/irc-logs / w3c / #html-wg / 2007-04-09 / end

Options:

# Session Start: Mon Apr 09 00:00:00 2007
# Session Ident: #html-wg
# [00:00] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
# [00:05] * Joins: gavin (gavin@74.103.208.221)
# [00:26] * Joins: myakura (myakura@60.239.122.32)
# [00:31] * Joins: mjs (mjs@64.81.48.145)
# [00:53] * Joins: asbjornu (asbjorn@84.48.116.134)
# [00:59] * Quits: dbaron (dbaron@71.198.189.81) (Quit: 8403864 bytes have been tenured, next gc will be global.)
# [01:17] * Parts: hasather (hasather@81.235.209.174)
# [01:22] * Quits: Hixie (ianh@129.241.93.37) (Ping timeout)
# [01:24] * Joins: Hixie (ianh@129.241.93.37)
# [01:46] * Quits: Zeros (Zeros-Elip@69.140.48.129) (Quit: Leaving)
# [02:07] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
# [02:11] * Joins: gavin (gavin@74.103.208.221)
# [02:22] * Joins: karl (karlcow@128.30.52.30)
# [02:38] <Philip> karl: I've been doing a bit of work on canvas tests at http://canvex.lazyilluminati.com/tests/tests/ - the "Annotated specification" part is possibly interesting
# [02:38] <Philip> That part of the spec seems almost entirely self-contained (except for some bits about parsing numbers, I think), so it looks like it would be easy to separate it out into its own specification
# [02:39] <Philip> (I don't expect that'd make it any easier to write/test/implement than with it being a self-contained section in the middle of the HTML5 spec, but presumably it would help in getting that part finished sooner and making the rest of HTML5 smaller)
# [02:40] <karl> interesting
# [02:41] <karl> very cool work
# [02:41] <karl> you just send it to the list
# [02:41] <Philip> I'm not on the list, unfortunately
# [02:42] <Philip> (Maybe I should try joining some time, but I don't have much time to be involved at the moment :-( )
# [02:55] * Joins: olivier (ot@128.30.52.30)
# [03:12] <Hixie> wow, microsoft really want to introduce a new rendering mode
# [03:12] <Hixie> can you imagine what life will be like in 10 years if they do this with every release
# [03:19] <mjs> even if they have the resources to do this for every release, I don't think anyone else does
# [03:19] <mjs> I guess they feel really burned by the IE7 experience
# [03:21] <Hixie> so every web page has to start with an "opt-in" to use the latest IE technology?
# [03:22] <Hixie> that's not gonna work
# [03:22] <Hixie> in fact it's directly against one of our principles, maybe even two
# [03:22] <mjs> using the HTML version as an IE-specific standards compliance opt-in is also abusive of other browser
# [03:24] <Hixie> the whole thing is ridiculous
# [03:24] <Philip> For fairness to the others, the version 'number' should be an infinite-dimensional vector, with a separate axis for each browser than does/will exist
# [03:44] <mjs> ok, I want to reply to Chris Wilson but I think my response may be too inflammatory
# [03:44] * Joins: foca (foca@190.64.4.252)
# [03:44] <mjs> would anyone like to preview it for me?
# [03:44] <karl> mjs send it to yourself, let it lie for 24h
# [03:45] <karl> do not work on sundays ;) and read it again on monday
# [03:45] <mjs> karl: I can save it in my drafts folder, but I don't think it will be more clear come Monday
# [03:46] <mjs> it wasn't written in anger or anything
# [03:46] <mjs> I just asked if Microsoft doesn't plan to ever conform to any spec the HTMLWG comes out with, why we should listen to their input
# [03:47] <mjs> which seems like a rude thing to say, but also kind of a valid question
# [03:48] <karl> mjs: I guess it's part of the consensus building. Chris is not alone to be in favor of version numbers.
# [03:49] <karl> the best in this kind of cases
# [03:49] <karl> is not to target products or people
# [03:49] <karl> but technology
# [03:49] <mjs> he's saying he wants version numbers because for each version, Microsoft will stop fixing conformance bugs at some point
# [03:49] <mjs> and won't try to make those bugs errata to the spec
# [03:49] <mjs> that seems like he's saying the spec should have versions because Microsoft will never conform to it
# [03:49] <karl> I think the debate is interesting and when I read emails, it gives interesting views on both sides.
# [03:50] <mjs> which seems like an invalid argument to me
# [03:50] <mjs> I don't even care that much about versioning, just his apparent claim that Microsoft plans to have an unbounded number of undocumented quirks modes
# [03:50] <karl> so the better is to make an argumentation with two columns and with the pros and cons for each side.
# [03:51] <mjs> I think the problem is that what seems like a pro to him (ability of Microsoft to lock in their particular set of bugs forever) seems like a con to other people
# [03:52] <karl> versioning existing in any language environment, in life, etc. We can't understand English from 500 years ago without a dictionary of this time, because semantics evolve. Or we have to be ready to do misunderstanding or even worse wrong interpretation.
