Options:
- # Session Start: Wed Feb 18 00:00:00 2009
- # Session Ident: #html-wg
- # [00:04] * Quits: adele (adele@17.203.15.145) (Ping timeout)
- # [00:19] * Joins: adele (adele@17.244.16.159)
- # [00:23] * Quits: adele (adele@17.244.16.159) (Quit: adele)
- # [00:30] * Quits: gavin_ (gavin@99.226.207.11) (Ping timeout)
- # [00:33] * Joins: adele (adele@17.244.16.159)
- # [00:33] * Joins: gavin_ (gavin@99.226.207.11)
- # [01:07] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Ping timeout)
- # [01:28] * Quits: Sander (svl@86.87.68.167) (Quit: And back he spurred like a madman, shrieking a curse to the sky.)
- # [01:29] * Joins: adele_ (adele@17.244.16.159)
- # [01:29] * Quits: adele (adele@17.244.16.159) (Connection reset by peer)
- # [02:02] * Quits: tH (Rob@129.11.83.14) (Ping timeout)
- # [02:10] * Quits: syp (syp@83.219.118.137) (Ping timeout)
- # [02:42] * Quits: adele_ (adele@17.244.16.159) (Quit: adele_)
- # [03:05] * Quits: tlr (tlr@128.30.52.30) (Ping timeout)
- # [03:33] * Quits: hober (ted@206.212.254.2) (Client exited)
- # [03:34] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
- # [04:18] * Joins: rking3 (rking3@99.27.42.97)
- # [05:10] * Quits: dbaron (dbaron@63.245.220.246) (Quit: 8403864 bytes have been tenured, next gc will be global.)
- # [05:18] * Joins: adele (adele@24.7.123.50)
- # [05:18] * Quits: adele (adele@24.7.123.50) (Client exited)
- # [05:19] * Joins: adele (adele@24.7.123.50)
- # [05:24] * Quits: rking3 (rking3@99.27.42.97) (Quit: rking3)
- # [05:39] * Joins: dbaron (dbaron@98.234.51.190)
- # [06:48] * Quits: adele (adele@24.7.123.50) (Quit: adele)
- # [06:50] * Joins: Zeros (Zeros-Elip@68.50.195.181)
- # [07:10] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Ping timeout)
- # [07:12] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
- # [07:25] * Quits: heycam (cam@61.88.219.66) (Ping timeout)
- # [07:32] * Joins: heycam (cam@61.88.219.66)
- # [07:48] * Quits: aroben (aroben@71.58.73.153) (Connection reset by peer)
- # [08:03] * Quits: shepazu (schepers@128.30.52.30) (Quit: shepazu)
- # [08:06] * Quits: heycam (cam@61.88.219.66) (Quit: bye)
- # [08:09] * Joins: shepazu (schepers@128.30.52.30)
- # [08:11] * Quits: jwatt (roslea@61.88.219.66) (Ping timeout)
- # [08:11] * Quits: shepazu (schepers@128.30.52.30) (Quit: shepazu)
- # [08:15] * Quits: dbaron (dbaron@98.234.51.190) (Quit: 8403864 bytes have been tenured, next gc will be global.)
- # [08:23] <MikeSmith> http://gyu.que.jp/jscloth/
- # [08:23] <pimpbot> Title: js cloth (at gyu.que.jp)
- # [08:27] <MikeSmith> http://d.hatena.ne.jp/gyuque/
- # [08:27] <pimpbot> Title: 最速チュパカブラ研究会 (at d.hatena.ne.jp)
- # [08:28] <MikeSmith> ャ腥銀
- # [08:30] <MikeSmith> "Fastest Chupacabra Research Club"
- # [08:30] <MikeSmith> scroll down on that page for English
- # [08:37] <Zeros> that's pretty impressive
- # [08:40] <MikeSmith> dude is a force
- # [08:41] <MikeSmith> as are most of the guys involved with this:
- # [08:41] <MikeSmith> http://shibuyajs.org/
- # [08:41] <pimpbot> Title: Shibuya.js (at shibuyajs.org)
- # [08:46] <MikeSmith> Shunsuke: have you met gyuque?
- # [08:46] <MikeSmith> trying remember if I met the guy
- # [09:04] * Joins: tH (Rob@129.11.83.14)
- # [09:14] * Quits: Zeros (Zeros-Elip@68.50.195.181) (Quit: Leaving)
- # [09:16] * Joins: laplink (link@78.26.2.125)
- # [09:22] * Quits: laplink (link@78.26.2.125) (Ping timeout)
- # [09:39] * Joins: billyjackass (MikeSmith@mcclure.w3.org)
- # [09:40] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Ping timeout)
- # [09:40] * billyjackass is now known as MikeSmith
- # [09:44] <pimpbot> planet: Bespin and Canvas <http://benzilla.galbraiths.org/2009/02/17/bespin-and-canvas/>
- # [10:00] * Joins: laplink (link@78.26.2.125)
- # [10:03] * Quits: laplink (link@78.26.2.125) (Ping timeout)
- # [10:46] * Quits: Lachy (Lachlan@85.196.122.246) (Quit: Leaving)
- # [10:46] * Joins: Lachy (Lachlan@85.196.122.246)
- # [10:50] * Joins: ROBOd (robod@89.122.216.38)
- # [11:10] * Quits: Lachy (Lachlan@85.196.122.246) (Quit: This computer has gone to sleep)
- # [11:47] * Joins: Lachy (Lachlan@213.236.208.22)
- # [12:05] * Joins: Sander (svl@86.87.68.167)
- # [12:21] * Joins: syp (syp@83.219.105.167)
- # [12:58] * Joins: Julian (chatzilla@217.91.35.233)
- # [13:01] * Joins: maddiin (mc@87.185.198.194)
- # [13:25] * Quits: ifette_GOOG (kvirc@216.239.45.4) (Ping timeout)
- # [13:33] * Joins: myakura (myakura@221.191.119.20)
- # [13:46] * Quits: annevk (opera@77.163.243.203) (Ping timeout)
- # [14:11] * Joins: annevk (opera@77.163.243.203)
- # [14:23] * Quits: annevk (opera@77.163.243.203) (Ping timeout)
- # [14:39] * Quits: Sander (svl@86.87.68.167) (Quit: And back he spurred like a madman, shrieking a curse to the sky.)
- # [14:40] * Joins: annevk (opera@77.163.243.203)
- # [15:01] * Joins: Sander (svl@86.87.68.167)
- # [15:21] * Quits: annevk (opera@77.163.243.203) (Ping timeout)
- # [15:39] * Quits: Sander (svl@86.87.68.167) (Quit: And back he spurred like a madman, shrieking a curse to the sky.)
- # [15:40] * Joins: billyjackass (MikeSmith@mcclure.w3.org)
- # [15:40] * billyjackass is now known as Mike
- # [15:44] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Ping timeout)
- # [15:47] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
- # [15:47] * MikeSmith is now known as MikeS
- # [15:50] * Joins: annevk (opera@77.163.243.203)
- # [15:51] * Quits: Mike (MikeSmith@mcclure.w3.org) (Ping timeout)
- # [15:52] * Quits: gsnedders (gsnedders@86.136.52.180) (Quit: Killin' teh intarwebs)
- # [15:57] * DanC checks the time in Japan.. 11:55pm...
- # [15:57] <DanC> hi MikeS ; are you in office mode?
- # [16:02] <MikeS> DanC: I'm on a call now
- # [16:02] <MikeS> but I'll be around after
- # [16:06] * Philip hopes the canvas accessibility thing won't get too ugly...
- # [16:08] <MikeS> Philip: did you see that gyuque links I posted earlier?
