Options:
- # Session Start: Sun May 06 00:00:00 2007
- # Session Ident: #html-wg
- # [00:04] * Quits: edas (edaspet@88.191.34.123) (Quit: http://eric.daspet.name/ et l'édition 2007 de http://www.paris-web.fr/ )
- # [00:24] * Joins: asbjornu (asbjorn@84.48.116.134)
- # [00:27] * Quits: sbuluf (hvtbcfc@200.49.140.172) (Ping timeout)
- # [00:39] * Quits: tH (r@87.102.20.82) (Quit: ChatZilla 0.9.78.1-rdmsoft [XULRunner 1.8.0.9/2006120508])
- # [00:39] * Quits: zcorpan_ (zcorpan@217.211.77.236) (Ping timeout)
- # [00:50] * Quits: asbjornu (asbjorn@84.48.116.134) (Ping timeout)
- # [00:59] * Joins: sbuluf (itqakvg@200.49.140.7)
- # [01:05] * Joins: hyatt (hyatt@24.6.91.161)
- # [01:31] * Quits: gavin_ (gavin@74.103.208.221) (Ping timeout)
- # [01:37] * Joins: gavin_ (gavin@74.103.208.221)
- # [02:17] * Joins: polin8 (polin8@24.184.204.6)
- # [02:21] * Joins: schepers (schepers@208.38.57.47)
- # [02:22] * Parts: hasather (hasather@81.235.209.174)
- # [02:53] * Joins: schepers_ (schepers@208.38.57.47)
- # [02:55] * Quits: schepers (schepers@208.38.57.47) (Ping timeout)
- # [02:56] * Joins: schepers (schepers@208.38.57.47)
- # [02:59] * Quits: schepers_ (schepers@208.38.57.47) (Ping timeout)
- # [03:01] * Quits: polin8 (polin8@24.184.204.6) (Quit: polin8)
- # [03:05] * Quits: hyatt (hyatt@24.6.91.161) (Connection reset by peer)
- # [03:05] * Joins: hyatt (hyatt@24.6.91.161)
- # [03:10] * Joins: schepers_ (schepers@208.38.57.47)
- # [03:12] * Joins: polin8 (polin8@24.184.204.6)
- # [03:13] * Quits: schepers (schepers@208.38.57.47) (Ping timeout)
- # [03:13] * Quits: schepers_ (schepers@208.38.57.47) (Ping timeout)
- # [03:32] * Quits: polin8 (polin8@24.184.204.6) (Quit: :wq)
- # [03:38] * Quits: gavin_ (gavin@74.103.208.221) (Ping timeout)
- # [03:43] * Joins: gavin_ (gavin@74.103.208.221)
- # [04:02] * Quits: hyatt (hyatt@24.6.91.161) (Quit: hyatt)
- # [04:19] * Joins: inimino (chatzilla@75.71.88.233)
- # [04:29] * Joins: myakura (myakura@125.194.247.196)
- # [05:15] * Quits: beowulf (carisenda@91.84.50.132) (Ping timeout)
- # [05:18] * Joins: beowulf (carisenda@91.84.50.132)
- # [05:21] * Quits: zdenko (zdenko@84.255.203.169) (Quit: zdenko)
- # [05:44] * Joins: Shunsuke (Shunsuke@219.110.80.235)
- # [05:45] * Quits: jdandrea (jdandrea@68.192.161.254) (Quit: jdandrea)
- # [05:45] * Quits: gavin_ (gavin@74.103.208.221) (Ping timeout)
- # [05:50] * Joins: gavin_ (gavin@74.103.208.221)
- # [06:00] * Joins: hyatt (hyatt@24.6.91.161)
- # [06:22] * Joins: schepers (schepers@208.38.57.47)
- # [06:28] * Joins: htmlr (htmlr@138.130.176.60)
- # [06:58] * Quits: schepers (schepers@208.38.57.47) (Quit: Free at last!)
- # [07:13] * Quits: Shunsuke (Shunsuke@219.110.80.235) (Ping timeout)
- # [07:23] * Joins: heycam (cam@203.214.12.71)
- # [08:11] * Quits: myakura (myakura@125.194.247.196) (Quit: Leaving...)
