/irc-logs / w3c / #html-wg / 2007-03-22 / end

Options:

# Session Start: Thu Mar 22 00:00:00 2007
# Session Ident: #html-wg
# [00:01] * Quits: kingryan (kingryan@66.92.180.242) (Connection reset by peer)
# [00:01] * Joins: kingryan (kingryan@66.92.180.242)
# [00:02] * Quits: Lachy_ (chatzilla@220.245.91.147) (Client exited)
# [00:03] * Joins: Lachy_ (chatzilla@220.245.91.147)
# [00:11] * Quits: billmason (billmason@69.30.57.156) (Quit: .)
# [00:36] * Quits: marcos (chatzilla@131.181.148.226) (Connection reset by peer)
# [00:39] * Quits: tylerr (tylerr@66.195.32.2) (Quit: Leaving)
# [00:43] * Quits: hober (ted@66.162.32.234) (Quit: ERC Version 5.1.3 (IRC client for Emacs))
# [00:49] * Joins: marcos (chatzilla@131.181.148.226)
# [00:52] * Joins: karl (karlcow@128.30.52.30)
# [00:56] * Quits: bfults (bfults@66.162.32.234) (Quit: |)
# [00:57] * Parts: hasather (hasather@81.235.209.174)
# [01:02] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
# [01:07] * Joins: gavin (gavin@74.103.208.221)
# [01:25] * Quits: bewest (ben@209.237.236.227) (Quit: Leaving.)
# [01:37] * Joins: chaals (chaals@212.181.250.84)
# [01:40] * Quits: hsivonen (hsivonen@130.233.41.50) (Ping timeout)
# [01:42] * Joins: hsivonen (hsivonen@130.233.41.50)
# [01:46] * Quits: kingryan (kingryan@66.92.180.242) (Client exited)
# [01:46] * Joins: kingryan (kingryan@66.92.180.242)
# [01:55] * Quits: chaals (chaals@212.181.250.84) (Ping timeout)
# [02:05] * Quits: kingryan (kingryan@66.92.180.242) (Client exited)
# [02:16] * Joins: olivier (ot@128.30.52.30)
# [03:10] * Quits: quaiz (quaiz@66.162.32.234) (Ping timeout)
# [03:10] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
# [03:15] * Joins: gavin (gavin@74.103.208.221)
# [03:29] * Joins: Bob_le_Pointu (mallory@80.248.208.232)
# [04:02] * Quits: mjs (mjs@17.255.99.134) (Quit: mjs)
# [04:11] * Joins: glazou (daniel@212.180.54.82)
# [04:25] * Joins: quaiz (quaiz@70.181.154.111)
# [04:34] * Quits: quaiz (quaiz@70.181.154.111) (Quit: quaiz)
# [04:39] * Joins: colin_lieberman (colin@24.5.197.118)
# [04:40] * Quits: glazou (daniel@212.180.54.82) (Quit: zzzz)
# [04:51] * Joins: billmason (billmason@71.236.194.106)
# [05:00] * Quits: colin_lieberman (colin@24.5.197.118) (Quit: colin_lieberman)
# [05:08] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
# [05:17] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
# [05:23] * Joins: gavin (gavin@74.103.208.221)
# [05:42] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Ping timeout)
# [05:56] * Quits: dbaron (dbaron@63.245.220.242) (Quit: 8403864 bytes have been tenured, next gc will be global.)
# [06:07] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
# [06:20] * Joins: mjs (mjs@64.81.48.145)
# [06:25] * Joins: dbaron (dbaron@71.198.189.81)
# [06:25] * Joins: bfults (bfults@70.95.237.98)
# [06:28] * Quits: marcos (chatzilla@131.181.148.226) (Ping timeout)
# [06:30] * Joins: marcos (chatzilla@131.181.148.226)
# [06:54] * Joins: Charl (charlvn@196.2.153.111)
# [06:54] * Joins: quaiz (quaiz@70.181.154.111)
# [06:59] * Joins: preston (chatzilla@70.181.71.135)
# [07:14] * Quits: Lachy_ (chatzilla@220.245.91.147) (Quit: Chatzilla 0.9.77 [Firefox 2.0.0.3/2007030919])
# [07:23] * Quits: bfults (bfults@70.95.237.98) (Quit: bfults)
# [07:25] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
# [07:30] * Joins: gavin (gavin@74.103.208.221)
# [07:35] * Quits: preston (chatzilla@70.181.71.135) (Quit: Chatzilla 0.9.77 [Firefox 2.0.0.2/0000000000])
# [07:41] * Quits: Hixie (ianh@129.241.93.37) (Ping timeout)
# [07:44] * Joins: Hixie (ianh@129.241.93.37)
# [07:50] * karl has approved a 30 something persons again today
# [08:05] * Joins: icaaq (icaaaq@85.228.53.245)
# [08:17] * Quits: dbaron (dbaron@71.198.189.81) (Quit: g'night)
# [08:19] * Quits: billmason (billmason@71.236.194.106) (Quit: billmason)
# [08:21] * Joins: anne (annevk@86.90.70.28)
# [08:30] * Quits: karl (karlcow@128.30.52.30) (Quit: Where dwelt Ymir, or wherein did he find sustenance?)
# [08:32] * Quits: olivier (ot@128.30.52.30) (Quit: This computer has gone to sleep)
# [08:48] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Quit: Get thee behind me, satan.)
# [09:21] <anne> RRSAgent, make logs public
# [09:21] <RRSAgent> I have made the request, anne
# [09:22] <anne> RRSAgent, pointer?
# [09:22] <RRSAgent> See http://www.w3.org/2007/03/22-html-wg-irc#T08-25-46
# [09:32] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
# [09:37] * Joins: gavin (gavin@74.103.208.221)
# [09:43] * Joins: schnitz (Miranda@85.181.10.132)
# [09:46] <schnitz> hi everyone :-)
# [09:47] * Joins: matt (fire4x@84.9.127.20)
# [09:48] <anne> morning
# [09:48] <schnitz> mornin' anne :-)
# [09:54] <mjs> hello
# [10:07] <anne> why are the W3C list archives offline?
# [10:09] <Bob_le_Pointu> Morning.
# [10:31] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
# [10:59] * Joins: ROBOd (robod@86.34.246.154)
# [11:39] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
# [11:44] * Joins: gavin (gavin@74.103.208.221)
# [11:57] * Quits: mjs (mjs@64.81.48.145) (Ping timeout)
# [11:58] <schnitz> anne, yeah, weird, can't remember this ever happening before...
# [11:59] <schnitz> I also sent a message to public-html, and it doesn't seem to get thru
# [12:04] <MikeSmith> schnitz - unfortunately the W3C list server has been down for several hours
# [12:04] <MikeSmith> no W3C mailing-list mail at all is going through right now
# [12:04] <anne> MikeSmith, any idea when it will get back online?
# [12:05] <MikeSmith> anne - nope
# [12:05] <MikeSmith> might not be much happen until US/East systeam people get online
# [12:05] * Joins: mjs (mjs@64.81.48.145)
# [12:06] * anne mumbles
# [12:46] * Parts: icaaq (icaaaq@85.228.53.245)
# [12:52] * Joins: glazou (daniel@80.118.184.70)
# [12:52] <glazou> bonjour
# [12:53] <glazou> DanC: ping
# [12:54] <schnitz> bonjour glazou :-)
# [12:54] <glazou> goodday schnitz
# [12:55] <glazou> long time no see
# [12:55] <glazou> last time in mandelieu right ?
# [12:55] <glazou> or ac in tokyo ?
# [12:56] <schnitz> mandelieu :-)
# [12:57] <schnitz> that was great
# [12:57] <anne> yeah, mandelieu was nice
# [12:57] <schnitz> yep :-)
# [12:58] <glazou> what I love really is arriving by car from the highway, open the car's windows and let the mimosa smell go through the car
# [12:58] <schnitz> aren't we having one of those dark and cold bostom TPs this fall again? :-)
# [12:58] <glazou> schnitz: I am freezing already only thinking about it :)
# [12:59] <schnitz> glazou, oh yeah, I've been driving to mandelieu all the way from munich, and in munich is was deep winter with lots of snow, and in mandelieu it was spring, that was amazing...
# [12:59] * glazou should suggest this excellent hotel full of meeting rooms in Tenerife
# [13:08] * glazou is now known as glazou_afk
# [13:08] <anne> lists are online again...
# [13:13] * Quits: Charl (charlvn@196.2.153.111) (Quit: Leaving)
# [13:13] <schnitz> ah
# [13:18] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Quit: Get thee behind me, satan.)
# [13:21] <schnitz> ok, lists are up again, I sent a message though earlier today, and it doesn't seem to get there... wondering whether I should re-send
# [13:21] <anne> dunno
# [13:22] <schnitz> hmm, I'll wait a little...
# [13:22] <schnitz> maybe some sys folks in boston are currently fishing for lost mail...
# [13:23] * Joins: citoyen (eira@195.139.204.228)
# [13:43] * Quits: matt (fire4x@84.9.127.20) (Quit: matt)
# [13:52] <@DanC> glazou_afk, pong. but I've got an 8am telcon, and I'm chairing.
# [13:56] * Joins: olivier (ot@128.30.52.30)
# [14:12] <schnitz> bbl
# [14:23] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
# [14:28] * Joins: gavin (gavin@74.103.208.221)
# [14:58] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
# [15:29] * Joins: bfults (bfults@70.95.237.98)
# [15:30] * Joins: bfults_ (bfults@70.95.237.98)
# [15:30] * Quits: bfults (bfults@70.95.237.98) (Connection reset by peer)
# [15:35] * Joins: billmason (billmason@69.30.57.156)
# [15:39] <anne> ah
# [15:39] <anne> e-mails are arriving on W3C lists now...
# [15:41] <Dave> apparently there was some problem with blogging that impacted other services, but I understand that this has now been fixed.