- # [16:08] <MikeS> would like to know what you think
- # [16:09] * Philip doesn't think releasing version 0.1 of an online text editor that doesn't support screenreaders (along with not supporting IE or Opera) is equivalent to racism
- # [16:09] <Philip> MikeS: I saw the links but I didn't dereference them
- # [16:11] <MikeS> OK
- # [16:12] * Philip dereferences now
- # [16:13] * Joins: tlr (tlr@128.30.52.30)
- # [16:13] <Philip> Looks quite neat, but the lack of canvas support for perspective transforms makes textures ugly :-(
- # [16:14] <MikeS> true that
- # [16:14] <MikeS> but still impressive what can be done within the constraints
- # [16:15] <Philip> The only way to get perspective-correct texturing is to draw each line of pixels individually, which is horribly slow
- # [16:15] <Philip> (which is why Canvex doesn't have floor textures enabled)
- # [16:15] <MikeS> aha
- # [16:16] <Philip> Given the limitations, it's certainly nicely done :-)
- # [16:17] * DanC sent "Gentle now... " mail re bespin and accessibility
- # [16:18] <Philip> http://canvex.lazyilluminati.com/misc/3d.html was my attempt at 3D triangles
- # [16:18] <Philip> but I didn't try to add texturing
- # [16:19] <Philip> (Ooh, 8fps in FF3.1 - that's not too awful)
- # [16:20] <Philip> (Well, yes it is, but relative to expectations it's okay)
- # [16:22] <annevk> wow, that thread is weird
- # [16:22] <jgraham> annevk: Which one?
- # [16:22] * jgraham wonders why he is missing a number of mails
- # [16:23] <jgraham> Ah, there they are
- # [16:23] * Joins: billmason (bmason@69.30.57.102)
- # [16:23] <annevk> me too, do you think it might be mail.opera.com?
- # [16:24] <jgraham> annevk: Seems reasonable. I was missing everything in public-html from about 15:30 - 16:00 and a few either side of that
- # [16:25] <annevk> I meant the <canvas> thread btw
- # [16:25] <MikeS> jgraham, annevk : that's a secret feature of Opera Mail
- # [16:25] <MikeS> the noise throttler
- # [16:25] <annevk> :)
- # [16:26] <jgraham> Awesome
- # [16:26] <MikeS> I think JL worked on that one
- # [16:27] <MikeS> DanC: I'm off my call and back on now for 30 minutes or so, until I need to go to the 羚
- # [16:27] * DanC looks that up..
- # [16:28] * DanC should have sent mail in the mean time
- # [16:30] <MikeS> DanC: I should have time to work more on the markup draft tomorrow. though I got a serious case of 援 aka cedar fever
- # [16:30] <MikeS> stupid cedar trees all over the place in Tokyo
- # [16:30] <DanC> MikeS, is there outstanding feedback on http://www.w3.org/html/wg/markup-spec/ ? or have you answered all the feedback you intend to answer?
- # [16:30] <pimpbot> Title: HTML 5: The Markup Language (at www.w3.org)
- # [16:30] <MikeS> not as bad as Austin, pbut close
- # [16:31] <MikeS> I'm sure there is outstanding feed back
- # [16:31] <MikeS> I have a number of messages I need to respond to
- # [16:31] <MikeS> I also opened issues in W3C bugzilla for others
- # [16:31] <DanC> ok
- # [16:31] <MikeS> the ones that Hixie brought up
- # [16:32] <MikeS> I'd really rather managed things in bugzilla
- # [16:32] * DanC is still hoping to avoid w3c bugzilla
- # [16:32] <MikeS> as much as we can
- # [16:32] <MikeS> whY?
- # [16:32] <MikeS> it works fine
- # [16:32] <MikeS> bugzilla is bugzilla
- # [16:32] <DanC> it's yet another tracking thingy to learn
- # [16:32] <DanC> learn, track, manage, etc.
- # [16:32] <MikeS> well, at least it's a semi-standard
- # [16:32] <MikeS> and not yet another NIH thing
- # [16:33] <DanC> yes, it's fairly widely used, but I have managed to avoid using it pretty much everywhere
- # [16:33] <MikeS> well, that's not much to be proud of
- # [16:33] <Philip> It's not that complicated once you've spent a couple of years using it
- # [16:33] <DanC> yes, it's that 2 years that I'm not looking forward to
- # [16:33] <MikeS> since several major browser projects use Bugzilla for tracking bugs
- # [16:33] * jgraham has yet to find a tracking system that is much worse than bugzilla
- # [16:33] * Joins: laplink (link@78.26.2.115)
- # [16:34] <MikeS> even Opera was using a Bugzilla for bug tracking
- # [16:34] <jgraham> Er, that's not waht I meant
- # [16:34] <MikeS> internally
- # [16:34] <MikeS> at least Opera as also until recently
- # [16:34] <jgraham> I mean Bugzilla is not much worse than the median bug tracking system
- # [16:34] <Philip> MikeS: Opera's wasn't really like proper Bugzilla
- # [16:35] <jgraham> Even though bugzilla has quite a bad reputation
- # [16:35] <MikeS> bugzilla sucks
- # [16:35] <MikeS> really
- # [16:35] <MikeS> mightily
- # [16:35] * Philip doesn't know whether it was originally derived from Bugzilla code, but it didn't really look like it
- # [16:35] <MikeS> it is total is total crap
- # [16:36] <MikeS> there are infinitely better bug tracking systems
- # [16:36] <jgraham> e.g.?
- # [16:36] <MikeS> they are just all proprietary
- # [16:36] <Philip> MikeS: Have you seen the median bug tracking system? :-)
- # [16:36] <jgraham> :p
- # [16:36] <MikeS> well, I'm not saying there's nothing worse
- # [16:36] <jgraham> For example trac is bad in different ways
- # [16:36] <DanC> for a community shaped like mozilla (zillions of users, hundreds of developers), I bet bugzilla is quite a reasonable match.
- # [16:37] <jgraham> Roundup is supposed to be good but I have never actually seen it used
- # [16:37] <DanC> roundup is happy-happy-joy-joy
- # [16:37] <DanC> at least: for a smallish project
- # [16:38] <DanC> I used roundup to manage a few students working with timbl on tabulator. http://dig.csail.mit.edu/issues/tabulator/
- # [16:38] <pimpbot> Title: List of issues - Tabulator Issue Tracker (at dig.csail.mit.edu)
- # [16:38] <DanC> anyway... I don't expect to change MikeS's mind about bugzilla
- # [16:39] <MikeS> I will keep using Bugzilla for the markup spec
- # [16:39] <MikeS> as long as I'm involved with it
- # [16:39] <MikeS> and I'm glad that Hixie and zcorpan and annevk and hsivonen and others have found it useful
- # [16:40] * DanC hunts for a "view source" link; finds one, but it doesn't point to a CVS deely
- # [16:40] <MikeS> one fundamental problem with W3C Tracker is that it's not open-access
- # [16:40] * Quits: laplink (link@78.26.2.115) (Ping timeout)
- # [16:40] <DanC> that's a feature
- # [16:40] <MikeS> nope
- # [16:40] <DanC> i.e. it's as open-access as we choose to make it
- # [16:41] <MikeS> it's not a feature if you want to avoid wasting the time of browser developers and library developers who only care about one particular issue at a time
- # [16:41] <MikeS> they certainly do not want to subscribe to public-html and enjoy the full grandness of that
- # [16:42] <DanC> I wonder why you haven't changed the way the HTML WG tracker works, if you think it's broken
- # [16:43] <DanC> the one change I remember you making was to make it *less* open
- # [16:43] <MikeS> because it's not my code
- # [16:43] <MikeS> yeah, I'm all for less open
- # [16:43] <DanC> the code supports as much openness as we like, no?
- # [16:44] <MikeS> I guess
- # [16:44] <DanC> indeed, the HTML WG tracker is optimized for a coherent WG tackling a set of issues as a group.