- # [08:12] * Joins: Shunsuke (Shunsuke@219.110.80.235)
- # [08:15] * Joins: zcorpan_ (zcorpan@217.211.77.236)
- # [08:16] * Joins: Lachy_ (Lachlan@203.214.143.196)
- # [08:18] * Quits: Lachy (Lachlan@124.168.27.56) (Ping timeout)
- # [08:50] * Quits: gavin_ (gavin@74.103.208.221) (Ping timeout)
- # [08:55] * Joins: gavin_ (gavin@74.103.208.221)
- # [09:41] * Joins: ROBOd (robod@86.34.246.154)
- # [10:27] * Quits: hyatt (hyatt@24.6.91.161) (Quit: hyatt)
- # [10:49] <anne> "Dean Jackson has joined the HTML Working Group sysbot+ipp@w3.org" yay!
- # [10:50] * Quits: zcorpan_ (zcorpan@217.211.77.236) (Ping timeout)
- # [10:52] <mjs> as an individual?
- # [10:53] * Quits: ROBOd (robod@86.34.246.154) (Quit: http://www.robodesign.ro )
- # [10:53] <mjs> looks like
- # [10:55] <anne> much better for him than Team contact I suppose. he can now actively advocate to break the web :)
- # [10:56] * anne reads through www-archive
- # [10:56] <mjs> do you think he would want to advocate that?
- # [10:57] <mjs> are you reading my little side discussion with John Boyer?
- # [10:57] <mjs> I'm waiting to hear what he thinks constitutes a compromise
- # [10:57] <anne> he would as a joke
- # [10:57] <anne> yeah
- # [10:57] * Quits: gavin_ (gavin@74.103.208.221) (Ping timeout)
- # [11:01] <mjs> I wasn't sure what to say to Gregory Rosmaita
- # [11:02] * Joins: gavin_ (gavin@74.103.208.221)
- # [11:03] <mjs> looks like Microsoft didn't vote by the time the survey closed
- # [11:03] <mjs> not even to abstain
- # [11:03] <anne> http://www.w3.org/2004/01/pp-impl/32107/instructions ?
- # [11:04] <mjs> lol
- # [11:05] * Quits: inimino (chatzilla@75.71.88.233) (Client exited)
- # [11:09] * Joins: inimino (chatzilla@75.71.88.233)
- # [11:10] <mjs> hmm, should I suggest that the HTML WG group page be changed from pseudo-XHTML to HTML?
- # [11:10] <mjs> or would that just start a useless flamewar?
- # [11:10] <anne> you could ask DanC I suppose
- # [11:11] <anne> Bjoern ranting about WHATWG: http://www.w3.org/mid/d7sq33hdh2cqqs576jr3516d33albui2p2@hive.bjoern.hoehrmann.de
- # [11:11] <anne> Although it seems to focus mostly on class names, which is just an idea and not really finished or anything
- # [11:12] <mjs> if he cannot tell that "<div class=copyright>" is not allowed, then he must not be a very good reader
- # [11:12] <mjs> "It must only be used on the following elements: p, span"
- # [11:12] <mjs> I think "must" is pretty clear
- # [11:12] * Joins: myakura (myakura@125.194.247.196)
- # [11:13] <mjs> also no additional content model restrictions are added
- # [11:13] <mjs> personally I think predefined classes should be allowed on as many elements as possible rather than as few as possible
- # [11:13] <mjs> the lists right now are way too restrictive
- # [11:13] <anne> I mostly want to know what class="copyright warning issue" means
- # [11:15] <mjs> it implies that your copyright notice is in dispute and is a warning or admonition
- # [11:15] <mjs> which seems like an odd combination
- # [11:15] <mjs> I think the predefined class names are weirdly biased to the ones a standards document would want to use
- # [11:17] <anne> much is biased towards Hixie's pov because there hasn't been much input yet
- # [11:17] <mjs> Bjoern has a really weird stance about web standards
- # [11:17] <mjs> here he is being picky about precision of conformance requirements that actually are stated pretty directly
- # [11:18] <mjs> yet for DOM 3 Events, he was unwilling to rewrite things to actually state conformance requirements
- # [11:18] <mjs> as for example proposed by me
- # [11:18] <mjs> I don't really understand his point of view on such things
- # [11:18] <anne> i think he believes that DOM 3 Events is already clear on that
- # [11:21] * Joins: hyatt (hyatt@24.6.91.161)
- # [11:21] <mjs> he seems to think it is ok for most conformance requirements not to use RFC2119 keywords
- # [11:21] <mjs> he doesn't even reference RFC2119
- # [11:21] <mjs> I don't know how he can think that matches current best practices for stating conformance requirements
