09:35 | <MikeSmith> | it mostly means that people are just busy and have not made time to review and comment on it |
| 09:35 | <annevk> | they went through multiple implementation cycles, have had developers play with them to some extent, etc. |
| 09:36 | <MikeSmith> | annevk: I think with Workers and SSE, that is more the case, yeah |
| 09:36 | <annevk> | and some have been revised in design several times; something that happens for most W3C specifications post Last Call |
| 09:36 | <MikeSmith> | yep |
| 09:36 | <annevk> | post first Last Call, that is |
| 09:37 | <MikeSmith> | and we can sit around and wring our hands worrying about having not gotten some comments we should have, or we can instead just progress it and say we've done as much as we can for now |
| 09:38 | <MikeSmith> | as has been pointed out, it's a lot like shipping software |
| 09:38 | <MikeSmith> | at some point, you need to just ship it |
| 09:43 | <annevk> | agreed |
| 09:44 | <MikeSmith> | hsivonen: (from comment I missed yesterday) about why a web app should know atmospheric pressure of a user |
| 09:44 | <MikeSmith> | I assume that would be so it could use that to determine elevation |
| 09:44 | <MikeSmith> | altitude |
| 09:46 | <MikeSmith> | annevk: totally unrelated question -- can you remind me what browser XML parsers do with named character references in docs that have a valid XHTML doctype? |
| 09:46 | <MikeSmith> | I remember discussion of this a while back, but don't remember the answer |
| 09:47 | <annevk> | well, the "entities" "work" |
| 09:47 | <MikeSmith> | OK |
| 09:47 | <annevk> | the exact details of how might very well differ between implementations |
| 09:47 | <MikeSmith> | that's only *if* you have a non-borkend doctype, right? |
| 09:48 | <annevk> | yes, we recognize several opaque identifiers |
| 09:48 | <MikeSmith> | yeah, that how is something I'm curious about |
| 09:48 | <annevk> | and the set each browser has is probably not identical to that of other browsers |
| 09:48 | <MikeSmith> | I assume no browser actually has DTD copies or something it reads |
| 09:48 | <MikeSmith> | OK |
| 09:49 | <annevk> | I think Mozilla uses special DTDs |
| 09:49 | <MikeSmith> | hmm |
| 09:49 | <annevk> | to make it work with their off the shelf parser |
| 09:49 | <MikeSmith> | ah |
| 09:49 | <annevk> | WebKit might do the same, they have an OTS parser too |
| 10:01 | <MikeSmith> | annevk: OTS parser? |
| 10:01 | <annevk> | off-the-shelf |
| 10:01 | <annevk> | so about WebSRT |
| 10:01 | <annevk> | "Otherwise, append escape followed by a U+003B SEMICOLON character (;) to result." should prolly be preceded by & |
| 10:01 | <annevk> | furthermore, WebSRT parsing does not seem to indicate errors |
| 10:02 | <annevk> | should it not be similar to the HTML parser in this respect? |
| 10:26 | <aho> | websrt? is there a proposal for that? |
| 10:27 | <annevk> | http://www.whatwg.org/specs/web-apps/current-work/complete/video.html#websrt-0 |
| 10:27 | <MikeSmith> | aho: Hixie working on it as we speak |
| 10:27 | <MikeSmith> | *continuing to work on it |
| 10:28 | <jgraham> | Unless he is sleeping |
| 10:28 | <jgraham> | Although possibly he is dreaming in WebSRT |
| 10:28 | <aho> | http://pastebin.com/hHce723e <- "small ASS" (hurr) |
| 10:28 | <MikeSmith> | I wonder how the work on coming up with a browser-implementible WebTTML spec is going |
| 10:29 | <aho> | so, how does websrt compare to currently used formats such as SSA, AAS, and well... SRT? |
| 10:29 | <MikeSmith> | aho: I voted for WebASS |
| 10:29 | <MikeSmith> | but I got overruled |
| 10:29 | <aho> | aww |
| 10:30 | <MikeSmith> | we missed a big opportunity on that one |
| 10:30 | <aho> | indeed |
| 10:30 | <MikeSmith> | maybe I can get Hixie to reconsider |
| 10:30 | <aho> | :> |
| 10:30 | <MikeSmith> | I think we should start prefixes specs with a "Bad" convention |
| 10:30 | <MikeSmith> | instead of "Web" |
| 10:30 | <MikeSmith> | and they we could have "BadASS" |
| 10:30 | <aho> | and BadVoodoo |
| 10:31 | <MikeSmith> | yes! |
| 10:31 | <aho> | mh. but wasn't srt very limited? |
| 10:32 | <MikeSmith> | Hixie is working on making it un-limited |
| 10:32 | <MikeSmith> | TurboSRT |
| 10:32 | <MikeSmith> | which argues for naming it SST instead |
| 10:33 | <MikeSmith> | aho: you might take a look at the spec... |
| 10:33 | <aho> | i'm taking a look right now |
| 10:33 | <MikeSmith> | the answers to all life's questions can be found within that spec |
| 10:33 | <MikeSmith> | if you look hard enough |
| 10:33 | <MikeSmith> | and long enough |
| 10:36 | <aho> | is there some example websrt file i can look at? |
| 10:36 | <annevk> | aho, it's basically SRT plus ruby |
| 10:37 | <annevk> | aho, on top of that the idea is to have CSS extensions to cover styling |
| 10:37 | <aho> | which is awful since css doesn't support text outlines yet |
| 10:38 | <aho> | there is text-stroke, but that's something different |
| 10:38 | <annevk> | http://wiki.whatwg.org/wiki/Timed_tracks has background information |
| 10:38 | <annevk> | I'm not sure how it's awful |
| 10:38 | <annevk> | currently we have nothing |
| 10:38 | <aho> | http://kaioa.com/svg/fatten.png <- stroke vs outline |
| 10:39 | <annevk> | sure, but what does that have to do with choice of format? |
| 10:39 | <aho> | #3 = light blue text with light blue stroke... and a clone below with an even fatter stroke |
| 10:39 | <aho> | nothing |
| 10:40 | <aho> | but it's bad if you can't display text in a way you need for subtitles |
| 10:40 | <aho> | your subtitles won't be any good if they are hard to read |
| 10:43 | <hsivonen> | MikeSmith: if pressure is for elevantion, why is elevation important? |
| 10:43 | <hsivonen> | MikeSmith: and if it is, shouldn't the browser expose elevation |
| 10:44 | <hsivonen> | MikeSmith: and leave it as an impl. detail whether it comes from pressure or GPS |
| 10:49 | <MikeSmith> | hsivonen: yeah, true |
| 10:49 | <MikeSmith> | browser should just expose elevation as an abstraction |
| 10:49 | <MikeSmith> | and it does with location |
| 10:49 | <hsivonen> | aho: I've been hoping CSS outlines to land before captions/subtitles :-/ |
| 10:50 | <MikeSmith> | and it's up to the browser or platform to determine where it gets the elevation info from |
| 10:50 | <hsivonen> | it's going to suck if they happen in the other order |
| 10:50 | <aho> | indeed |
| 10:50 | <aho> | well, one could already create a websrt parser/renderer with js |
| 10:50 | <annevk> | aho, I don't see why CSS extensions would not cover that |
| 10:51 | <MikeSmith> | on many mobile devices, I think the exact method used to determine location is not even necessarily exposed to the OS |
| 10:51 | <aho> | annevk, they currently don't afaik |
| 10:51 | <annevk> | aho, CSS extensions for WebSRT are not defined at all |
| 10:51 | <annevk> | premature panicking imo |
| 10:52 | <jgraham> | aho: The CSS requirements are independent of the choice of format |
| 10:52 | <hsivonen> | annevk: I think panicing about CSS outlines isn't premature |
| 10:52 | <aho> | not really... it's more of a requirements check |
| 10:53 | <aho> | jgraham, yes.<annevk> sure, but what does that have to do with choice of format? <aho> nothing |
| 10:53 | <MikeSmith> | many chipsets for mobile devices just have location-sensing built into the chipset, and the hardware does the switching among location-sensing mechanisms it is capable of doing (wifi triangulation, gps, whatever) and then just reports back the position data without saying how it determined it |
