/irc-logs / w3c / #html-wg / 2007-04-13 / end

Options:

# Session Start: Fri Apr 13 00:00:00 2007
# Session Ident: #html-wg
# [00:02] * Joins: karl (karlcow@128.30.52.30)
# [00:12] * Quits: heycam (cam@203.214.79.176) (Ping timeout)
# [00:13] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Quit: Get thee behind me, satan.)
# [00:43] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
# [00:57] * Joins: heycam (cam@130.194.72.84)
# [01:01] * Joins: marcos__ (chatzilla@203.206.31.102)
# [01:01] * marcos__ is now known as marcos
# [01:05] * Joins: DanC_lap (connolly@128.30.52.30)
# [01:08] * Joins: sbuluf (kzxdkz@200.49.140.77)
# [01:27] * Quits: Zeros (Zeros-Elip@67.154.87.254) (Quit: Leaving)
# [01:32] * Joins: Shunsuke (kuruma@219.110.80.235)
# [01:36] * Quits: Shunsuke (kuruma@219.110.80.235) (Quit: shutdown)
# [01:39] * Parts: hasather (hasather@81.235.209.174)
# [01:40] * Quits: DanC_lap (connolly@128.30.52.30) (Ping timeout)
# [01:44] * Quits: Philip (excors@80.177.163.133) (Quit: Philip)
# [01:47] * Joins: Mazzie (cucme@194.109.21.4)
# [01:52] * Quits: Voluminous (Voluminous@66.195.32.2) (Quit: Leaving)
# [02:23] <mjs> this versioning essay is pretty long
# [02:24] <sbuluf> the chris wilson one? http://lists.w3.org/Archives/Public/public-html/2007Apr/0612.html ?
# [02:24] <mjs> yes
# [02:24] <sbuluf> i agree, i'm still halfway through it
# [02:26] <Lachy> that essay is nonsese! I just finished reading it
# [02:26] <mjs> finally got through it
# [02:26] <sbuluf> how would you summarize it, lachy?
# [02:27] <mjs> I'm not sure he actually answered my question that partially triggered it
# [02:27] <mjs> the answer is sort of implied, but I would like him to state it clearly
# [02:27] <Lachy> sbuluf, politely or honestly?
# [02:27] <sbuluf> honestly
# [02:27] <Lachy> then his suggestions are BS!
# [02:27] <sbuluf> (this is logged, so beware..PM if you wish)
# [02:27] <sbuluf> oops, late
# [02:28] <Lachy> like his suggestion for using <!DOCTYPE html5> won't work because that triggers quirks mode
# [02:28] <Dashiva> He suggests we version the spec, when what he really argues for is versioning pages to a IE version
# [02:29] <Lachy> and his suggestion to upgrade that to <!DOCTYPE html6>, <!DOCTYPE html7> and into infinity totally misses the point of Don't Break the Web
# [02:29] <Dashiva> There have been several suggestions like <meta ie-version-support="7"> already, which should suit his needs just fine
# [02:29] <Lachy> Dashiva, those suggestions are equally silly
# [02:30] <mjs> I can see how spec versioning would be convenient for his IE versioning plan, but that doesn't seem like a sufficient justification in itself for versioning
# [02:30] <Dashiva> But they at the very least attempt to solve the right problem
# [02:30] <mjs> I'll have to read the replies and think about it
# [02:30] <Lachy> must we be doomed to repeat the mistakes of browser sniffing?
# [02:31] <Lachy> his example of the child selector html>body breaking the web when they implemented it is based on the fact that they fixed the filter, but failed to fix the significant limitations that relied on it
# [02:31] <Lachy> the same applies to * html
# [02:35] <zcorpan> Lachy: authors shouldn't rely on one bug in order to fix another, as it might break in browser upgrades. but they do. ms cannot fix all bugs at the same time. it is problematic but it's not a reason to introduce another rendering mode
# [02:35] <zcorpan> the real problem is that microsoft stopped developing ie for a few years
# [02:36] <Dashiva> They're willing to ship bugs, but not fixes. A cruel process for us all.
# [02:37] <Lachy> zcorpan, they didn't even fix all the major well documented limitations, which is the problem...
# [02:37] <zcorpan> yeah
# [02:38] <edas> Dashiva: <meta ie-version-support="7"> does not solve the problem. This problem is not IE-centric. Mozilla, Opera and Safari have bugs. All of them have or will have a "big bug" that will break the web if corrected. They will have to define a similar mecanism for opt-in (they all have already done it one time with doctype switching, they will do many time more in the future).