# [15:44] <anne> over 150 IE
# [15:44] <anne> s
# [15:46] <Dave> we should conduct a survey to get a feeling for their backgrounds
# [15:47] <Lachy> what kind of information do we need to know about them?
# [15:48] * Quits: olivier (ot@128.30.52.30) (Quit: Leaving)
# [15:49] <Dave> it would be nice to know where they are based in the world, and what they do, e.g. role in development team, what kinds of websites they work on etc.
# [15:50] <anne> for statistics?
# [15:53] * MikeSmith notes posting to public-html from Murray Maloney
# [15:53] <Dave> yes, preferably as pretty graphs
# [15:54] <Lachy> perhaps, but I find that kind of information isn't really useful till you have an idea of their personality, which is determined by the way the express themselves in their e-mails
# [15:54] <anne> MikeSmith, hmm, the latest HTML WD?
# [15:54] <anne> that's an awful document :)
# [15:54] <MikeSmith> anne, yeah that
# [15:54] <MikeSmith> that message
# [15:55] <MikeSmith> I'm wondering if Murray is aware of the WHATWG work at all
# [15:55] <Lachy> I don't appear to have received that e-mail yet. What's the subject line?
# [15:56] <Dave> It would be unlikely that he isn't aware of it.
# [15:56] <anne> Lachy, refresh
# [15:56] <Lachy> I did
# [15:56] <billmason> Lachy, "Straw man proposal to build an agenda/issue list"
# [15:56] <Lachy> I'll check the list archive\
# [15:57] <MikeSmith> Dave, what is the URI for the latest draft?
# [15:57] <anne> w3.org/tr/html/
# [15:57] <Dave> Good question, in theory it should be on the WG page if Karl has done his bit
# [15:57] <anne> or maybe w3.org/tr/html4/
# [15:58] <Lachy> /TR/html/ should change to HTML5 when it moves to W3C, /html4/ should stay as is
# [15:59] <anne> yeah, prolly
# [15:59] <Dave> I am not sure where we are in respect to being able to make HTML5 into a formal WD though.
# [15:59] <MikeSmith> hmm, by "drafts", does Murray mean the public RECs?
# [15:59] <anne> a lot further than last year
# [15:59] <Dave> good question
# [15:59] <anne> MikeSmith, there's not much else
# [16:01] <Lachy> Dave, we need to find out the exact conditions under which Apple, Mozilla, Opera and Hixie would agree to publish it at the W3C
# [16:01] <Lachy> One of them would probably be to ensure that the spec on whatwg.org and w3c.org are the same spec!
# [16:01] <Dave> They could make a formal contribution subject to the Patent Policy, that's easy enough
# [16:01] <Dave> but what about the other contributors?
# [16:02] <Dave> I am sure that Ian could generate the WD format easily enough.
# [16:02] <anne> what about other significant text?
# [16:02] * anne is only aware of the HTML5 proposal so far
# [16:02] <Dave> e.g. ?
# [16:02] <Lachy> have most of the invited experts come across from the whatwg?
# [16:02] <anne> dunno
# [16:02] <anne> you're bringing it up :)
# [16:03] <MikeSmith> Lachy - I don't know about "most" but there are many that have come from elsewhere
# [16:03] <anne> s/significant text/proposals/
# [16:03] <Dave> If there is some kind of record of who contributed to the spec then we can ask every such contribute to agree to W3C PP in regards to their contribution.
# [16:03] <anne> maybe significant proposals
# [16:04] <Lachy> hmm. looks like the subscriber list for whatwg is no longer available, so I can't compare the list of e-mails with the names in the HTMLWG
# [16:04] <Dave> Current memebers of HTML WG have already made such a commitment.
# [16:04] <MikeSmith> I guess the WHATWG work could be seen being already existing work that attempts to accomplish the first part of what Murray suggests
# [16:04] <anne> http://www.whatwg.org/specs/web-apps/current-work/#acknowledgements
# [16:05] <anne> has all the names
# [16:05] <anne> but I wonder if that's really needed
# [16:05] <Lachy> only those who have contributed to the spec, none of the lurkers are included
# [16:05] <Dave> Lurkers don't matter if they don't contribute to the spec.
# [16:06] <Lachy> I suppose the number of lurkers on the HTMLWG would be comparitively few
# [16:06] <Lachy> at the moment, anyway
# [16:06] <Dave> Lurkers on WhatWG list that is.
# [16:10] <Dave> The list of people given in the acknowledgements should be good enough if it is deemed to be reasonably accurate.
# [16:10] <anne> it still seems silly to me as you don't know who influenced those people, etc.
# [16:11] <anne> (and the same goes for current REC track documents, btw)
# [16:11] <Dave> There are limits to what we can practically do to protect implementers (and end users)
# [16:11] <Dave> W3C Patent Policy is the best we have been able to come up with so far.
# [16:12] <Dave> I have been involved in several Patent Advisory Groups and so far things have worked out pretty good with exception of Eoalas case.
# [16:12] <anne> and that has been pretty bad
# [16:13] <Dave> yes, it wasn't pretty
# [16:13] <Dave> but going forward the work arounds aren't too bad.
# [16:19] * Joins: kingryan (kingryan@64.81.240.149)
# [16:21] * Joins: N-K (N-K@85.68.10.172)
# [16:22] <Lachy> I have the lists of names prepared from each group, does anyone know a quick method of comparing them?
# [16:22] * Quits: N-K (N-K@85.68.10.172) (Quit: Chatzilla 0.9.77 [Firefox 2.0.0.2/2007021917])
# [16:23] * Quits: bfults_ (bfults@70.95.237.98) (Quit: bfults_)
# [16:24] * glazou_afk is now known as glazou
# [16:29] * Joins: NicolasLG (me@86.220.181.223)
# [16:29] <NicolasLG> Hi everybody!
# [16:30] <anne> allo
# [16:30] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
# [16:31] * @DanC finishes a telcon, tunes in, not sure what priority he can put on IRC office hours today...
# [16:36] * Joins: gavin (gavin@74.103.208.221)
# [16:37] * Joins: tylerr (tylerr@66.195.32.2)
# [16:37] <tylerr> Morning all.
# [16:40] <anne> Murray seems to live in the future
# [16:41] <Dave> that's where we are all heading ...
# [16:41] <Dave> He's right to focus on agenda and plans.
# [16:42] <@DanC> I'd rather he didn't actually.
# [16:42] * anne meant the e-mail date
# [16:42] * DanC changes topic to 'W3C HTML WG http://www.w3.org/html/wg/ - http://www.w3.org/2007/03/22-html-wg-irc (logged)'
# [16:42] <@DanC> grumble... the log from yesterday is very incomplete. I sent a sysreq
# [16:43] <anne> DanC, http://krijnhoetmer.nl/irc-logs/
# [16:43] <anne> (has all the important logs)
# [16:43] <anne> with respect to HTML anyway
# [16:43] * tylerr reading up on the Apple audio/video submission.
# [16:44] <@DanC> ah... thanks, anne
# [16:53] <glazou> hi DanC
# [16:54] <glazou> my potential host for the ftf pinged me this morning, they'd like to have more details (dates, number of attendees) to give their formal agreement
# [16:55] <tylerr> Hi Dan, have a minute to answer a question or two in PM?
# [16:55] * Quits: quaiz (quaiz@70.181.154.111) (Quit: quaiz)
# [16:56] <tylerr> DanC rather.
# [16:58] * Joins: icaaq (icaaaq@85.228.53.245)
# [17:04] * glazou notes that dan rather is retired and never standardized html (sorry could not resist)
# [17:05] * tylerr smiles.
# [17:05] * Dave thinks that after chairing this WG, Dan may well want to retire! :-)
# [17:06] <glazou> ROTFL
# [17:06] <glazou> we need a fortunes page in this HTML WG...
# [17:10] * Quits: NicolasLG (me@86.220.181.223) (Ping timeout)
# [17:10] <tylerr> Well folks, just got accepted into the working group, glad to be here and to be working with all of you!