| 10:53 | <aho> | websrt requires text outlines for being usable |
| 10:53 | <annevk> | hsivonen, what is the doom scenario here? |
| 10:53 | <annevk> | hsivonen, I don't see it |
| 10:53 | <jgraham> | aho: Right, so the problem with CSS is more urgent because we have a use case |
| 10:54 | <aho> | yes |
| 10:54 | <jgraham> | but it is nothing to do with WebSRT vs other formats |
| 10:54 | <aho> | (we already had some beforehand though) :> |
| 10:54 | <jgraham> | because current browsers don;t support outlined text anyway |
| 10:54 | <hsivonen> | annevk: WebSRT happens before outlines and as a result, the default WebSRT styling as well as explicit styling that wants to be compatible with outlineless browsers will look bad compred to TV subtitles |
| 10:55 | <aho> | thing is, websrt = css styled text... this is what adds that extra requirement |
| 10:55 | <jgraham> | aho: From a browser point of view it is irrelevant |
| 10:55 | <hsivonen> | annevk: the basic Scenario of DOOM is that we'll be stuck with to kind of appearance youtube captions have now |
| 10:55 | <jgraham> | You need to add support in your gfx layer either way |
| 10:55 | <aho> | yes |
| 10:56 | <hsivonen> | a solid box and normally filled unstoked unoutline glyph over the box |
| 10:56 | <aho> | it still adds that extra checkpoint on the roadmap though |
| 10:56 | <hsivonen> | right, the big deal is gfx support cross-platform. not the style system side |
| 10:57 | <jgraham> | hsivonen: (fwiw teletext-provided subtitles in the UK have had that appearance for years) |
| 10:57 | <hsivonen> | jgraham: teletext subtitles are rubbish! that's the DOOM! |
| 10:57 | <aho> | so, is there some proposal for text outlines already? |
| 10:58 | <jgraham> | hsivonen: Well maybe, but they work |
| 10:58 | <jgraham> | So it is not really about being better than tv |
| 10:58 | <aho> | text-shadow could have taken care of that actually... if it would have used the same parameters as box-shadow |
| 10:58 | <aho> | i still don't know why they didn't do that |
| 10:58 | <aho> | (that optional outset thing) |
| 10:59 | <hsivonen> | jgraham: my baseline for "tv" is the look we have here |
| 11:00 | <hsivonen> | jgraham: wouldn't you agree the Swedish burned-in or DVB subtitles are better than UK teletext subtitles? |
| 11:00 | <annevk> | hsivonen, hmm |
| 11:00 | <jgraham> | hsivonen: Better, yes |
| 11:01 | <annevk> | hsivonen, not particularly convinced, but maybe |
| 11:02 | <jgraham> | hsivonen: I don't see that people are likely to go to any lengths to get the browser-that-doesn't-support-outlines look in browsers that do |
| 11:02 | <hsivonen> | annevk: you get outlines on Dutch TV, right? don't you agree that the black background box is a terrible alternative? |
| 11:03 | <hsivonen> | jgraham: how do you write backwards-compatible CSS? |
| 11:03 | <aho> | black background = 90s tech :> |
| 11:03 | <hsivonen> | jgraham: coloc: white; outline-color: black; background-color: transparent; doesn't exactly degrade gracefully |
| 11:04 | <hsivonen> | aho: 50s tech outline FTW! |
| 11:04 | <jgraham> | hsivonen: Honestly if the difference is that big people will UA sniff |
| 11:04 | <jgraham> | The fact that is hard to do in CSS is unfortunate |
| 11:05 | <aho> | most people would probably use js to detect support... and than add some extra class if it isn't supported |
| 11:06 | <hsivonen> | that would still leave bad default styling, assuming that browsers will be as shy to change default styling as they are now |
| 11:07 | <aho> | one can also use 9 (nine) text-shadows to get some kind of pseudo outlines :> |
| 11:07 | <annevk> | hsivonen, oh, I'm not convinced we'd get stuck so soon as you think we will |
| 11:07 | <annevk> | hsivonen, I agree outlines are nicer |
| 11:07 | <annevk> | hsivonen, I hardly watch Dutch TV though |
| 11:07 | <hsivonen> | aho: yeah, text-shadow will probably save us |
| 11:08 | <aho> | they really should have added that outset parameter |
| 11:08 | <aho> | text-shadow and box-shadow would be more uniform then |
| 11:16 | <aho> | http://kaioa.com/k/outline.html |
| 11:16 | <aho> | pretty crude, really |
| 11:17 | <aho> | also, 8 shadows... not 9 ;> |
| 11:17 | <aho> | (there isn't one for the center position) |
| 11:20 | <annevk> | btw, http://www.w3.org/TR/css3-text/#text-outline |
| 11:23 | <aho> | does any browser support it right now? |
| 11:23 | <Peter`> | WebKit has text-outline |
| 11:24 | <Peter`> | text-stroke, not outline, n/m |
| 11:25 | <Peter`> | same idea though |
| 11:25 | <aho> | Peter`, no... stroke and outline are *very* different things |
| 11:25 | <MikeSmith> | Peter`: btw, cool to see <mark> stuff poised to go in |
| 11:26 | <Peter`> | aho: ok, my bad |
| 11:26 | <Peter`> | MikeSmith: Yeah, hope my testcase is alright |
| 11:28 | <MikeSmith> | Peter`: I guess Maciej will let you know if it's not |
| 11:28 | <aho> | annevk, well... i don't understand those specs. it doesn't even bother mentioning which length thingy does what |
| 11:29 | <mikekelly> | am I correct in thinking that forms don't have an @rel attribute? |
| 11:29 | <Peter`> | MikeSmith: Indeed, I'm still waiting for three build-bots to succeed but don't expect any problems. We'll know tomorrow morning |
| 11:35 | <drclue> | hey jagoffs |
| 12:44 | <roc> | text-outline doesn't sound that hard |
| 12:45 | <roc> | we wouldn't need per-platform support. We could do it via a alpha buffer the way we do text-shadow |
| 12:46 | <annevk> | hmm, did someone just spam the WHATWG list? |
| 12:46 | <annevk> | blargh |
| 12:47 | <MikeSmith> | webkit nightlies now how .dataset love - |
| 12:48 | <MikeSmith> | http://trac.webkit.org/export/62376/trunk/LayoutTests/fast/dom/dataset.html |
| 12:48 | <annevk> | MikeSmith, your English is getting worse |
| 12:48 | <MikeSmith> | what English? |
| 12:49 | <annevk> | right |
| 12:50 | <annevk> | dataset is nice |
| 12:50 | <annevk> | would be cool to have that in Opera too |
| 13:09 | <roc> | MikeSmith: you know we landed a <mark> implementation a few days ago? |
| 13:10 | <MikeSmith> | roc: no, didn't know that |
| 13:10 | <MikeSmith> | link for the changeset or bug? |
| 13:11 | <roc> | https://bugzilla.mozilla.org/show_bug.cgi?id=485377 |
| 13:12 | <MikeSmith> | thanks |
| 13:46 | <mut> | boo |
| 13:48 | <MikeSmith> | Peter`: |
| 13:48 | <MikeSmith> | FAIL getWeight("mark2") should be bold. Was bolder. |
| 13:48 | <MikeSmith> | FAIL getWeight("span1") should be bold. Was bolder. |
| 13:48 | <MikeSmith> | ...when I run you test case in Minefield |
| 13:49 | <Peter`> | MikeSmith: could you try running the header test? |
| 13:50 | <Peter`> | it's in the same directory |
| 13:50 | <MikeSmith> | same directory? |
| 13:50 | <Peter`> | http://trac.webkit.org/export/62376/trunk/LayoutTests/fast/html/header-element.html |
| 13:50 | <MikeSmith> | ah |
| 13:50 | <Peter`> | Same result on minefield there |
| 13:51 | <Peter`> | so my guess is that Minefield sets font-weight: bolder; on <b> tags |
| 13:51 | <Peter`> | where WebKit uses "bold" |
| 13:51 | <MikeSmith> | yeah |
| 13:51 | <MikeSmith> | [[ |
| 13:51 | <MikeSmith> | Residual style: |
| 13:51 | <MikeSmith> | FAIL getWeight("header4") should be bold. Was bolder. |
| 13:51 | <MikeSmith> | FAIL getWeight("span1") should be bold. Was bolder. |
| 13:51 | <MikeSmith> | FormatBlock: |
| 13:51 | <MikeSmith> | FAIL successfullyParsed should be true (of type boolean). Was undefined (of type undefined). |