# [03:52] <karl> it is specifically we have a live language that it makes it difficult :)
# [03:52] <karl> on another side,
# [03:52] <mjs> but Shakespeare's plays are not tagged with a version number
# [03:52] <karl> yes
# [03:52] <karl> they are
# [03:52] <karl> definitely
# [03:52] <karl> it is called a year
# [03:53] <karl> :)
# [03:53] <karl> I think on both side on HTML versioning there are good arguments
# [03:53] <karl> It would be worthwhile to not take positions, but more to write down the pros/cons.
# [03:54] * Parts: foca (foca@190.64.4.252)
# [03:54] <karl> for me I haven't made up my mind. :) and it doesn't matter what I think ;) I must stay neutral. that is my burden ;)
# [03:54] <Philip> Is his desire for "100.000% backward-compatible" actually possible (or nearly so), with the approach HTML5 has taken?
# [03:55] * Quits: myakura (myakura@60.239.122.32) (Quit: Leaving...)
# [03:56] <Philip> (I don't quite understand his comments about deprecated features - that seems like a purely author-oriented issue (even if the deprecated feature is removed entirely), and the spec can (and must, if it's going to be compatible with old content) still specify the UA behaviour for those deprecated/removed features)
# [04:01] <Hixie> mjs: sure, i can check your mail for appropriateness
# [04:08] <nickshanks> shouldn't UA behaviour be contained in a separate, kept-up-to-date document from the HTML5 spec
# [04:09] <Dashiva> mjs: I sent a very short mail along some of those lines just now, although it might not be keen enough to make the point (punny)
# [04:14] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
# [04:19] * Joins: gavin (gavin@74.103.208.221)
# [04:20] * Quits: Philip (excors@80.177.163.133) (Quit: Philip)
# [04:22] * Joins: Shunsuke (kuruma@133.27.63.42)
# [05:02] <Lachy> I really don't understand Chris. I think he's just being stubborn and ignorant
# [05:06] * Quits: Shunsuke (kuruma@133.27.63.42) (Ping timeout)
# [05:12] * Joins: myakura (myakura@60.239.122.32)
# [05:17] * Joins: htmlr (htmlr@203.206.237.84)
# [05:18] * Joins: Zeros (Zeros-Elip@69.140.48.129)
# [05:18] * Quits: Zeros (Zeros-Elip@69.140.48.129) (Quit: Leaving)
# [05:23] * Hixie hopes chris isn't one of lachlan's friends on twitter
# [05:23] <Lachy> nope, he's not
# [05:24] <Hixie> he's only one step removed then :-)
# [05:24] <Lachy> I didn't call him names, just insulted his ideas
# [05:30] * Quits: htmlr (htmlr@203.206.237.84) (Quit: htmlr)
# [05:32] <Hixie> i've found people in the htmlwg from the w3c side to be especially thin-skinned, fwiw
# [05:42] * karl would recommend everyone to keep a correct tone on a group channel.
# [05:43] <Lachy> karl: agreed. That's why I used twitter
# [06:12] * Quits: myakura (myakura@60.239.122.32) (Quit: Leaving...)
# [06:22] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
# [06:27] * Joins: gavin (gavin@74.103.208.221)
# [07:34] * Joins: myakura (myakura@60.239.122.32)
# [07:45] * Quits: Yudai (Yudai@59.147.29.149) (Quit: SIGTERM received; exit)
# [08:00] * Joins: Zeros (Zeros-Elip@69.140.48.129)
# [08:23] * Joins: Yudai (Yudai@59.147.29.149)
# [08:28] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
# [08:33] * Joins: gavin (gavin@74.103.208.221)
# [08:49] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
# [09:30] * Joins: anne (annevk@83.82.206.111)
# [09:37] * Quits: olivier (ot@128.30.52.30) (Quit: Leaving)
# [09:44] * Quits: karl (karlcow@128.30.52.30) (Quit: Where dwelt Ymir, or wherein did he find sustenance?)
# [10:20] * Joins: ROBOd (robod@86.34.246.154)
# [10:35] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
# [10:37] * Joins: erik (erik@84.29.164.153)
# [10:40] * Quits: Zeros (Zeros-Elip@69.140.48.129) (Quit: Leaving)
# [10:40] * Joins: gavin (gavin@74.103.208.221)
# [10:54] * Joins: marcos_ (chatzilla@203.206.31.102)
# [10:57] * Quits: marcos_ (chatzilla@203.206.31.102) (Ping timeout)
# [11:00] * Joins: hasather (hasather@81.235.209.174)
# [11:04] * Quits: erik (erik@84.29.164.153) (Quit: Bye bye)
# [11:16] * Joins: marcos__ (chatzilla@203.206.31.102)
# [11:16] * marcos__ is now known as marcos
# [11:22] * Quits: marcos (chatzilla@203.206.31.102) (Ping timeout)
# [11:32] <anne> 7 new IEs so far
# [11:32] <anne> today
# [11:35] <MikeSmith> anne - any names you recognize
# [11:36] <anne> nope
# [11:51] <MikeSmith> so have now reached the 300 member mark in one month
# [11:52] <MikeSmith> I would guess it'd be at least 400 by this time next month
# [11:52] * Joins: marcos__ (chatzilla@203.206.31.102)
# [11:52] * marcos__ is now known as marcos
# [11:54] <hsivonen> I don't see new Microsoft Corporation representatives on the participant list. are you talking about something else?