# [17:12] * Quits: kingryan (kingryan@64.81.240.149) (Quit: kingryan)
# [17:14] <icaaq> tylerr: congrats!
# [17:15] <tylerr> Thank you icaaq. :)
# [17:15] <@DanC> hi glazou . It would help if you would suggest a date
# [17:17] <@DanC> the survey shows 39 people interested in a meeting, though not all of them are interested/available for a meeting in france. http://www.w3.org/2002/09/wbs/40318/ftf07/results
# [17:18] <anne> 21-05 / 23-05, Paris
# [17:18] * @DanC struggles to decode those dates
# [17:18] <schnitz> gna, re-sent my mail and still not getting thru, I guess it will pop up twice then, like tomorrow
# [17:18] <@DanC> that's 21 and 23 May?
# [17:18] <anne> 21 May to 23 May
# [17:19] <anne> (if it's your typical three-day meeting)
# [17:19] <@DanC> I don't see anything typical about this meeting
# [17:19] <anne> :)
# [17:19] <schnitz> ;-)
# [17:19] * schnitz is thining: humor?
# [17:19] <anne> it's the first Monday after XTech (which ends on Friday)
# [17:20] <schnitz> ups, s/thining/thinking :-)
# [17:20] <tylerr> I would need my company to fund my travel. Unfortunately we run a tight ship. **chuckles**
# [17:20] <@DanC> hmm... I'd like to go to xtech, and not stay over the weekend.
# [17:20] <schnitz> AH
# [17:20] <schnitz> now it came thru
# [17:20] <@DanC> maybe May 14-15?
# [17:21] <anne> that works too, although I believe some bits of XTech start on 15
# [17:21] <@DanC> yeah... I'm checking the XTech schedule... http://2007.xtech.org/public/schedule/grid
# [17:22] * @DanC wonders if he lost glazou
# [17:22] <glazou> I'm here
# [17:22] <glazou> but have to go home in 10 minutes
# [17:22] <glazou> my host and xtech are not in the same part of paris
# [17:23] <glazou> xtech is west
# [17:23] <glazou> host is south
# [17:23] <glazou> both inside city limits
# [17:23] <anne> Paris has a reasonable metro system...
# [17:23] <@DanC> well, it's the same airport, yes?
# [17:23] <glazou> DanC: rofl
# [17:23] <anne> heh
# [17:23] <Dave> schnitz, I've got your email now
# [17:23] <@DanC> er... I was serious; NY has more than one airport, as does the SFO area
# [17:23] <glazou> probably 20 minutes by subway between both
# [17:24] <@DanC> ah... 20 minutes by subway means people don't even have to change hotels
# [17:24] <glazou> DanC: sure
# [17:24] <glazou> DanC: paris subway network is excellent
# [17:24] <@DanC> I'm OK with overlapping XTech on 15 May; that's the tutorial day
# [17:24] <glazou> yeah, I think that day is reasonably available for us
# [17:25] <@DanC> I don't see any WG members on the tutorial schedule, though Steven P. might be interested to attend.
# [17:25] <Dave> Hmm, I am chairing xtech ubiweb track on Tuesday May 15
# [17:25] <anne> Dave Raggett is on there
# [17:25] <@DanC> all day, dave? or just onen day?
# [17:25] <@DanC> phph. or just one half?
# [17:26] <Dave> Let me check
# [17:26] <glazou> DanC: are you suggesting 14 and 15 may ?
# [17:26] <@DanC> I can see when it starts from http://2007.xtech.org/public/schedule/detail/19 but not when it ends
# [17:26] <@DanC> yes, I'm thinking about May 14-15
# [17:26] <glazou> ok
# [17:26] * glazou will ping the host
# [17:26] <glazou> I have to go on daddy duty
# [17:26] <glazou> bbl
# [17:26] <@DanC> enjoy
# [17:27] * Quits: glazou (daniel@80.118.184.70) (Quit: glazou)
# [17:27] <Dave> xtech ubiweb track is whole day, see http://2007.xtech.org/public/schedule/topic/7
# [17:27] <@DanC> ah. I see.
# [17:28] <@DanC> counter-proposal? or are you ok with attending just the 1st of 2 days?
# [17:29] <Dave> I could be, depends on the agenda
# [17:29] <Dave> MikeSmith is presenting on Wednesday 16th at 11am on the future of HTML
# [17:29] <Dave> followed by Molly
# [17:29] <@DanC> yeah; I'd like to be there for that
# [17:32] <Dave> Hey I see that Antoine Quint now works for Joost and is talking on the 17th
# [17:33] <Dave> Henri Sivonen is talking on HTML5 conformance on morning of 18th
# [17:33] <MikeSmith> Dave - actually I'm not presenting that "Future of HTML" session, but just moderating. Because it's panel discussion.
# [17:33] <Dave> Thx
# [17:35] <anne> MikeSmith, who's on the panel?
# [17:39] <anne> schnitz, how's that concrete?
# [17:39] * anne ponders
# [17:40] <schnitz> well, we all know where the spec is?
# [17:40] <schnitz> :-)
# [17:40] <schnitz> I will expand, of course
# [17:40] * Dave thinks of marine concrete that sets even when underwater :)
# [17:40] <anne> what exactly does XForms Transitional address that HTML5 doesn't, etc.
# [17:40] <schnitz> anne, I'm a fan of small emails
# [17:41] <anne> well, that doesn't really help here
# [17:41] <schnitz> anne, the expressions that Dave went to in his last mail?
# [17:41] <Dave> also separation between presentation and value stored in DOM
# [17:41] <schnitz> anne, come on, don't be so negative, I'm certainly not, I will expand, that was just the short answer for this minute
# [17:42] <MikeSmith> anne - haven't set final lineup for that panel
# [17:42] <schnitz> XHTML Modularization is the answer to the statement that HTML5 is heavily intertwined and cannot be modularized, not true, we did it in M12N
# [17:42] <schnitz> and XHTML M12N is very close to HTML 4.01
# [17:43] <schnitz> so no magic XHTML2 stuff here
# [17:43] <anne> FYI: XHTML Modularization has zero web browser implementations
# [17:43] <anne> (if something like that is even possible)
# [17:43] <schnitz> anne, FYI: M12N is not MEANT to be in the browser :-)
# [17:43] <schnitz> its a way of organizing a markup language
# [17:43] <anne> well, we need to pass some CR criteria in due course
# [17:43] <schnitz> CR?
# [17:44] <schnitz> what has that to do with M12N?
# [17:44] <anne> it means you need impl
# [17:44] <schnitz> did u read? M12N is not about being implemented in a UA, its about the ML itself
# [17:44] <@DanC> indeed, schnitz , I don't see how M12N addresses intertwining of parsing and document.write(), as detailed in http://lists.w3.org/Archives/Public/public-html/2007JanMar/0172.html
# [17:45] * Joins: bewest (ben@209.237.236.227)
# [17:46] * Joins: preston (chatzilla@70.181.71.135)
# [17:49] <Dave> I am not sure I understand DanC, for document.write it seems to be question of what markup you have before and after the write operation.
# [17:49] <anne> schnitz, if you're talking about just grouping some related elements that's already done
# [17:50] <schnitz> anne, well, ok, but, I am generally asking the question, and thats my message, whether the work of this group will divide over time with the rest of W3C, or not
# [17:50] <anne> i've no idea what that means
# [17:50] <anne> HTML5 is compatible with the DOM, CSS specifications and is largely compatible with HTML4 and XHTML1 where these are compatible with the web...
# [17:51] <schnitz> anne, so when you say grouping has been done, has it been done consistent with M12N for example? Can it be done? Does it make sense? Maybe? Should we?
# [17:51] <anne> I don't think M12N makes sense
# [17:51] <anne> (fwiw)
# [17:51] <schnitz> anne, ok thats fine
# [17:51] <schnitz> anne, still, its a valid question to ask
# [17:51] <anne> (or the whole XHTML Mod stuff for that matter)
# [17:52] <schnitz> its a valid as a directional question, and of course we can dive away into details, but thats not the point, what I am saying here, being explicitly general, whether we have a chance, for the benefit of the overall W3C story, that those technologies grow together over time, or not
# [17:53] <anne> i think it would be better if you reviewed HTML5 and said what would need to change in order for it to be more in line
# [17:53] <schnitz> thats the next step
# [17:53] <schnitz> first, this is brainstorming time, I would like to know whether the group feels at all whether this is a goal worth persueing
# [17:53] <anne> having everyone agree into something from which it's not clear what the consequences are seems silly
# [17:54] * anne much rather has concrete suggestions than architectural thoughts
# [17:54] <Dave> surely thinking about how thinks could pan out makes sense and avoids walking blindfold into dead ends.
# [17:54] <Dave> s/thinks/thinking/
# [17:55] <Dave> s/how thinking/things/
# [17:55] * Joins: kingryan (kingryan@66.92.180.242)
# [17:55] <@DanC> the TAG discussions of tagSoupIntegration touch on modularization...
# [17:55] <anne> WHATWG has been thinking about HTML for over two years now...
# [17:55] <bewest> except that future vision is much worse than hindsight
# [17:55] <anne> (and many on the WHATWG for a longer period)
# [17:56] <anne> why do we need modularization anyway?
# [17:56] <anne> on the web you don't want profiles etc.
# [17:56] <@DanC> in particular... http://www.w3.org/mid/1171386787.7497.1030.camel@dirk XHTML modularization and substitution groups (tag issue XMLVersioning-41, TagSoupIntegration-54, RDFinXHTML-35)
# [17:56] <Dave> bewest, hindsight is like a hangover when you wake up with a headache having walked into something hard
# [17:56] <schnitz> DanC, interesting, thank you
# [17:58] * Joins: Deeder (Deeder@86.198.252.223)
# [17:58] * Quits: kingryan (kingryan@66.92.180.242) (Ping timeout)
# [17:58] <@DanC> on the WG homepage, http://www.w3.org/html/wg/ , I have a few lists that are ordered, but I want bullets rather than numbers. how do I do that?