| 13:51 | <MikeSmith> | ]] |
| 13:52 | <Peter`> | That succeeds on WebKit |
| 13:54 | <MikeSmith> | Peter`: I was wanting to make a simple test case that'd work cross-browser |
| 13:55 | MikeSmith | reads through js source |
| 13:55 | <Peter`> | I poked @rakaz (owner of html5test.com) that his tests on mark/ruby/rt/rp relied on styling as well |
| 13:55 | <Peter`> | which is, to my knowledge, one of the few public test-cases around |
| 13:56 | <Peter`> | default style in Minefield and my WebKit patch both have a yellow background-color and black text however, that is testable |
| 13:56 | <MikeSmith> | yeah |
| 13:57 | <MikeSmith> | just not a super exciting test :( |
| 13:57 | <Peter`> | No, and unreliable as well, since the spec does not state any default styling |
| 13:57 | <MikeSmith> | true dat |
| 13:58 | <annevk> | Peter`, it doesn't? |
| 13:59 | <annevk> | per HTML5 it's b { font-weight:bold } |
| 13:59 | <Peter`> | mark { background: yellow; color: black; } |
| 13:59 | <Peter`> | it does |
| 13:59 | <annevk> | but that might just be based on WebKit |
| 13:59 | <Peter`> | You're right :) |
| 13:59 | <annevk> | bolder seems better |
| 14:03 | <Peter`> | so yes, checking for the default background color probably is the best test |
| 14:34 | <Lachy> | does anyone know how accurate are Amazon's delivery estimates generally are? The site orignally said 7 to 10 days for products shipped to Norway, but now that it's despatched, they're telling me 19 days |
| 14:35 | <jgraham> | Lachy: They are usually made up by untrained armadillos |
| 14:35 | <Lachy> | I didn't want to have to wait nearly 3 weeks :-( |
| 14:35 | <doublec> | their estimates for delivery to NZ are often pretty bad |
| 14:35 | <doublec> | they'll say 3 weeks and I get it in 8 days |
| 14:35 | <Lachy> | ok |
| 14:35 | <doublec> | so 'bad' in a good way |
| 14:35 | <jgraham> | Or the opposite |
| 14:36 | <jgraham> | I have had "it will be (small number of days) and it was actually (large number of days)" |
| 14:36 | <jgraham> | er, the second " was in quite the wrong place there |
| 14:40 | <mut> | hey, can someone point me in the right direction, i have some lines rendered on canvas |
| 14:40 | <mut> | and i want to be able to measure them |
| 14:40 | <mut> | with the mouse ; |
| 14:40 | <mut> | just wondered if anyonecould point me at some recommened reading>? |
| 14:44 | <Lachy> | mut, in what unit do you want to measure them? Number of pixels? length in mm as rendered on the screen? Other? |
| 14:45 | <Lachy> | in any case, I think the basic concept of doing that is to find the position of line's end points, and then use a bit of trig to calculate the length |
| 14:48 | <Lachy> | or, actually, a^2 = b^2 + c^2 should be enough to calculate the length, where b and c are abs(x1-x2) and abs(y1-y2) |
| 14:49 | <Philip`> | (You don't need abs) |
| 14:50 | <Lachy> | yeah, I realised that after I wrote it cause they're squared. But whatever |
| 14:50 | <TelFiRE> | If I use HTML5 can I use <script> tags or will I still need to include the text/javascript for the time being? |
| 14:50 | <TelFiRE> | looking to support IE7 on up |
| 14:51 | <Philip`> | TelFiRE: You mean writing <script> instead of <script type="text/javascript">? |
| 14:51 | <TelFiRE> | correct |
| 14:51 | <Philip`> | That works and has always worked and will always work |
| 14:51 | <Philip`> | so there's no compatibility problem |
| 14:51 | <TelFiRE> | ah ok cool, but wasn't it invalid with html4? |
| 14:51 | <mut> | Lachy pixles |
| 14:51 | <TelFiRE> | ok awesome, thanks |
| 14:51 | <Philip`> | TelFiRE: It may have been |
| 14:51 | <Lachy> | mut, ok, then that's really simple |
| 14:52 | <Philip`> | TelFiRE: but that's independent of what browsers support (and have to support, since lots of people already write <script>) |
| 14:52 | <mut> | with xy coords of the mouse? |
| 14:52 | <mut> | any way to lock on corners/endpoints? |
| 14:52 | <Lachy> | do you mean, when the user clicks a point on the canvas, how do you get those co-ordinates? |
| 14:53 | <Lachy> | or do you want to snap to the nearest end-point of an existing line? |
| 14:53 | <mut> | well mouse XY will do i suppose, but it would be great ifi could snap to existing endpoints |
| 14:54 | <Lachy> | do you have a collection with all the line end point co-ordinates, that you can search through? |
| 14:55 | <mut> | a collection? |
| 14:55 | <mut> | errm |
| 14:56 | <mut> | well i guess i could make one |
| 14:56 | <mut> | like stick all the endpoints in an array or something |
| 14:56 | <Lachy> | yeah |
| 14:57 | <Lachy> | so then when you capture the mouse event, you get the x,y coords from that event object, adjusting for the element's position (can't remember the exact code for this off hand) |
| 14:58 | <mut> | ok that doesnt sound too hard |
| 14:58 | <Lachy> | then when you have those coords, you have to search through the array to find the nearest point. |
| 14:58 | <mut> | heh CAD in browser |
| 14:58 | <Lachy> | that requires some math that I'm not personally familiar with, but you should be able to find it through a web search |
| 14:58 | <mut> | im pretty good with the maths |
| 14:59 | <mut> | but the code escapes me sometimes! |
| 14:59 | <Lachy> | it's probably something simple like finding the lowest values of (mouseX - pointX) and (mouseY - pointY), for each point |
| 14:59 | <jgraham> | It's just the same maths as before |
| 14:59 | <Lachy> | oh, yeah, find the shortest hypotenuse will do it |
| 14:59 | <mut> | well I wouldnt have thourght of storing all of the end points, so thats good |
| 15:00 | <mut> | this way, i can scale the canvas down, and the measuring will still work |
| 15:03 | <mut> | thanks alot guys |
| 17:00 | <cyberix> | What happened to the peer-to-peer communication spec? |
| 17:00 | <cyberix> | is there still one |
| 17:00 | <cyberix> | some parts of the communication stuff got moved to websocket spec |
| 17:01 | <cyberix> | and some to postmessage |
| 17:01 | <cyberix> | but where did the peer-to-peer part go |
| 17:01 | <cyberix> | more specifically, how do I code a clientside websocket server in javascript |
| 17:02 | <annevk> | it didn't go anywhere |
| 17:02 | <annevk> | it's simply not done |
| 17:03 | <cyberix> | but are the place holders still somewhere? |
| 17:03 | <annevk> | sure |
| 17:03 | <cyberix> | I cannot find them |
| 17:03 | <annevk> | http://www.whatwg.org/specs/web-apps/current-work/complete/commands.html#peer-to-peer-connections |
| 17:05 | <cyberix> | I'm thinking that maybe letting javascript listen on an IPv4 interface is not a good idea |
| 17:05 | <cyberix> | but some discardable virtual addresses might work |
| 17:06 | <cyberix> | I'm trying to research that kind of approaches as part of my work |
| 17:06 | <cyberix> | and I am just wondering, if others have similar ideas |
| 17:07 | <annevk> | the WHATWG email archives have some additional information |
| 17:07 | <annevk> | other than that I think it is ok |
| 17:07 | <annevk> | I mean, not much |
| 17:08 | <cyberix> | in my opinion browser-to-browser communication should be IPv4, but the actual js-to-js communication should probably use some new addressing model |
| 17:08 | <jgraham> | annevk: Aren't you supposed to out partying on the streets or something |
| 17:08 | <jgraham> | ? |
| 17:11 | <annevk> | jgraham, already did |
| 17:12 | <annevk> | jgraham, but I'll be out again soon :) |
| 17:12 | <Peter`> | It was quite an unexpected win.. I didn't give our team a very big chance. |