- # [11:21] <mjs> hello hyatt
- # [11:22] <hyatt> hi
- # [11:22] <anne> his argument was that the conformance criteria section says when an implementation is conforming
- # [11:22] <anne> iirc
- # [11:22] <anne> but i agree that the dom specs are not very clear
- # [11:23] <mjs> if he defined what it means to "support an interface and its associated semantics" that might be an argument
- # [11:24] <mjs> and certainly the language binding sections are broken
- # [11:24] <mjs> especially the definition of EventListener function
- # [11:25] <mjs> which is one of the very first issues I raised and which he has ignored the whole time
- # [11:25] <hyatt> anne: you should tell hakon that i got @font-face working
- # [11:25] <hyatt> anne: in webkit
- # [11:25] <hyatt> anne: i haven;'t landed it yet though
- # [11:25] <hyatt> anne: but i know downloadable fonts are an area he feels strongly about :)
- # [11:25] * Quits: anne (annevk@213.236.208.22) (Connection reset by peer)
- # [11:26] * Joins: anne (annevk@213.236.208.22)
- # [11:29] <hyatt> i drove anne away
- # [11:29] <mjs> he came back
- # [11:29] <mjs> anne: so why is http-equiv Content-Type missing?
- # [11:30] <mjs> it's a pretty common way to indicate the charset
- # [11:30] <anne> it was replaced with <meta charset=utf-8>
- # [11:30] <mjs> it should be there at least as a UA requirement
- # [11:30] <anne> that's covered, yes
- # [11:30] <anne> in the parsing section
- # [11:30] <anne> although the encoding detection algorithm is currently fairly restrictive and might not work
- # [11:31] <anne> (it requires to only seek within the first 512 bytes)
- # [11:31] <mjs> is it?
- # [11:31] <mjs> the only values I see defined for http-equiv are refresh and default-style
- # [11:31] <mjs> http://www.whatwg.org/specs/web-apps/current-work/#http-equiv0
- # [11:31] <anne> hyatt, i'll tell him :)
- # [11:32] <anne> mjs, the UA criteria are in the parsing section
- # [11:32] <anne> mjs, specifically, the input stream subsection
- # [11:32] <anne> http://www.whatwg.org/specs/web-apps/current-work/#the-input0
- # [11:33] <mjs> I was just looking there and I don't see mention of http-equiv in that section
- # [11:34] <anne> "If it is 'content'"
- # [11:35] <anne> It doesn't actually check the http-equiv attribute. Just the content= and charset= attributes on meta elements
- # [11:35] <mjs> ah
- # [11:36] <mjs> does that actually match what current browsers do? will they respect a content= that appears to be a MIME type with a charset parameter even if there is no http-equiv=?
- # [11:36] <anne> ask Hixie, I was not that closely involved
- # [11:36] <anne> the main thing that might be problematic is that it requires UAs to only sniff within the first 512 bytes where some UAs just sniff like forever
- # [11:37] <mjs> certainly that part of the parsing spec could use an informative note
- # [11:37] <mjs> and I agree, the 512 limit probably won't work in the real world
- # [11:37] <mjs> some UAs will reparse even if the <meta> is all the way at the end of the document, I think
- # [11:37] <anne> Firefox
- # [11:38] <anne> The question is of course what content requires...
- # [11:38] <anne> although current implementations might be a good indicator :)
- # [11:38] <anne> s/might be/are/
- # [11:39] <mjs> we found that having a small limit did not cut it for some real cases
- # [11:39] <mjs> olliej recently made changes to this area
- # [11:39] <anne> iirc the spec also says to use the first charset you find
- # [11:40] <anne> which is not what some impl do
- # [11:42] * anne has to leave for a few secs again
- # [11:42] * Parts: anne (annevk@213.236.208.22)
- # [11:43] * Joins: anne (annevk@213.236.208.22)
- # [11:44] * Quits: anne (annevk@213.236.208.22) (Connection reset by peer)
- # [11:45] * Joins: anne (annevk@213.236.208.22)
- # [11:45] <gavin_> hrm, my inbox and http://lists.w3.org/Archives/Public/www-html/2007May/ don't agree
- # [11:46] <gavin_> I suppose that may have something to do with http://lists.w3.org/Archives/Public/www-html/2007May/ having the same days split up for some reason?