- # [16:44] <DanC> it seems there's room for both tracker and bugzilla
- # [16:44] <MikeS> the only thing I really find exceptional about Tracker is the IRC integration
- # [16:45] <MikeS> there is indeed room for both Tracker and bugzilla
- # [16:45] * DanC goes back to trying to check out H:TML
- # [16:45] <DanC> I'm at http://dev.w3.org/cvsweb/ ; which link do I follow?
- # [16:45] <pimpbot> Title: W3C Public CVS Repository (at dev.w3.org)
- # [16:45] <Lachy> MikeS or DanC, is it possible to edit the title of Action 105 to fix the spelling mistake in the title? http://www.w3.org/html/wg/tracker/actions/105
- # [16:45] <pimpbot> Title: ACTION-105 - HTML Weekly Tracker (at www.w3.org)
- # [16:46] <Lachy> s/insure/ensure/ (in the title of that action)
- # [16:46] <DanC> yes, you can edit actions, lachy
- # [16:46] <MikeS> "Tracker and bugzilla both have their place" is a point I tried to make 6 months ago
- # [16:46] <Lachy> ah, ok. Done
- # [16:47] <annevk> and s/XHTML WG/XHTML2 WG/
- # [16:47] <annevk> oh well
- # [16:47] <Lachy> annevk, fixed
- # [16:48] * Joins: gsnedders (gsnedders@86.136.52.180)
- # [16:49] <DanC> MikeS, I've looked in all the places I can think to look, and I can't find the CVS repository for http://www.w3.org/html/wg/markup-spec/ ; help?
- # [16:49] <pimpbot> Title: HTML 5: The Markup Language (at www.w3.org)
- # [16:50] <MikeS> DanC: WWW/html/wg/markup-spec
- # [16:50] <gsnedders> Mike is working in secret!
- # [16:50] <MikeS> DanC: I want to move it all to dev.w3.org
- # [16:50] <gsnedders> It's a whole new cabal of one!
- # [16:50] * Quits: myakura (myakura@221.191.119.20) (Quit: Leaving...)
- # [16:50] <MikeS> but systeam has told me they are going to chop me up into little pieces if I put more load on dev.w3.org
- # [16:50] <DanC> ah
- # [16:51] <DanC> if there's moving to be done, let's move to hg
- # [16:51] <MikeS> because apparently, dev.w3.org is, like, a sewing machine with a web server on it
- # [16:51] <MikeS> or something
- # [16:51] * DanC LOL
- # [16:51] <MikeS> my answer to them has been, buy a computer
- # [16:51] <MikeS> or two
- # [16:51] <MikeS> or there
- # [16:51] <MikeS> three
- # [16:52] <DanC> cue re-run of me telling W3C management how much resource should be allocated to the HTML WG, followed by them (a) chartering 2 WGs in stead of one and (b) cutting my job in half.
- # [16:52] <MikeS> amen to that
- # [16:53] <MikeS> preaching to the choir
- # [16:54] <annevk> I wonder what would've happened with one WG
- # [16:54] <MikeS> anyway, I got to step out for now
- # [16:54] <DanC> ok. hasta.
- # [16:54] <MikeS> I'll be back before the hour
- # [16:54] <MikeS> or so
- # [16:54] <DanC> I still hope we'll find out, annevk
- # [16:57] * Joins: laplink (link@78.26.2.115)
- # [16:58] <DanC> ah... a spot-check `xmlwf markup-spec/src/syntax.html ` give thumbs-up
- # [17:00] * DanC studies the Makefile... wow...
- # [17:01] * DanC wanders off to grab breakfast...
- # [17:10] * Quits: Julian (chatzilla@217.91.35.233) (Ping timeout)
- # [17:12] * Joins: Julian (chatzilla@217.91.35.233)
- # [17:16] * Joins: Sander (svl@86.87.68.167)
- # [17:18] * Philip can't work out whether http://lists.w3.org/Archives/Public/public-html/2009Feb/0444.html means "Why use canvas for Bespin?" or "Why have canvas in HTML5?"
- # [17:18] <pimpbot> Title: Re: Example canvas element use - accessibility concerns from Janina Sajka on 2009-02-18 (public-html@w3.org from February 2009) (at lists.w3.org)
- # [17:30] * Lachy wonders why the bespin thread has a high proportion of annoying top posting
- # [17:33] <DanC> ACTION-98 due next week
- # [17:33] * trackbot attempting to change due date on ACTION-98.
- # [17:33] <trackbot> ACTION-98 Discuss missing-alt with the WAI CG and report back due date now next week
- # [17:33] <Philip> Ooh, how useful - bz asked exactly the thing I was wondering about, and for each case gave exactly the same answers that I was thinking of but was too lazy to post
- # [17:34] * Joins: MichaelC (Michael@128.30.52.30)
- # [17:42] * MikeS wanders back in
- # [17:42] <MikeS> DanC: that makefile will give you hallucinations if you stare at it long enough
- # [17:42] <MikeS> or nightmares
- # [17:43] <DanC> I tried `make clean; make` and lost thusly:
- # [17:43] <DanC> Traceback (most recent call last):
- # [17:43] <DanC> File "./xhtmltohtml", line 16, in <module>
- # [17:43] <DanC> from html5lib import html5parser, liberalxmlparser
- # [17:43] <DanC> ImportError: No module named html5lib
- # [17:43] <DanC> ah... I can prolly fix that myself...
- # [17:44] <DanC> darn:
- # [17:44] <DanC> ~/projects/html5lib$ python setup.py install --home=~
- # [17:44] <DanC> python: can't open file 'setup.py': [Errno 2] No such file or directory
- # [17:45] <DanC> aha...
- # [17:45] <DanC> TEST FAILED: /home/connolly/lib/python/ does NOT support .pth files
- # [17:45] <DanC> oh no... level 15 of setuptools hell
- # [17:46] <Philip> You can use html5lib without installing it
- # [17:46] <MikeS> DanC: would be good to have a local install of html5lib regardless
- # [17:47] <Philip> Just point PYTHONPATH at html5lib/python/src or whatever it's called
- # [17:47] * Philip has never intentionally installed html5lib anywhere
- # [17:47] <DanC> wild... the Makefile has FOO_FLAGS for umpteen values of FOO, but not PYTHON
- # [17:48] <DanC> html/wg$ PYTHONPATH=~/projects/html5lib/python/src make
- # [17:48] <DanC> ok, that got me past the html5lib error...
- # [17:49] <DanC> File "/home/connolly/projects/html5lib/python/src/html5lib/inputstream.py", line 270, in detectBOM
- # [17:49] <DanC> self.rawStream.seek(encoding and seek or 0)
- # [17:49] <DanC> IOError: [Errno 29] Illegal seek
- # [17:49] * jgraham often installs html5lib but maybe I alays have a raqndom setup.py file lying around
- # [17:49] <jgraham> *random
- # [17:49] <Philip> DanC: I know nothing about that bit of code; blame jgraham ;-)
- # [17:49] <DanC> I found the setup.py file in the python directory
- # [17:49] <MikeS> DanC: cause it don't run python directly (it just runs modified html5lib parse.py script)
- # [17:50] <jgraham> DanC: Is it trying to read from stdin?
- # [17:50] <DanC> yes, I think so, jgraham (make output includes: ./xhtmltohtml - )
- # [17:50] <jgraham> Ah. That might be broken
- # [17:50] <DanC> MikeS, what version of html5lib works for you?
- # [17:50] * jgraham goes to file a bug on himself
- # [17:51] <MikeS> DanC: the source from html5lib svn
- # [17:51] <DanC> what svn rev?
- # [17:51] <MikeS> 1171
- # [17:51] <DanC> I seem to be at Revision: 1258 Last Changed Date: 2009-01-13 11:16:10 -0600 (Tue, 13 Jan 2009)
- # [17:52] <DanC> ah. ETOONEW
- # [17:52] <MikeS> I've not synced up in a long time
- # [17:52] <DanC> are you game to try an svn up? or shall I downgrade?
