Options:
- # Session Start: Wed Dec 05 00:00:00 2007
- # Session Ident: #html-wg
- # [00:02] * Joins: Zeros (Zeros-Elip@129.2.175.74)
- # [00:05] * Joins: mjs (mjs@17.255.98.46)
- # [00:07] * Quits: timbl_ (timbl@128.30.6.53) (Quit: timbl_)
- # [00:10] * Quits: laplink (link@85.19.196.186) (Quit: Leaving)
- # [00:10] * Joins: laplink (link@85.19.196.186)
- # [00:10] * Joins: gsnedders (gsnedders@86.145.188.131)
- # [00:12] * Joins: sbuluf (cavgya@200.49.132.115)
- # [00:16] * Quits: jgraham_ (james@81.86.218.70) (Quit: This computer has gone to sleep)
- # [00:26] * Quits: gsnedders (gsnedders@86.145.188.131) (Quit: 404: Not Found)
- # [00:31] * Joins: jgraham_ (james@81.86.218.70)
- # [00:31] * Quits: Julian (chatzilla@130.129.20.163) (Quit: ChatZilla 0.9.79 [Firefox 2.0.0.11/2007112718])
- # [00:59] * Quits: laplink (link@85.19.196.186) (Ping timeout)
- # [01:04] * Joins: laplink (link@85.19.196.186)
- # [01:04] * Joins: hyatt (hyatt@17.255.110.53)
- # [01:11] * Joins: timbl (timbl@209.6.134.246)
- # [01:16] * Quits: hyatt (hyatt@17.255.110.53) (Quit: hyatt)
- # [01:19] * Quits: aaronlev (chatzilla@66.30.196.151) (Ping timeout)
- # [01:20] * Quits: billmason (billmason@69.30.57.156) (Quit: .)
- # [01:21] <Hixie> MikeSmith: fyi i'm starting work on the server side
- # [01:28] * Quits: jgraham_ (james@81.86.218.70) (Quit: This computer has gone to sleep)
- # [01:33] * Quits: Zeros (Zeros-Elip@129.2.175.74) (Quit: Leaving)
- # [01:40] * Joins: Zeros (Zeros-Elip@129.2.175.74)
- # [01:40] * Quits: Zeros (Zeros-Elip@129.2.175.74) (Quit: Leaving)
- # [01:46] <MikeSmith> Hixie - cool
- # [01:46] <MikeSmith> Sorry for not having gotten to it
- # [01:46] <Hixie> no worries
- # [01:47] <Philip> "Passed: 2452. Failed: 249. Todo: 5" - hooray, not too many more tests to delete before Mozilla can pass 100%
- # [01:48] <mjs> heh
- # [01:53] * Quits: timbl (timbl@209.6.134.246) (Connection reset by peer)
- # [01:53] * Joins: timbl (timbl@209.6.134.246)
- # [02:18] * Joins: hyatt (hyatt@17.255.110.53)
- # [02:22] <Philip> "Passed: 2444. Failed: 91. Todo: 73"
- # [02:25] * Joins: DougJ (djones4@74.76.28.112)
- # [02:35] * gavin cheers
- # [02:37] * Quits: DougJ (djones4@74.76.28.112) (Quit: DougJ)
- # [02:38] * Joins: DougJ (djones4@74.76.28.112)
- # [02:42] * Quits: hyatt (hyatt@17.255.110.53) (Quit: hyatt)
- # [02:43] * Quits: dbaron (dbaron@63.245.220.241) (Quit: 8403864 bytes have been tenured, next gc will be global.)
- # [02:48] <Philip> "Passed: 2440. Failed: 0. Todo: 151" - whoo
- # [02:48] <gavin> have you filed the bug yet? :)
- # [02:48] <Philip> Assuming not too many of these break on non-Linuxes, I guess they should be useful as regression tests
- # [02:49] <Philip> gavin: Not yet, and I need to go to sleep an hour ago so it'll have to wait until tomorrow :-)
- # [02:49] <gavin> alright
- # [02:49] <gavin> be sure to CC me when you do
- # [02:50] <Philip> Okay, will do
- # [02:53] * Quits: timbl (timbl@209.6.134.246) (Connection reset by peer)
- # [02:53] * Joins: timbl (timbl@209.6.134.246)
- # [03:02] * Joins: dbaron (dbaron@71.204.145.103)
- # [03:10] * Joins: aaronlev (chatzilla@209.6.168.245)
- # [03:13] * Quits: heycam (cam@203.217.91.18) (Ping timeout)
- # [03:13] <MikeSmith> can anybody recommend good blog search engines that you use?
- # [03:13] <MikeSmith> I mostly use google blogsearch these days
- # [03:14] <MikeSmith> and don't know what all else is out there
- # [03:14] <mjs> I use google blog search and technorati
- # [03:14] <mjs> the latter when I don't mind seeing even the trash posts on a topic
- # [03:14] <Hixie> i use google alerts
- # [03:14] <mjs> well yeah, google alerts mainly
- # [03:14] <Hixie> have one for whatwg and one for html5
- # [03:15] <mjs> I only have an alert for "webkit"
- # [03:15] <mjs> that consumes enough of my time
- # [03:15] <Hixie> hah
- # [03:15] <MikeSmith> mjs, Hixie - how often do you have them sent? once a day?
- # [03:15] <mjs> yeah daily
- # [03:16] <mjs> when bored I also do technorati and google blogsearch for "webkit" and "safari"
- # [03:16] * Joins: heycam (cam@203.217.91.18)
- # [03:17] <Hixie> i have them sent in realtime
- # [03:21] <MikeSmith> Hixie - I'm using the following search terms with Google to look for HTML5-related things. Wondering if it's overkill to do, e.g., "html5" OR "html 5" as far a google goes
- # [03:21] <Hixie> i just search for "html5"
- # [03:21] <Hixie> dunno if i'm missing stuff
- # [03:21] <Hixie> i might be
- # [03:21] <gavin> google will return resuls for "html 5" if you search for "html5", afaik
- # [03:21] <gavin> maybe not ranked as highly
- # [03:22] <MikeSmith> OK
- # [03:22] <Hixie> gavin: yeah but that's not the same as alerts
- # [03:22] <gavin> ok, I don't know what alerts does
- # [03:22] <gavin> I thought it just did a google query and sent you the new results every day
- # [03:23] <gavin> I suppose what it really does is top secret :)
- # [03:23] * Quits: timbl (timbl@209.6.134.246) (Connection reset by peer)
- # [03:23] * Joins: timbl (timbl@209.6.134.246)
- # [03:42] * Quits: DougJ (djones4@74.76.28.112) (Quit: DougJ)
- # [03:48] <hober> Sometimes I fear that "vast browser-wing conspiracy" will be my only long-lasting contribution to HTML5.