- # [11:47] <mjs> we really shouldn't have a voting process that automatically makes any no vote a Formal Objection
- # [11:47] <mjs> most WG's I've worked with give you much more ability to note disagreement without pulling out the big FO stick
- # [11:47] <anne> in that case more people might have voted no?
- # [11:47] * anne ponders
- # [11:48] <mjs> no, some of the people who voted no might not have chosen to escalate their disagreement to a Formal Objection that must be reviewed by the Director personally if the question carries
- # [11:48] * Quits: Shunsuke (Shunsuke@219.110.80.235) (Ping timeout)
- # [11:49] <mjs> like the vote of "no" for the name HTML5 that suggests HTML5.01 instead
- # [11:49] <mjs> I remember in other groups there are often straw poll votes where if there is rough consensus apparent, you then ask the dissentors if they would like to register their objection
- # [11:50] * anne likes (X)HTML5 as name
- # [11:50] <mjs> similarly, giving a comment on a spec, even if you disagree, does not automatically make it an FO as far as I can tell
- # [11:50] <anne> btw, will it be HTML 5 or HTML5?
- # [11:50] <anne> oh well
- # [11:50] <anne> mjs, yeah
- # [11:50] <anne> most voting processes in other WGs I've dealt with have yes / no
- # [11:51] <anne> "majority wins"
- # [11:51] <anne> if you really disagree you can raise a FO during LC or something
- # [11:53] <hyatt> it should be called XFOHTMLSVGUL23
- # [11:53] <hyatt> otherwise. formal objection.
- # [11:53] <hyatt> from me.
- # [11:53] <sbuluf> gavin_, might, yes. i said here a couple of times the archives vieww was broken, but it is still like that
- # [12:00] <anne> mjs, so what's the current proposal for forms? Set up a task force that defines requirements for the web forms technology and require the specs to address those?
- # [12:01] <anne> I think that would work for me
- # [12:01] <mjs> anne: that's what I proposed
- # [12:02] <mjs> anne: John Boyer is presumably opposed but has not really addressed my proposal directly
- # [12:02] <mjs> other than at first mistakenly claiming that it violated our charter and the vision document
- # [12:02] <anne> argh
- # [12:03] <mjs> after I pointed out he was wrong on that, he proposed that I take a "chill pill"
- # [12:03] <anne> citing charters is a very bad way to make progress (though I just did it :))
- # [12:03] <mjs> I'm not sure if that is a controlled substance or what
- # [12:03] <anne> especially when groups are in disagreement over what the charter actually "implies"
- # [12:03] <mjs> btw http://lists.w3.org/Archives/Public/www-archive/2007May/0037.html
- # [12:03] <mjs> I think what I proposed is sensible independent of the charters
- # [12:04] <mjs> it will lead to better cooperation / alignment without putting groups in a situation where they are likely to fight a lot
- # [12:04] <mjs> and without the weird situation of one group co-editing another group's spec, which afaict is not normally done
- # [12:06] * Quits: hyatt (hyatt@24.6.91.161) (Quit: hyatt)
- # [12:06] <anne> one of the things John Boyer seems to want is automatic conversion from HTML forms to XForms and maybe vice versa
- # [12:06] <anne> It's not clear to me how that works with scripting
- # [12:07] <anne> Halting problem and all that
- # [12:07] <mjs> unless both include identical scripting APIs and identical event names
- # [12:07] <mjs> though to be compatible with both existing HTML Forms and existing XForms, that would require duplication of many things
- # [12:07] <mjs> so I assume that's not indended
- # [12:08] <mjs> XForms that don't use script could probably be translated to HTML
- # [12:08] <mjs> btw when we say "halting problem" it would probably be more appropriate to cite Rice's Theorem
- # [12:08] <mjs> http://en.wikipedia.org/wiki/Rice's_theorem
- # [12:08] <mjs> which is more general
- # [12:09] <hsivonen> anne: XForms theory starts falling apart anyway once you introduce imperative scripting, so perhaps it would be enough to only cover scriptless cases. :-)
- # [12:10] <mjs> hsivonen: but... but... John Boyer said the strength of XForms is that it supports both imperative and declarative models!