- # [17:53] * DanC hears the ghost of Jim Gettys droning on about error-22, i.e. the use of a new technology in the development of a new technology
- # [17:54] * MikeS syncs up now to latest html5lib svn
- # [17:54] <Philip> Syncing is probably dangerous (just ask the Titanic) - old known-to-work versions are better than the latest barely-tested version
- # [17:55] <DanC> "if it jams, force it; if it breaks, it needed replacing anyway"
- # [17:55] <MikeS> amen
- # [17:55] <MikeS> bash it
- # [17:55] * DanC Updated to revision 1268.
- # [17:55] <Philip> (I think I noticed a while ago that html5lib stopped reading from stdin, but I don't know if I even mentioned the problem on IRC)
- # [17:55] <MikeS> me to r1268 too
- # [17:56] * DanC doesn't get the connection between svn sync issues in the Titanic... wonders if now is the time to fill gaps in his pop-culture education...
- # [17:57] <MikeS> setup.py build and install worked find for me
- # [17:57] <DanC> did you install as root? I try not to do that
- # [17:57] <Philip> DanC: ("sync" vs "sink")
- # [17:57] * Philip doesn't know anything about pop-culture
- # [17:57] <jgraham> DanC: I just checked in a quick hack to make reading from stdin maybe work
- # [17:57] <DanC> ah. a nice shallow rabit trail. thanks, Philip
- # [17:57] * MikeS now runs make clean && make on markup-spec sources
- # [17:59] <MikeS> GNU make is an abomination
- # [17:59] <DanC> I'm learning to transition from terminal windows to emac's shell mode; in a terminal, ctrl-p gets me the previous command to re-run or edit; surely there's an analog in emacs; anybody have the relevant clue?
- # [17:59] * MikeS is a Vim user
- # [17:59] <DanC> heathen.
- # [17:59] <MikeS> Emacs makes baby Jesus cry
- # [18:00] <jgraham> ctrl + up?
- # [18:00] <DanC> bingo!
- # [18:00] <DanC> ~/projects/html5lib/python$ python setup.py install
- # [18:00] <DanC> running install
- # [18:00] <DanC> error: can't create or remove files in install directory
- # [18:01] <Philip> My supervisor uses cut-and-paste to repeat shell commands in Emacs
- # [18:01] <DanC> did you install as root, MikeS ? I try not to do that; I try to let ubuntu manage the software in /usr
- # [18:01] <MikeS> DanC: I installed as root
- # [18:02] <MikeS> I don't use Ubuntu
- # [18:02] <DanC> ok. I'll stick with PYTHONPATH
- # [18:02] <jgraham> DanC: That sounds like setuptool's fault, not mine :)
- # [18:02] <MikeS> I run Debian unstable
- # [18:02] <MikeS> Ubuntu is for sissies
- # [18:02] <DanC> debian unstable is fun, as long as you have a 2nd machine on stand-by for when it horks up
- # [18:02] <Philip> DanC: Use --home (or --prefix or something) to install, perhaps?
- # [18:03] <DanC> I tried --home; see above for resulting poetry
- # [18:03] <MikeS> I would run Gentoo but I'm not quite that insane yet
- # [18:03] <MikeS> I need to sustain some further major head injuries before being ready for Gentoo
- # [18:03] <DanC> actually, the 2nd machine bit goes for all operating systems. they all hork up now and again.
- # [18:03] <gsnedders> MikeS: I'm sure that can be arranged
- # [18:03] <Philip> DanC: I thought the above with --home was because you weren't in the right directory for setup.py?
- # [18:03] <Philip> MikeS: I run Gentoo
- # [18:04] <MikeS> uh oh
- # [18:04] <Philip> MikeS: (I'm not sure whether that supports your point)
- # [18:04] <MikeS> heh
- # [18:04] <DanC> hmm... could be, Philip ...
- # [18:04] * jgraham have never quite got setuptools to work right
- # [18:04] <Philip> (Probably it does)
- # [18:04] <jgraham> Or rather it did once but doesn't anymore
- # [18:04] <MikeS> Philip: I didn't know you had so much free time on you hands
- # [18:04] <Philip> DanC: Oh, or was that the error about .pth files?
- # [18:04] <MikeS> it's like building Qt
- # [18:05] <MikeS> start the build and come back a week later to check on its progress
- # [18:05] <Philip> MikeS: It's nowhere near that slow on a modern machine :-)
- # [18:05] <DanC> ~/projects/html5lib/python$ python setup.py install --home=~
- # [18:05] <DanC> TEST FAILED: /home/connolly/lib/python/ does NOT support .pth files
- # [18:05] <jgraham> And I did once manage to get a nice configuration where each user installed packages to their home directory on the local machine and the root user installed to a custom location on the server and got various 32 vs 64 bit things right working
- # [18:05] <Philip> MikeS: though if you're compiling the entirety of KDE then you probably want to leave it running overnight
- # [18:06] <jgraham> But I have no idea how
- # [18:06] * Joins: dbaron (dbaron@63.245.220.246)
- # [18:06] <Philip> DanC: Does ~/lib/python exist?
- # [18:06] <DanC> yes
- # [18:06] <Philip> MikeS: (but compiling the entirety of KDE is very rare, and you can still run everything while it's building in the background)
- # [18:07] * DanC is back to IOError: [Errno 29] Illegal seek
- # [18:07] * DanC checks to see if mike's make worked
- # [18:07] <Philip> DanC: Hmm, okay, I've got no idea what .pth files are so I can't help much
- # [18:07] <MikeS> Philip: speaking of.. one of the really great things about KDE is that even if you want to make one really minor hack to a KDE app -- like Konsole -- you the have to recompile the entirety of KDE to actually get a build of the app with your change
- # [18:07] <MikeS> geniuses
- # [18:07] <Philip> DanC: Did you update to r1269?
- # [18:07] <DanC> no; MikeS, did you?
- # [18:07] <Philip> MikeS: Not in Gentoo :-)
- # [18:07] <Philip> MikeS: It's all modularised nowadays
- # [18:08] <MikeS> yippee for Gentoo
- # [18:08] <DanC> MikeS, are you using r1269?
- # [18:08] <MikeS> I use Debian for religious reasons
- # [18:08] <Philip> DanC: r1269 is a few minutes old
- # [18:08] <MikeS> DanC: just synced up and rebuilding now
- # [18:09] * DanC Updated to revision 1269
- # [18:09] <Philip> and hopefully fixes the stdin thing
- # [18:09] <Philip> (and if it doesn't then it's jgraham's fault)
- # [18:09] <DanC> indeed, no more Illegal seek
- # [18:09] <jgraham> I think there is no doubt that it is all my fault
- # [18:09] <jgraham> DanC: :(
- # [18:09] <Philip> jgraham: Why ":("?
- # [18:10] <jgraham> Oh, I misread that
- # [18:10] <jgraham> As no, more illegal seek
- # [18:10] <DanC> ok... so... the main product is 5.html , MikeS ?
- # [18:10] <Philip> MikeS: (http://www.gentoo.org/proj/en/desktop/kde/kde-split-ebuilds.xml - it's so exciting to read all these things when trying to install Linux)
- # [18:10] * DanC dons dunce cap
- # [18:10] <MikeS> r1269 work fine for me
- # [18:10] <pimpbot> Title: Gentoo Linux Documentation -- The KDE Split Ebuilds HOWTO (at www.gentoo.org)
- # [18:11] <Philip> jgraham: Ah, right :-)
- # [18:11] <DanC> I've been running make in html/wg , not html/wg/markup-spec
- # [18:11] <MikeS> Philip: "exciting" is one way top put it
- # [18:11] <MikeS> DanC: ah
- # [18:11] <MikeS> I wrote that html/wg Makefile about 20 year ago, I think
- # [18:11] <DanC> current hurdle: make: trang: Command not found
- # [18:11] <MikeS> I have no idea what it actually does
- # [18:11] * Philip is always glad to have a choice between three cron daemons when following the Gentoo installation manual
- # [18:12] <MikeS> trang makes an orange drink for you, I think
- # [18:12] <MikeS> or a transvestite
- # [18:12] <MikeS> I think there's an option for choosing which
- # [18:13] * DanC prays to the apt gods for trang...