- # [03:49] <hober> heh. meant that for #whatwg
- # [03:53] * Quits: trackbot-ng (trackbot-n@128.30.52.30) (Client exited)
- # [03:53] * Quits: timbl (timbl@209.6.134.246) (Connection reset by peer)
- # [03:53] * Joins: timbl (timbl@209.6.134.246)
- # [03:55] * Parts: timbl (timbl@209.6.134.246)
- # [03:57] * Joins: trackbot-ng (trackbot-n@128.30.52.30)
- # [03:57] * trackbot-ng is loading HTML Issue Tracking data...
- # [03:57] * trackbot-ng found 16 users
- # [03:57] <trackbot-ng> Tracking ISSUEs and ACTIONs from http://www.w3.org/html/wg/tracker/
- # [04:02] * Quits: trackbot-ng (trackbot-n@128.30.52.30) (Client exited)
- # [04:02] * Quits: aroben (aroben@17.203.12.72) (Quit: Leaving)
- # [04:03] * Quits: dbaron (dbaron@71.204.145.103) (Ping timeout)
- # [04:07] * Joins: trackbot-ng (trackbot-n@128.30.52.30)
- # [04:07] * trackbot-ng is loading HTML Issue Tracking data...
- # [04:07] * trackbot-ng found 16 users
- # [04:07] <trackbot-ng> Tracking ISSUEs and ACTIONs from http://www.w3.org/html/wg/tracker/
- # [04:12] * Quits: drry (drry@222.225.140.32) (Ping timeout)
- # [04:13] * Quits: Shunsuke (Shunsuke@123.176.107.50) (Ping timeout)
- # [04:14] * Joins: Shunsuke (Shunsuke@123.176.107.50)
- # [04:23] * Joins: timbl_ (timbl@209.6.134.246)
- # [04:53] * Quits: timbl_ (timbl@209.6.134.246) (Connection reset by peer)
- # [04:53] * Joins: timbl (timbl@209.6.134.246)
- # [05:49] * Joins: drry (drry@222.225.140.32)
- # [05:53] * Quits: timbl (timbl@209.6.134.246) (Connection reset by peer)
- # [05:54] * Joins: timbl (timbl@209.6.134.246)
- # [06:09] * Quits: drry (drry@222.225.140.32) (Ping timeout)
- # [06:18] * Quits: mjs (mjs@17.255.98.46) (Quit: mjs)
- # [06:21] * Joins: drry (drry@222.225.140.32)
- # [06:23] * Quits: timbl (timbl@209.6.134.246) (Connection reset by peer)
- # [06:24] * Joins: timbl (timbl@209.6.134.246)
- # [07:02] <Hixie> "HTML5 is not backwards compatable" - http://warpspire.com/features/html5-css3/
- # [07:09] * Joins: dbaron (dbaron@71.204.145.103)
- # [07:11] * Joins: mjs (mjs@64.81.48.145)
- # [07:11] * Joins: anne-mac (annevk@212.17.132.42)
- # [07:13] * Joins: aroben (aroben@67.160.250.192)
- # [07:17] <MikeSmith> so "HTML5 is not backwards compatable" because...
- # [07:17] <MikeSmith> "... there is no HTML 4 or XHTML 1 equivelent of a <section />"
- # [07:17] <mjs> oh no
- # [07:17] <mjs> time to explain the compatibility Design Principles again
- # [07:18] <MikeSmith> time to explain basic logic maybe
- # [07:20] <MikeSmith> since instead of mentioning section specifically, he might just as well say, "there is no HTML 4 or XHTML 1 equivalent for any new element that might possibly be added"
- # [07:23] * Quits: timbl (timbl@209.6.134.246) (Connection reset by peer)
- # [07:24] * Joins: timbl (timbl@209.6.134.246)
- # [07:24] * MikeSmith wanders away for a bit
- # [07:24] * Quits: aaronlev (chatzilla@209.6.168.245) (Quit: ChatZilla 0.9.79 [Firefox 3.0b2pre/2007120105])
- # [07:41] * anne-mac wonders if the Forms TF is still alive
- # [07:42] <Hixie> you're one sixth of it :-)
- # [07:43] <Hixie> you could just resolve that xforms and wf2 are both perfect from the point of view of arch consistency with xforms transitional, and see who complains
- # [07:43] <mjs> heh
- # [07:43] <mjs> I didn't reply to your email cause I agree with the charter
- # [07:43] <mjs> but I guess I should have replied
- # [07:43] <anne-mac> maybe that's it
- # [07:53] * Quits: timbl (timbl@209.6.134.246) (Connection reset by peer)
- # [07:54] * Joins: timbl (timbl@209.6.134.246)
- # [07:59] * Quits: anne-mac (annevk@212.17.132.42) (Ping timeout)
- # [08:12] * Joins: Lachy (Lachlan@212.17.132.42)
- # [08:21] * Joins: anne-mac (annevk@212.17.132.42)
- # [08:26] <Lachy> DanC, yt? regarding your comment http://www.alistapart.com/comments/previewofhtml5?page=3#26
- # [08:28] <Lachy> how do you possibly expect us to finish the spec, write a full testsuite and achieve 2 interoperable implementations within 3 years?
- # [08:32] <mjs> I would be surprised if WebKit has a complete, fully conforming implementation by then
- # [08:32] <mjs> but I think CR is a more interesting milestone than REC
- # [08:33] <Hixie> for the spec, CR is the only milestone that really matters.
- # [08:34] <mjs> well, I can see how for content authors, "when can we expect to have this widely available" is an interesting question
- # [08:34] <mjs> it's too bad that it's so hard to answer that when it comes to open standards
- # [08:34] <Hixie> that answer will differ for each section
- # [08:34] <Hixie> some sections will have the answer "never" (and will in due course be dropped)
- # [08:34] <Hixie> others have the answer "yesterday"
- # [08:34] <Hixie> e.g. much of the stuff that was in HTML 3.2
- # [08:36] <mjs> Still, I wish we could cite some number smaller than 10-15 years
- # [08:37] <Lachy> mjs, next year we can say 9-14 :-)
- # [08:37] <mjs> clearly that doesn't play well with the audience when cited out of context
- # [08:37] <Hixie> http://www.whatwg.org/specs/web-apps/current-work/TIMETABLE says CR in 2012
- # [08:38] <Hixie> and last call in 2009
- # [08:38] <Hixie> so you could cite those numbers
- # [08:38] <Hixie> i changed the faq a few hours ago to point to the 2012 number first
- # [08:39] * mjs is reading over the listapart comments with interest
- # [08:39] <mjs> why the 3 years from LC to CR?
- # [08:39] <mjs> (I know, I'm being *such* an engineering manager right now - clearly I've lost sight of my roots)
- # [08:40] <Lachy> I tried to explain that the 10-15 years was due to the requirement for implementation and test suite, but some people seem to miss that point
- # [08:40] <mjs> I could see 1 year to process LC feedback and achieve at least initial CR, but 3?