# [17:59] * Joins: kingryan (kingryan@66.92.180.242)
# [17:59] <Dave> the example in your email reminds me of the HTML+ mechanism for importing names which said treat this tag like that one.
# [17:59] <@DanC> <ol type="none"> achieves "no numbers" but doesn't give a bulliet
# [17:59] <schnitz> list-style-type?
# [17:59] <anne> list-style:circle
# [17:59] <schnitz> yup
# [17:59] <anne> (rtfm: http://www.w3.org/TR/CSS21/generate.html#lists ;))
# [18:00] <schnitz> :-)
# [18:00] <@DanC> <ol style="list-style: circle"> works. thanks.
# [18:02] <@DanC> grumble... changing to <ol class="event_list"> doesn't seem to work. Is this wrong? ol.event_list { list-style: circle }
# [18:03] <Dave> sounds like a buggy implementation
# [18:03] <anne> the underscore is likely an issue
# [18:03] <@DanC> hm... seems the problem was # comment syntax
# [18:03] <anne> that'd be another one :)
# [18:04] <anne> I suppose underscore now works in most browsers
# [18:04] <@DanC> why did karl grey out <cite>? or is it visited links?
# [18:05] <Dave> I find the grey on white text hard to read
# [18:06] <Dave> BTW I would be happy to contribute HTML5 tutorials when we get a bit further along
# [18:06] <schnitz> Dave, always nice :-)
# [18:06] <Dave> and could give some DOM/scripting stuff too.
# [18:06] <anne> http://wiki.whatwg.org/wiki/HTML5_Tutorial
# [18:07] <Dave> different people have different needs for tutorials, so having more than one is good
# [18:07] <anne> yeah, definitely
# [18:08] <anne> (as opposed to specs, come to think of it)
# [18:08] <schnitz> anne, well.... ;-)
# [18:08] <Dave> but lots of test cases is a must! :)
# [18:09] <anne> yeah, we should have a requirement for a number of testcases in our charter
# [18:09] <anne> so we won't slack off halfway
# [18:09] <@DanC> committed http://www.w3.org/html/wg/ Revision: 1.22
# [18:10] <@DanC> no more @@s
# [18:10] <anne> you can prolly drop everything after CR...
# [18:10] <Dave> can you fix the grey text problem while you are at it?
# [18:10] <anne> oh nm
# [18:10] <anne> it already mentions where it comes from
# [18:10] <@DanC> I can fix things more easily if you tell me how. I seem to be late for a meeting meanwhile
# [18:11] <Dave> let's leave it for now, as Karl isn't here
# [18:11] <anne> http://www.w3.org/html/css/screen.css
# [18:11] <anne> change the lines with :visited
# [18:11] <anne> but I'm not sure what a better color would be...
# [18:12] <Dave> something with better contrast
# [18:12] <anne> purple
# [18:12] <Dave> the light blue isn't great either
# [18:22] * Joins: quaiz (quaiz@66.162.32.234)
# [18:31] <anne> http://groups.google.com/group/comp.infosystems.www.authoring.html/browse_frm/thread/45ccb1fa7847fea1/39f7e00f8481e524
# [18:32] * schnitz is now known as schnitzAw
# [18:33] <anne> http://groups.google.com/group/comp.infosystems.www.authoring.html/browse_frm/thread/68cc062324b71173/50049eff53cbb5f2 (never say never)
# [18:34] <gavin_> haha
# [18:34] <billmason> heh
# [18:35] <Deeder> on May 2001
# [18:37] <Deeder> now we're in 2007 and there is a lot of people who believe more in HTML than in XHTML
# [18:38] <preston> Yep
# [18:38] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
# [18:40] <tylerr> I apologize for the lack of participation today in the conversation, I'm horribly busy at work.
# [18:41] <anne> you don't need to apologize for that
# [18:41] <anne> it's stated somewhere that you don't need to make time commitment or anything
# [18:42] <tylerr> Oh sure anne, read that today. :) I'm just rather excited to get involved and I wish I had more time today.
# [18:44] * Joins: gavin (gavin@74.103.208.221)
# [18:54] * Joins: hasather (hasather@81.235.209.174)
# [18:57] * Dave is now known as Dave-off
# [19:00] * Joins: CWilso (cwilso@131.107.0.75)
# [19:01] * Joins: bfults (bfults@66.162.32.234)
# [19:01] * Quits: CWilso (cwilso@131.107.0.75) (Quit: CWilso)
# [19:06] <anne> woha
# [19:06] <anne> lol @ glazou
# [19:09] * Quits: anne (annevk@86.90.70.28) (Ping timeout)
# [19:11] * Quits: hasather (hasather@81.235.209.174) (Quit: hasather)
# [19:13] * Joins: hasather (hasather@81.235.209.174)
# [19:48] * Joins: anne (annevk@81.68.67.12)
# [20:01] <@DanC> phpht. ff crashed. now the big decision: restore session or just garbage collect?
# [20:03] <anne> use Opera?
# [20:04] * Quits: icaaq (icaaaq@85.228.53.245) (Ping timeout)
# [20:09] <gsnedders> anne: now, is that a bias opinion or not? :)
# [20:10] <gavin_> why is that a hard decision?
# [20:10] <gavin_> maybe I don't understand what you mean by "garbage collect"
# [20:11] * Joins: icaaq (icaaaq@85.228.53.245)
# [20:11] <anne> gsnedders, professional
# [20:11] <gsnedders> anne: obviously
# [20:12] <@DanC> I use browser windows to represent tasks. but if a task is critical, I start a mail message. I count on evolution to keep drafts even if it crashes. but browser state is not critical.
# [20:13] <@DanC> so far
# [20:13] <@DanC> so sometimes, when I have a zillion windows and the browser crashes, it's better to start over than to restore the session and re-evaluate the relevance of each of 30 to 50 windows/tabs.
# [20:14] <@DanC> it's especially inconvenient that the windows all get restored on one workspace, where they were on separate workspaces before the crash
# [20:15] <@DanC> I pretty much stick to open source, anne. Release opera as open source, and I'll try it out again.
# [20:34] * Quits: icaaq (icaaaq@85.228.53.245) (Ping timeout)
# [20:37] <Hixie> Dave-off: is there a cvs log somewhere of the changes to the draft?
# [20:38] <Dave-off> I should be able to create that from CVS but don't quite have the skill to do as yet. What do you do to get such diff marked docs?
# [20:39] <Hixie> dunno, whatwg uses subversion
# [20:39] <Dave-off> Essentially the change is to clarify that external functions shouldn't access form fields other than those passed as arguments (and hence subject to the dependency analysis).
# [20:39] <Hixie> (anne and others built http://html5.org/tools/web-apps-tracker for html5)
# [20:40] <Dave-off> Hmm I am hoping W3C moves on from CVS which is far too painful in practice.
# [20:40] <Hixie> Dave-off: how does that change the ua requirements?
# [20:40] <Hixie> i guess i still don't understand how it's supposed to be implemented
# [20:41] <Dave-off> well you could look at my script if you want to see how I did it.
# [20:41] <Hixie> does the calculate="" field take JavaScript?
# [20:41] <Hixie> your script uses regular expressions to mutate the string into JS, which i assume isn't the "official" way
# [20:42] <Hixie> i don't really understand how you go from the value of a calculate="" field to a list of dependencies
# [20:42] <Dave-off> specs should impose requirements but not overconstrain implementation, so I tried to do just that, maybe my wording isn't perfect but it can be improved.
# [20:43] <Dave-off> okay, you use the syntax of JS expressions to pick out identifiers
# [20:43] <Hixie> e.g. if the field is calculate="document.forms[0][document.forms[0].field.value] + 1"
# [20:43] <Hixie> how do you know which fields that depends on?
# [20:45] <Hixie> it's also not clear to me how you're supposed to handle doing a topological sort on cyclic graphs
# [20:45] <Dave-off> I think that the spec needs to be tighter than that, as there you are picking the name from the value of a field and that could indeed cause problems
# [20:46] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
# [20:46] <Dave-off> so it is a matter of narrowing the valdity for expressions.
# [20:46] * Joins: dbaron (dbaron@63.245.220.242)
# [20:47] <Dave-off> Essentially, fields should be referenced by name.
# [20:48] <Hixie> i'm not making any statements regarding authoring requirements here
# [20:48] <Dave-off> I will have a go at tightening the wording to eliminate corner cases.
# [20:48] <Hixie> my concern is only with ua requirements.
# [20:48] <Dave-off> that's what we are talking about - what constitutes a valid expression
# [20:48] <anne> no
# [20:48] <Hixie> no
# [20:48] <bewest> I would greatly prefer change control over wiki
# [20:49] <Hixie> i'm talking about what the ua must do
# [20:49] <bewest> svn or mercurial would be good
# [20:49] <anne> what's a valid expression is an authoring requirement...
# [20:50] <Dave-off> essentially the UA must be able to analyse the expression to identify the names of fields used in calcuate expressions in order to figure out in which order to apply then when one calculated field depends on another.