- # [18:13] <Philip> It translates between different flavours of fruit drinks
- # [18:13] <DanC> next hurdle:
- # [18:13] <MikeS> trang is packaged for debian as a gcj binary
- # [18:13] <DanC> trang schema/html5full-aria.rnc html5full-aria.rng
- # [18:13] <DanC> Exception in thread "main" java.lang.NullPointerException
- # [18:13] <DanC> *** Got java.lang.NullPointerException while trying to print stack trace
- # [18:13] <MikeS> I think
- # [18:13] <MikeS> contact Mr. James Clark please
- # [18:14] <MikeS> trying to debug JC code is futile
- # [18:14] <Philip> An exception while printing the stack trace? That sounds pretty weird
- # [18:14] <DanC> I suspect java horkage
- # [18:15] <MikeS> yah
- # [18:15] <MikeS> gcj stuff is a total mess, basically
- # [18:15] <MikeS> failure points all over
- # [18:15] <DanC> $ java -version
- # [18:15] <DanC> java version "1.6.0_0"
- # [18:15] <DanC> IcedTea6 1.3.1 (6b12-0ubuntu6.1) Runtime Environment (build 1.6.0_0-b12)
- # [18:15] <DanC> OpenJDK Server VM (build 1.6.0_0-b12, mixed mode)
- # [18:15] <MikeS> what the hell is that
- # [18:15] <MikeS> ?
- # [18:15] <MikeS> install the open-java
- # [18:16] <MikeS> or whatever they call call
- # [18:16] <MikeS> it
- # [18:16] <MikeS> those other VMs are a total joke
- # [18:16] <DanC> `apt-cache search open-java` loses
- # [18:16] <Philip> openjdk-6-jre I think
- # [18:17] * DanC sees openjdk-6-jre - OpenJDK Java runtime, using Hotspot JIT
- # [18:17] * Quits: Lachy (Lachlan@213.236.208.22) (Quit: This computer has gone to sleep)
- # [18:17] <DanC> "openjdk-6-jre is already the newest version"
- # [18:17] <MikeS> DanC: what Philip said
- # [18:18] <MikeS> openjdk-6-jdk
- # [18:18] <Philip> The JRE ought to be enough, I'd have thought
- # [18:18] <MikeS> and apt-get burn "other java VM junk"
- # [18:18] <DanC> "openjdk-6-jdk is already the newest version"
- # [18:18] <Philip> Maybe there's two separate versions of Java installed?
- # [18:19] <DanC> $ dpkg -S `type -p java`
- # [18:19] <DanC> dpkg: /usr/bin/java not found.
- # [18:19] <MikeS> ls /etc/alternatives/java
- # [18:19] <DanC> $ ls /etc/alternatives/java
- # [18:19] <DanC> /etc/alternatives/java
- # [18:19] <Philip> DanC: update-java-alternatives -s java-6-openjdk
- # [18:19] <Philip> perhaps
- # [18:20] <Philip> or run /usr/lib/jvm/java-6-openjdk/bin/java directly
- # [18:20] <MikeS> fugging idiot debian java toads
- # [18:20] <DanC> `sudo update-java-alternatives -s java-6-openjdk` produced several screenfulls of something... let's see...
- # [18:20] <DanC> but `java -version` gives the same IcedTea foo
- # [18:21] <Philip> Oh
- # [18:21] <Philip> Hmm
- # [18:21] <Philip> http://www.cyberciti.biz/faq/howto-ubuntu-linux-install-configure-jdk-jre/ says more stuff but it's too long so I didn't read it
- # [18:21] <DanC> it does include "OpenJDK Server VM (build 1.6.0_0-b12, mixed mode)"; that's not what we want?
- # [18:21] <pimpbot> Title: Ubuntu Linux Install Sun Java Development Kit ( JDK ) and Java Runtime Environment ( JRE ) (at www.cyberciti.biz)
- # [18:22] <DanC> MikeS, what does `java -version` show for you?
- # [18:23] <DanC> $ file /usr/bin/trang
- # [18:23] <DanC> /usr/bin/trang: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV), for GNU/Linux 2.6.8, dynamically linked (uses shared libs), stripped
- # [18:23] <DanC> is trang actually java?
- # [18:23] <Philip> Maybe it's compiled with gcj
- # [18:24] <Philip> On Ubuntu Hardy I get:
- # [18:24] <Philip> java version "1.6.0"
- # [18:24] <Philip> OpenJDK Runtime Environment (build 1.6.0-b09)
- # [18:24] <Philip> OpenJDK Client VM (build 1.6.0-b09, mixed mode, sharing)
- # [18:24] <DanC> aside from client vs server, that's pretty close
- # [18:24] * Joins: dsinger (dsinger@17.202.35.52)
- # [18:25] <MikeS> mike@yarukinai
- # [18:25] <MikeS> $ java -version
- # [18:25] <MikeS> java version "1.6.0_0"
- # [18:25] <MikeS> OpenJDK Runtime Environment (build 1.6.0_0-b11)
- # [18:25] <MikeS> OpenJDK Client VM (build 1.6.0_0-b11, mixed mode, sharing)
- # [18:25] <Philip> DanC: And apart from the whole IcedTea bit
- # [18:25] <MikeS> DanC:
- # [18:26] <Philip> DanC: (Client vs server is probably because you're on 64-bit and only have the server VM)
- # [18:26] <DanC> hmm... I think I'm running 32 bit ubuntu. (though this is a 64bit machine, but I went back to 32 bits for quicken)
- # [18:26] <Philip> But anyway it sounds kind of like Ubuntu compiled trang with GCJ so the JVM is irrelevant
- # [18:27] <Philip> DanC: (Oh, okay then, it might be something else :-) )
- # [18:27] <DanC> MikeS, where do you get trang from? do you build your own?
- # [18:28] <MikeS> apt-cache show trang
- # [18:29] <MikeS> DanC: fwiw, I find that those Debian f*ucked up java alternatives don't work with validator.nu local install either
- # [18:30] <Philip> https://bugs.launchpad.net/ubuntu/+source/trang/+bug/141447
- # [18:30] <pimpbot> Title: Bug #141447 in trang (Ubuntu): trang segfaults on gutsy (x86 and AMD64) (at bugs.launchpad.net)
- # [18:30] <Philip> The last comments have NullPointerExceptions
- # [18:30] <MikeS> well, that sucks
- # [18:30] <MikeS> my build needs trang
- # [18:30] <Philip> Should be easy enough to use the .jar instead
- # [18:31] <MikeS> yeah
- # [18:31] <Philip> Download trang.jar and run 'java -jar trang.jar com.thaiopensource.relaxng.translate.Driver ...' or something like that, I guess?
- # [18:31] <MikeS> TRANG="java <something> et.c" make
- # [18:32] <MikeS> TRANG="java -jar trang.jar com.thaiopensource.relaxng.translate.Driver ...'" make
- # [18:32] <Philip> Only without the ... or the ' :-)
- # [18:33] <MikeS> problem with debian devs is that they really are convinced they are so much goddamn smarter than everybody else
- # [18:33] <MikeS> e.g, the openssl patch thing
- # [18:34] <MikeS> constantly coming up with non-solutions to non-problems
- # [18:35] <MikeS> i.e., in this case, why do I need gcj-monstrosity version of trang that introduces new bug vectors?