- # [08:40] <Hixie> mjs: it's based on what i can optimistically do as a single person compared to what the csswg did as a committee
- # [08:41] <Hixie> mjs: and the expected volume of comments on html5 vs css2.1
- # [08:41] <Hixie> mjs: (i.e. i expect many more comments, but expect to go quicker per comment, so what took about 6 years in css should be doable in 3)
- # [08:42] <mjs> did css2.1 take 6 years to the first CR or to the most recent (hopefully final) CR?
- # [08:42] <mjs> (wondering which one you have in mind)
- # [08:43] <Hixie> most recent
- # [08:43] * Quits: Lachy (Lachlan@212.17.132.42) (Quit: Leaving)
- # [08:43] * Joins: Lachy (Lachlan@212.17.132.42)
- # [08:43] <mjs> so first CR might be sooner? (or is HTML5 comment stream expected to be more steady so there's no premature illusion of being done?)
- # [08:44] <Hixie> the timetable only includes the "real" milestones
- # [08:45] <Hixie> we could easily have many other milestones in there
- # [08:45] <Hixie> the "cr" in that timetable is when i'd expect to stop editing regularly
- # [08:47] <mjs> sadly it's hard to turn all this into a clear message
- # [08:47] <Hixie> see the faq for my best effort attempt
- # [08:50] <mjs> not a bad attempt
- # [08:50] <mjs> but not a sound bite either
- # [08:50] <Hixie> feel free to make it better :-)
- # [08:50] <Hixie> i'm a spec weenie, not a marketting guy :-)
- # [08:51] * Quits: dbaron (dbaron@71.204.145.103) (Quit: 8403864 bytes have been tenured, next gc will be global.)
- # [08:52] <mjs> I'm too honest to make a good marketing guy
- # [08:53] <Hixie> "OTOH, construction of The Cathedral of St. John The Divine was started in 1892 and it is still not yet completeted, yet it functions fully. Maybe HTML5 will be the St. John’s of the web." -- http://www.alistapart.com/comments/previewofhtml5?page=4#40
- # [09:00] <mjs> definitely a mixed bag of responses in the comments
- # [09:00] <mjs> some positive responses
- # [09:01] <mjs> negative-ish ones basically fall into "omg 10 years is too long", "I like div soup" and "I wish Microsoft supported standards better"
- # [09:03] <mjs> linking blog posts seem more positive on average than comments on the site afaict
- # [09:03] <mjs> anyway
- # [09:03] <Lachy> DanC, the markup on the html wg page is broken. See http://www.alistapart.com/comments/previewofhtml5?page=5#43
- # [09:03] <mjs> back to the grindstone
- # [09:05] <mjs> the comments tell me that Lachy's article was a much-needed piece of educational outreach and that more may be needed
- # [09:06] <Lachy> Hixie, is Vlad wrong about this being conforming? http://www.alistapart.com/comments/previewofhtml5?page=5#43
- # [09:07] <Lachy> IIRC, the spec says headings need to be used in order.
- # [09:07] <Hixie> it's certainly defined what that means
- # [09:07] <Hixie> i forget if it's conforming, hang on
- # [09:08] <Hixie> "Sections may contain headers of any rank, but authors are strongly encouraged to either use only h1 elements, or to use elements of the appropriate rank for the section's nesting level."
- # [09:08] <Hixie> so it's currently conforming
- # [09:08] <Hixie> i don't mind making it non-conforming, send mail
- # [09:11] <Lachy> ok. I thought it was already, it definitely should be.
- # [09:24] * Joins: timbl_ (timbl@209.6.134.246)
- # [09:24] * Quits: timbl (timbl@209.6.134.246) (Connection reset by peer)
- # [09:49] <Hixie> i don't understand karl's e-mail
- # [09:50] <Hixie> i agree that those use cases are things we should handle, that's why the spec is designed the way it is
- # [09:50] <anne-mac> I guess he wonders how you can ensure the correct order in face of copy and paste, etc.
- # [09:51] <Hixie> simply by always using <h1>...
- # [09:51] <anne-mac> I know
- # [09:52] <Hixie> you can't copy and paste an <h6>-headed-block and an <h1>-headed-block next to each other without checking anyway, since you don't know what order they'll be in (i.e. <h1>-then-<h6> means something different than <h1>-then-<h6>)
- # [09:52] <Hixie> so...
- # [09:52] <Hixie> oh well
- # [09:52] <Hixie> nn
- # [09:52] <MikeSmith> Lachy - problem with HTML WG page as my fault
- # [09:53] <MikeSmith> and it wasn't really a markup problem, it was just a borked URL in the link to the Design Principles WD
- # [09:55] <Lachy> MikeSmith, ok
- # [10:04] * Joins: jgraham_ (james@81.86.218.70)
- # [10:07] * Quits: Lachy (Lachlan@212.17.132.42) (Quit: This computer has gone to sleep)
- # [10:08] * Quits: anne-mac (annevk@212.17.132.42) (Ping timeout)
- # [10:14] * Joins: Lachy (Lachlan@212.17.132.42)
- # [10:20] * Quits: laplink (link@85.19.196.186) (Quit: Leaving)
- # [10:22] * Quits: Lachy (Lachlan@212.17.132.42) (Quit: This computer has gone to sleep)
- # [10:22] * Joins: anne-mac (annevk@212.17.132.42)
- # [10:23] * Joins: Lachy (Lachlan@212.17.132.42)
- # [10:24] * Quits: timbl_ (timbl@209.6.134.246) (Connection reset by peer)
- # [10:24] * Joins: timbl (timbl@209.6.134.246)
- # [10:25] * Quits: Lachy (Lachlan@212.17.132.42) (Quit: This computer has gone to sleep)
- # [10:29] * Quits: jgraham_ (james@81.86.218.70) (Quit: This computer has gone to sleep)
- # [10:30] * Joins: ROBOd (robod@89.122.216.38)
- # [10:33] * Joins: Lachy (Lachlan@212.17.132.42)
- # [10:35] * Quits: anne-mac (annevk@212.17.132.42) (Ping timeout)
- # [10:42] <Lachy> wow, articles like this annoy me http://24ways.org/2007/my-other-christmas-present-is-a-definition-list
- # [10:46] * Joins: anne-mac (annevk@212.17.132.42)
- # [10:47] * Quits: olivier (ot@128.30.52.30) (Quit: Leaving)
- # [10:54] * Quits: timbl (timbl@209.6.134.246) (Connection reset by peer)
- # [10:54] * Joins: timbl_ (timbl@209.6.134.246)
- # [11:03] * Quits: anne-mac (annevk@212.17.132.42) (Ping timeout)
- # [11:04] * Joins: anne-mac (annevk@212.17.132.42)
- # [11:11] <MikeSmith> hsivonen - running http://www.w3.org/html/wg/ through validator.nu against HTML5, I get:
- # [11:11] <MikeSmith> "Bad value irc://irc.w3.org:6665/html-wg for attribute href on element a: UNREGISTERED_IANA_SCHEME in SCHEME."