- # [12:11] * Joins: hyatt (hyatt@24.6.91.161)
- # [12:11] * Joins: hasather (hasather@81.235.209.174)
- # [12:12] <anne> mjs, ah ok, so "halting problem" is only whether a script stops running or not and "Rice's theorem" is about it not being possible to tell what a script will do?
- # [12:14] <mjs> anne: well, the Halting Problem indirectly proves that you couldn't determine many kinds of properties of scripts, by the method of reduction
- # [12:14] <mjs> if determining a particular script property could be used to solve the halting problem, then it is not decidable
- # [12:14] <mjs> but Rice's Theorem is already more general
- # [12:16] * mjs highly recommends taking a class on complexity and computability theory if your university offers one
- # [12:16] * Quits: inimino (chatzilla@75.71.88.233) (Quit: ChatZilla 0.9.78.1 [Firefox 2.0.0.3/2007031001])
- # [12:16] <mjs> not that you use that stuff much in everyday life, but it will make you feel smarter
- # [12:17] <anne> heh
- # [14:17] * Disconnected
- # [14:17] * Attempting to rejoin channel #html-wg
- # [14:17] * Rejoined channel #html-wg
- # [14:17] * Topic is 'HTML WG http://www.w3.org/html/wg/ logged: http://krijnhoetmer.nl/irc-logs/'
- # [14:17] * Set by Zeros on Mon Apr 30 23:38:28
- # [16:22] * Disconnected
- # [16:22] * Attempting to rejoin channel #html-wg
- # [16:22] * Rejoined channel #html-wg
- # [16:22] * Topic is 'HTML WG http://www.w3.org/html/wg/ logged: http://krijnhoetmer.nl/irc-logs/'
- # [16:22] * Set by Zeros on Mon Apr 30 23:38:28
- # [16:44] * Quits: htmlr (htmlr@138.130.176.60) (Quit: htmlr)
- # [17:07] <Lachy> does anyone understand what Philip is talking about and what relevance it has to HTML http://lists.w3.org/Archives/Public/public-html/2007May/0699.html
- # [17:08] <Lachy> oh my gosh, 705 messages in just 6 days! At this rate, we'll be well over 3000 this month. :-/
- # [17:13] <Philip`> http://canvex.lazyilluminati.com/misc/copyright.html - how some sites are using the class 'copyright'
- # [17:14] <Philip`> based on a list of the top n Yahoo search results for the word "the", where n is however many I'd got around to downloading when I generated that file
- # [17:14] <Lachy> were they all used for copyright notices?
- # [17:15] <Lachy> I couldn't see any that didn't after I quickly scanned the results
- # [17:15] <Philip`> Ah, it's about n=300 (with ~30 copyrights over ~20 sites)
- # [17:16] <Lachy> you should also look for pages that use id="copyright" and values that are not exact matches
- # [17:16] <Philip`> I'm not sure, I haven't actually read the list :-)
- # [17:16] <Lachy> like, for instance, class="copyrightline" and "footCopyright" were 2 that I found
- # [17:16] <Lachy> (see my e-mail to www-html for my results)
- # [17:36] * Lachy thinks the whole debate about class=copyright is becoming a bikeshed
- # [17:37] <Philip`> id=copyright seems to be about a third as frequent as class=copyright
- # [17:37] <Philip`> It takes a surprisingly long time to parse 500 sites :-(
- # [17:40] <Philip`> See http://canvex.lazyilluminati.com/misc/copyright.html again, now with lots of iframes
- # [17:40] <Philip`> and summary table at the bottom
- # [17:40] <Philip`> (42 class=copyright, 14 id=copyright, and only 1 or 2 each of anything else)
- # [17:40] <Lachy> were you using html5lib for the survey?
- # [17:40] <Philip`> Yes
- # [17:41] <Philip`> (The current SVN version, in particular)
- # [17:41] <Lachy> cool. Are you going to post that to the list?