# [20:50] <Dave-off> The sort algorithm can detect cyclic dependencies and report that as an error.
# [20:50] <anne> but how?
# [20:51] <Dave-off> the algorithm?
# [20:51] <Hixie> _how_ must the ua analyse the expression to identify the names of fields used in calcuate expressions in order to figure out in which order to apply then when one calculated field depends on another?
# [20:51] * Joins: gavin (gavin@74.103.208.221)
# [20:51] <Dave-off> topological sort - http://en.wikipedia.org/wiki/Topological_sorting
# [20:52] <Hixie> topological sort requires static analysis to greate an acyclic graph
# [20:52] <Hixie> here we have turing complete code that may create a cyclic graph
# [20:52] <Hixie> it is therefore impossible to apply a topological sort
# [20:52] <Hixie> it reduces to the halting problem followed by a mis-use of the topological sort algorithm
# [20:53] <mjs> I think there is an implicit requirement that the UA must ignore anything but "obvious" references to fields
# [20:53] <Dave-off> ECMA 262 defines the expression syntax, and as you rightly point out this requires a static analysis which thereby constrains what constitutes valid expressions.
# [20:53] <mjs> for some definition of obvious
# [20:53] <Hixie> the ECMA 262 expression syntax cannot be statically analysed to determine without code execution what the references are.
# [20:53] <Hixie> e.g. the following code:
# [20:54] <Hixie> myObject[getFieldNameByCallingRemoteServer()]
# [20:54] <Hixie> ...may reference a different field each time it is called
# [20:54] <Hixie> and you cannot tell which it will be without actually running the code
# [20:54] <Dave-off> A subset of ECMA 262 expressions certainly can be so analysed and this subset constitutes the set of valid expressions.
# [20:54] <gavin_> where is that subset defined?
# [20:55] <Hixie> that subset would not be able to use any external functions, which you explicitly allow.
# [20:55] <Dave-off> I think we are discussing how to define the subset
# [20:55] <anne> you can propotype all kinds of objects which would even make the simplest "words" not trustable
# [20:55] <Dave-off> yes, you are right, using functions to compute names wouldn't be statically analysable.
# [20:55] <Hixie> i'm not convinced it is possible to define such a subset while making it still be ECMAScript in any useful sense of the word
# [20:56] <anne> i think you'd have to define your own pseudo syntax
# [20:56] <Dave-off> I am not sure what you mean by useful sense?
# [20:56] <Dave-off> A grammar would be perhaps the best route.
# [20:56] <mjs> it would be so restricted that knowing JavaScript wouldn't let you know how to write it
# [20:56] <Hixie> something would be ECMAScript in a useful sense of the word if it could be passed to a JS compiler without risk, e.g.
# [20:56] <mjs> and having a JavaScript implementation would be not nearly sufficient to execute it
# [20:56] <mjs> you would want a custom parser
# [20:57] <Hixie> yeah
# [20:57] <Hixie> at which point you have a new language
# [20:57] <Dave-off> Only in the sense that a subset is a new language.
# [20:57] <mjs> and if you can't use "real" JS, then you still need extended event listener attributes, so that people can escape to doing imperative things at validation time etc
# [20:57] <anne> so how is this much better than simply using event handlers?
# [20:57] <Hixie> a true subset would not require a new parser and compiler. this would.
# [20:57] <Dave-off> It permits round tripping of semantics which is a huge win.
# [20:58] <Dave-off> The preprocessor is only a few lines of code, as I have shown.
# [20:59] <anne> it has also been pointed out that it breaks in all kinds of ways...
# [20:59] <Dave-off> the preprocessor could check for validity in a straightforward way using regular expressions.
# [20:59] <mjs> I'm out, back later
# [21:00] <mjs> ECMAScript's grammar is not a regular language
# [21:00] <mjs> so no, you can't check it using regular expressions
# [21:00] <anne> calculate="x + y" is obviously executed in some scope right?
# [21:00] * Quits: mjs (mjs@64.81.48.145) (Quit: mjs)
# [21:00] <Dave-off> mjs, wrong, since we are talking about a subset
# [21:00] <anne> what if x and y are somehow getters that return random fields?
# [21:01] <anne> how would you know that in advance?
# [21:01] <Dave-off> anne right.
# [21:01] <Dave-off> It is in the context of the field owning the expression and the form it belongs to.
# [21:02] <Dave-off> so in my implementation form is a local variable initialized from the field.
# [21:03] <Dave-off> when eval is called the expression has been rewritten to use access fields via that variable.
# [21:04] <Dave-off> I will have a go at defining the expression syntax subset via grammar or such like tomorrow, but it's late here so cheerio for today.
# [21:04] <anne> this sounds very much like a new language to me...
# [21:05] <anne> custom parsing, rewriting expressions, etc.
# [21:08] <Hixie> it really seems like way too much work for very small gain, given that you can do this using events and js and authors aren't clamouring for declarative expressions
# [21:09] <anne> yeah, maybe you should just define a valid subset of that and use that in some authoring language or something...
# [21:09] * Quits: bewest (ben@209.237.236.227) (Ping timeout)
# [21:09] <Dave-off> The authors in question don't know about HTML tags, CSS or JavaScript, so I think you are thinking of the wrong group of developers - the ones who are script hackers and can figure out what the scripts do.
# [21:10] <Hixie> i am pretty sure we have a pretty big cross-section of authors on the whatwg list by now
# [21:11] <anne> well in that case I'm not sure what the problem is Dave-off
# [21:11] <Hixie> and i'm pretty sure i'd have heard if google spreadsheets people or users needed this in html
# [21:11] <anne> they wouldn't ever touch the code...
# [21:11] <Dave-off> I doubt very much that that includes non-techie authors who don't care and don't want to learn HTML.
# [21:13] <Dave-off> If I went to a local college and asked the students there, I bet they would have different ideas as to what they found important.
# [21:13] <Dave-off> anyway I really have to leave now, cheers!
# [21:13] <Hixie> as far as i can tell, college students are well represented in the whatwg demographic
# [21:13] <tylerr> Later.
# [21:13] <Hixie> and they haven't been asking for this
# [21:13] <Hixie> later
# [21:16] <tylerr> I was going to ask how the day's been going for you Hixie?
# [21:16] <Hixie> fine so far
# [21:16] <tylerr> Good good.
# [21:16] <preston> If you want to make HTML/CSS/Javascript easier to learn, you probably want to think more about taking things out, rather than adding new alternatives.
# [21:22] <tylerr> That would be ideal, but I think there have been needs that haven't been met with the current state of HTML/CSS/JavaScript that only additional specs can address.
# [21:24] <preston> Sounds like fodder for a working list of topics. What are those needs (and are they really needed)?
# [21:25] <anne> semantics for web applications is one
# [21:25] * tylerr nods.
# [21:25] <anne> richer semantics for documents is another (such as <header>, <section>, <footer>, etc.)
# [21:26] <tylerr> Also alternative media types.
# [21:26] <Hixie> alternative media types?
# [21:26] <anne> I suppose he means <video> and maybe <audio>?
# [21:26] <tylerr> Yes.
# [21:26] <tylerr> Rather than text and graphic content.
# [21:27] <preston> Semantics means what? What does a <header>, <section>, etc mean - outside the scope of a particular web application?
# [21:28] <preston> The semantics of a web application are not general.
# [21:29] <anne> <header> and <section> are document semantics
# [21:30] <anne> for applications you want "semantics" for device independence and such
# [21:31] <tylerr> anne: For universal UA usage?
# [21:31] <preston> If you want to embed externally meaningful semantics into a web document, what you are talking about is an "interface" (in the C++/Java sense). Interfaces - if implemented - might best be described by something like microformats, not generic HTML.
# [21:32] <anne> not if you want to tie javascript and such too it
# [21:33] <anne> if you want to provide a date picker for instance
# [21:33] <preston> (The <video> and <audio> embeds are a different can of worms.)
# [21:33] * Joins: icaaq (icaaaq@85.228.53.245)
# [21:33] <anne> <input type=date> is the current "proposal" for that
# [21:34] <anne> or take <progress> for instance: http://www.whatwg.org/specs/web-apps/current-work/#the-progress
# [21:35] * Parts: icaaq (icaaaq@85.228.53.245)
# [21:35] <anne> preston, they are :)
# [21:36] <preston> New <input> types do make a bit more sense to me ... to a degree.
# [21:37] <preston> Of course, folk have managed to write custom widgets using existing HTML. Would the browser built-ins be superior?