- # [11:13] <Philip> Is that incorrect, or is it just unhelpful?
- # [11:13] <MikeSmith> does the spec really say that only registered schemes in values are conformant?
- # [11:13] <MikeSmith> Philip - helpful
- # [11:13] <MikeSmith> just that the irc scheme is supported in browsers
- # [11:14] <MikeSmith> and in use in existing content
- # [11:14] <Lachy> MikeSmith, irc:// is not registered http://www.iana.org/assignments/uri-schemes.html
- # [11:14] * Quits: aroben (aroben@67.160.250.192) (Ping timeout)
- # [11:14] <Lachy> it really should be registered, since it's widely deployed already
- # [11:15] <MikeSmith> yeah
- # [11:16] <Lachy> MikeSmith, there's a draft for it, but AFAICT, it never got published as an RFC http://www.w3.org/Addressing/draft-mirashi-url-irc-01.txt
- # [11:16] <MikeSmith> but in the mean time, it would be nice to not have conformance checkers flag it as an error
- # [11:17] <MikeSmith> I don't think HTML5 itself says that href values must only use registered schemes
- # [11:18] <Lachy> it would be a useful feature for checking typos, like if you typed href="htp://..." instead of http:, but I'm not sure if it should only use the IANA registry\
- # [11:19] <MikeSmith> the fact that validator.nu is flagging it as an error and not a warning makes me believe Henri has probably got a reason for it
- # [11:19] <MikeSmith> HTML5 just says "The href attribute on a hyperlink element must have a value that is a URI (or IRI)."
- # [11:21] <MikeSmith> As far as I remember, the URI/IRI spec doesn't say that they can only contain registered schemes.
- # [11:30] * Quits: anne-mac (annevk@212.17.132.42) (Ping timeout)
- # [11:37] * Joins: beowulf (beowulf@194.74.230.217)
- # [11:49] * Quits: sbuluf (cavgya@200.49.132.115) (Quit: sbuluf)
- # [11:53] * Quits: timbl_ (timbl@209.6.134.246) (Connection reset by peer)
- # [11:54] * Joins: timbl (timbl@209.6.134.246)
- # [12:11] * Quits: Lachy (Lachlan@212.17.132.42) (Quit: This computer has gone to sleep)
- # [12:15] * Joins: anne-mac (annevk@212.17.132.42)
- # [12:17] <Hixie> Each URI begins with a scheme name that refers to a specification for
- # [12:17] <Hixie> assigning identifiers within that scheme.
- # [12:17] <Hixie> The scheme registry maintains the mapping between scheme names and
- # [12:17] <Hixie> their specifications.
- # [12:17] <Hixie> -- http://www.ietf.org/rfc/rfc3986.txt (3.1:1 and 3.1:3)
- # [12:23] * Quits: timbl (timbl@209.6.134.246) (Connection reset by peer)
- # [12:24] * Joins: timbl (timbl@209.6.134.246)
- # [12:26] * Joins: myakura (myakura@221.191.115.31)
- # [12:27] <hsivonen> MikeSmith: IRI schemes must be registered or must start with "x-", IIRC
- # [12:27] <hsivonen> MikeSmith: or perhaps it was that the unregistered ones need to contain a '-'
- # [12:28] <hsivonen> MikeSmith: anyway, if there's a list of Support Existing Content de facto IRI schemes that I should whitelist in defiance of the RFCs, I could do that
- # [12:28] <hsivonen> MikeSmith: javascript: is already allowed
- # [12:28] <MikeSmith> about the registration constraint, according to RFC 3986 and 3987, you mean?
- # [12:29] <MikeSmith> irc: should be whitelisted, if it comes to that
- # [12:29] <MikeSmith> I think
- # [12:30] <MikeSmith> but I can't find any assertion in RFC 3986 that schemes must be registered
- # [12:30] <MikeSmith> that UNREGISTERED_IANA_SCHEME thing is a Java error, right?
- # [12:31] <MikeSmith> from a Java library of some kind, I mean?
- # [12:31] <MikeSmith> that is, not some constraint check that you have added to the code yourself?
- # [12:32] <hsivonen> MikeSmith: http://tools.ietf.org/html/bcp35 says non-IETF schema names need to have a hyphen in them
- # [12:32] <hsivonen> MikeSmith: the error comes from the Jena IRI lib
- # [12:32] <hsivonen> MikeSmith: and in general, it appears that the Jena IRI guy read the RFCs very carefully
- # [12:32] <hsivonen> s/schema/scheme/
- # [12:35] <MikeSmith> as far I can see, BCP 35 defines a process for registering scheme names
- # [12:35] <MikeSmith> it does not define a what a conformant unregistered URI is
- # [12:35] <MikeSmith> but I'm reading it right now anyway
- # [12:36] * Quits: anne-mac (annevk@212.17.132.42) (Ping timeout)
- # [12:36] <hsivonen> "Each URI begins with a scheme name that refers to a specification for assigning identifiers within that scheme.
- # [12:36] <hsivonen> "
- # [12:36] <hsivonen> "The scheme registry maintains the mapping between scheme names and their specifications.
- # [12:36] <hsivonen> "
- # [12:37] <hsivonen> MikeSmith: it's quite possible that an IRI5 spec is needed in due course. for example, the restrictions that IRI compliance places on javascript: are unhelpful
- # [12:38] <MikeSmith> hsivonen - yeah
- # [12:40] <MikeSmith> very much a case of the RFCs being broken if they do in fact make actual use of irc: and javascript: in existing content non-conformant
- # [12:42] * Quits: heycam (cam@203.217.91.18) (Quit: bye)
- # [12:43] <hsivonen> http://wiki.whatwg.org/wiki/De_facto_IRI_schemes
- # [12:43] <hsivonen> please fill in other cases you might find
- # [12:43] <MikeSmith> hsivonen - OK, will do
- # [12:45] <hsivonen> I'll fix irc: after lunch
- # [12:50] <MikeSmith> hsivonen - as I read it at least, BCP 35 is not asserting that any given unregistered scheme name that lacks a dash is non-conformant; it's instead saying (1) that IETF-registered scheme names must not contain a dash and (2) that scheme names for "schemes registered in alternative trees" must be identified by a unique prefix, followed by a dash, followed by some identifier unique within that alternative.
- # [12:50] <hsivonen> Hixie: have you analyzed Web content for IRI schemes by any chance?