- # [18:35] <MikeS> anyway
- # [18:36] <Philip> Because GCJ is Free
- # [18:36] <DanC> wait... first we worship debian devs ("I use debian for religious reasons") then we villify them? ;-) i'm sure they have good days and bad, just like the rest of us
- # [18:36] <Philip> and it's native code so it must be faster than those silly JITs
- # [18:37] <DanC> I'm using intrepid (ubuntu 8.10), not hardy. I wonder if bug 141447 is still there
- # [18:37] <MikeS> Free has become a euphemism for FUBAR
- # [18:37] <Philip> DanC: Last comment - "I've just tried trang under Intrepid (version 20030619-6.1), and got the same NullPointerException as seen on Hardy."
- # [18:38] <DanC> trang from debian works for you, MikeS ? and it's native code?
- # [18:38] <DanC> ah. I see, Philip
- # [18:38] <MikeS> DanC: gcj trang in Debian unstable works for me in my environment
- # [18:39] <MikeS> "I use debian for religious reasons" was spoken with a shovelfull of irony
- # [18:40] <jgraham> To remind you of the problems with religion?
- # [18:40] <DanC> oh, really? ;-)
- # [18:41] <MikeS> maybe what the free software movement needs is more brutality.. more Pinochet, less flute-playing fat RMS
- # [18:42] <MikeS> hmm, on 2nd thought I take that back
- # [18:42] * Joins: asbjornu (asbjorn@84.48.117.191)
- # [18:42] <MikeS> except for the the "less flute-playing fat RMS" part
- # [18:43] <Philip> Indeed, we should firebomb the houses of developers of immoral non-Free software
- # [18:43] <asbjornu> hi, what's the current status on extensibility in HTML5? is the proposed way forward to include foreign grammars into the main specification on a need for need basis or is there a more general mechanism at place (or in work)?
- # [18:44] * jgraham isn't sure he likes that idea
- # [18:44] <MikeS> asbjornu: case-by-case
- # [18:44] <asbjornu> can I propose an alternative?
- # [18:44] <jgraham> asbjornu: Suggestions are evaluated on a case-by-case basis. Other extension points include class, rel, etc.
- # [18:45] <jgraham> asbjornu: I bet you can
- # [18:45] <MikeS> asbjornu: though there already exist a number of mechanisms for extensibility in html4/html5
- # [18:45] * jgraham maybe didn't quite sound right there
- # [18:45] <asbjornu> nothing as generic as xmlns afaik
- # [18:45] <Julian> indeed
- # [18:45] <asbjornu> so I want to propose something as generic, but compatible with the HTML5 parser
- # [18:46] <Philip> (There's nothing that lets you extend HTML5 with new elements, or with new attributes other than data-*)
- # [18:46] <asbjornu> indeed
- # [18:46] <gsnedders> (In a conforming way)
- # [18:46] <MikeS> asbjornu: we love namespaces
- # [18:46] <gsnedders> http://wiki.whatwg.org/wiki/FAQ#HTML5_should_support_a_way_for_anyone_to_invent_new_elements.21
- # [18:46] <pimpbot> Title: FAQ - WHATWG Wiki (at wiki.whatwg.org)
- # [18:46] <MikeS> asbjornu: if you plan doesn't use namespaces, we will reject it out of hand
- # [18:46] <asbjornu> yea, I've read the FAQ
- # [18:47] <asbjornu> my plan doesn't involve namespaces :)
- # [18:47] <Philip> MikeS: Your shovel is seeing a lot of use today
- # [18:47] <jgraham> also note that using : anywhere is basically not allowed
- # [18:47] <gsnedders> Also, using any characters that gsnedders can't easily type is basically not allowed.
- # [18:47] <asbjornu> it involves a generic mechanism for creating a well-specified DOM out of any foreign or known markup
- # [18:47] <MikeS> asbjornu: oK, at least it *must* use colons, and not, say, dashes. because colons are better
- # [18:47] <jgraham> asbjornu: I suggest proposing the proposal rather than talking about proposing the proposal
- # [18:48] <asbjornu> yea, I'm trying ;)
- # [18:48] <gsnedders> Otherwise I'm sure we'll just become ridiculous :)
- # [18:48] <gsnedders> (As if we weren't already.)
- # [18:48] <Philip> gsnedders: I think we're a long way past that point already
- # [18:48] <gsnedders> Philip: I thought it was a line and not a pont.
- # [18:48] <gsnedders> *point
- # [18:49] <Philip> gsnedders: Well, they're the same thing from a certain perspective
- # [18:49] <asbjornu> if we, at the parser level, say how a given DOM is to be constructed given a set of characters defined as an element with attributes, etc, that is understood at a semantic level in the user agent before-hand, can't this exact same parsing mechanism also be used to create a DOM for completely foreign (at the time, but perhaps not in the future) markup as well?
- # [18:49] <jgraham> gsnedders: Not even 鏝?
- # [18:49] <gsnedders> Philip: Only one very specific perspective though
- # [18:49] <gsnedders> jgraham: Not even.
- # [18:49] <Philip> gsnedders: Actually two
- # [18:49] <gsnedders> Philip: Oh shut up.
- # [18:49] * Joins: Lachy (Lachlan@85.196.122.246)
- # [18:50] <jgraham> asbjornu: Yes
- # [18:50] <gsnedders> asbjornu: Yeah, sure. At the current time you can create any element name you want.
- # [18:50] <asbjornu> in that way, any foreign markup can be declared as such in the DOM and in a validator, but can still be styled, hooked up with scripts, etc., even though the browser has no idea about the semantic meaning and function of the given markup?
- # [18:50] <gsnedders> (except for image)
- # [18:50] <asbjornu> exactly
- # [18:50] <gsnedders> (which has really whacky behaviour for compat)
- # [18:50] <asbjornu> so the deal is basically just to allow any kind of foreign markup and not mark it as non-conforming or invalid
- # [18:51] <Philip> asbjornu: HTML5 already says "<foo><bar>x</foo>" is parsed into a foo element containing a bar element containing a text node; is your proposal just to make that conforming?
- # [18:51] <gsnedders> asbjornu: What happens if we want to add an element "foo" in the future? What do we do about all the people who already use it?
- # [18:51] <Philip> (That seems really bad from a conformance-as-a-form-of-error-checking perspective, since it'll never notice when you misspell tag names)
- # [18:51] <asbjornu> Philip: my proposal is at least to not make it invalid and if possible; conforming, yes
- # [18:52] <DanC> asbjornu, there's quite a diversity of opinion in the HTML WG on how extensibility should work on HTML. We try to track it as issue 41 http://www.w3.org/html/wg/tracker/issues/41
- # [18:52] <pimpbot> Title: ISSUE-41 - HTML Weekly Tracker (at www.w3.org)
- # [18:52] * jgraham was going to say that having <igm> be valid would be annoying
- # [18:52] <DanC> ugh... the description in http://www.w3.org/html/wg/tracker/issues/41 is pretty slanted.
- # [18:52] <pimpbot> Title: ISSUE-41 - HTML Weekly Tracker (at www.w3.org)
- # [18:52] <asbjornu> I assume we will have the same or even better data harvesting techniques for HTML6 as we have now, given the decision to drop certain elements and attributes defined in HTML4 from HTML5 based on that data
- # [18:53] <asbjornu> that same technique can then be used when HTML6 is on the table to survey which elements are safe to define and which are not
- # [18:54] <asbjornu> my assumption is that, mostly, extensions will be defined by working groups and initiatives like microformats and not in a large scale by ad hoc HTML authors, so my bet is that this won't give us a headache down the line
- # [18:55] <Philip> asbjornu: There's already some difficulty with conflicts (e.g. some people wrote pages with <footer> years ago, and Opera had problems with people using WF2 input attribute names for unrelated purposes), so it seems dangerous to encourage more conflicts because there won't be any decent names left for HTML6 to use
- # [18:55] <asbjornu> true
- # [18:56] <Julian> sounds like paritioning the set of valid names would make sense, so HTML* can alyways draw from a reservoir not allowed to be used by anybody else
- # [18:56] <Philip> asbjornu: There seems to be quite a lot of evidence of ad hoc making-stuff-up, e.g. http://blog.unto.net/web/a-survey-of-rel-values-on-the-web/comment-page-1/ found 1.8 million distinct 'rel' values
- # [18:56] <pimpbot> Title: A Survey of Rel Values on the Web DeWitt Clinton (at blog.unto.net)
- # [18:57] <gsnedders> Let's prefix all HTML* elements with "html-" and we'll be saved!