# [21:38] <anne> well, they would prolly be more accessible
# [21:39] <preston> An exercise (for "free" time) : could <input type="date"> be rewritten by Javascript to invoke a custom widget?
# [21:39] <anne> it's already implemented like that in a library for IE6
# [21:40] <anne> (another advantage of <input type=data> is that you get a single submission format, simpler client side validation, etc.)
# [21:40] <preston> So if you have the library, do you need it built into the browser? :)
# [21:41] <anne> if you want it to become more accessible, more integrated with the default UI and simpler to use for people who can't use or don't have access to the library, yes
# [21:43] <anne> it also makes for cleaner code, etc.
# [21:45] <preston> Validation is another issue. There is per-field validation, inter-field validation, and server-based validation. Does it make sense to try and capture that all in declarative markup?
# [21:46] <anne> server-based validation not
# [21:46] <anne> at least not in HTML
# [21:48] <preston> So does mixing declarative and scripted validation make the model clearer, or harder to learn?
# [21:48] <anne> most of the things can be addressed declaratively i think
# [21:49] <anne> see http://www.whatwg.org/specs/web-forms/current-work/
# [21:50] <preston> Saw it :(
# [21:52] <preston> Can't say I've studied the whole thing in detail. The phrase
# [21:52] <preston> "Ran away screaming" comes to mind.
# [21:53] <preston> Seems near to another whole new programming language.
# [21:53] <anne> it's just an extension to HTML4
# [21:54] <anne> (well, and XHTML1 and DOM2HTML)
# [21:54] <anne> programming language?!
# [21:55] <preston> In the sense that the C++ templates extension are (almost) a programming language.
# [21:56] <anne> Unfortunately I'm not familiar with C++
# [21:57] <anne> btw, if you have substansive comments I encourage you to raise them on whatwg@whatwg.org
# [22:01] <preston> Anne, I am a minimalist. Much of what I found reading the WHATWG stuff ... baroque. Too big a difference in philosophy?
# [22:02] <Hixie> what would you propose we do with HTML5?
# [22:03] <preston> As little as possible? :) I found the charter encouraging.
# [22:03] <Hixie> as little as possible would be "nothing".
# [22:03] <Hixie> in which case, we're done: HTML4
# [22:03] <billmason> Drinks for all.
# [22:03] <jgraham> Indeed. The idea is to make things better
# [22:03] <jgraham> not maintain the status quo
# [22:03] <Hixie> which isn't hard, given the poor state that is html4
# [22:04] * anne notes that the charter contains much of what's in the WHATWG specs: http://www.w3.org/2007/03/HTML-WG-charter.html#deliverables (if not all)
# [22:04] <Hixie> unsurprising given that it's almost a direct copy and paste from something i sent months ago that was itself basically a copy and paste of the wa1 contents list at the time
# [22:05] <preston> Heh :) In fact there is a lot of cruft left over from the browser wars. HTML would be a lot easier to learn (and use) if the disused bits were partitoned off (if not deprecated).
# [22:05] <Hixie> the whatwg html5 does exactly that
# [22:05] <bfults> whatwg's work was done with an eye to what is and isn't used
# [22:05] <Hixie> huge chunks of stuff are thrown out
# [22:05] <Hixie> (or rather, not added)
# [22:05] <Hixie> (since html5 started from zero)
# [22:06] <preston> The thrown out stuff sounds great. I tripped over all the additions.
# [22:06] * tylerr thinks of writing a specification consisting solely of all the thrown out stuff. ;)
# [22:07] <anne> preston, it's still not clear what your goal is then to me
# [22:08] <preston> A clear programming model for web applications (in this context the client side HTML/CSS/Javascript) going forward.
# [22:08] <anne> but without anything new?
# [22:08] <jgraham> preston: what do mean "programming model"?
# [22:09] * bfults is now known as h3h
# [22:10] <preston> Fair question. Structure in HTML. Style in CSS. Behavior in Javascript.
# [22:11] <jgraham> So why do you think WebApps doesn't fit this general mold?
# [22:11] <preston> ? WebApps ?
# [22:11] <jgraham> WHATWG HTML5
# [22:11] <preston> Ah
# [22:11] <tylerr> Most WebApps follow the same concepts when using MVC.
# [22:12] <preston> Mixing behavior into declarative HTML.
# [22:12] <jgraham> Oh so you think <input type="date"> is bad because you'd rather everyone wrote their own js datepicker?
# [22:13] <anne> that would make <a> bad too
# [22:13] <anne> btw
# [22:13] <h3h> it isn't mixing custom behavior into declarative HTML. it's indicating the semantic content of the elements in the markup and letting the UA provide an appropriate interface
# [22:14] <h3h> and that doesn't (and shouldn't) preclude custom behavior specified via JS
# [22:14] <preston> "Bad" is a bit strong. What happens when you need a customized data picker?
# [22:14] <anne> you style it?
# [22:14] <h3h> you write custom JS+CSS :)
# [22:15] <anne> you'd still use <input type=date> just with different styles (and if you need different behavior too you'd use something like XBL I suppose)
# [22:15] <preston> ... with a library ...
# [22:15] <h3h> it's about solving the common use cases to solidify a base
# [22:15] <h3h> based on the semantics of the data elements on the page
# [22:16] <gsnedders> the question is how wide we make the base
# [22:17] <h3h> right.
# [22:17] <gsnedders> if we make it too wide the spec becomes hard to properly know to write without referencing it constantly
# [22:17] <h3h> I think "bigger than HTML4 made it" is obvious
# [22:17] <gsnedders> if we make it too narrow we're going to be very reliant on JS+CSS
# [22:17] <h3h> "somewhere short of insane" seems an appropriate upper-bound
# [22:18] <h3h> or "something that can be spec'd in a reasonable amount of time"
# [22:18] <h3h> like...hey! the whatwg work!
# [22:18] <gsnedders> I'd say somewhere where you can remember it all as an author
# [22:18] <gsnedders> without difficulty, that is
# [22:18] <h3h> I'd say that's pretty arbitrary
# [22:18] <h3h> admirable for some maybe, but not a practical measure
# [22:18] <h3h> most languages (programming, markup, styling) can't be kept in a single author's head in their entirety
# [22:19] <preston> Hey! There's a metric! Good = spec + additions - cleanup < HTML4 :)
# [22:19] <h3h> and certainly aren't in the majority case
# [22:19] <gsnedders> HTML's success is based on the fact anyone can write it because it is limited and can be remembered
# [22:19] <preston> Well....
# [22:19] <Hixie> i bet nobody here can remember all of HTML4
# [22:19] <h3h> HTML's success is based on the fact that anyone can write it and it generates instant results
# [22:19] <@DanC> I just sent mail about the issues list/agenda I've been building. http://www.w3.org/html/wg/il16
# [22:19] * Quits: ROBOd (robod@86.34.246.154) (Quit: http://www.robodesign.ro)
# [22:19] <h3h> fact is most people don't use (or know about) large chunks of HTML4
# [22:20] <gsnedders> maybe not all, but large parts.
# [22:20] <h3h> and that trend will continue
# [22:20] <anne> yay, issue 1 is mine
# [22:20] <h3h> the useful parts are remembered
# [22:20] <preston> Remember - with lots of reference to Goodman's book (in my case), and lots of filtering to ignore chunks.
# [22:20] <h3h> so the whatwg's goals were/are to spec the useful bits
# [22:20] <h3h> while remembering that different authors have different uses for HTML
# [22:21] <h3h> whatwg's HTML is about solving real problems for real HTML authors
# [22:22] <h3h> and if this WG's output is anything less than that, I fear it will fail miserably
# [22:22] <gsnedders> WA is also about defining how to parse it
# [22:22] <h3h> which is essential for the interoperability necessary to build platforms and reliable apps upon
# [22:23] <h3h> the uses of HTML 4 and below should be ample evidence for that claim
# [22:23] <h3h> or abuses
# [22:24] <gsnedders> I was having to explain NETs to a friend as to why his document wasn't validating, though rendering fine
# [22:24] <anne> DanC, http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2007-March/010417.html (guy from Dirac on timeline)
# [22:26] <@DanC> interesting. but not exactly in scope for the HTML WG
# [22:26] <anne> just mentioning it because Dirac was mentioned there
# [22:27] <@DanC> yes; I apprecaite the pointer.
# [22:27] <anne> I agree that standardizing a codec is not in scope, requiring a baseline format may be...
# [22:28] <@DanC> as if chairing this WG weren't enough of a challenge?!?! video codec standardization is about the most politically charged thing I can think of.
# [22:28] <Hixie> we can't just not do something because it would be inconvenient for us...
# [22:29] <jmb> if interoperability is the (/a key) goal, then I see no alternative to requiring a baseline format, tbh
# [22:29] <@DanC> I can see a path to success that doesn't involve solving world hunger nor video codec standardization
# [22:30] <jmb> heh. I'm not quite convinced those are both at the same level of difficulty ;)
# [22:30] <Hixie> well fwiw whatwg isn't solving world hunger but is having to work out what baseline codecs to require
# [22:30] <@DanC> if this WG wants to make choosing a video codec a goal, I'll have to go back to the AC and the Director, since it'll involve so much more staff resource.