- # [12:50] <hsivonen> MikeSmith: yeah, I can't find a clear sentence requiring registration, either
- # [12:51] <hsivonen> MikeSmith: in general, for HTML5 validation whether to require IANA registration is a practical judgment call
- # [12:51] <hsivonen> MikeSmith: it is feasible to require language tags and charsets to be registered
- # [12:52] <hsivonen> MikeSmith: it isn't feasible to require MIME types to be registered
- # [12:52] <hsivonen> MikeSmith: apparently, it isn't feasible for IRI schemes, either
- # [12:52] <hsivonen> but I'm hoping I could get away with a whitelist instead of opening it to everything
- # [12:53] <MikeSmith> hsivonen - seems like maybe not, since neither BCP 35 or RFC 3986 seem to say that only registered names are conformant. BCP 35 does not seem to say anything at all about unregistered schemes. It seems to only be defining rules for registered names.
- # [12:54] <MikeSmith> hsivonen - whitelist + emitting a warning for others (rather than error) would seem appropriate
- # [12:55] <MikeSmith> Is there a way you can emit a warning instead of an error for this particular check?
- # [12:55] <hsivonen> the bigger problem that registration as such is that the draft registrations for irc and javascript languish instead of making it through the system swiftly
- # [12:55] <hsivonen> MikeSmith: the infrastructure does not support warnings for these checks
- # [12:55] <MikeSmith> OK
- # [12:55] <hsivonen> MikeSmith: changing that would not be pretty :-(
- # [12:56] <MikeSmith> yeah, I figured that was probably so
- # [13:00] <MikeSmith> hsivonen - Another thing I've been wanting to ask Hixie: If <meta@http-equiv=Content-Type...> is not conformant according to HTML5, then the mass of content out there that has instances of that in it are all going to fail conformance-checking, right?
- # [13:00] <hsivonen> MikeSmith: yes, changing the doctype is not enough to become conferming ATM, you also need to change the meta charset thingy
- # [13:01] <hsivonen> conforming even
- # [13:02] <hsivonen> MikeSmith: I've argued that for compat with legacy markup generation code we should make the old-style long form conforming
- # [13:02] <hsivonen> MikeSmith: Hixie has disagreed so far
- # [13:02] <MikeSmith> Yeah, I would argue for that too
- # [13:03] <MikeSmith> Otherwise, what I fear conformance checkers will do is just add an option 'Ignore error messages about <meta http-equiv="Content-Type">'
- # [13:04] <MikeSmith> rather than the maintainer getting 500 messages a week from users bitching about the checker
- # [13:05] <hsivonen> MikeSmith: well, the #1 contender issue for that is the requirement to have wrapper <p>s in more places than in HTML 4.01 Transitional
- # [13:05] <MikeSmith> I've not run into that problem so far with pages I've checked with validator.nu
- # [13:06] <MikeSmith> So I guess in instances I author, it's something I'm already doing right :)
- # [13:06] <hsivonen> me too :-)
- # [13:07] <MikeSmith> so to hell with all the other people who mark stuff up in such a way that it causes them to run into that
- # [13:07] <MikeSmith> sucks for them
- # [13:07] <MikeSmith> too bad
- # [13:07] * MikeSmith is kidding of course
- # [13:08] <MikeSmith> the best measure of some of those cases is going to be the maintainers of conformance checkers, really
- # [13:10] <Philip> Should the HTML4/XHTML1/etc doctypes be conforming HTML5 too? Then people wouldn't have any necessary changes to make before being conforming HTML5
- # [13:12] <hsivonen> Philip: yeah, that's the obvious next question. and politically, it wouldn't fly. at least not easily
- # [13:12] <MikeSmith> But there is a finite set of them
- # [13:13] <MikeSmith> Though I guess it comes down to the supported vs. conformant distinction
- # [13:14] <Philip> An anonymous developer should set up a politically incorrect fork of the conformance checker, and see if authors like it more
- # [13:14] <MikeSmith> true
- # [13:17] <MikeSmith> but it does make sense to have automated checkers flagging instances of this stuff -- markup cluttered with unnecessary and arcane information that browsers just discard or ignore anyway
- # [13:17] <MikeSmith> there should be some means to discourage use of that markup
- # [13:17] <MikeSmith> and if conformance checkers don't flag it, there won't be
- # [13:21] <Philip> There are far more significant errors on most web pages, so perhaps more benefit would come from a conformance checker that people will use more often because it doesn't complain at them so much about minor things with little practical benefit, rather than a conformance checker which flags more issues and drives away more users
- # [13:23] <MikeSmith> Philip - aye
- # [13:23] * Quits: myakura (myakura@221.191.115.31) (Quit: Leaving...)
- # [13:23] * Quits: timbl (timbl@209.6.134.246) (Connection reset by peer)
- # [13:24] * Joins: timbl (timbl@209.6.134.246)
- # [13:29] <hsivonen> http://en.wikipedia.org/wiki/URI_scheme#Unofficial_but_common_URI_schemes
- # [13:34] <hsivonen> well, it looks like trying to enforce the IANA registry is pointless
- # [13:35] <hsivonen> and using wikipedia as the registry would be arbitrary
- # [13:35] <hsivonen> sign
- # [13:35] <hsivonen> sigh even
- # [13:37] <hsivonen> MikeSmith: what's your take on backslashes in file: URLs?
- # [13:38] <hsivonen> it seems to me I have to do IRI checking the other way round
- # [13:40] <MikeSmith> I have seen enough questions on mailing lists about backslashes in file: URLs to make me think that it might be another case of be liberal rather that trying to educate
- # [13:40] <MikeSmith> hsivonen
- # [13:41] <hsivonen> MikeSmith: I see
- # [13:43] * Joins: billyjack (MikeSmith@mcclure.w3.org)
- # [13:43] * Quits: billyjack (MikeSmith@mcclure.w3.org) (Quit: Less talk, more pimp walk.)
- # [13:47] * Quits: timbl (timbl@209.6.134.246) (Quit: timbl)
- # [13:52] <hsivonen> mjs: regarding WebKit video selection:
- # [13:52] <hsivonen> bug #1: video/mp4 is skipped if it declares codecs: http://hsivonen.iki.fi/test/moz/video-selection/source-mp4-params.html
- # [13:54] <hsivonen> bug #2: Ogg in <source> never plays even with XiphQT
- # [13:54] <hsivonen> http://hsivonen.iki.fi/test/moz/video-selection/source-ogv.html
- # [13:54] <hsivonen> http://hsivonen.iki.fi/test/moz/video-selection/source-ogg.html
- # [13:54] <hsivonen> those two don't play when
- # [13:54] <hsivonen> http://hsivonen.iki.fi/test/moz/video-selection/src-ogg.html
- # [13:54] <hsivonen> and
- # [13:54] <hsivonen> http://hsivonen.iki.fi/test/moz/video-selection/src-ogv.html
- # [13:54] <hsivonen> do
- # [14:07] * Joins: myakura (myakura@221.191.115.31)
- # [14:15] * Joins: Lachy (Lachlan@212.17.132.42)
- # [14:18] * Quits: Lachy (Lachlan@212.17.132.42) (Quit: This computer has gone to sleep)
- # [14:19] * Joins: Lachy (Lachlan@212.17.132.42)
- # [14:19] * Quits: Lachy (Lachlan@212.17.132.42) (Client exited)
- # [14:19] <MikeSmith> hsivonen - another issue with <meta http-equiv="Content-Type" ...> being nonconformant is that the XSLT spec says:
- # [14:19] * Joins: Lachy (Lachlan@212.17.132.42)
- # [14:20] <MikeSmith> says that for xsl:output method="html" that:
- # [14:20] <MikeSmith> [[If there is a HEAD element, then the html output method should add a META element immediately after the start-tag of the HEAD element specifying the character encoding actually used. For example,
- # [14:20] <MikeSmith> <HEAD>
- # [14:20] <MikeSmith> <META http-equiv="Content-Type" content="text/html; charset=EUC-JP">
- # [14:20] <MikeSmith> ...]]