- # [17:41] <Philip`> I think that suggests that if class=copyright is handled specially, id=copyright probably should be too
- # [17:42] <Lachy> I think we should just stick with predefined class values
- # [17:43] <Lachy> defining special id values would just open an even bigger can of worms
- # [17:43] <Philip`> I'm not particularly intending to, since it takes too much work trying to find a relevant place to reply and to fix the archive's reply mailto links and I really ought to be working on other stuff instead of this :-)
- # [17:44] <Philip`> but anyone else could feel free to post a link if they want, since that'd let me be lazy ;-)
- # [17:44] <Lachy> I'll post it
- # [17:44] <Philip`> (though this is obviously far from a perfect survey)
- # [17:45] <Philip`> (but it's probably better than just choosing a few arbitrary sites to look at by hand)
- # [17:48] * Philip` wonders why the character encoding got messed up for w3.org but not for any other site
- # [17:50] <Philip`> (Oh, maybe because they set it in a content-type header but I didn't bother passing that into html5lib)
- # [17:57] <Philip`> Oh, and some of the bits in the iframe are broken because I serialised to XML but it's sent as text/html, but I don't think it matters enough to bother fixing
- # [18:42] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
- # [19:04] * Joins: schepers (schepers@208.38.57.47)
- # [19:06] * Joins: zcorpan_ (zcorpan@217.211.77.236)
- # [19:08] * Quits: gavin_ (gavin@74.103.208.221) (Ping timeout)
- # [19:13] * Joins: gavin_ (gavin@74.103.208.221)
- # [19:18] <Philip`> I think the issues with copyright depend largely on the use cases for extracting and processing the data, and the balance between the cost of false positives and the value of true positives - without knowing that, it seems unreasonable to argue that class="copyright" is either good or bad. (Or has someone already suggested uses that I've missed?)
- # [19:24] <xover> I would imagine the main concern is that it violates the “contract” made by previous specifications that class='…' imparts no particular meaning above providing an identification string to hang, e.g., CSS off.
- # [19:26] <gsnedders> @role (in the XHTML2 WD) will have similar issues if any future revisions to the spec want to add values in the global namespace
- # [19:27] <Philip`> HTML4 says it's for "For general purpose processing by user agents" in addition to being a style hook, which I'd interpret as saying it can be used for whatever arbitrary things the UA wants
- # [19:28] <gsnedders> @profile directly references it, does it not?
- # [19:28] <Philip`> (though it'd be valid to ignore what HTML4 says and only consider the implied agreement between real UAs and authors, that class is just used for styles)
- # [19:28] <xover> That may be a point for disagreement then. Worth exploring I think.
- # [19:29] <Philip`> FF3's microformat support would also be a non-styling use of class
- # [19:31] <gsnedders> and that won't use @profile, as too many documents with µf don't use the @profile
- # [19:39] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Ping timeout)
- # [19:50] * Quits: zcorpan_ (zcorpan@217.211.77.236) (Ping timeout)
- # [20:04] <Bob_le_Pointu> \o/
- # [20:04] <Bob_le_Pointu> Hail to the new French president.
- # [20:27] <Lachy> I can't believe both Terge and Jukka did a falty analysis of those class=copryight stats!
- # [20:27] <Lachy> s/Terge/Terje/
- # [20:35] <xover> Lachy: I honestly think you need to step back and reconsider a little.
- # [20:35] <xover> From where I'm sitting you seem to have gotten so caught up in “winning” that you're no longer objectively evaluating the facts.
- # [20:36] <Lachy> xover, what?
- # [20:36] <Lachy> I am looking at the facts. See my last email!@
- # [20:36] <xover> I will happily accept if we just end up disagreeing on something, but I'm left with the impression that that's not what's going on here.
- # [20:36] <xover> It's that last email I'm talking about (among others).
- # [20:36] <sbuluf> class, afaik, is unrelated to styling, at all. it is a general purpose specification mechanism.
- # [20:36] <xover> This isn't an accusation btw.
- # [20:37] <Lachy> precisely what was wrong with my analysis?
- # [20:37] <Lachy> I looked at the evidence closely, presented statistics and reached a conclusion
- # [20:37] <xover> I'm just... That's the impression I get right now, so I'm suggesting you try to mull it over a bit.