# [02:38] <edas> We shouldn't have a proprietary ie-specific mecanism
# [02:38] <Dashiva> edas: Mozilla, Opera, Safari have all demonstrated will to fix bugs
# [02:38] <Lachy> they need to completely remove hasLayout from their standards mode, it's nonsense. They need to fix the bugs that made use of: * html { height: 1%; }, which they didn't (yet they fixed * html)
# [02:39] <zcorpan> edas: the other browser vendors don't want to introduce new rendering modes
# [02:40] <edas> zcorpan, what they will do if some correction breaks the web (and not only a dozen of pages) ?
# [02:40] <zcorpan> edas: change the spec
# [02:40] <edas> 1- leave that bug ; 2- break the web ; 3- require an opt-in mecanism
# [02:40] <Lachy> I know you can't always rely on them fixing both bugs (the one that let the filter work and the limitation), but the major, well documented ones should have been fixed
# [02:40] <zcorpan> 1
# [02:40] <edas> zcorpan, well change the spec is possible if all have the same bugs
# [02:41] <edas> zcorpan, if Mozilla has an Opera doesn't, change the spec will introduce the opposite bug in Opera
# [02:41] <zcorpan> edas: that's what we're aiming for (if all agree on the same set of bugs, we have interop)
# [02:42] <zcorpan> edas: content generally doesn't rely on things where all browsers disagree
# [02:42] <Lachy> if one browser does one thing and another does the opposite, then we standardise that one that is most compatible with the web.
# [02:42] <edas> zcorpan, you're saying that what will be in the future will correspond exactly to what they want (is to say they won't do bug withou willing it)
# [02:42] <Dashiva> The whole SGML business showed some browsers are willing to break for fixes, even when they have no benefit whatsoever
# [02:42] <mjs> edas: representatives from Mozilla, Opera and Apple have all said that we don't want a switch for opt-in to web-breaking standards
# [02:42] <Dashiva> *SGML comment
# [02:43] <mjs> edas: so I think you are out on a limb saying we will want such a mechanism in the future
# [02:43] <mjs> edas: our preference is to fix the spec if something it requires would BTW
# [02:43] <edas> mjs, I understand that, but I don't see what they will do if they are in front of a "oups, we did somehting incompatible and it's used by many authors"
# [02:44] <Dashiva> It doesn't get that far
# [02:44] <zcorpan> Dashiva: the intended benefit was increased interop. although there was a different way to achieve interop we found out later
# [02:44] <mjs> edas: most sites basically have one mode for IE and another for Firefox+Opera+Safari
# [02:44] <Dashiva> Relevant intercompatabilities are noticed and reported
# [02:44] <mjs> edas: since we are already closer to each other, it's less likely for browser-specific bugs to get widely locked in
# [02:44] <edas> you have faith in future, I don't ;)
# [02:44] <mjs> anyway, most of these web breakage issues come up for CSS and DOM
# [02:45] <mjs> using an HTML version to trigger CSS and DOM changes seems like bad layering
# [02:45] * zcorpan would like to get interop with ie on css and dom, even if that meant changing the css and dom specs
# [02:45] <edas> I agree
# [02:45] <mjs> I have never heard an example of a site depending on a browser-specific *HTML* bug (as opposed to DOM, JS or CSS)
# [02:46] <Dashiva> Conditional comments ;)
# [02:46] <edas> you may be right with this on
# [02:47] * zcorpan should study the differences between almost standards mode and full standards mode a bit more, then propose to the css wg to spec almost standards mode so full can be dropped
# [02:47] <mjs> not saying it hasn't happened, but it seems less common, and I think there are structural reasons why that is so
# [02:48] <mjs> zcorpan: I think it would probably be good to have almost standards mode be the standard model
# [02:48] * Quits: kingryan (rking3@66.92.187.33) (Quit: kingryan)
# [02:48] <zcorpan> indeed
# [02:49] <Dashiva> As long as quirks mode is the default, every new document will be quirks mode if the author doesn't take care...
# [02:49] <zcorpan> Dashiva: indeed. new documents will be produced that rely on quirks mode
# [02:50] <zcorpan> today, i think 50% or so doesn't have a doctype at all, and 90% or so use quirks mode
# [02:51] <zcorpan> the rest use almost standards mode with only very few using full standards mode
# [02:51] <zcorpan> according to the studies i've seen anyway
# [02:51] * Joins: marcos_ (chatzilla@131.181.148.226)
# [02:52] <zcorpan> perhaps css 2.1 should spec what is required for quirks mode too
# [02:52] <edas> wen I think about it I'm just scared about having multiple proprietary versionning mecanisms which aim specific versions of specific browsers, having to explicitly opt-in a version for each browser, and in consequence to always aim a series of version/brower and never the spec. (I know, I'm pessimist, I had bad experiences with all CSS hacks, doctype switching and conditionnal comments)