- # [18:58] <gsnedders> Seriously though, if we do really allow any element, I'd rather we just followed what xml-names already does, just using a hyphen instead of a colon
- # [18:58] <Philip> and if you scroll down http://philip.html5.org/data/attr-count-pages.txt then there's lots of typos and lots of what look like large scale ad hoc making-stuff-up (and this is despite validators complaining about it)
- # [18:59] <gsnedders> Philip: That means we can't create a @cat :(
- # [19:00] <gsnedders> How am I going to live in a world without semantic lolcats?
- # [19:00] <Philip> Julian: That partitioning seems to be the approach taken by data-*
- # [19:00] <Philip> Julian: but it seems a bit ugly :-(
- # [19:01] <Philip> (particularly for anything that's meant to be used more widely than a single site)
- # [19:01] <Julian> Yes, I would recommend another way to do that.
- # [19:01] <asbjornu> hm
- # [19:01] <asbjornu> I'll think a bit about it over dinner
- # [19:01] * asbjornu is now known as asbjornu|dinner
- # [19:01] <Julian> I think it's used in some other markup languages.
- # [19:01] <gsnedders> Dinner is always a good idea.
- # [19:01] <MikeS> dinner+wine is better
- # [19:02] <Philip> A while ago I proposed reserving any attribute names whose MD5 sum started with an even digit, which would allow plenty of scope for choosing good names on either side of the partition
- # [19:02] <gsnedders> MikeS: I disagree.
- # [19:06] <Philip> gsnedders: I think lolcats are inherently presentational, so they should be handled by CSS
- # [19:09] <MikeS> DanC: so have you managed to actually build successfully?
- # [19:09] * Joins: hober (ted@206.212.254.2)
- # [19:10] * Quits: dsinger (dsinger@17.202.35.52) (Quit: dsinger)
- # [19:12] <MikeS> DanC: I got to drop off relatively soon
- # [19:13] * MikeS is on 4 times the recommended dosage of 援/allergy meds, plus two tall cans of Yebisu: The Hop
- # [19:13] * Joins: adele (adele@17.255.108.202)
- # [19:14] * Quits: adele (adele@17.255.108.202) (Client exited)
- # [19:14] * Joins: adele (adele@17.255.108.202)
- # [19:15] <gsnedders> MikeS: Drop off voluntarily or will you be forced to by the meds?
- # [19:17] <MikeS> trying to do the math.. 3am here, and I have a 10am meeting
- # [19:17] <MikeS> so that gives me 13 hours to slepp
- # [19:17] <MikeS> sleeeeep
- # [19:17] <MikeS> excellent
- # [19:17] * Quits: maddiin (mc@87.185.198.194) (Quit: maddiin)
- # [19:18] <Philip> MikeS: http://www.google.com/search?q=10-3 may help you
- # [19:18] <pimpbot> Title: 10-3 - Google Search (at www.google.com)
- # [19:21] <MikeS> Philip: that pages appears to be saying I have will either get 7 days of sleep or will need to catch the 3:10 to Yuma
- # [19:22] * MikeS is away: 3:10toYuma
- # [19:23] * Joins: smedero (smedero@192.223.6.251)
- # [19:25] * Joins: ChrisWilson (cwilso@131.107.0.74)
- # [19:34] * asbjornu|dinner is now known as asbjornu
- # [19:36] <asbjornu> ok, so perhaps a large scale "everyone's invited" extension mechanism isn't the best way forward, but how about not requiring more than an easy IANA registration or something similar instead?
- # [19:36] <asbjornu> head@profile could be the place to put the extension information, in the form of an URI (if the extensions aren't registered) or a keyword (if the extensions are registered)
- # [19:37] <asbjornu> I know head@profile is dropped from HTML5 at the moment, but it could be reintroduced with a larger and more important purpose than it has in HTML4
- # [19:38] <asbjornu> how about it? I do, however, have to say that extending based on reserved prefixes of MD5 sums sounds like a very intriguing and well thought out idea! ;)
- # [19:39] * DanC tunes back in after lunch, wondering if MikeS is still around
- # [19:39] <DanC> no, I haven't built yet...
- # [19:40] <DanC> didn't chase down trang (did some WAI coordination email )
- # [19:40] * Quits: adele (adele@17.255.108.202) (Quit: adele)
- # [19:40] <smedero> http://groups.google.com/group/mozilla.dev.platform/browse_thread/thread/0c2bbb6ed726800b
- # [19:40] <pimpbot> Title: An alternate take on HTML5 - mozilla.dev.platform | Google Groups (at groups.google.com)
- # [19:55] * Joins: rking3 (rking3@99.27.42.97)
- # [19:55] <pimpbot> planet: Bespin and Canvas: Part 1 <http://benzilla.galbraiths.org/2009/02/17/bespin-and-canvas/>
- # [20:04] * Joins: aroben (aroben@69.142.95.189)
- # [20:16] <Philip> asbjornu: Why would people bother with the registration process and adding a profile attribute, when they could just use the new elements/attributes without all that work (like they already do today)?
- # [20:16] * Quits: laplink (link@78.26.2.115) (Quit: This computer has gone to sleep)
- # [20:22] * Joins: laplink (link@78.26.2.115)
- # [20:40] <asbjornu> because the profile could have an effect on how the element is interpreted?
- # [20:40] <gsnedders> asbjornu: But nobody bothers with @profile.
- # [20:41] <gsnedders> asbjornu: So at least pre-existing things need to cope with it not being there
- # [20:41] <Philip> asbjornu: Interpreted by what?
- # [20:41] <Philip> gsnedders: That's irrelevant since we're talking about new things, not pre-existing things
- # [20:42] <asbjornu> yes, Philip is right, this is for forward compatibility, not backwards
- # [20:43] <asbjornu> I mean interpreted by the user agent, e.g. given functionality and a semantic meaning
- # [20:44] <asbjornu> for a given registered profile, elements defined in that profile and their underlying semantics and functionality can be implemented by a UA in a distributed but still controlled manner
- # [20:44] <Philip> asbjornu: Would <link rel=stylesheet> and <script src> not be sufficient for providing the new features' functionality in UAs?
- # [20:45] <asbjornu> I'm talking about the whole smack you get form a defined element and/or attribtue(s) today, including semantics
- # [20:45] <asbjornu> If all I wanted was extensible functionality, then yes, scripts and perhaps stylesheets might have fit the bill
- # [20:46] <Philip> asbjornu: What do you mean by "semantics"? It seems a fairly vague term :-)
- # [20:46] <asbjornu> but I want HTML extensible in all directions, so a future MathML, SVG or whatever might be integrated into user agents long before the HTML WG have convened and thrown a new draft of HTML on the table
- # [20:46] * Quits: MichaelC (Michael@128.30.52.30) (Ping timeout)
- # [20:47] <asbjornu> "semantics" is a fairly vague term on a semantical level :P
- # [20:47] <gsnedders> asbjornu: But that realistically needs namespaces and not what you're proposing :P
- # [20:48] <asbjornu> I don't think it does although I agree that namespaces have solved that problem for XML quite nicely (although not very elegantly)
- # [20:48] <gsnedders> asbjornu: MathML and SVG both define themselves as being in x namespace, so we'd need a way to get an element in that namespace
- # [20:48] <asbjornu> I think <head profile="mathml svg"> could work
- # [20:49] <asbjornu> that's the work of the profile
- # [20:49] <gsnedders> asbjornu: But if I don't know of the SVG profile, I can't put it in the namespac
- # [20:49] <gsnedders> asbjornu: Which means I get a different DOM in one that knows the profile and one that doesn't, which sucks.