| 17:12 | <annevk> | I'm trying to tweet #worldcup #ned #win but it's not responding |
| 17:28 | <TabAtkins> | re: CSS on captions: A plain text-shadow with a small blur is enough to render captions readable on any background. |
| 17:28 | <TabAtkins> | You don't need to stack 9 atop each other, just do something like "color:white;text-shadow:0 0 .2em black;" |
| 17:30 | <TabAtkins> | At least, that works for my gradient-demonstration page that I made a long time ago, and keeps the text visible on any gradient I've ever done. |
| 17:32 | <karlushi> | http://www.joindiaspora.com//2010/07/01/one-month-in.html |
| 17:33 | <karlushi> | "We have also started to build the latest and greatest in web standards into Diaspora. Websockets are already in the Diaspora core, and any Diaspora plugins will be able to have rapid two-way functionality, (think chat, games etc) almost effortlessly!" |
| 17:33 | <miketaylr> | thats awesome |
| 17:35 | <TabAtkins> | Agreed! |
| 17:37 | <AryehGregor> | Too bad WebSockets aren't remotely stable yet? |
| 17:38 | micheil | uses websockets + heaps of other companies |
| 17:38 | <micheil> | the spec may not be stable, but the technology seems it |
| 17:40 | <AryehGregor> | Chrome's implementation just changed completely incompatibly in the dev channel, didn't it? |
| 17:40 | <micheil> | uhh.. I hope not |
| 17:40 | <micheil> | Chromium is currently draft 76+, chrome is 75 |
| 17:40 | <micheil> | if, that's what you mean |
| 17:41 | <TabAtkins> | Yeah, 75-76 has some breaking changes, iirc. |
| 17:41 | <micheil> | yeah |
| 17:41 | <micheil> | but there's been no other changes since, right? |
| 17:42 | <AryehGregor> | Since like a month or two ago, no. |
| 17:42 | <micheil> | ah, good good |
| 17:42 | <micheil> | I'm on top of the spec then |
| 17:42 | <AryehGregor> | But more incompatible changes are likely, from what I've heard. |
| 17:45 | <micheil> | I hope they don't go with that SSL type thing that was being concocted |
| 17:45 | <micheil> | that'd be really bad for the protocol, imho. |
| 18:42 | <Hixie> | micheil: why? |
| 18:42 | <Hixie> | micheil: (i agree with you, just curious what your reasons are) |
| 18:42 | <micheil> | because it'll make implementation heaps harder |
| 18:42 | <micheil> | and it'll probably deter people from using what is currently a fairly open protocol |
| 18:43 | <micheil> | it's current relative easy of implementation is really awesome |
| 18:43 | <micheil> | I mean, a few hours after I pushed support for draft76 in my node.js websocket server, someone was already starting work on a draft76 node.js client. |
| 18:44 | <Hixie> | yeah that was my thinking too |
| 18:44 | <Hixie> | adam seemed to think it was relatively easy to use ssl libraries |
| 18:44 | <micheil> | not at all. |
| 18:45 | <micheil> | in node.js alone, we went about 3 months without ssl, purely because we couldn't put the resources to writing the bindings for it |
| 18:45 | <Hixie> | yeah that was my impression too |
| 18:45 | <Hixie> | are you on the hybi list? |
| 18:45 | <micheil> | not only that, but licensing on both openssl and gnupg is odd, so opensource inclusion is difficult |
| 18:45 | <micheil> | no |
| 18:46 | <micheil> | I keep forgetting to subscribe to it |
| 18:46 | <Hixie> | k |
| 18:46 | <Hixie> | might be good to have you on there just in case that gets traction, so i'm not hte only one arguing against it :-) |
| 18:46 | <micheil> | yeah, fair point |
| 18:46 | <micheil> | what was the address for the membership page of that mailing list? |
| 18:47 | <Hixie> | i forget, by google for hybi will get it |
| 18:47 | <Hixie> | gotta go |
| 18:47 | <micheil> | okay, thanks |
| 18:47 | <micheil> | :) |
| 18:59 | <AryehGregor> | Is there any way to get CSS from your HTML document to work on non-inline SVG? |
| 18:59 | <AryehGregor> | Like "color" and "height". |
| 19:01 | <micheil> | uhh.. I think so |
| 19:01 | <micheil> | inline style attribute, but some things are specially named |
| 19:02 | <micheil> | check the spec would be the best bet. |
| 19:02 | <Philip`> | AryehGregor: <iframe seamless src=example.svg></iframe> |
| 19:02 | <AryehGregor> | Nice. Is that actually supported yet? |
| 19:02 | <Philip`> | Otherwise I don't think styles can ever propagate into an external document, if that's what you're wanting |
| 19:02 | <Philip`> | No |
| 19:02 | <AryehGregor> | I'll wait, then. |
| 19:14 | <Philip`> | About image resolutions: How about using progressive JPEG / interlaced PNG / etc, and teaching UAs to only download enough of the image to match their DPI? |
| 19:14 | <AryehGregor> | That's a sneaky idea. |
| 19:14 | <AryehGregor> | Does it have obvious drawbacks? I don't know much about image formats. |
| 19:17 | <Philip`> | If you download the entire image then it's slightly less efficient compression than a plain full-res version |
| 19:17 | <Philip`> | but on the other hand you could download half the image data to display the detail you need, and if the user zooms in then you only need to download the other half and not the entire full-res image |
| 19:18 | <AryehGregor> | That would be very cool. |
| 19:18 | <Philip`> | The interaction with <canvas> drawImage would be weird if you intentionally stop after downloading half the image, though |
| 19:19 | <AryehGregor> | If you need to draw to canvas, you could just download the rest of the image before doing so. |
| 19:19 | <AryehGregor> | Same as if the user wants to save the image, etc. |
| 19:20 | <Philip`> | That sounds synchronous |
| 19:20 | <Philip`> | and nobody likes synchronous APIs |
| 19:20 | <AryehGregor> | I guess normally you'd avoid the problem by having the script wait for onload before drawing the image? |
| 19:21 | <AryehGregor> | This would mess up onload semantics in that respect, clearly. |
| 19:21 | <Philip`> | When would onload occur, if the UA downloads and renders half the image data and not the rest? |
| 19:21 | <AryehGregor> | Also, I wonder if the quality is really controllable enough for users' needs. |
| 19:21 | <Philip`> | Either it blocks onload forever, which is bad, or onload fires before it's fully loaded, which is bad if you then try to draw the image onto a canvas |
| 19:22 | <AryehGregor> | Oh well, the idea was a hack anyway. |
| 19:22 | <AryehGregor> | (although a cool hack) |
| 19:22 | <Philip`> | But you could probably do it automatically for images used in CSS |
| 19:22 | <AryehGregor> | Which is exactly where the guy said it's not needed. |
| 19:23 | <Philip`> | and you could add a flag like <img async> to say you don't want to block the load event and don't mind if it never completes downloading |
| 19:25 | <AryehGregor> | Sounds very cool, and also unlikely to be a good idea. |
| 19:25 | <AryehGregor> | Too elegant to be practical. |
| 19:26 | <AryehGregor> | No, what we need is a total kludge, like: <img src="foo.jpg" altsrcs="x2:foox2.jpg x4:foox4.jpg"> |
| 19:26 | <Philip`> | <img src="foox4.jpg" lowsrc="foox2.jpg" lowersrc="foo.jpg"> |
| 19:27 | <AryehGregor> | That doesn't give the magnification values. |
| 19:28 | <Philip`> | They're defined to be 100%, 50%, 25% |
| 19:28 | <AryehGregor> | That's too inflexible. |
| 19:28 | <Philip`> | Sure |
| 19:28 | <AryehGregor> | What if you want the default to be the low-res version? That's what the use-case asked for. |
| 19:28 | <Philip`> | I thought the idea was to be a total kludge |
| 19:28 | <Philip`> | s/ / / |