# [22:31] <@DanC> whatwg is _choosing_ to do that, no?
# [22:31] <Hixie> anything the html wg doesn't want to do the whatwg will be happy to do; like i said, i expect the whatwg spec to be a superset of the html wg spec :-)
# [22:31] <preston> Good :)
# [22:31] <Hixie> well we're "choosing" to do it in so far as we are "choosing" to have a spec that can be baseline-interoperable
# [22:32] <preston> My hope is that standardizing <video> and/or <audio> is outside the scope of this WG. Too thorny.
# [22:32] <@DanC> is flash in or out?
# [22:33] <Hixie> what do you mean by "flash"? whatwg has already specified <embed> and <object> and fixed the various things html4 got wrong about them with respect to flash integration
# [22:33] <Hixie> if that's what you mean
# [22:33] * Joins: Dashiva (noone@129.241.151.35)
# [22:34] <@DanC> if baseline-interoperable involves choosing video codecs, it must involve choosing image formats (GIF, JPEG, PNG?) and plug-in formats too... java, flash.
# [22:35] <Hixie> yup
# [22:35] <Hixie> we haven't decided on baseline image and plugin formats yet
# [22:35] <preston> Could that be another WG. Please?
# [22:36] <Hixie> i would expect the baseline image formats to be GIF87, GIF89, JPEG JFIF, PNG, and APNG
# [22:36] <Hixie> i would expect the baseline plugin formats to be none
# [22:36] <@DanC> why not flash? it's expected by the typical web user
# [22:36] <Hixie> it's proprietary
# [22:36] <Hixie> same reason we're not requiring mpeg4
# [22:37] <Hixie> it's also undocumented
# [22:37] <Hixie> and plugin wise, you'd be requiring the shipping of a binary blob
# [22:37] <Hixie> which is hardly acceptable
# [22:37] <Hixie> overall it's an easy decision
# [22:37] <@DanC> hmm... so this is more of a batteries-included approach than I'm used to.
# [22:38] <Dashiva> Has there been discussion on MNG vs APNG, if so is it archived?
# [22:38] <anne> yes, bugzilla.mozilla.org
# [22:38] <Hixie> Dashiva: MNG is the most bloated image format since the invention of image formats, and isn't even worth talking about
# [22:38] * @DanC hopes we the W3C HTML WG is willing to declare victory without including quite so many batteries
# [22:38] <Hixie> MNG makes SVG Tiny 1.2 look well designed
# [22:38] <gavin_> you do not want to read the mozilla mng bug
# [22:38] <Hixie> and that's not a compliment to SVG
# [22:38] <gavin_> or maybe you do...
# [22:39] <gavin_> it's "interesting"
# [22:39] <Dashiva> I take the point
# [22:39] <Hixie> DanC: well like i said. anywhere where html wg drops the ball (or doesn't pick it up), whatwg will be there to catch it :-)
# [22:40] <Hixie> it would be sad if the w3c wasn't the final word on html, though
# [22:41] <kingryan> Hixie: I don't see anything wrong with a difference between "HTML" and "web applications", where html-wg defines the former and what-wg the latter and the latter is a superset of the former.
# [22:42] <Hixie> kingryan: it wouldn't be "wrong", but it would beg the question of what the w3c's role was
# [22:42] <kingryan> Hixie: I'm just saying there are different layers here. 1. the language. 2. browser implementations
# [22:43] <Hixie> i don't really follow
# [22:43] <@DanC> my goal is to take some set of features and develop a test suite within a couple years and put a bow on it.
# [22:43] <kingryan> #1 includes how to parse, produce and validate the syntax of the language
# [22:44] <kingryan> #2 includes dom, css, xbl, svg, etc. interaction
# [22:44] <gavin_> parsing and DOM aren't neatly separable like that
# [22:44] <Hixie> king: a spec that doesn't define how it interacts with other specs is a poor spec indeed
# [22:44] <kingryan> I think I see things this way because my work is on a non-browser user-agent
# [22:45] <kingryan> Hixie: but not all user-agents of html will implement css, dom, etc.
# [22:45] <kingryan> (like search engine spiders)
# [22:45] * @DanC is pretty nervous about the lack of separation beteen parsing and scripting
# [22:45] <Hixie> DanC: ah. i'm fully expecting this to be a multi-decade effort. i don't think you can advance the web just like that. but *shrug*.
# [22:45] <@DanC> I tend to get bored after a couple years
# [22:45] <kingryan> DanC: this work may not be for you then :D
# [22:45] <Hixie> yeah
# [22:46] <Hixie> this isn't a 2 year stint
# [22:46] <Hixie> this is a career
# [22:46] <@DanC> this _is_ a 2 year thing. maybe 3. argument by assertion works here too ;-)
# [22:46] <Hixie> good luck with that
# [22:46] <kingryan> however, there might be a 2-5 year portion (my #1 above)
# [22:46] <@DanC> I'll just keep repeating it until it becomes true. 1/2 ;-)
# [22:48] <kingryan> DanC: I have a feeling you'll be repeating that for more than 2 years
# [22:48] <preston> (Thinking the DanC might be *exactly* the guy for the job ... :) )
# [22:49] <Hixie> i don't see why we'd be stopping the development of the web in two years
# [22:49] <Hixie> the web is going to exist for decades at least, maybe hundreds or thousands of years.
# [22:49] <hsivonen> kingryan: WA10 says that if you don't do scripting, you don't need to have a DOM inside your box if the outside of your box looks as if you did have DOM inside
# [22:50] <hsivonen> (I don't use the DOM for conformance checking, for example)
# [22:51] * Joins: an (chatzilla@85.118.224.254)
# [22:52] <kingryan> hsivonen: I'm aware that I don't have to implement the DOM to implement wa10
# [22:52] <@DanC> I don't intend to stop work on the web; just to declare victory on an HTML spec that's better than HTML 4
# [22:52] <kingryan> DanC: we could declare that victory today
# [22:52] <Hixie> yeah
# [22:53] <Hixie> that's not hard :-)
# [22:53] <hsivonen> kingryan: ok. I guess I misread your comment above
# [22:53] <@DanC> well, I want a test suite
# [22:53] <Hixie> yeah, that's the tough one
# [22:53] <Hixie> a test suite for html is going to have to be on the order of about 10,000 to 50,000 tests, i'd say (based on several years of doing qa for html professionally)
# [22:54] <@DanC> I want to make it so that W3C has a spec where if you implement it, you'll have a chance of interoperating with some interesting fraction of the web.
# [22:54] * Joins: mjs (mjs@17.255.100.126)
# [22:54] <Hixie> 50000 tests is 68 tests a day if we start today and work non-stop for two years
# [22:54] <kingryan> hsivonen: I was trying to make the point that the non-dom, non-css, non-... level of implementation could be seen as a separate (lower) layer than the rest and maybe a separate spec.
# [22:54] <Hixie> (we better get cracking)
# [22:55] <Dashiva> Hixie: We all know you can do 680 easily ;)
# [22:55] <mjs> hi everyone
# [22:55] <@DanC> I can imagine signicantly enhancing the state-of-the-art with fewer than 10,000 tests.
# [22:55] <Hixie> well yeah, you could enhance it with 1 test, if you write the right test
# [22:55] <@DanC> but xquery has around that many; it's not unprecedented
# [22:55] <Hixie> witness, e.g., acid2
# [22:55] <mjs> we want to advance not just the state-of-the-art but also interoperability
# [22:56] <Hixie> DanC: i'm not arguing against the test suite, on the contrary. i absolutely agree we need one, and i seriously do think it should be about 50,000 tests, and i don't think we can make a quality test suite that big in such a short time.
# [22:57] * schnitzAw is now known as schnitz
# [22:57] * Quits: an (chatzilla@85.118.224.254) (Quit: Chatzilla 0.9.77 [Firefox 2.0.0.3/2007030919])
# [22:57] <kingryan> building the test suite will probably follow the 90/10 rule, too (the last 10% will take 90% of the time)
# [22:57] <@DanC> the W3C process document talks about tests for each feature; I wonder what's a reasonable way to count features in HTML. Let's say there's 150 of them. ~10 tests each would be ~1500. I'd be fairly happy with that. But if we can get more before we all get bored, very well.
# [22:58] <kingryan> so we'd need to be doing more than 68 test/day now :D
# [22:58] <Hixie> anyone who talks about "a test for each feature" has never done real testing
# [22:58] <@DanC> hixie, that's either false or insulting
# [22:58] <Hixie> it's likely insulting
# [22:58] <Hixie> but still true.
# [22:59] <kingryan> it should be a "test for each state and state transition"
# [22:59] <Hixie> i have 68 tests for margin collapsing in css2.1 alone, and that's highly inadequate for that feature.
# [22:59] <Hixie> there are features that alone could need thousands of tests for any sort of decent interoperability coverage.