- # [20:37] <sbuluf> which is the the exact reason i kept it in my off topic html replcamente, which should not be discussed here
- # [20:37] <sbuluf> *replacement
- # [20:38] <Lachy> xover, it would help if you would express yourself more clearly
- # [20:38] <xover> My apologies. I strive for clarity, but I seem to do quite badly at it.
- # [20:39] <Lachy> what is it about presenting a statistical analysis that gives you that impression?
- # [20:39] <xover> That I do not draw the same conclusions from the same statistical data that you do, to a remarkable degree.
- # [20:40] <xover> And neither, apparently, does Jukka.
- # [20:40] <Lachy> how does 54/58 sites using it for copyright info let you reach a different conclusion?
- # [20:40] <xover> Which of course may be an indication that we are both wrong.
- # [20:41] <Lachy> 2 out of the remaing 4 were clearly just a mistake, their intention was clear from looking at the site
- # [20:41] <Lachy> the remaining 2 were elimiated because they didn't used a variation of the class
- # [20:42] * Quits: MrNaz (Naz@203.214.95.166) (Ping timeout)
- # [20:42] <Lachy> it appears that Jukka actually didn't investigate it thoroughly enough
- # [20:42] <Philip`> Are you both considering "using it for copyright info" in the same way?
- # [20:42] * Joins: MrNaz (Naz@203.214.95.166)
- # [20:42] <Philip`> e.g. if something contains copyright info plus other stuff, does that count?
- # [20:43] <Lachy> I counted it, yes
- # [20:43] <Philip`> That's where I think it depends on what the data's going to be used for, and whether false positives (or half-false positives) are a problem or not
- # [20:44] <Lachy> I wonder if we did a study on the use of <em>, whether we'd actually get stats that showed it was abused far more than class="copyright" appears to be, relatively speaking
- # [20:44] <jgraham> Ironically very little existing "semantic" markup on the web is actually used for semantics by general purpose UAs anyway
- # [20:45] <jgraham> Lachy: Exactly
- # [20:45] <Lachy> with HTML, you have to be a little bit lenient. You cannot possibly draw a hard line on the exact semantics of an element.
- # [20:46] <Lachy> cf. the discussion I had with tantek about the correct use of <p> a week or so ago
- # [20:46] <jgraham> Lachy: Who was arguing what about <p>?
- # [20:46] <Lachy> tantek
- # [20:47] <Philip`> (Incidentally, the empty <p class="copyright"></p> was caused by the page writing <p class="copyright"><p>copyright stuff</p></p> so at least they were trying to get it right)
- # [20:47] <Lachy> was arguing that it should only be used for paragraphs, based on a strict dictionary defintion of the term
- # [20:47] * jgraham realises the last sentence of his latest email makes no sense
- # [20:47] <Lachy> Philip`, yes, I mentioned that in my last email
- # [20:48] <Lachy> and I explicitly counted that, and another that made a simlar mistake, separately
- # [20:48] <jgraham> Lachy: Well the HTML4 spec uses it as a container for form elements and authors use it for almost anything (I guess) so neither prescriptivism or descriptivism seem to support that view
- # [20:49] <Lachy> yeah, I know. I agree with what HTML5 says about paragraphs, tantek was complaining bitterly
- # [20:50] <Philip`> Lachy: Ah, okay, don't think I read that
- # [20:51] <jgraham> Unless you consider the OED applied to normative text more authoritative than the non-normative examples I suppose
- # [20:51] <Lachy> jgraham, thanks for responding to Jukka. I couldn't think of how to address that statement of his politely
- # [20:52] * Philip` parallelises his markup-searching program, but sadly it doesn't seem to go much faster than before
- # [20:52] <jgraham> Philip`: It's using html5lib? Have you tried running it with psyco?
- # [20:54] <Philip`> jgraham: It is; I haven't
- # [20:54] * Philip` tries that
- # [20:55] <jgraham> Philip`: In my entirely "gut feeling" tests it seemed to speed things up noticeably
- # [20:56] <jgraham> It certianly ought to work well, if I understand what it is supposed to do
- # [21:01] <Philip`> I think that's looking faster now, though I have no way to actually measure, and it's blocked on network traffic quite a lot
- # [21:01] <Philip`> But I just found a site that causes a NotImplementedError...