# [02:54] <Dashiva> The future we're offering contains a single <!DOCTYPE html> and eternal bliss ;)
# [02:58] <zcorpan> nn
# [03:03] * Parts: zcorpan (zcorpan@84.216.43.114)
# [03:04] * Quits: Mazzie (cucme@194.109.21.4) (Quit: My damn controlling terminal disappeared!)
# [03:05] * Quits: edas (edaspet@88.191.34.123) (Quit: http://eric.daspet.name/ et l'édition 2007 de http://www.paris-web.fr/ )
# [03:05] * Quits: adele (adele@67.170.236.225) (Quit: adele)
# [03:06] * Parts: asbjornu (asbjorn@84.48.116.134)
# [03:09] * Quits: mjs (mjs@17.255.97.103) (Quit: mjs)
# [03:27] * Joins: DougJ (djones4@74.76.23.86)
# [03:41] * Joins: adele (adele@67.170.236.225)
# [03:56] * Quits: adele (adele@67.170.236.225) (Quit: adele)
# [03:57] * Parts: DougJ (djones4@74.76.23.86)
# [04:16] * Joins: Shunsuke (kuruma@133.27.63.42)
# [04:18] * Joins: DanC_lap (connolly@128.30.52.30)
# [04:25] * Quits: Shunsuke (kuruma@133.27.63.42) (Ping timeout)
# [04:25] * Joins: htmlr (htmlr@203.206.237.84)
# [04:34] * Joins: foca (foca@190.64.9.145)
# [04:37] * Joins: Shunsuke (kuruma@133.27.63.42)
# [04:40] * Quits: Shunsuke (kuruma@133.27.63.42) (Ping timeout)
# [04:41] * Joins: myakura (myakura@60.239.122.32)
# [04:46] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Quit: Get thee behind me, satan.)
# [05:02] <DanC_lap> hmm... I'm drafting a "empty +1/-1 messages are a poor use of time and reputation capital" message. I wonder if it's worth sending. meta-discussion backfires as often as it helps.
# [05:05] * DanC_lap opts to not send it
# [05:05] * karl suggests DanC to put it in the Weblog
# [05:05] <Lachy> DanC_lap, put it on the wiki or blog
# [05:08] <DanC_lap> which weblog?
# [05:08] <Lachy> yours?
# [05:09] <Lachy> you have one, don't you?
# [05:09] <DanC_lap> I have several, none particularly focussed on HTML WG stuff
# [05:09] <karl> yours or QA or a Web page on W3C web site. It's just a matter of something on the Web that people can point at
# [05:09] <DanC_lap> one for research (dig.csail.mit.edu) one for open source (advogato). the closest one is the QA weblog
# [05:10] <DanC_lap> I'm having a hard time with the latency in the QA weblog; I'm too impatient to wait 15 minutes for things to post
# [05:11] <karl> well if you find someone to create a perl plug-in which does CVS Commit in MT
# [05:11] <DanC_lap> well, I'm also not interested in customizing the QA weblog while it's based on MT, since MT isn't open source.
# [05:13] <karl> then put it somewhere else on the Web. No troubles.
# [05:14] <DanC_lap> a "welcome to the HTML WG" FAQ would be a good place for it.
# [05:14] <DanC_lap> I have been thinking about it, but haven't made much progress.
# [05:19] * Quits: htmlr (htmlr@203.206.237.84) (Quit: htmlr)
# [05:27] * Joins: Zeros (Zeros-Elip@67.154.87.254)
# [05:35] * Joins: adele (adele@67.170.236.225)
# [05:46] * Joins: Shunsuke (kuruma@133.27.53.98)
# [06:00] * Joins: Shunsuke_ (kuruma@133.27.53.98)
# [06:01] * Quits: Shunsuke (kuruma@133.27.53.98) (Ping timeout)
# [06:03] * Quits: DanC_lap (connolly@128.30.52.30) (Quit: Leaving)
# [06:06] * Quits: mw22 (chatzilla@63.245.220.228) (Ping timeout)
# [06:06] <myakura> http://i-yudai.blogspot.com/ nice :)
# [06:11] * Quits: Shunsuke_ (kuruma@133.27.53.98) (Quit: reboot)
# [06:16] * Parts: foca (foca@190.64.9.145)
# [06:18] <marcos_> myakura, hope s/he can keep it up
# [06:19] <myakura> yeah
# [06:19] <marcos_> Annevk is also posting weekly updates about what is happing on the list... with that nice Anne slant on things :)