| 19:29 | <Philip`> | You could do it with highsrc but I wanted to reuse existing attributes |
| 19:34 | <AryehGregor> | lowsrc and lowersrc are existing attributes? |
| 19:40 | <Philip`> | AryehGregor: lowsrc was |
| 19:41 | <Philip`> | http://www.ross.net/webhints/image_lowsrc.html etc |
| 19:42 | <sonja_> | does it look like there will be a <video> codec that IE and FF and Chrome and Opera and Safari will all support? |
| 19:42 | <sonja_> | or are there still codec wars? |
| 19:42 | <AryehGregor> | Already answered you in #mediawiki. VP8 is a plausible contender, but it's not there yet. |
| 19:43 | <AryehGregor> | (= WebM, specifically) |
| 19:43 | <sonja_> | :) |
| 19:45 | <sonja_> | do the latest versions of the big 5 browsers currently support webm? |
| 19:45 | <TabAtkins> | 3 of them do, if by "latest" you mean "dev versions, not yet publicly released but available if you download them manually". |
| 19:45 | <AryehGregor> | No, only Chrome, Firefox, and Opera. Safari and IE9 only support it if the user happens to have the codec installed (dunno if there's a QuickTime codec yet for WebM). |
| 19:45 | <sonja_> | hmm |
| 19:47 | <sonja_> | so maybe we just have to wait a few more monhs |
| 19:47 | <sonja_> | and regular "download GA version of browser" versions will work |
| 19:51 | <TabAtkins> | AryehGregor: I agree with Philip` - I think using a progressive format is the best idea. There are some weird interactions, but once we deal with them it's all super-elegant and nice. |
| 19:56 | <AryehGregor> | TabAtkins, that's why I'm suspicious. It sounds too elegant. But hey, if implementers think it's a good idea, I'm all for it. |
| 19:56 | <TabAtkins> | You are too cynical. ^_^ |
| 20:16 | <AryehGregor> | You're too optimistic. |
| 20:17 | <AryehGregor> | This is a feature that's almost impossible for users or authors to notice, and very hard for them to control, so there's little incentive for browsers to bother implementing it. |
| 20:18 | <othermaciej> | what feature? |
| 20:24 | <AryehGregor> | othermaciej, some feature that would allow different-resolution images to be served to different browsers by just serving progressive JPEG/interlaced PNG, and having browsers stop downloading once they have enough of the image to display at their full resolution. |
| 20:24 | <AryehGregor> | Presumably needing support like an "async" attribute so that it doesn't delay the load event, or whatever. |
| 20:27 | <othermaciej> | I see |
| 20:28 | <othermaciej> | I think at least some mobile browsers will do that automatically |
| 20:28 | <AryehGregor> | It was pointed out that this would cause bugs if, for instance, the image was drawn to a canvas. |
| 20:29 | <othermaciej> | if you want to serve multiple image resolutions to account for high-DPI and non-high-DPI displays, the best current way to do it is to use CSS media queries |
| 20:29 | <othermaciej> | at least for images referenced from CSS |
| 20:29 | <AryehGregor> | Which doesn't work for <img>, thus the whatwg thread. |
| 20:29 | <othermaciej> | http://webkit.org/blog/55/high-dpi-web-sites/ |
| 20:29 | <othermaciej> | http://webkit.org/blog/56/high-dpi-part-2/ |
| 20:30 | <othermaciej> | that's what the WebKit team recommended for this, a couple of years ago |
| 20:30 | <othermaciej> | I guess the best we had to suggest for <img> was always serving an oversize image |
| 20:31 | <othermaciej> | and using width/height to scale it down |
| 20:34 | <othermaciej> | AryehGregor: I think using the content: property on an image might be a way to change the resolution with CSS from a media query, but then you couldn't prevent the original from being loaded |
| 20:34 | <othermaciej> | AryehGregor: I am not sure magically solving it in the image format is a good solution |
| 20:35 | <othermaciej> | if the tradeoff is image quality vs. bandwidth - if you request the magic auto-sizing image, even if you cut off the download early the server has probably already sent the data and filled your pipe |
| 20:36 | <othermaciej> | if bandwidth is not a concern, you can just always serve a high-rez version |
| 20:37 | <othermaciej> | we also thought about making a multi-resolution version of PNG that would degrade gracefully, in case scaling down doesn't result in a good-looking image |
| 20:38 | <AryehGregor> | Do all browsers support non-cruddy image shrinking these days? As of a few years ago, images shrunk by the browser would look awful. |
| 20:38 | <othermaciej> | Safari has good looking image scaling in either direction |
| 20:39 | <othermaciej> | as for other browsers, I do not know |
| 20:39 | <AryehGregor> | I suspect they might all have it these days. |
| 20:39 | <othermaciej> | I would expect 2x shrink to not look that bad in any browser, but non-integer scale factors might not look so hot |
| 20:47 | <jgraham> | TabAtkins: Opera 10.60 shipped |
| 20:48 | <KaOSoFt> | Arrgh, and I just installed Opera a week ago. |
| 20:48 | <KaOSoFt> | ._. |
| 20:49 | <jgraham> | Hixie: (I will also argue against mandatory-TLS websockets if it is pursued) |
| 20:49 | <jgraham> | Hixie: (for the same high-barrier-to-entry reasons) |
| 20:55 | <AryehGregor> | By the way, is anyone but MS and Mozilla working much on GPU-accelerated rendering for video/canvas/etc.? |
| 20:55 | <AryehGregor> | MS is making it sound like The Future of the Web. |
| 20:57 | <jgraham> | AryehGregor: I think others typically don't make forward looking statements or whatever the apple lingo is |
| 20:57 | <jgraham> | (except when they do of course) |
| 21:00 | <AryehGregor> | I was asking about current work. |
| 21:00 | <jgraham> | Doesn't current work typically mean future products? |
| 21:00 | <jgraham> | (I don't know would have been a shorter answer I guess) |
| 21:10 | <jgraham> | AryehGregor: http://my.opera.com/core/blog/2009/02/04/vega?id=2953968 might interest you |
| 21:20 | <MikeSmith> | Peter`: I added your http://peter.sh/ feed to Planet HTML5 |
| 21:21 | <Peter-> | Thank you! Shall I create a seperate category/tag for it so random photography blogs don't get re-published? |
| 21:33 | <Peter-> | MikeSmith: Mind updating the feed URL to http://peter.sh/category/tech/feed ? |
| 21:36 | <MikeSmith> | Peter-: sure |
| 21:36 | <Peter-> | Cheers :-) |
| 21:37 | <MikeSmith> | and btw, the aggregator uses a very simple, ham-fisted means for filtering posts |
| 21:38 | <Peter-> | Ah, ok. Just to be sure, wouldn't want random non-tech things to accidentially get published. |
| 21:39 | <AryehGregor> | MikeSmith, what means is that? |
| 21:39 | <MikeSmith> | Peter-: it basically filters just on the string "html5" anywhere in the rendered text of a blog entry |
| 21:39 | <MikeSmith> | or the source |
| 21:39 | <MikeSmith> | that is, anywhere it in source (whether it's in the rendered part or not) |
| 21:40 | <AryehGregor> | What if it's spelled "HTML 5"? |
| 21:40 | <MikeSmith> | that'll work too |
| 21:40 | <Peter-> | MikeSmith: Ok, very clear |
| 21:40 | <MikeSmith> | I think "HTML¥s*WG" and "WHATWG" and maybe a few other things will work too |
| 21:40 | <AryehGregor> | ¥? |
| 21:41 | AryehGregor | bets \ |
| 21:41 | AryehGregor | loves what insight you can get into people's keyboard layout from typos. |
| 21:41 | <MikeSmith> | AryehGregor: heh, yeah :) |
| 21:41 | <AryehGregor> | One guy I know regularly types entire English sentences in Greek characters by mistake. |
| 21:41 | <AryehGregor> | Touch-typing without paying attention. |
| 21:41 | MikeSmith | chuckles |