# [11:55] <hasather> hsivonen: anne meant Invited Experts by IE
# [11:55] <hsivonen> oh
# [12:18] <MikeSmith> I hope some developers from KDE/KHTML/Konqueror will join the WG
# [12:42] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
# [12:47] * Joins: gavin (gavin@74.103.208.221)
# [12:49] * Quits: sbuluf (du@200.49.140.20) (Ping timeout)
# [12:52] * Joins: Philip (excors@80.177.163.133)
# [14:12] * Quits: myakura (myakura@60.239.122.32) (Ping timeout)
# [14:13] * Joins: myakura (myakura@60.239.122.32)
# [14:32] * Quits: myakura (myakura@60.239.122.32) (Quit: Leaving...)
# [14:40] <MikeSmith> just read Art Barstow's "Shaping the future of secure Ajax mashups" posting to the public-appformats list
# [14:41] <MikeSmith> http://www.w3.org/mid/44451BAE-376D-4E7F-B956-22DF70845B13@nokia.com
# [14:42] <MikeSmith> was there ever any discussion on the whatwg list of Douglas Crockford's proposal for a <module> tag?
# [14:42] <MikeSmith> element
# [14:42] <MikeSmith> http://json.org/module
# [14:42] <anne> yeah, I think so
# [14:43] <anne> but it seems that cross-site XHR and cross document messaging cover both...
# [14:43] <MikeSmith> ah, yeah
# [14:43] <anne> no need for JSONRequest and <module>
# [14:43] <MikeSmith> yeah, I found the thread and reading it now
# [14:47] <MikeSmith> anne - Web API WG is working on spec'ing (or adopting spec for) cross-document messaging? or planning to?
# [14:50] * Joins: foca (foca@190.64.13.153)
# [14:50] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
# [14:53] * Parts: foca (foca@190.64.13.153)
# [14:55] * Joins: gavin (gavin@74.103.208.221)
# [15:18] <anne> nope
# [16:19] <mjs> what is <module>?
# [16:21] * Joins: h3h (bfults@66.162.32.234)
# [16:35] <MikeSmith> mjs - something that Douglas Crockford wrote a proposal for
# [16:35] <MikeSmith> http://json.org/module
# [16:36] <MikeSmith> http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2006-October/007522.html
# [16:51] <mjs> sounds like cross-document messaging will work just as well w/o having to introduce a new element
# [16:54] <mjs> I wish Doug Crockford would follow HTML standards work more closely if he wants to make proposals like that
# [16:55] <anne> he made that proposal on the whatwg list ages ago
# [16:56] <anne> i think since then he's sort of convinced that cross-doc is the way to go
# [16:56] <anne> but i'm not entirely sure
# [16:57] <mjs> October 2006 isn't all that long ago
# [16:57] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
# [16:57] <anne> fair enough
# [17:02] * Joins: gavin (gavin@74.103.208.221)
# [17:06] * Quits: mjs (mjs@64.81.48.145) (Quit: mjs)
# [17:29] * Joins: tylerr (tylerr@66.195.32.2)
# [17:29] <tylerr> G'day everyone.
# [17:34] * anne just made http://wiki.whatwg.org/wiki/Changes_from_HTML4#New_Attributes
# [17:42] <MikeSmith> anne - have you or anybody else written up a brief description of the html5 parsing algorithm?
# [17:42] <MikeSmith> briefer than reading the whole wepapps1.0 spec, I mean
# [17:42] <anne> how would that work?
# [17:42] <hsivonen> MikeSmith: I have written a very short goal-level description twice
# [17:43] * hsivonen looks them up
# [17:43] <MikeSmith> hsivonen - also wanted to ask if your PDF version of the webapps1.0 spec is linked to on the site anywhere
# [17:43] <hsivonen> http://hsivonen.iki.fi/thesis/html5-conformance-checker.xhtml#p122
# [17:44] <hsivonen> (unstable fragment id)
# [17:44] <hsivonen> MikeSmith: it is linked to on my site. not on whatwg.
# [17:44] <MikeSmith> hsivonen - OK, thanks
# [17:45] <nickshanks> what is the reasoning for allowing <font> when inserted by WYSIWYG editors. there are no such editors for HTML5 yet, so any ones that come along can be made without using <font>
# [17:46] <anne> <font> isn't done yet
# [17:46] <hsivonen> MikeSmith: the other one is in http://groups.google.fi/group/comp.infosystems.www.authoring.html/msg/963cbf7f54567bff
# [17:46] * Joins: dbaron (dbaron@71.198.189.81)
# [17:47] <hsivonen> MikeSmith: I think it would make sense to generate Letter and A4 PDFs is a post-commit hook on the Dreamhost server
# [17:47] <hsivonen> MikeSmith: but that requires the appropriate Prince license
# [17:48] <MikeSmith> hsivonen - yeah, that would be good. I think Mike Day would probably be happy to give you license
# [17:48] <anne> nickshanks, " Usage varies widely by language." sounds like commentary
# [17:48] <MikeSmith> would be good to have the PDF linked to straight from the online (HTML) spec
# [17:49] <anne> nickshanks, the quotes from those elements (<b> too) is taken literally from the HTML5 spec...
# [17:49] <nickshanks> well i am trying to point out that "ship names" is just English
# [17:49] <anne> s/quotes/description/
# [17:49] <nickshanks> oh right. shame that isn't a wiki too
# [17:50] * anne shrugs
# [17:50] <anne> MikeSmith, arrange it :)
# [17:54] <MikeSmith> anne - I'd be happy to. Is there currently any kind of build that runs when the spec is updated (using make or ant or whatever)?