- # [20:50] <Philip> MathML and SVG require years of work to implement their rendering and DOM interfaces in browsers, so the syntax is a pretty trivial issue
- # [20:50] <asbjornu> for HTML, namespaces are non-existent, so that's in fact a non-issue
- # [20:50] <gsnedders> asbjornu: For SVG and MathML they are existant, so that is an issue
- # [20:50] <asbjornu> for XHTML, that might be a problem if the namespace is required
- # [20:51] <Philip> Namespaces do exist in HTML - you can write createElementNS etc
- # [20:51] <gsnedders> asbjornu: Their namespace is required by the SVG and MathML specifications.
- # [20:51] <Philip> (There's just no syntax for them)
- # [20:51] * Joins: karl (karlcow@70.81.32.202)
- # [20:51] <asbjornu> I know, but I don't think the namespace URI has a purpose in (X)HTML
- # [20:52] <gsnedders> For embedding anything like SVG and MathML is certainly does
- # [20:52] <asbjornu> at least not when you get the grammar defined through a profile
- # [20:52] <gsnedders> asbjornu: How is @profile better?
- # [20:52] <asbjornu> it's compatible with HTML
- # [20:52] <asbjornu> namespaces isn't
- # [20:52] <asbjornu> @profile is a layer of indirection; an abstraction
- # [20:52] <gsnedders> asbjornu: But SVG isn't compatible with HTML, so that's not an issue
- # [20:52] <pimpbot> asbjornu: Huh?
- # [20:53] <gsnedders> pimpbot: Shut up you stupid bot.
- # [20:53] <pimpbot> gsnedders: Huh?
- # [20:53] <asbjornu> SVG can be made fully compatible with HTML
- # [20:53] <gsnedders> Just make the svg element magic and make desendent-or-self go in the SVG namespace, and all works fine
- # [20:54] <asbjornu> it's basically just to drop the 'xmlns' attribute, define an 'svg' profile and do <head profile="svg">
- # [20:54] <gsnedders> asbjornu: The problem is then the UA needs a mapping of @profile to namespace URI
- # [20:54] <Philip> asbjornu: And then make the profile attribute optional, and it'll be like HTML5 :-)
- # [20:54] <gsnedders> From a DOM POV the svg elements have to be in the svg namespace
- # [20:55] <Philip> (Well, forbidden rather than optional)
- # [20:55] * gsnedders goes back to doing schoolwork due in yesterday
- # [20:55] <asbjornu> only if it needs that namespace URI for anything. what purpose does the namespace URI have other than define the elements within it belonging to a specified grammar?
- # [20:55] <asbjornu> from a DOM POV, the namespace is only relevant because SVG defines it to be relevant. HTML5 can circumvent this, imho
- # [20:56] <asbjornu> if you introduce foreign markup (markup defined out of band of the HTML5 specification), a @profile attribute could be made required
- # [20:57] * Joins: MichaelC (Michael@128.30.52.30)
- # [20:58] <asbjornu> back to the namespace URI; the grammar a given namespace defines the meaning of can just as well be defined by a keyword in @profile
- # [20:59] <asbjornu> an unknown namespace URI yields the exact same result as an unknown @profile keyword would, so in that respect it's the same
- # [20:59] <asbjornu> a known namespace URI yields the same result in XHTML as a @profile keyword, but the @profile approach has the advantage of being portable and compatible with HTML as well
- # [21:00] <asbjornu> what I'm thinking of is something like this for registered values of @profile: http://www.iana.org/assignments/link-relations/link-relations.xhtml
- # [21:00] <pimpbot> Title: Atom Link Relations (at www.iana.org)
- # [21:04] <asbjornu> ...
- # [21:06] * Joins: sryo (sryo@190.245.204.198)
- # [21:07] <Philip> asbjornu: I'm still not sure what you're suggesting the profile attribute should do... If I have an HTML document that uses SVG like you're suggesting, and then remove the profile attribute, what would change?
- # [21:12] <asbjornu> the SVG elements would be rendered as unknown markup
- # [21:13] <asbjornu> only with the correct @profile value in place would the SVG elements be given proper meaning to the UA
- # [21:13] * Joins: adele (adele@17.244.16.159)
- # [21:13] <Philip> asbjornu: Why would that be an advantage compared to a design where SVG elements are always rendered as SVG elements, regardless of the profile?
- # [21:14] <asbjornu> they aren't in HTML
- # [21:15] <asbjornu> tbh, the 'xmlns' could in fact be kept as is, but would give absolutely no meaning to an HTML UA
- # [21:22] <Philip> Whether they are "in HTML" seems a fairly arbitrary distinction - if a UA implements SVG, it could just consider the SVG elements to be "in HTML" (like how HTML5 already handles MathML)
- # [21:35] <asbjornu> sorr if I was unclear, what I meant was that the SVG elements aren't rendered as SVG elements "in HTML" serialization
- # [21:36] <asbjornu> the problem with including support for a grammar on a case per case basis is that it doesn't scale, and isn't what I'd call an extension mechanism, but rather the quite opposite
- # [21:37] <asbjornu> having support for MathML in HTML should not be up to HTML for anything else than having defined a consistent and clear way to extend the language so user agents can implement it regardless of what the HTML WG thinks about the grammar and whether it deserves the WG's blessing for inclusion in HTML6 in 2014
- # [21:44] <karl> http://blog.unto.net/web/a-survey-of-rel-values-on-the-web/
- # [21:44] <pimpbot> Title: A Survey of Rel Values on the Web DeWitt Clinton (at blog.unto.net)
- # [21:52] * Quits: MikeS (MikeSmith@mcclure.w3.org) (Ping timeout)
- # [22:11] <hober> DanC, re: your C-p / C-<up> Emacs shell question from a few hours ago, M-p/M-n are (usually) the way to cycle through history, in shell modes, the minibuffer, and elsewhere.
- # [22:12] * DanC gives it a try...
- # [22:12] <DanC> nifty. thanks.
- # [22:21] * Quits: tlr (tlr@128.30.52.30) (Quit: tlr)
- # [22:23] * Quits: ROBOd (robod@89.122.216.38) (Quit: http://www.robodesign.ro )
- # [22:29] <DanC> MikeS, you were going to raise an issue about mime types... I thought you did that, but I don't see it in http://www.w3.org/html/wg/tracker/issues
- # [22:29] <pimpbot> Title: Issues - HTML Weekly Tracker (at www.w3.org)
- # [22:29] * DanC hunts for paper trail...
- # [22:31] * DanC finds the internal action...
- # [22:32] <DanC> ah. http://www.w3.org/html/wg/tracker/issues/53
- # [22:32] <pimpbot> Title: ISSUE-53 - HTML Weekly Tracker (at www.w3.org)
- # [22:37] * Joins: shepazu (schepers@128.30.52.30)
- # [22:42] * Joins: maddiin (mc@87.185.200.240)
- # [22:49] * Quits: maddiin (mc@87.185.200.240) (Quit: maddiin)
- # [22:50] * Philip wishes (in the context of the canvas thread) there were fewer "I think accessibility is important" posts and more "I propose this feature that I think will improve accessibility" posts, because the latter could lead to something productive
- # [22:53] * Joins: heycam (cam@61.88.219.66)
- # [23:12] * Joins: jwatt (roslea@61.88.219.66)
- # [23:32] * Quits: laplink (link@78.26.2.115) (Quit: This computer has gone to sleep)
- # [23:36] * Quits: MichaelC (Michael@128.30.52.30) (Quit: ChatZilla 0.9.84 [Firefox 3.0.6/2009011913])
- # Session Close: Thu Feb 19 00:00:01 2009
The end :)