| 21:42 | <AryehGregor> | I occasionally start typing something with all lowercase letters in Hebrew, but I've always noticed in time so far, I think. |
| 21:42 | <AryehGregor> | (oddly, if you hold shift in a standard Hebrew keyboard layout, you get QWERTY caps . . . not much else to do with it, I guess, since Hebrew has no letter casing) |
| 22:05 | <MikeSmith> | AryehGregor: ¥ and \ are the main problematic chars as far as Japanese and JP keyboards go |
| 22:05 | <AryehGregor> | "Japanese and JP"? |
| 22:06 | <TabAtkins> | jgraham: Why are you telling me about Opera? |
| 22:06 | <MikeSmith> | Japanese encodings |
| 22:06 | <MikeSmith> | I actually don't have a "\" label anywhere on my keyboard |
| 22:06 | AryehGregor | realizes he has no idea what Japanese keyboards actually looks like, Googles them |
| 22:06 | <MikeSmith> | the key I need to type is labeled "¥" |
| 22:07 | <MikeSmith> | and when I type that in my console, I get "\" |
| 22:07 | MikeSmith | tries to remember how he types "\" in other apps |
| 22:07 | <AryehGregor> | Hmm, so they have kana on them? I thought there are like 72 kana characters. Is some special input method used even for typing kana? |
| 22:08 | <jgraham> | TabAtkins: Because it is teh awesome. Or because it is a stable release with WebM |
| 22:08 | AryehGregor | reads http://en.wikipedia.org/wiki/Japanese_input_methods |
| 22:08 | <MikeSmith> | symbol and number keys are aloe used for typing kana |
| 22:08 | <jgraham> | At least one of those reasons |
| 22:08 | <TabAtkins> | jgraham: Ah, ok. I suppose that's relevant to my interests, then. |
| 22:08 | <MikeSmith> | but most people in Japan do not use type in kana directly |
| 22:08 | TabAtkins | needs to go find some webm video then. |
| 22:09 | <MikeSmith> | they type in romaji |
| 22:09 | <MikeSmith> | into the IME |
| 22:09 | TabAtkins | is going to do a half-hour talk on Wednesday, and is opting for <video>/<canvas> shenanigans. |
| 22:09 | <MikeSmith> | and the kana appears first in the IME after that, then use space to cycle through kanji if needed |
| 22:10 | <jgraham> | TabAtkins: Didn't you say earlier that only dev versions of browsers had WebM? That was why I told you anyway |
| 22:10 | <TabAtkins> | jgraham: Yeah, I figured. |
| 22:10 | <AryehGregor> | You're aware that CJK writing systems are completely insane, and East Asians are all nuts for not just switching to romaji/pinyin/etc., right? :) |
| 22:11 | <MikeSmith> | AryehGregor: :) |
| 22:11 | <Hixie> | jgraham: good to know |
| 22:11 | TabAtkins | needs to go open up Opera on his win machine so he can be reminded to update again. |
| 22:12 | AryehGregor | isn't sure whether that ":)" is like "haha, yeah :)" or "you're being a jerk but I'll be polite and just won't say anything, instead I'll say ':)'" |
| 22:13 | <Hixie> | MikeSmith would tell you if you were being a jerk, don't worry |
| 22:13 | <MikeSmith> | :) |
| 22:13 | <Hixie> | oh dear, now i don't know if i was wrong and he's telling _me_ that i'm being a jerk or if he's just laughing! |
| 22:13 | <MikeSmith> | see how I do that ? :) |
| 22:14 | <Hixie> | :) |
| 22:14 | <MikeSmith> | <chuckle><snort> |
| 22:14 | <MikeSmith> | AryehGregor: we need to keep kanji for microblogging purposes |
| 22:14 | <AryehGregor> | Hmm. That's a good point. |
| 22:15 | <Philip`> | You could switch to milliblogging and wouldn't need to be so concise |
| 22:15 | <TabAtkins> | 140k characters! |
| 22:15 | <Hixie> | or go to nanoblogging and kanji would be too big! |
| 22:15 | <AryehGregor> | We should devise some way to mash multiple Latin letters into one Unicode character to get the benefit of that. Doesn't hangul do something like that? |
| 22:15 | <jgraham> | https://heatmap.mozillalabs.com/mozmetrics/?data=perc&os=all&colorscheme=hsl is pretty interesting |
| 22:16 | <TabAtkins> | That wouldn't be too hard. You'd need, hm, a block of about 19k characters to get every possible combination of three latin letters. Much less if you do a trivial analysis over text from a couple languages and just use the triples that show up a non-trivial amount of time. |
| 22:17 | <AryehGregor> | Big surprise, no one clicks the EV button! |
| 22:17 | <AryehGregor> | TabAtkins, unfortunately, three Latin letters is still much less than one kanji character's worth. But you can optimize on a per-language basis. |
| 22:17 | <AryehGregor> | If Han unification works for CJK, why can't we use it for other languages too? |
| 22:17 | <jgraham> | AryehGregor: really |
| 22:17 | <TabAtkins> | Yeah. |
| 22:18 | <Philip`> | jgraham: Looks like people scroll up and left more than down and right, which makes no sense to me |
| 22:18 | <AryehGregor> | So have dedicated characters for common words like "whatever". |
| 22:18 | <TabAtkins> | Philip`: Most people have scrollwheels? |
| 22:18 | <AryehGregor> | Philip`, maybe they use scroll wheel or scroll bar for scrolling down, but not up? |
| 22:18 | AryehGregor | isn't sure why |
| 22:19 | <TabAtkins> | Wait, no, Philip`, where are you getting that? |
| 22:19 | MikeSmith | considers preparing to trot out his pro-EV hobbyhorse, to point out that if the UI works the say it should, nobody should need to click on the EV button |
| 22:19 | <TabAtkins> | up/down are both green, while the up/down bar is red. |
| 22:19 | <jgraham> | Philip`: Maybe they use the scroll wheel or keyboard to go in the more natural directions |
| 22:19 | <jgraham> | Philip`: But resort to the buttons for more unusual activities |
| 22:20 | <AryehGregor> | What are "Paste Button" and "Cut Button"? I don't see them in the picture, only in the list. |
| 22:20 | TabAtkins | sees the actual numbers now, so nevermind. |
| 22:20 | <jgraham> | AryehGregor: Non-default |
| 22:20 | <AryehGregor> | I figured. |
| 22:21 | <MikeSmith> | about kanji, the expressiveness of kanji for Japanese seeme essential to the written language, to the point where it is impossible for me at least to imagine there ever being a switch to phonetic-alphabet-only writing |
| 22:26 | <AryehGregor> | You mean because of too many homophones that have different kanji? If so, how are those distinguished in speech? |
| 22:29 | <MikeSmith> | AryehGregor: yeah, in part because of the homophones |
| 22:29 | <MikeSmith> | they are distinguished in speech by context |
| 22:30 | <AryehGregor> | But in writing, I guess it's harder to discern context, which is why written text tends to be much more precise and thought-out in any language. |
| 22:30 | <AryehGregor> | Interesting. |
| 22:30 | <MikeSmith> | and by inflections |
| 22:30 | <MikeSmith> | is "inflections" the right word? |
| 22:30 | <AryehGregor> | Tone? |
| 22:31 | <MikeSmith> | by the part of the multi-syllable word that is emphasized |
| 22:31 | <MikeSmith> | AryehGregor: not Tone in the same sense as Chinese |
| 22:31 | <MikeSmith> | in that regard Japanese is more atonal |
| 22:32 | <MikeSmith> | word "stress" is the term, I guess |
| 22:32 | <AryehGregor> | It doesn't have stress like in English, though, does it? Or at least the stress is much weaker? |
| 22:33 | <MikeSmith> | no, it does have stress pretty much just like in English |
| 22:33 | <MikeSmith> | I think |
| 22:33 | <AryehGregor> | It would be really interesting to learn more Japanese than the few words I know, but unfortunately, that's approximately item #136 on my List of Things to Learn if I Get Vast Quantities of Free Time. |
| 22:34 | <TabAtkins> | I don't even have a hundred things on my list. |