- # [21:01] <Lachy> I just realised I did make a mistake in my analysis, oh well
- # [21:02] <Lachy> I should have only counted the sites that used teh exact value class="copyright", of which there are 27 unique sites
- # [21:02] * Quits: myakura (myakura@125.194.247.196) (Quit: Leaving...)
- # [21:02] * Quits: MrNaz (Naz@203.214.95.166) (Quit: Leaving)
- # [21:04] <Philip`> Aha, that page used <footer>, which kills html5lib
- # [21:08] <Philip`> (See http://www.classesusa.com/featuredschools/programs/featured_health.cap for <footer>, plus exciting comments like "<!-- removed depreciated font, b and div not closed -->")
- # [21:15] * Quits: gavin_ (gavin@74.103.208.221) (Ping timeout)
- # [21:19] * Quits: Shunsuke (Shunsuke@219.110.80.235) (Ping timeout)
- # [21:20] * Joins: gavin_ (gavin@74.103.208.221)
- # [21:25] * Quits: mw22 (chatzilla@213.51.191.252) (Ping timeout)
- # [21:28] * Joins: MrNaz (Naz@203.214.95.166)
- # [21:38] <jgraham> Philip`: So if we find a site that uses <footer> are we blocked from putting it in the spec? ;)
- # [21:40] * Joins: inimino (chatzilla@75.71.88.233)
- # [21:42] <Philip`> From looking at 2521 pages, <footer> is 134 times less common than class="copyright", and that one use is kind of compatible with the HTML5 definition anyway, so I hope not many people would complain about that :-)
- # [21:43] * Parts: schepers (schepers@208.38.57.47)
- # [21:43] * Quits: MrNaz (Naz@203.214.95.166) (Ping timeout)
- # [21:44] <jgraham> Ah but in principle someone might use it when they're talking about "[the] difficulty of a particular song in the video game Dance Dance Revolution" (http://en.wikipedia.org/wiki/Footer). Can't be too careful, you understand ;)
- # [21:49] <Philip`> I suppose you'll get exactly the same problem with <header> on football sites
- # [22:11] * Quits: ROBOd (robod@86.34.246.154) (Quit: http://www.robodesign.ro )
- # [22:19] * Joins: zcorpan_ (zcorpan@84.216.42.227)
- # [22:39] * Joins: schepers (schepers@72.29.239.177)
- # [22:59] * Joins: edas (edaspet@88.191.34.123)
- # [23:01] * Quits: zcorpan_ (zcorpan@84.216.42.227) (Ping timeout)
- # [23:01] <Hixie> i don't understand why bjoern, whom i have frequently asked for feedback on the html5 spec, feels the need to be so dismissive about the html5 spec, instead of giving the same feedback in a constructive fashion
- # [23:02] * Joins: zcorpan_ (zcorpan@84.216.42.227)
- # [23:06] * Joins: hyatt (hyatt@24.6.91.161)
- # [23:07] * Joins: mw22 (chatzilla@84.41.169.151)
- # [23:11] * Joins: asbjornu (asbjorn@84.48.116.134)
- # [23:19] * Quits: schepers (schepers@72.29.239.177) (Ping timeout)
- # [23:20] <gsnedders> Hixie: because XHTML2 is the future!
- # [23:24] * Joins: schepers (schepers@72.29.239.177)
- # [23:31] <hsivonen> from the discussions one might think that ship names are a common use case
- # [23:39] <Philip`> jgraham: html5lib seems to take roughly quadratic time to parse when you have a very long sequence of number entities (like http://www.archive.org/details/DuckandC1951) - is that a feature or a known bug or something unexpected?
- # [23:39] * Joins: cying (cying@75.41.252.252)
- # [23:48] * Quits: schepers (schepers@72.29.239.177) (Connection reset by peer)
- # [23:50] <Dashiva> I wonder if the whole i/em emphasis debate will lead to anything. Even if it does, it won't change how they're being used right now as we debate...
- # [23:50] <Philip`> (Uh, I think I mean more like exponential than quadratic. In any case, it's really slow - ~24 seconds for 600KB, ~44 seconds for 800KB, etc)
- # Session Close: Mon May 07 00:00:00 2007
The end :)