# [17:54] <anne> if you arrange some script that makes a PDF I'm sure Hixie can integrate it
# [17:54] <MikeSmith> OK
# [17:55] <MikeSmith> hsivonen - you want to contact Prince/Mike Day and ask about the license? Or you want me to?
# [17:56] <anne> lol, Dao is funny
# [17:56] <MikeSmith> hsivonen - and I don't find any link for the PDF on your home page
# [17:56] <anne> he suggests that XSLT can be used but then he never used it...
# [17:57] <Dashiva> Is that the html60 guy?
# [17:57] <anne> no
# [17:57] <anne> dao@design-noir.de
# [18:00] <MikeSmith> ah, "Printing Web Apps 1.0" . I see it now
# [18:01] <hsivonen> MikeSmith: should be under blogish notes
# [18:02] <MikeSmith> hsivonen - yeah, got it
# [18:02] <hsivonen> MikeSmith: it would be nice if you contacted him
# [18:03] <MikeSmith> OK, will do
# [18:03] <MikeSmith> do you guys have shell access to the Dreamhost server?
# [18:03] <hsivonen> I don't.
# [18:03] <anne> only Hixie I guess
# [18:04] <hsivonen> I think Hixie is the only one with the required level of access
# [18:04] <anne> heh, people did indeed want to remove <div> (re: last post)
# [18:04] <anne> and <b> and <i> are back in
# [18:05] <anne> and all for reasons of more "semantic" markup
# [18:05] <hsivonen> to my knowledge, Lachy and zcorpan are sandboxed away from the spec virtual host
# [18:05] <MikeSmith> well, I guess we don't necessarily need to have the PDF build script run on the Dreamhost server
# [18:06] <anne> just give the script to Hixie; the main thing is the license for Prince
# [19:11] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
# [19:16] * Joins: gavin (gavin@74.103.208.221)
# [19:19] * Joins: mjs (mjs@17.255.99.15)
# [19:20] <mjs> good morning everybody
# [19:20] <tylerr> Hey there mjs, how was the weekend?
# [19:20] <mjs> it was fun, I went to a rodeo
# [19:20] <mjs> the bull-riding event reminded me a lot of working in a standards committe
# [19:20] <mjs> e
# [19:24] <tylerr> Haha nice!
# [19:28] <tylerr> Well hopefully everything will get wrangled in here pretty soon.
# [19:28] <tylerr> I'm waiting for things to get a bit more organized.
# [19:36] * Joins: kingryan (rking3@66.92.187.33)
# [19:56] * Quits: dbaron (dbaron@71.198.189.81) (Ping timeout)
# [20:08] * anne is curious about Microsoft's versionining pov
# [20:08] <anne> versioning*
# [20:10] * mjs hopes the essay won't just be an attempt to avoid being clear on the issue
# [20:10] <anne> yeah...
# [20:11] <anne> I thought your questions already framed it enough...
# [20:11] <anne> guess not
# [20:12] <mjs> I thought it was a pretty simple question, yeah
# [20:18] * Joins: dbaron (dbaron@63.245.220.242)
# [20:24] <tylerr> We work with MS pretty closely, everything is about keep things foggy. :-)
# [20:24] <tylerr> s/keep/keeping
# [20:25] <tylerr> But for the work we do it needs to be since everything changes so much on the fly, so it works to our advantage.