| 22:34 | AryehGregor | finds this on Google: http://japanese.about.com/blqow25.htm |
| 22:34 | <MikeSmith> | usual examples is two-syllable word like "hana", which is either "nose" or "flower", depending on where you put the stress -- or "hashi", which is either "chopsticks" or "bridge", depending on which syllable is stressed |
| 22:34 | <AryehGregor> | TabAtkins, a wise optimization, since on such a list you're unlikely to get to item #1. |
| 22:35 | <AryehGregor> | Until you retire, maybe, but it will probably be different then, so lazy initialization is a wise strategy. |
| 22:35 | <MikeSmith> | AryehGregor: http://japanese.about.com/blqow25.htm seems completely wrong to me |
| 22:35 | <MikeSmith> | it is not "pitch" accent at all |
| 22:36 | <MikeSmith> | it seems like whoever wrote that is actually thinking about Chinese |
| 22:36 | <MikeSmith> | " Japanese sentences are made up like a melody with rising and falling pitches. Unlike English's uneven rhythm, Japanese sounds like a steadily flowing stream." ... um, no |
| 22:37 | <AryehGregor> | How a language sounds is ridiculously subjective. It depends on what you're used to. |
| 22:37 | <AryehGregor> | Every language sounds like a fairly steady stream if you don't understand it well enough to mentally insert word breaks. |
| 22:38 | <AryehGregor> | (probably not *that* subjective, in that there are likely similarities between native speakers of the same language) |
| 22:38 | <AryehGregor> | Well, this says that English is stress-timed and Japanese is mora-timed, for what that's worth: http://en.wikipedia.org/wiki/Stress_(linguistics)#Timing |
| 22:38 | <AryehGregor> | Which actually sort of agrees with your last quote. |
| 22:39 | <Hixie> | TabAtkins: i gave up having a list at all when i realised i had enough things on my "important urgent" list that i'd never have free time :-) |
| 22:39 | <MikeSmith> | like anywhere else, spoken Japanese sounds radically different depending on who's speaking, and where they are from |
| 22:39 | <TabAtkins> | I've mostly given up, too, and instead just maintain my "video games to play in my free time" list, which I can *barely* keep up with. |
| 22:39 | <Hixie> | BC2 |
| 22:40 | <Hixie> | (and P2 and HL3E3 when they finally comes out, but valve hates me so they might never release them) |
| 22:40 | <Hixie> | (and GTA5, if that ever exists) |
| 22:40 | <TabAtkins> | OMG P2 |
| 22:40 | TabAtkins | gushes. |
| 22:40 | <AryehGregor> | BC2? |
| 22:40 | <TabAtkins> | Bad Company 2 |
| 22:40 | <AryehGregor> | Never heard of it. |
| 22:40 | <Hixie> | it's pretty fun |
| 22:40 | <AryehGregor> | (why does Valve hate you?) |
| 22:40 | <Hixie> | they hate us all! |
| 22:40 | <MikeSmith> | people from northern Japan, when they speak in dialect, I have absolutely no idea what they are saying -- it sounds like a completely different language but they are in fact speaking Japanese with mostly normal grammar, but some different words and a very different way of pronouncing and rhythm |
| 22:40 | <Hixie> | if they don't hate us, why haven't they released ep3 yet!!! |
| 22:40 | Hixie | cries |
| 22:40 | <TabAtkins> | Valve Time is a cross between marketing time and infinity. |
| 22:41 | MikeSmith | don't know what "mora-timed" is and goes to read |
| 22:41 | <AryehGregor> | MikeSmith, kind of like a Scottish brogue? |
| 22:41 | <Hixie> | MikeSmith: i have that sometimes with both french and english accents |
| 22:41 | <MikeSmith> | AryehGregor: yeah, kinda like that |
| 22:42 | <MikeSmith> | Hixie: you have an accent when you speak French? |
| 22:42 | MikeSmith | wondering what Hixie sounds like in French |
| 22:42 | <Hixie> | MikeSmith: no i mean i have that problem understanding other people's french and english when they speak with some of the more obscure accents |
| 22:42 | <Hixie> | there's some parts of switzerland and france where i'll be damned if i can understand a word they're saying |
| 22:43 | <Hixie> | same with english and some parts of the UK or US |
| 22:43 | <MikeSmith> | yeah |
| 22:43 | jgraham | assumes Hixie has *some* accent |
| 22:43 | <Hixie> | (and i'm not even talking about the weird dialects like cockney rhyming slang) |
| 22:43 | <AryehGregor> | http://en.wikipedia.org/wiki/Japanese_pitch_accent |
| 22:43 | <AryehGregor> | It sounds like Japanese stress is pitch-based, not loudness-based as in English? |
| 22:43 | <MikeSmith> | my accent varies based on who I'm talked to and where I am |
| 22:43 | <hober> | there's a public-html joke in here somewhere... |
| 22:44 | <Hixie> | what should i call the event that fires when a cue becomes active? |
| 22:44 | <hober> | cuechange? |
| 22:45 | <TabAtkins> | So, hey, I'm going to be giving a 30min tech talk on Wednesday about HTML5, focusing on hackable stuff. Planning on doing <video>/<canvas> stuff. What are the coolest demos of things that J. Random Hacker would like to do with those, but are simple enough to explain in 30 min? |
| 22:45 | <jgraham> | oncue not at all confusing |
| 22:45 | <AryehGregor> | TabAtkins, the IE9 page has tons of them! |
| 22:45 | <Hixie> | i need an event for when the cue becomes active and an event for when the cue becomes inactive |
| 22:45 | <Hixie> | to fire on the cue itself |
| 22:45 | <TabAtkins> | AryehGregor: Ooh, I'll check. |
| 22:45 | <Hixie> | and then an event to fire when the active cues change, to fire on the media element |
| 22:45 | <AryehGregor> | (particularly, it has one when IE9 performs great and all other browsers behave like molasses) |
| 22:45 | <Hixie> | onactive, oninactive, oncuechange? |
| 22:46 | <jgraham> | Hixie: seems good |
| 22:46 | <hober> | cue{,de}activate & cuechange? |
| 22:46 | <MikeSmith> | AryehGregor: hmm, yeah, it is also about pitch, now that I think more -- it's partly about drawing vowel sounds out longer and rising tone .. but that seems somehow quite different to me than how pitch is in Chinese .. maybe it's not so different actually |
| 22:46 | <Hixie> | oncuedeactivate is long |
| 22:46 | <hober> | true |
| 22:47 | <AryehGregor> | MikeSmith, it's practically impossible to actually notice these things yourself. It just seems natural, it's hard to think about. But when you think about stuff that linguists say, you realize that they're right. |
| 22:47 | <AryehGregor> | (in my experience with English, at least) |
| 22:48 | <AryehGregor> | Phonology is practically unnoticeable. Our brain only gets passed the phonemes once the lower levels sort out what the phones map to. |
| 22:48 | <AryehGregor> | Or something like that. |
| 22:48 | <AryehGregor> | (are you a native Japanese speaker?) |
| 22:48 | <MikeSmith> | no, not a native speaker |
| 22:49 | <AryehGregor> | My mother was born in Japan to American parents, after the war. Lived there till she was four. |
| 22:49 | <MikeSmith> | but most of my Japanese learning has come more from just listening and speaking |
| 22:49 | <AryehGregor> | I learned a bunch about Japan when I was young mainly for that reason. |
| 22:49 | <jgraham> | There can't be many native Japanese called Mike Smith |
| 22:49 | <MikeSmith> | where did she live? do you know? maybe Tachikawa area? |
| 22:49 | hober | minored in Japanese, but I've lost virtually all of it since then, from lack of use |
| 22:50 | <AryehGregor> | jgraham, yeah, I thought that too. |
| 22:50 | <MikeSmith> | well, my daughter is a Smith also, and doesn't look very Japanese, but she's a native speaker (in fact, unfortunately so far, she doesn't speak much English at all) |