- # [14:20] * Quits: Lachy (Lachlan@212.17.132.42) (Quit: Leaving)
- # [14:21] <hsivonen> MikeSmith: method='html' is not appropriate for HTML5 anyway
- # [14:21] <MikeSmith> hsivonen - why not?
- # [14:22] * Joins: Lachy (Lachlan@212.17.132.42)
- # [14:22] <hsivonen> MikeSmith: It cannot generate <!DOCTYPE html>. I don't remember if there were other issues.
- # [14:23] <hsivonen> MikeSmith: oh yeah, now I remember
- # [14:23] <hsivonen> MikeSmith: method='html' wants the elemnts to be in no namespace as opposed to the XHTML namespace
- # [14:25] <hsivonen> the XSLT4HTML5 sample program that comes with the Validator.nu parser solves both problems
- # [14:28] <MikeSmith> I do the following to deal with the doctype problem:
- # [14:28] <MikeSmith> <xsl:template match="/">
- # [14:28] <MikeSmith> <xsl:text disable-output-escaping="yes"><!DOCTYPE html></xsl:text>
- # [14:28] <MikeSmith> <xsl:text> </xsl:text>
- # [14:28] <MikeSmith> <xsl:apply-templates/>
- # [14:28] <MikeSmith> </xsl:template>
- # [14:28] <MikeSmith> ugly hack but it works
- # [14:28] <MikeSmith> learnt that from Sam Ruby
- # [14:30] <MikeSmith> hsivonen - are you saying that for method="html", one would want for some reason for the output to be in the XHTML namespace?
- # [14:36] * Joins: timbl (timbl@128.30.6.53)
- # [14:36] <Philip> You would want the input to be in the XHTML namespace, assuming you're using XHTML5 and not some other undefined XMLisation of HTML5
- # [14:36] <Philip> Don't know if that's actually the problem here, though
- # [14:40] <hsivonen> MikeSmith: I'm saying that one would want the tree that results from the transformation to have its elements in the XHTML namespace
- # [14:40] <hsivonen> MikeSmith: and then serializing to HTML5 should be a serializer-level issue
- # [14:40] <hsivonen> MikeSmith: your transformation program shouldn't have to deal with serialization details
- # [14:42] <MikeSmith> hsivonen - I see. But when I run method=html transformations on XHTML files, XSLT engines do seem to preserve the namespace in the result tree
- # [14:43] <MikeSmith> and I need run a namespace-stripping template match to strip the namespace in fact
- # [14:43] <hsivonen> MikeSmith: do you mean that if you change to method='xml', the namespace is the XHTML namespace?
- # [14:43] <hsivonen> XSLT4HTML5 makes things work
- # [14:44] * Parts: timbl (timbl@128.30.6.53)
- # [14:44] <hsivonen> the output is HTML5 and you write the transformation to target the HTML5-compatible subset of XHTML5
- # [14:46] <MikeSmith> hsivonen - If I run an identity transform on an XHTML-namespaced source file, the result tree is in the XHTML namespace both when method=xml and method=html
- # [14:46] <hsivonen> MikeSmith: and what does the method=html serialization look like?
- # [14:46] <hsivonen> (IIRC, Gecko doesn't comply with the spec here to make things work)
- # [14:49] <MikeSmith> hsivonen - for a minimal input doc like this:
- # [14:49] * Quits: Lachy (Lachlan@212.17.132.42) (Quit: This computer has gone to sleep)
- # [14:50] <MikeSmith> <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"
- # [14:50] <MikeSmith> "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
- # [14:50] <MikeSmith> <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
- # [14:50] <MikeSmith> <head>
- # [14:50] <MikeSmith> <title>FOO</title>
- # [14:50] <MikeSmith> </head>
- # [14:50] <MikeSmith> <body>
- # [14:50] <MikeSmith> <p>FOO</p>
- # [14:50] <MikeSmith> </body>
- # [14:50] <MikeSmith> </html>
- # [14:50] <MikeSmith> that gets serialized for output=html like this:
- # [14:51] <MikeSmith> <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
- # [14:51] <MikeSmith> <head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
- # [14:51] <MikeSmith> <title>FOO</title>
- # [14:51] <MikeSmith> </head>
- # [14:51] <MikeSmith> <body>
- # [14:51] <MikeSmith> <p>FOO</p>
- # [14:51] <MikeSmith> </body>
- # [14:51] <MikeSmith> </html>
- # [14:51] <hsivonen> MikeSmith: looks broken as designed to me
- # [14:51] <MikeSmith> hsivonen - how so?
- # [14:52] <hsivonen> MikeSmith: we want this:
- # [14:52] <hsivonen> <!DOCTYPE html>
- # [14:52] <hsivonen> <html lang='en'>
- # [14:52] <hsivonen> <head><meta charset='utf-8'>
- # [14:52] <hsivonen> etc.