# [20:31] <tylerr> Anyway, any progress made this weekend? I've been away for a few days.
# [20:39] * Quits: kingryan (rking3@66.92.187.33) (Quit: kingryan)
# [20:45] * Joins: kingryan (rking3@66.92.187.33)
# [20:47] * Quits: kingryan (rking3@66.92.187.33) (Quit: kingryan)
# [20:48] * Joins: kingryan (rking3@66.92.187.33)
# [21:10] * Quits: tylerr (tylerr@66.195.32.2) (Quit: Leaving)
# [21:15] * Joins: Zeros (Zeros-Elip@67.154.87.254)
# [21:18] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
# [21:21] <Hixie> in Priority of Constituencies, i'd like to suggest an addition
# [21:21] <Hixie> it currently says "In case of conflict, consider users over authors over implementors over specifiers"
# [21:21] <Hixie> I'd like to add "over theoretical purity" at the end there.
# [21:24] * Joins: gavin (gavin@74.103.208.221)
# [21:24] <Hixie> to prevent people from splitting the spec when all it would do is make more work for the editor(s)
# [21:25] <Hixie> e.g. innerHTML, or setTimeout or Window...
# [21:26] <Zeros> So you're saying just add innerHTML to the spec and be done with it?
# [21:26] <Hixie> for instance
# [21:26] <Hixie> instead of making two specs
# [21:33] * Parts: asbjornu (asbjorn@84.48.116.134)
# [21:36] <anne> yes, same for <canvas>
# [21:39] <Zeros> A separate spec for the canvas API would be beneficial since they could update that separately and leave HTML alone
# [21:39] <Philip> Same for 3D canvas too?
# [21:39] <Zeros> We wouldn't need HTML5.01 to get extra canvas features
# [21:40] <anne> it would just be part of HTML6
# [21:41] <Zeros> The larger the spec the more difficult it is to modify it later and get it pushed out
# [21:41] <Zeros> How many years did it take to get HTML5 going?
# [21:42] <Zeros> We've had 3 versions of CSS, 3 DOM levels
# [21:42] <anne> Get going where?
# [21:42] * Joins: jgraham (jgraham@81.178.84.187)
# [21:43] <Zeros> anne, anywhere, HTML4.01 was the last version of HTML that was actually implemented since 1999
# [21:43] <anne> HTML5 started in 2004
# [21:43] <Philip> People seem happy to implement and use features from heavily-under-development specs, so I suppose it wouldn't matter if new features are in an HTML6 working draft that won't be a recommendation until a decade later because it's being held up by all the other unrelated features - if it's good, people will use it regardless of the official status
# [21:43] <Zeros> anne, WHATWG started a group in 2004, the W3 was not working on it
# [21:43] <anne> people will use interoperable implemented features
# [21:44] <anne> Zeros, so?
# [21:44] <Zeros> anne, I fail to see how making the canvas API separate so it can evolve faster and be implemented separately is a problem
# [21:44] <anne> it started in 2004 and not earlier because the W3C didn't do anything about it, btw
# [21:44] <Zeros> anne, clearly, and that's till 5 years later
# [21:44] <anne> Zeros, I don't see it evolving faster separately
# [21:44] <anne> do you have an editor?
# [21:45] <Zeros> Maybe Apple should take it, since its their project, heh.
# [21:45] <anne> right...
# [21:45] <Zeros> And there's already talk about Canvas3D
# [21:45] <anne> next topic
# [21:45] <Zeros> anne, that was sarcasm, do you have an editor for HTML6?
# [21:45] <anne> yeah, both Mozilla and Opera have experimental implementations
# [21:46] <anne> Zeros, yes, Hixie, by the looks of it
# [21:46] <Zeros> anne, you're advocating that Canvas should be left alone and the spec not changed
# [21:46] * Quits: ROBOd (robod@86.34.246.154) (Quit: http://www.robodesign.ro )
# [21:46] <Zeros> so I don't why we need a separate editor now anyway
# [21:46] <anne> (of Canvas 3D)
# [21:46] <anne> Zeros, I'm not
# [21:46] <anne> Zeros, Hixie stated he doesn't want to edit several specs, which makes sense, as it's a lot more work
# [21:47] <Zeros> Last I checked Canvas wasn't hixie's responsibility
# [21:47] <anne> ?
# [21:47] <Zeros> "Regarding <canvas>. <canvas> is mostly done. There are three implementations in browsers that are becoming more interoperable each release. " is what you said
# [21:47] <anne> yes
# [21:48] <Zeros> maciej said that the spec is pretty much finalized
# [21:48] <anne> yes
# [21:48] <Zeros> so how is it more work for Hixie?
# [21:48] <Zeros> To put the canvas in a separate API spec that can be implemented outside HTML5 support and modified later
# [21:48] <Philip> There would be some more work where the <canvas> definition depends on other parts of the HTML5 spec
# [21:49] <anne> you need a whole new template for your new spec, all the definitions need to be redone to be independent of the HTML5 doc, etc.
# [21:49] <Zeros> Philip, where? Canvas doesn't depend on it now.
# [21:49] <anne> anyway, not worth discussing until someone finds an editor
# [21:50] <Philip> I think it's just in the width/height attributes, where it interacts with the DOM
# [21:50] <Philip> (so that's only a pretty small area)
# [21:50] <Zeros> Philip, so reference the relevant spec like the HTML4 spec does?
# [21:51] <Zeros> Both the DOM and CSS evolved independently of the HTML spec. If they were one giant spec we'd have HTML6.25 or some such
# [21:51] <Philip> I'd assume it can't normatively reference the HTML5 spec until the HTML5 spec is done. I have no idea if there's some other spec which defines the same things that could be used instead
# [21:52] <anne> CanvasRenderingContext2D.canvas would be a problem
# [21:52] <anne> the drawImage() methods would be problematic
# [21:53] <Zeros> how so?
# [21:53] <anne> etc.
# [21:53] <Zeros> What makes it easier to put them inside the HTML5 spec?
# [21:53] <anne> also, see above
# [21:54] <Philip> Ah, the .canvas would be a problem if the <canvas> bit is separate from the Context2D bit, since the Context2D needs to know about the existence of HTMLCanvasElement
# [21:56] <Zeros> Philip, its just a rendering context, it can be specified to reference the DOM node. Like the DOM spec already does
# [21:56] <Philip> It seems to me like it'd just add unnecessary administrative overhead to maintain separate somewhat-parallel specs
# [21:56] <Zeros> Philip, HTMLSelectElement is in a separate spec from the <form> tag which it references with .form
# [21:57] <Zeros> This isn't a new idea
# [21:58] <anne> yes, and that spec is totally broken
# [21:58] <Zeros> What's broken about the DOM spec exactly?
# [21:58] <anne> lots of things
# [21:58] <Zeros> like what? :)
# [21:59] <Philip> It looks like HTMLSelectElement references HTMLFormElement with its .form, and that is in the same spec
# [21:59] <Zeros> Philip, So Context2D can't reference HTMLCanvasElement?
# [22:00] <Philip> (and HTMLFormElement appears to only vaguely say it's related to the <form> element from HTML)
# [22:00] <Zeros> If it was a separate spec it could be implemented for other things that don't necessarily apply directly to HTML as well.