| 22:50 | <AryehGregor> | I don't know where she lived. Her father taught in a university somewhere. He taught about Japanese art to the Japanese, in English, with a heavy German accent. |
| 22:50 | <MikeSmith> | heh |
| 22:51 | <Hixie> | MikeSmith: how old is she? |
| 22:51 | <MikeSmith> | she is 12 now |
| 22:51 | <Hixie> | ah |
| 22:51 | <MikeSmith> | I learn a lot of Japanese from her |
| 22:51 | <MikeSmith> | she's a good explainer |
| 22:51 | <AryehGregor> | My sister moved to Israel, and my parents made it very clear that the grandkids had better speak fluent English, or else. |
| 22:51 | <MikeSmith> | but the side effect is that when I speak Japanese, I sound like a 12-year-old Japanese girl |
| 22:51 | <AryehGregor> | Heh. |
| 22:51 | <jgraham> | MikeSmith: Sure there could be some. But historically (and presently?) the Japanese have had rather strict immigration policies |
| 22:52 | <MikeSmith> | when my daughter was 8, I spoke more like an 8-year-old Japanese girl... so, I'm making progress |
| 22:53 | <jgraham> | I heard a story the other day about some (English? Swedish?) parents who moved to Portugal, put their kid in daycare a lot and ended up with a kid that spoke only Portugese, whilst they knew no Portugese |
| 22:53 | <jgraham> | Kinda sad |
| 22:53 | <AryehGregor> | MikeSmith, so eventually you'll sound exactly like an adult Japanese woman? |
| 22:53 | <AryehGregor> | jgraham, that's why you keep them at home a lot and speak exclusively in your native language there. |
| 22:53 | <AryehGregor> | Only way to do it. |
| 22:54 | <MikeSmith> | jgraham: immigration policies in Japan are slowly being relaxed.. Japanese birthrate is such that everybody is realizing increased immigration is going to be needed in order to sustain things |
| 22:54 | <AryehGregor> | At the age of four, my mother apparently talked to the maid in Japanese, but she forgot it all. |
| 22:54 | <jgraham> | AryehGregor: It seems unfortunate not to end up with bilingual children in that situation |
| 22:55 | <AryehGregor> | It's like one child per mother average, right? Population halving every generation. |
| 22:55 | <MikeSmith> | AryehGregor: yeah, eventually :) but we are a few years away from that yet |
| 22:55 | <jgraham> | Since being bilingual is supposed to be generally awesome |
| 22:55 | <MikeSmith> | (about sounding like a grown-up) |
| 22:55 | <AryehGregor> | jgraham, they learn the local language in school or whatever, and your language at home. Everyone wins. |
| 22:55 | jgraham | is far too monolingual |
| 22:55 | <Hixie> | MikeSmith: i left switzerland when i was 10, so i speak french like a ten year old kid, if that makes you feel better :-) |
| 22:55 | <jgraham> | AryehGregor: Exactly |
| 22:56 | <MikeSmith> | AryehGregor: the other thing is that people on average live a very long time in Japan, and remain healthy, so the population is really aging |
| 22:56 | <MikeSmith> | Hixie: heh :) |
| 22:56 | <AryehGregor> | Longest average lifespan in the world. |
| 22:56 | <jgraham> | My Swedish is marginally worse than my three-year-old not-really-niece |
| 22:57 | <AryehGregor> | Not-really-niece? |
| 22:57 | <jgraham> | s/marginally// |
| 22:57 | <jgraham> | AryehGregor: Well she is my g/f's niece |
| 22:57 | <AryehGregor> | Ah. |
| 22:57 | <Hixie> | how about onentercue and onexitcue, rather than onactive and oninactive |
| 22:57 | <AryehGregor> | Possible-eventual-niece-in-law? |
| 22:57 | <jgraham> | (her sister in law is Swedish, but lives in England) |
| 22:58 | <Hixie> | or just onenter and onexit |
| 22:58 | <jgraham> | AryehGregor: Nah, don't believe in marriage |
| 22:58 | <Hixie> | any opinions? |
| 22:58 | <AryehGregor> | :/ |
| 22:58 | <jgraham> | (unless it becomes a legal necessity) |
| 22:58 | <MikeSmith> | Hixie: onenter and onexit are easier to read than onentercue and onexitcue, I think |
| 22:58 | <AryehGregor> | It's totally unclear that they're cue-related, though. |
| 22:58 | <MikeSmith> | true |
| 22:58 | <Hixie> | MikeSmith: k |
| 22:58 | <AryehGregor> | I'd go for something with "cue" in it. |
| 22:59 | <Hixie> | should i fire oncuechange on each TimedTrack whose cues changed (and maybe on the <track> element also?), or should I just fire it on the <video> element? |
| 22:59 | <Hixie> | (there's no details data with the event, if you want data you have to hook into each cue you care about) |
| 23:00 | <Hixie> | i guess per-track is what you need to do slides and stuff |
| 23:00 | <TabAtkins> | Individual cues, and per-track. |
| 23:00 | <TabAtkins> | Do they bubble? |
| 23:01 | <Hixie> | no |
| 23:01 | <TabAtkins> | Good. |
| 23:01 | <Hixie> | nothing to bubble to, these aren't DOM nodes |
| 23:01 | <TabAtkins> | Oh, right. |
| 23:01 | <TabAtkins> | I'm probably dumb, but how do you hook into a cue to grab events if it's not a DOM node? |
| 23:02 | <Hixie> | video.tracks[n].cues[m].onenter = mycallback; |
| 23:02 | <TabAtkins> | Got it. |
| 23:02 | <Hixie> | and video.tracks[n].oncuechange = mycallback; |
| 23:02 | <TabAtkins> | Then yeah, on the tracks/<track>, and on the cue itself. |
| 23:03 | <Hixie> | (or trackElement.track.cues[n].onenter = mycallback) |
| 23:03 | <TabAtkins> | Though... Without any details data in the event, not quite sure what you'd do with the event on the track. |
| 23:03 | <Hixie> | well you can easily get the list of active tracks |
| 23:03 | <Hixie> | so if you know there's only one active one at a time (e.g. the active slide) then you just do track.activeCues[0].whatever |
| 23:03 | <TabAtkins> | Ok, but it doesn't help you do slides, since a single slide can easily have multiple cues. |
| 23:04 | <Hixie> | by slides i mean if you have a track (kind=metadata) whose data is just the IDs of the slides to show for each segment of the audio track |
| 23:05 | <TabAtkins> | Oh, I see. And would *that* information be gettable from the event fired at the track? |
| 23:05 | MikeSmith | finds that all this talk of ’cues is making him hungry, steps out to grab some breakfast |
| 23:05 | <Hixie> | <track kind=metadata src="myslides.srt" oncuechange="changeToSlide(track.activeCues[0].getCueAsSource())"> |
| 23:06 | <Hixie> | assuming the text of each of your cues is whatever changeToSlide() is expecting as an argument |
| 23:06 | <TabAtkins> | Got it. (I don't know enough about the api you're defining right now to ask intelligent questions.) |
| 23:06 | <Hixie> | heh |
| 23:06 | <Hixie> | neither do i :-P |
| 23:06 | <Hixie> | so dumb questions will have to suffice |
| 23:07 | <Hixie> | right now i'm porting the text i'd commented out for cue ranges to work with the new API |
| 23:07 | <Hixie> | i knew i'd reuse it eventually, glad i didn't delete it! |
| 23:13 | <AryehGregor> | http://www.imperialviolet.org/2010/06/25/overclocking-ssl.html |
| 23:13 | <AryehGregor> | Yay for SSL that takes zero extra round trips in the common case. |
| 23:19 | <TabAtkins> | Cool! |
| 23:19 | <TabAtkins> | Now, if only browsers weren't retarded about self-signed certs. |
| 23:21 | <AryehGregor> | I'm waiting for DNSSEC on that score. |
| 23:26 | <AryehGregor> | What sites allow all Flash embeds? |
| 23:26 | <AryehGregor> | That sounds crazy. |
| 23:27 | <AryehGregor> | Or maybe not totally crazy? I don't know the rules for how <object> works, actually. |
| 23:59 | <MikeSmith> | kennyluck: http://www.readwriteweb.com/archives/facebook_the_semantic_web.php |
| 23:59 | <MikeSmith> | if you've not found it already |