# [22:59] <@DanC> I've done real testing, and I just used the words "a test for each feature".
# [22:59] <Hixie> (and note that the margin collapsing tests can each take up to two or three hours to write)
# [23:00] <Hixie> DanC: you've done real interoperability testing for a web browser product?
# [23:00] * Joins: AnPol (chatzilla@85.118.224.254)
# [23:00] <@DanC> no, but I've done QA for other software products.
# [23:01] <Hixie> browsers aren't like most software products.
# [23:01] <Hixie> at all.
# [23:01] <Hixie> they're more like operating systems.
# [23:01] <Hixie> and operating system APIs
# [23:01] <Hixie> both in scope and complexity, as well as in testing resources needed
# [23:01] <@DanC> ok, a third possibility: false, insulting, or unclear. I wasn't clear that "real testing" was exclusively browser testing.
# [23:01] <Hixie> oh sorry, i assumed that's what we were talking about
# [23:01] <Hixie> yes i meant for browsers.
# [23:02] <Dashiva> By the way, is there a link path from the html-wg page to the relevant polls, or will they just be announced on the mail list or similar?
# [23:02] <@DanC> this comes back to kingryan's point; I'm not trying to test browsers. I'm aiming for a test suite that makes the language spec clear.
# [23:02] <@DanC> indeed, there should be such a link, Dashiva ...
# [23:03] <@DanC> I also think per-feature tests are of limited value. I tend to focus on per-issue tests.
# [23:03] <tylerr> DanC, is #swig on this server? I tried joining it but it seems no one was in it.
# [23:03] <@DanC> no, #swig is on irc.freenode.net
# [23:03] <tylerr> Ah righto.
# [23:04] <tylerr> Thanks.
# [23:04] <Dashiva> I've been watching the f2f one, and the number of unanswered members grows faster than answers. Thought it might merit attention
# [23:04] <@DanC> the f2f one is linked.
# [23:04] <@DanC> if you're not seeing it, that suggests a usability problem. I'm not much of a web designer, unfortunately.
# [23:05] * @DanC re-reads... "I've been watching the f2f one"
# [23:05] <Dashiva> Ah, it's been added since I last checked
# [23:05] <Dashiva> The pale green is gone too, excellent
# [23:06] <@DanC> is it? I didn't do anything about the colors
# [23:07] <@DanC> the list of all HTML WG surveys is http://www.w3.org/2002/09/wbs/40318/all ; I'm not sure if it's world-readable; can you check for me, please?
# [23:07] <Dashiva> unauthorized
# [23:07] <kingryan> I get an auth request
# [23:07] <mjs> DanC: depends on what the scope of "a test" is
# [23:07] <Hixie> DanC: if you're not aiming to test browsers... what are you aiming to test? i'm confused
# [23:08] <mjs> DanC: but a thorough test suite needs to cover all edge and common cases of a feature, likely implementation pitfalls, known past failures in various implementations, and interaction with other features
# [23:08] <mjs> DanC: that's likely to be more than one test per feature
# [23:08] * mjs is now known as mjs_afk
# [23:08] <mjs_afk> back later
# [23:09] <@DanC> good question, Hixie . I think I'm too hungry to answer coherently, though.
# [23:11] <@DanC> hmm... I think my dog just got out of the yard. bbl.
# [23:12] <Dashiva> Hixie: Could be the difference between "making tests based on how browsers handle and mess up things" and "making tests based on just how the spec defines it"
# [23:14] <kingryan> Dashiva: shouldn't those be the same thing?
# [23:14] <Hixie> the second would be a superset of the first
# [23:14] <Dashiva> I would expect the latter to be simpler since it's more naive, sort of
# [23:15] <Hixie> the first is a huge job, which nobody has ever completed successfully, in the 17 years of web browsers existing.
# [23:15] <Hixie> no, testing everything in the spec would automatically test everything that was in the spec but implemented badly
# [23:16] <Dashiva> But any test suite we write would probably miss out on some way the browsers could get it wrong
# [23:18] <kingryan> Dashiva: to completely test any software you have to test every state and every state transition, which is usually infinite
# [23:18] <kingryan> so there'll always be things left out. the goal is to prioritize and approach the limit (even if we can't get there)
# [23:27] <@DanC> a test can answer questions about a spec, independent of implementations. e.g. "this is a good HTML document" or "this HTML document has 4 links".
# [23:28] <Hixie> i'm not sure i understand
# [23:28] <Hixie> could you show an example?
# [23:28] * Joins: Lachy_ (chatzilla@58.105.240.232)
# [23:29] <@DanC> well, I guess http://lists.w3.org/Archives/Public/public-html/2007JanMar/att-0172/_t.html is an example, or part of one
# [23:29] <@DanC> attached to http://lists.w3.org/Archives/Public/public-html/2007JanMar/0172.html
# [23:29] <@DanC> http://www.w3.org/TR/owl-test/ has a few hundred examples
# [23:29] <Hixie> that looks like what i'd call a "demo", not a test
# [23:30] <Hixie> it's not a conformance test of any kind, at least
# [23:30] <Hixie> right?
# [23:30] <@DanC> indeed, to make it into a test, we'd have to be more clear about the test hypothesis and the output of the test
# [23:30] <Hixie> such demo documents are certainly useful when reverse engineering products to write specs, but i don't think they make much sense as a group's output
# [23:31] <Hixie> i mean, they're not really useful outside of writing the spec
# [23:31] <@DanC> sure they are...
# [23:31] <Hixie> and they don't really usually result in any discussion, since the answer is fixed by existing browsers
# [23:31] <Hixie> so they're not really useful to the group either, other than to the editor who's writing the spec
# [23:31] <Hixie> imho
# [23:31] <Hixie> can you elaborate on who else they would help?
# [23:31] <@DanC> we discuss an issue... we build a document that we all agree characterizes the issue; then we decide the issue, and we record the results of the test that agree with the decision...
# [23:32] <Hixie> good lord, if we go through that process for every issue we'll be here for centuries
# [23:32] <@DanC> ... and then we can run that test against software, and if it fails, we know the software doesn't agree with the spec.
# [23:33] <@DanC> indeed, the number of issues one can look at in this much detail is limited.
# [23:33] <@DanC> but for OWL, we did it for several hundred tests, covering a few dozen issues.
# [23:33] * Quits: AnPol (chatzilla@85.118.224.254) (Quit: Chatzilla 0.9.77 [Firefox 2.0.0.3/2007030919])
# [23:33] <Hixie> well, we'll see
# [23:33] <Hixie> i'm not sure i really understand
# [23:33] <kingryan> DanC: but this is a much larger group with existing implementations
# [23:33] <Hixie> but it will presumably become clear when we do it
# [23:34] <@DanC> yes, it'll surely get more clear when we do it. Who knows if the style I'm used to will happen.
# [23:34] <Dashiva> For less "controversial" issues, reviewing larger sets of tests at a time could speed it up
# [23:35] <@DanC> yes, it's typical to have a dozen or 50 issues, e.g. the SPARQL punctuation syntax issue has about 40 or 50 tests to go with it.
# [23:35] <@DanC> phpht
# [23:35] <@DanC> a dozen or 50 tests to prepare before closing an issue.
# [23:37] <@DanC> meanwhile, for parts of the spec that aren't controversial, we might not ever build any tests.
# [23:37] <@DanC> hmm... well, we should.
# [23:37] <@DanC> but if the implementations just magically match up with the spec and everybody's happy, then everybody's happy.
# [23:38] <Dashiva> Until one comes along and says "You're doing it wrong" (cough sgml comments ;)
# [23:38] <@DanC> magically as in: without the chair doing any work. clearly the editor and the implementors have to do work in any case.
# [23:39] <@DanC> "you're doing it wrong" doesn't likely fall under "not controversial"
# [23:52] <@DanC> kingryan, the test process I'm talking about is pretty close to what we did for hCard. How do you think that went?
# [23:52] <kingryan> DanC: it went well. the domain was much smaller and we had a stable spec, though.
# [23:54] <@DanC> yes, the domain was smaller...
# [23:54] <kingryan> and by stable spec I mean vcard, not hcard
# [23:54] <@DanC> I hope some interesting fraction of the 190 HTML WG participants are willing to help with the tests. we only had 3 of us working on hCard tests.
# [23:55] * Quits: Deeder (Deeder@86.198.252.223) (Client exited)
# [23:55] <kingryan> yeah, but agreeing on and reviewing the tests isn't that parallelizable
# [23:55] <@DanC> I look at HTML as fairly stable, at least in that it's really expensive to change.
# [23:56] * mjs_afk is now known as mjs
# [23:56] <@DanC> reviewing will be fairly serialized, yes, but generating test materials is highly parallelizable
# [23:56] <kingryan> agreed
# [23:57] * kingryan is reminded that he has hcal tests to review
# [23:57] * @DanC dispairs of ever finding time for all the hCalendar stuff he wants to do
# [23:59] <mjs> writing good tests is hard
# [23:59] <mjs> sometimes harder than the actual implementation work
# Session Close: Fri Mar 23 00:00:00 2007

The end :)