# [22:00] <Zeros> The same way CSS applies to arbitrary XML and structured documents
# [22:00] <anne> atm it's pretty much tied to HTML
# [22:00] <anne> see my earlier references
# [22:00] <Zeros> If CSS was lumped in with the entire HTML spec that'd a be a problem
# [22:00] <anne> depends on how it was defined
# [22:01] <anne> s/was/is/
# [22:01] <Zeros> Then the same is true of placing the Canvas API in a separate spec
# [22:04] <Philip> I think I was thinking that if Context2D references HTMLCanvasElement, HTMLCanvasElement has to be defined somewhere; and if that's in the HTML5 spec (which seems the proper place for it), the Context2D spec would be held up by HTML5, so it couldn't be done any faster than if it was still part of HTML5
# [22:04] <anne> and vice versa
# [22:04] <anne> besides that it makes for editorial overhead
# [22:06] <Philip> The same problems probably wouldn't apply for a far-future Context3D spec, since HTML5 and HTMLCanvasElement will already be defined, and since HTML5 doesn't refer to Context3D (whereas it does refer back to Context2D, since that's a required part of <canvas>)
# [22:06] <anne> in fact, Opera already specced an extension and submitted that and based on that and other feedback <canvas> moved to v2
# [22:06] <Zeros> Philip, Canvas doesn't really need to refer to HTML at all
# [22:07] <Zeros> If the spec was written properly it would be implemented for anything outside HTML
# [22:07] <anne> why do you need it there?
# [22:07] <Zeros> anne, The same reason you need SVG or CSS outside HTML
# [22:07] <Philip> I'm not sure why it'd be useful outside HTML, since everything else already has 2D graphics APIs and OpenGL wrappers
# [22:08] <anne> yeah, it's very much about the element too
# [22:08] <Zeros> its just a 2d rendering context
# [22:08] <anne> which can be used by applications to provide bitmap drawing
# [22:08] <anne> or for games, etc.
# [22:08] <Zeros> Its slow as hell and a real processor hog for games atm.
# [22:09] <Zeros> or any realtime changes
# [22:09] <anne> what browser are you testing in?
# [22:09] <Philip> I can get ~20fps 320x240 3D out of <canvas>, which isn't bad compared to Doom (at least when Doom is on a processor that's ~30 times slower)
# [22:10] <Philip> (and I even have bilinear texture filtering, which Doom doesn't, though I don't actually want bilinear texture filtering because it's kind of slow :-( )
# [22:10] <Zeros> anne, Safari and Firefox, all things considered it uses about 80% of the processor to render a raster world.
# [22:11] <Zeros> Flash would be a much better solution than canvas, and has much better support
# [22:11] <Zeros> I don't imagine canvas is very fast in IE where it had to be emulated with JS either
# [22:12] <Philip> It's actually emulated with VML (using a JS wrapper to construct it), so the rendering is probably not much slower than in other browsers
# [22:12] <Philip> (though it totally doesn't follow the canvas spec at all, so it's unlikely any non-trivial games would work in it)
# [22:13] <Zeros> http://developer.mozilla.org/samples/raycaster/RayCaster.html gets me 75% of the processor when walking around
# [22:13] <Philip> Flash is probably a much better solution that <video> too - it has much better support, it has no problems with codecs, etc
# [22:13] <Zeros> Firefox 2
# [22:13] <Zeros> Philip, funny that
# [22:14] <Philip> except it's proprietary technology, which is why HTML people want to provide alternatives :-)
# [22:14] <Philip> That's a really rubbishly implemented raycaster :-p
# [22:14] <Philip> (http://canvex.lazyilluminati.com/ is mine)
# [22:14] <anne> I think that demo relies on some old canvas quirks
# [22:14] <anne> too
# [22:16] <anne> pretty cool Philip
# [22:16] <Philip> (It takes ~100% CPU, because it loops as fast as it can - it could intentionally reduce the framerate to save a bit)
# [22:18] <Philip> Hmm, looks like that RayCaster.html isn't doing any kind of beginPath so it's going to be stroking longer and longer paths all the time, which will kill performance