- # [14:52] <hsivonen> but that's from the future
- # [14:52] <hsivonen> the broken as designed part is
- # [14:53] <hsivonen> xmlns="http://www.w3.org/1999/xhtml" xml:lang="en"
- # [14:53] <hsivonen> that stuff has no place in XSLT 1.0-era HTML
- # [14:54] <hsivonen> I think we should just concede that method='html' sucks and people should use an HTML5 serializer instead if they want HTML5
- # [14:54] <MikeSmith> OK, I see. Yeah, when processing it with XSLT, it is possible to strip all that during the transform, and to add the <!DOCTYPE html>
- # [14:55] <MikeSmith> I can imagine that some developers might want to use XSTL 1.0 tools to manage transformations to HTML5
- # [14:55] <MikeSmith> at least until the more appropriate tools are widely available
- # [14:55] <MikeSmith> e.g. packaged for Linux distributions
- # [14:55] <MikeSmith> and such
- # [14:56] <hsivonen> MikeSmith: well, if they want to use the XSLT tool to produce the serilization, they are in the land of hurt
- # [14:56] <hsivonen> MikeSmith: what they should do is to take DOM or SAX output from the XSLT 1.0 engine and give it to an HTML5 serializer
- # [14:56] <MikeSmith> most developers already find themselves deep into the land of hurt already
- # [14:57] <Philip> You can't do much with DOM or SAX output if you're using command line processing tools
- # [14:57] * Philip guesses an xhtml2html command-line tool would be handy in that case
- # [14:57] <MikeSmith> yeh
- # [14:57] <MikeSmith> that would be the best case
- # [14:57] <MikeSmith> a simple standalone tool
- # [14:58] <MikeSmith> for the simple case, I mean
- # [14:58] <Philip> I guess it should preferably be non-Java
- # [14:59] <Philip> because Java isn't really great at doing quick command-line pipeline tools
- # [14:59] <MikeSmith> simple being if you don't have any need to manipulate the result tree before serializing
- # [14:59] <Philip> If you do need to, you do "xsltproc foo.xslt bar.xml | xhtml2html > bar.html"
- # [15:03] * Philip supposes xhtml2html would be far easier than html2xhtml
- # [15:08] <hsivonen> eventually, libxml2 needs to get an HTML5 parser and a serializer
- # [15:20] * Joins: aaronlev (chatzilla@66.30.196.151)
- # [15:27] <MikeSmith> hsivonen - yeah
- # [15:31] * Joins: Lachy (Lachlan@212.17.132.42)
- # [15:31] * Joins: matt (matt@128.30.52.30)
- # [15:33] * Quits: Lachy (Lachlan@212.17.132.42) (Quit: Leaving)
- # [15:33] <MikeSmith> hsivonen - in the mean time, I guess I should just write a simple-case xhtmltohtml tool like Philip describes, using html5lib
- # [15:33] * Joins: Lachy (Lachlan@212.17.132.42)
- # [15:37] <hsivonen> MikeSmith: good idea
- # [15:42] <MikeSmith> starting from http://html5lib.googlecode.com/svn/trunk/python/parse.py
- # [15:45] <Philip> Does html5lib have a serialiser now?
- # [15:46] <Philip> (I just wrote my own for the spec-splitter, since I couldn't find one elsewhere)
- # [15:47] <Philip> Oh, looks like HTMLSerializer probably does that
- # [16:06] * Quits: myakura (myakura@221.191.115.31) (Quit: Leaving...)
- # [16:08] * Quits: Lachy (Lachlan@212.17.132.42) (Quit: This computer has gone to sleep)
- # [16:12] * Joins: laplink (link@193.214.121.76)
- # [16:15] * Joins: Lachy (Lachlan@212.17.132.42)
- # [16:34] * Quits: Lachy (Lachlan@212.17.132.42) (Quit: This computer has gone to sleep)
- # [16:37] * Joins: Lachy (Lachlan@212.17.132.42)
- # [16:39] * Quits: Lachy (Lachlan@212.17.132.42) (Quit: Leaving)
- # [16:40] * Joins: Lachy (Lachlan@212.17.132.42)
- # [16:51] <Philip> Are svn:// and svn+ssh:// official/unofficial URI schemes?
- # [16:52] <Philip> (They're not in the Wikipedia list, but I can't find other details either)
- # [16:54] <hsivonen> Philip: dunno. Do they conform to the generic IRI syntax?
- # [16:56] <Philip> hsivonen: I have no idea whatsoever
- # [16:58] * Quits: laplink (link@193.214.121.76) (Quit: This computer has gone to sleep)
- # [17:01] <hsivonen> for now I am assuming that only mailto:, data: and javascript: are exempt from the generic syntax
- # [17:31] <MikeSmith> hsivonen - I remember one more that's supported here in Japan at least, in browsers running on handsets in the the KDDI/Au mobile network
- # [17:31] <MikeSmith> http://www.au.kddi.com/ezfactory/tec/spec/eznavi.html
- # [17:32] <MikeSmith> it's device:
- # [17:33] <MikeSmith> device:location in Au's browser causes the browser to initiate a query to the handset to get its GPS location
- # [17:34] <mjs> hsivonen: thanks for the details, I'll pass those test cases on to our <video> folks
- # [17:34] <hsivonen> MikeSmith: it isn't obvious to me that the device scheme were incompatible with the generic syntax
- # [17:35] <MikeSmith> hsivonen - I'm not sure that I understand what the generic syntax is
- # [17:35] <hsivonen> MikeSmith: the examples seem to pass as generic IRIs
- # [17:35] <hsivonen> MikeSmith: I'm not sure, either. :-)
- # [17:35] <hsivonen> MikeSmith: I just trust the lib
- # [17:35] <hsivonen> mjs: thanks
- # [17:45] <hendry> MikeSmith: that location stuff interests me. more of it. :) in English
- # [17:45] <hendry> if possible
- # [17:46] <MikeSmith> hendry - I have no info on it in English. But it's dead simple, really. Just relies on browser support.
- # [17:46] <MikeSmith> take a look at the source for this page:
- # [17:46] <MikeSmith> http://sideshowbarker.net/gps/
- # [17:47] <MikeSmith> that's a minimal demo page I set up a while back
- # [17:47] * hendry wonders how to disable caching in apache and/or browser for JS dev
- # [17:47] <MikeSmith> no security around it at all
- # [17:48] <hendry> though i guess the user must iniatiate this url right?
- # [17:49] <MikeSmith> yeah, user has to click on a link
- # [17:49] <MikeSmith> but except for preferences you set on the device about whether you want to be informed each time a location, that's it
- # [17:50] <MikeSmith> if the browser supported scripting and XHR and user had device preferences set to not inform them about location query is initiated, they would not know
- # [17:52] <MikeSmith> but the browser does not support scription, so it's a moot point
- # [17:52] <MikeSmith> this is Openwave v6
- # [17:52] <MikeSmith> anyway, I got to wander off for a while
- # [17:52] <MikeSmith> back later
- # [17:52] <hendry> MikeSmith: thanks for the info
- # [18:03] * Quits: mjs (mjs@64.81.48.145) (Quit: mjs)
- # [18:17] * Joins: aroben (aroben@17.203.12.236)
- # [18:17] * Quits: aroben (aroben@17.203.12.236) (Quit: Leaving)
- # [18:17] * Joins: aroben (aroben@17.203.12.72)
- # [18:26] * Joins: jgraham_ (james@81.86.218.70)
- # [18:40] * Joins: Sander (svl@86.87.68.167)
- # [18:43] * Quits: ChrisWilson (cwilso@131.107.0.101) (Ping timeout)
- # [18:49] * Joins: ChrisWilson (cwilso@131.107.0.104)
- # [19:04] * Joins: mjs (mjs@17.255.98.46)
- # [19:06] * Quits: Lachy (Lachlan@212.17.132.42) (Quit: This computer has gone to sleep)
- # [19:12] <mjs> good morning folks
- # [19:12] <ChrisWilson> good morning
- # [19:15] * Joins: dbaron (dbaron@63.245.220.241)
- # [19:18] <mjs> hi ChrisWilson
- # [19:18] <mjs> hello dbaron
- # [19:24] * Joins: Julian (chatzilla@130.129.20.163)
- # [19:44] * Joins: gsnedders (gsnedders@86.145.188.131)
- # [19:49] * Quits: mjs (mjs@17.255.98.46) (Quit: mjs)
- # [19:52] * Joins: mjs (mjs@17.255.98.46)
- # [19:53] * Quits: Julian (chatzilla@130.129.20.163) (Quit: ChatZilla 0.9.79 [Firefox 2.0.0.11/2007112718])
- # [19:55] * Joins: timbl_ (timbl@128.30.6.53)
- # [20:10] * DanC returns to the office after lunch... kinda low on sleep...