# [22:19] <Hixie> good lord let's not put canvas into another spec
# [22:19] <Hixie> i have enough work as it is
# [22:19] <Zeros> No one said you need to be the sole editor
# [22:20] <Hixie> unless you have another volunteer...
# [22:20] <Hixie> or i should say, until you have another volunteer
# [22:20] <Philip> It seems rather hard finding other good editors, unfortunately
# [22:20] <Zeros> Philip, I get redirected to http://canvex.lazyilluminati.com.nyud.net:8080/83/play.xhtml which gets me a server not found error
# [22:20] <Zeros> oh, i see the little uncached link
# [22:21] <Philip> Oh, that caching system doesn't really seem to work at the moment...
# [22:22] <Zeros> At least the other one is smart enough to idle the processing while you're not moving around
# [22:22] <Zeros> Yours chews 76% standing still, 82% moving around
# [22:22] <Zeros> 5fps on your counter too
# [22:22] <Philip> That's because I've got animated sprites and moving doors, and it's not clever enough to detect that nothing is visibly changing
# [22:24] <Zeros> still a pretty cool doom clone
# [22:24] <Philip> ((Removed the cached link, since it's not very useful))
# [22:24] <anne> didn't work for me either
# [22:24] <Zeros> Maybe when Tamarin is added to Firefox we'll see that get more practical
# [22:24] <Philip> In the starting position, I get ~15fps in Firefox 3
# [22:24] * Joins: mw22 (chatzilla@63.245.220.228)
# [22:24] <anne> Philip, you should use 'new Audio()' for shooting sounds :)
# [22:25] <Philip> It's spending almost all its time drawing bits of image onto the screen, not in the JS code
# [22:26] <anne> that should go for every browser methinks
# [22:27] <Philip> anne: Then I'd have to add guns and things to shoot at - I had enough problems getting the collision detection to roughly work, so I decided not to spend ages adding more gameplay :-)
# [22:27] <anne> see http://people.opera.com/howcome/2006/canvasdemo/canvascape/shoot.html for shooting sounds and such
# [22:28] <anne> actually, it's prolly a jumping sound :)
# [22:28] <anne> (copy from http://www.abrahamjoffe.com.au/ben/canvascape/ with added Audio() stuff)
# [22:29] <Philip> (I get ~15fps in FF3, ~13fps in FF2, and ~11fps in Opera 9)
# [22:30] <anne> in your game I get 20fps in Opera 9
# [22:33] <Philip> The audio seems to work nicely
# [22:34] <Zeros> oh I'm getting 13fps in Opera 9
# [22:34] <Zeros> much smoother too
# [22:34] <Zeros> Firefox 2 must just be slow
# [22:34] <Philip> (Can't tell from that example whether it can mix multiple sounds nicely, though)
# [22:37] <Zeros> oh
# [22:37] <Zeros> that must be a bug
# [22:37] <Zeros> I walked through a wall in Safari
# [22:37] <Zeros> and now I'm getting 52fps
# [22:38] <Philip> (Hmm, it seems Opera doesn't mix sounds, even if I construct a new Audio object each time I play it)
# [22:38] <Philip> (Oh, wait, it mixes sounds but only if they have different filenames)
# [22:39] <Zeros> Guess I need to do a reduction on that :/
# [22:39] <Zeros> Something with Safari, walking backwards in the blue area in the middle of the room makes you jump outside the game world
# [22:40] <anne> Philip, interesting...
# [22:41] <Philip> Hmm - in theory, it shouldn't be possible to jump outside the world
# [22:41] <Zeros> Philip, almost looks as if I'm falling through the floor when I see other things move a little bit
# [22:41] <Zeros> most of the graphics get frozen when it happens
# [22:41] <Zeros> and Safari reports "Undefined Value on line 1" with no file reference
# [22:42] <Philip> I've never tested my code with Safari, so I'm quite possibly doing something broken that it doesn't like :-)
# [22:42] <Zeros> Yeah, I feel through the floor
# [22:42] <Zeros> fell*
# [22:42] <Zeros> I can see the world above me trough a little crack and I can still hit walls
# [22:42] <Zeros> weird
# [22:43] <Philip> Does that happen at the point where you walk over the small raised bit onto the reddish surface?
# [22:43] <Philip> And does it happen if you jump over that bit?
# [22:44] <Philip> (I guess it dislikes the bit where you move from one of the level's sectors into another, in which case it'd still happen if you try jumping through the division)
# [22:44] <Zeros> Looks it happens when you walk over that little step and you're turning
# [22:44] <Zeros> straight over it and its fine
# [22:45] <Philip> Urgh, sounds like it's issues with floating point inaccuracies
# [22:45] <Philip> (There are problems in other browsers if you stand directly in the corner near the box, since the maths gets confused and it was too hard to want to fix)
# [22:46] <Philip> (*gets confused by being precisely on a dividing line in the world)
# [22:46] <Zeros> ah okay
# [22:47] <Zeros> Philip, http://enfinitystudios.thaposse.net/world.png
# [22:48] <Zeros> Depending on how you do it you may or may not get a little sliver in the view that still changes as you move around
# [22:48] <Zeros> That's what that grey line in the middle is
# [22:49] <Zeros> The footer dropping is another issue entirely, that I should look into as well since both Opera and Firefox don't do it
# [22:49] <Philip> Hmm, do you not see barrel sprites in the corners? (It looks like they're missing from that image)
# [22:50] <Zeros> Philip, I do when I'm in the world. Once it breaks all the sprites disappear and you're left with a view that's frozen in one position, sometimes pretty mangled with various parts of different walls
# [22:50] <Philip> Ah, okay
# [22:51] <Zeros> Either way Safari is getting 7-10fps when its not broken and Opera is getting 13fps average here
# [22:51] <Zeros> Firefox 2 is painfully slow at 5fps
# [22:51] <Zeros> Whatever they did to the mac version I guess...
# [22:52] <Philip> If you select the "large" size while it's in that broken state, does the screen go mostly black?
# [22:52] <Zeros> oh
# [22:52] <Zeros> let me test
# [22:52] <Philip> That ought to clear all the stuff it's rendered in previous frames
# [22:53] <Philip> FF2 has always been a bit faster than O9 for me on Windows, though FF1.5 was a bit slower - maybe the new renderer in FF3 will be better on Macs, or maybe that'll be just as broken...
# [22:54] <Zeros> Yeah it fixes it
# [22:54] <Zeros> Are you returning false from the onkeypress or onkeydown events you're using?
# [22:54] <Zeros> Safari is ignoring it and scrolling the page
# [22:56] <Philip> No - I'm just calling evt.preventDefault() in keypress, and doing nothing interesting in keydown/keyup, and returning nothing from either
# [22:57] <Zeros> which file is that in?
# [22:57] <Philip> game.js
# [22:58] <Zeros> Alright, I'll reduce it and add a ticket later
# [22:58] <Philip> (preventDefault was enough to prevent Opera stealing the keys, if I remember correctly)
# [22:58] <Zeros> Found a nasty crash bug with plugins sleeping in Safari too I need to submit tonight :/
# [22:58] <Philip> Unfortunately this kind of code isn't the nicest thing to extract bug reports from :-(
# [23:00] <Philip> (I found several rendering glitches in Firefox, and one security flaw that let me read 28KB of random heap memory on Linux...)
# [23:00] <Zeros> Nice
# [23:00] <Zeros> There's a rendering issue with that JS UI library (the one that named like quxdoo or something?) in Safari too
# [23:01] <Zeros> Can't say I have a huge desire to reduce their code though to figure out what's going on
# [23:07] <Zeros> Browsers should really implement a wildcard character for the search feature.
# [23:35] * Joins: hober (ted@69.45.6.105)
# [23:38] * Joins: asbjornu (asbjorn@84.48.116.134)
# [23:45] <Philip> Zeros: How about regular expressions?
# [23:45] <Philip> Try javascript:void(document.body.innerHTML=document.body.innerHTML.replace(/(pat.*ern)/g, '<blink>$1</blink>'))
# [23:46] <Zeros> You'd get attributes and parts of tags too
# [23:46] <Zeros> the regexp would need to be pretty complex to be useful
# [23:46] <Philip> Hmm, how about implementing the HTML5 parsing algorithm in regexps...
# [23:46] <Zeros> ouch
# [23:47] <Zeros> Philip, The only reason I thought of it was because I have a co-worker with some special letters in his name, and I couldn't remember how to type the g or u with accent marks so I couldn't search for it in the document
# [23:47] <anne> DOM2Range might be usable
# [23:47] <Zeros> being able to throw a * in the search box of the browser to get anything would be nice
# [23:50] <mjs> I am so curious to see Chris Wilson's promised essay now
# [23:51] <Hixie> yes
# [23:51] * anne too
# [23:52] <hober> mjs: yeah
# [23:52] * h3h four
# [23:52] <h3h> or five if I could count
# [23:53] <mjs> so far, I think he's actually increased opposition to versioning
# [23:53] <mjs> with his choice of arguments in favor
# [23:53] <Dashiva> He hasn't been very clear on what he really wants, so I think many are responding to the worst case scenario
# [23:53] <h3h> I don't really even see his arguments
# [23:54] <h3h> it just seems to me that he wants to say "there's police tape around IE and we're not going to touch it."
# [23:54] <Dashiva> I think it boils down to "People don't just make pages to a standard, they make pages to a standard and specific UAs with specific quirks"
# [23:54] <Hixie> i've been talking with microsoft people about this for some time, and as far as i can tell, the worst case scenario is exactly what they want.
# [23:56] <h3h> why? I'm assuming there has to be some overarching logical reason for it, even if it is corporate and conservative...
# [23:57] <mjs> my impression is that they feel burned by the IE7 experience
# [23:57] <hober> mjs: so don't we all! :/
# [23:58] <mjs> I don't think it's meant to be anti-competitive, although it may seem that way in effect
# [23:58] <h3h> somehow I can't believe that they're listening to the proposed solution
# [23:58] <h3h> Chris seems to be saying "we can't change how IE does things"
# [23:58] <h3h> when that's not the proposed solution
# [23:58] <Philip> Perhaps it's partly because they have to worry about lots of IE-only intranet sites, whereas other browsers only have to worry about the public web (which is already in various states of brokenness across different browsers, and is easier to survey)?
# [23:59] <Hixie> other browsers have their own problems, believe me
# [23:59] <Zeros> h3h, Sure it is. Right now <!DOCTYPE HTML> causes standards mode in IE. He wants way to cause "really standards mode".
# [23:59] <Hixie> e.g. safari has all kinds of legacy problems with dashboard widgets.
# Session Close: Tue Apr 10 00:00:00 2007

The end :)