- # [20:29] * Parts: hendry (hendry@89.16.172.32)
- # [20:39] * Quits: gsnedders (gsnedders@86.145.188.131) (Quit: 404: Not Found)
- # [20:40] * Joins: gsnedders (gsnedders@86.138.194.118)
- # [20:45] * Joins: tH (Rob@83.100.250.48)
- # [20:55] * Quits: mjs (mjs@17.255.98.46) (Quit: mjs)
- # [20:55] * Quits: tH (Rob@83.100.250.48) (Ping timeout)
- # [20:55] * Joins: Lachy (Lachlan@212.17.132.42)
- # [20:56] * Quits: Lachy (Lachlan@212.17.132.42) (Quit: Leaving)
- # [20:56] * Joins: Lachy (Lachlan@212.17.132.42)
- # [20:57] * Joins: tH (Rob@83.100.250.48)
- # [21:13] * Joins: mjs (mjs@17.255.98.46)
- # [21:14] * Quits: mjs (mjs@17.255.98.46) (Quit: mjs)
- # [21:19] * Quits: Lachy (Lachlan@212.17.132.42) (Quit: This computer has gone to sleep)
- # [21:24] * Quits: dbaron (dbaron@63.245.220.241) (Quit: 8403864 bytes have been tenured, next gc will be global.)
- # [21:31] * Joins: dbaron (dbaron@63.245.220.241)
- # [21:59] * Joins: mjs (mjs@17.255.98.46)
- # [22:07] * Quits: ROBOd (robod@89.122.216.38) (Quit: http://www.robodesign.ro )
- # [22:22] * Quits: dbaron (dbaron@63.245.220.241) (Quit: 8403864 bytes have been tenured, next gc will be global.)
- # [22:29] * Joins: dbaron (dbaron@63.245.220.241)
- # [22:30] <Hixie> hsivonen, MikeSmith: the main reason i don't want an htp-equiv="Content-Type" is to avoid redundant ways of doing the same thing -- if we want to deprecate it in due course, we should just ban it now
- # [22:32] <mjs> it does fall slightly afoul of making existing documents non-conforming for a slightly gratuitous reason
- # [22:32] <hsivonen> Hixie: what's your take on the validation results Sam Ruby posted?
- # [22:32] <mjs> even if they do what is currently a best practice
- # [22:34] <hsivonen> Hixie: a quick glance at the errors suggests that there are 2: meta charset in XHTML and lots and lots of inline stuff without paragraph wrappers
- # [22:35] <Hixie> well you know my take on mixing blocks and inlines without a block-level element
- # [22:36] <Hixie> i don't know how to handle the http-equiv=content-type case
- # [22:39] <gsnedders> I'd allow http-equiv=content-type, and require that the value MUST be text/html, with a charset parameter
- # [22:40] <hsivonen> gsnedders: that would still leave the question: what to do with Sam's case
- # [22:41] * gsnedders hasn't actually looked at that
- # [22:41] <mjs> what does Sam's case say?
- # [22:41] <hsivonen> his meta charset in non-conforming because he is trying to write a hybrid document so that the same bytes get served as text/html and application/xhtml+xml
- # [22:41] <Philip> For <meta charset> in XHTML, is it possible to say the value must be equal to whatever character encoding the XML layer chose to use?
- # [22:42] <hsivonen> mjs: it says <meta charset='utf-8'/> in application/xhtml+xml
- # [22:42] <Hixie> oh right
- # [22:42] <Hixie> his problem is the xml problem, not the legacy problem
- # [22:42] <Hixie> <-- was confoosed
- # [22:42] <hsivonen> Philip: possible but very bad from a clean layering point of view
- # [22:43] <mjs> ah, I see
- # [22:43] <Hixie> in sam's case he should just strip the meta charset
- # [22:43] <gsnedders> Philip: authors will easily get the wrong idea
- # [22:43] <Hixie> and use the mime type to set the charset
- # [22:43] <hsivonen> Philip: bad spec layering translates into bad software layering
- # [22:43] <hsivonen> Hixie: true
- # [22:51] <Hixie> "When reading the “differences” between version 4 and 5 I feel like we are going to lose a lot of time.
- # [22:51] <Hixie> There is no “semantic“, there is nothing that would make the useless union between contents and layout being split." -- http://www.fchouse.com/archives/html5-is-this-the-improvement
- # [22:51] <Hixie> that's especially funny given the people who are arguing that we're going too far into the theoretical
- # [22:53] <Hixie> "Let's hope that the new HTML 5 media nodes utilise Flash otherwise I feel that will be the end of Flash as we know it" -- http://www.talkphp.com/showthread.php?t=1615
- # [22:54] <mjs> Hixie: I... don't really understand that sentiment
- # [22:58] * Quits: timbl_ (timbl@128.30.6.53) (Quit: timbl_)
- # [23:00] <hsivonen> I identified four classes of errors in Sam's document
- # [23:03] <Dashiva> Hixie: Isn't he just suggesting using an XML toolchain instead of HTML in that html-must-die post he links to?
- # [23:06] <Hixie> it's a common suggestion
- # [23:07] <Dashiva> Yeah, but I don't see how that has to be and either/or with developing HTML (that is, the end format)
- # [23:07] <Hixie> wow, www-archive has 40 subscribers
- # [23:08] <Hixie> Dashiva: don't ask me :-)
- # [23:08] <Dashiva> True
- # [23:11] * Quits: matt (matt@128.30.52.30) (Client exited)
- # [23:22] * Joins: Zeros (Zeros-Elip@129.2.175.74)
- # [23:25] * Philip gets fed up with canvas tests again
- # [23:27] <Philip> (but after some mild practical achievement, so I feel satisfied for now)
- # [23:48] * Joins: timbl (timbl@209.6.134.246)
- # [23:58] * DanC looks around for ChisWilson
- # [23:58] <DanC> er... ChrisWilson
- # [23:59] <DanC> MikeSmith, I don't think (missed) is useful in the schedule. The schedule is subject to negotiation, as it says; that milestone was renegotiated
- # [23:59] * DanC goes to fix...
- # Session Close: Thu Dec 06 00:00:01 2007
The end :)