# [06:19] <marcos_> http://annevankesteren.nl/
# [06:21] * Joins: Shunsuke (kuruma@133.27.53.98)
# [06:30] <Yudai> oops
# [06:32] * Quits: claudio (claudioc@89.97.35.74) (Ping timeout)
# [06:32] * Joins: claudio (claudioc@89.97.35.74)
# [06:33] <Yudai> myakura: please send comments if you find something to add :)
# [06:33] <myakura> Yudai: k :)
# [06:36] <myakura> Yudai: seems you've pointed the wrong url for the slashdot article ("And I found a topic...")
# [06:36] <Yudai> oh really?
# [06:39] <myakura> yeah, two anchors point to the same url (http://lists.w3.org/Archives/Public/public-html/2007Apr/0611.html)
# [06:40] <Yudai> thanks, ive corrected it
# [06:49] * Quits: claudio (claudioc@89.97.35.74) (Ping timeout)
# [06:49] * Joins: claudio (claudioc@89.97.35.74)
# [06:50] * Quits: adele (adele@67.170.236.225) (Quit: adele)
# [07:03] * Quits: claudio (claudioc@89.97.35.74) (Ping timeout)
# [07:03] * Joins: claudio (claudioc@89.97.35.74)
# [07:29] * Quits: heycam (cam@130.194.72.84) (Ping timeout)
# [08:16] * Joins: mjs (mjs@64.81.48.145)
# [08:24] * Joins: anne (annevk@83.82.206.111)
# [08:25] * Quits: marcos (chatzilla@203.206.31.102) (Ping timeout)
# [08:40] * Joins: loic (loic@90.41.1.61)
# [08:40] * Quits: Shunsuke (kuruma@133.27.53.98) (Ping timeout)
# [08:41] * Joins: adele (adele@67.170.236.225)
# [08:45] <sbuluf> off topic: if anyone is awake, and feels like answering (even privately) please do. else, disregard, please.
# [08:45] <sbuluf> assuming the whole web moves to html5/xhtml5
# [08:45] <sbuluf> question 1: to move the part in html5 to xhtml5 could be done programatically, automatically, trivially, right?
# [08:45] <Lachy> plenty of peole are awake
# [08:45] <Lachy> *people
# [08:45] <sbuluf> question 2: if we have all the web in xhtml5...could it be moved to xhtml2? if so how much more easily? even totally programatically? (i.e, by providing a routine that would do it alone, perfectly in all cases, in an unsupervised way)
# [08:45] <anne> 1: wrong
# [08:46] <Lachy> I don't understand question 1
# [08:46] <sbuluf> lachy, any doc in html5 could be automatically translated into xhtml5?
# [08:46] <Lachy> there are some small incompatibilites between the HTML and XHTML syntaxes
# [08:47] <Lachy> e.g. <noscript> can't be used in XHTML
# [08:47] <anne> The problem is mostly scripting
# [08:47] <sbuluf> i see
# [08:47] <anne> document.write() for instance
# [08:47] <anne> and various other things
# [08:47] <mjs> there are CSS differences as well
# [08:47] <Lachy> In the HTML serialsation, comments containing "--" can be represented (though theyr'e not valid) e.g. . In XHTML, that's fatal
# [08:48] <anne> mjs, that should be fixed though
# [08:48] <mjs> a converted documents would not behave identically
# [08:48] <anne> we should still be able to fix that
# [08:48] <mjs> the CSS thing is probably fixable, yes
# [08:48] <mjs> fortunately IE doesn't support XHTML yet so we don't have to worry about their bugs
# [08:48] <Lachy> the CSS differences are the way body { background: xxx; } is painted to the canvas and the way overlapping cells in tables are handled
# [08:48] <Lachy> both are going to be removed from the CSS spec
# [08:48] <anne> and overflow:
# [08:49] <Lachy> overflow? How is that different?
# [08:49] <anne> same as background
# [08:49] <Lachy> really? I'll look it up
# [08:49] <anne> except that it propagates to the viewport and not the canvas
# [08:49] <sbuluf> it would, however, at the very least, be easier right? (easier then from today tagsoup, i mean) and if so, significantly so?
# [08:50] <mjs> HTML5 aims to be able to parse today's tag soup
# [08:50] <Lachy> oh, you mean body { overflow: x; } applies to the canvas in HTML?
# [08:50] <anne> no, to the viewport
# [08:50] <Lachy> anne, same thing
# [08:50] <anne> no
# [08:51] <Lachy> the way I was using it, it is
# [08:51] <anne> the canvas is infinite
# [08:51] <Lachy> oh, yes
# [08:51] <anne> the viewport is finite
# [08:51] <Lachy> sorry
# [08:51] <sbuluf> thanks for the answers, everyone.
# [08:51] <anne> XHTML2 is btw way less well defined then HTML5/XHTML5
# [08:52] <anne> so I'm not sure what the advantage would be here
# [08:53] <anne> (the tables thing is btw not implemented)
# [08:53] <sbuluf> in case anyone is wondering, i was thinking if (x)html5 was good from the POV of somebody who wants the whole web's content in some xml-based flavour
# [08:54] * Joins: edas (edaspet@88.191.34.123)
# [08:54] <sbuluf> (more particularly, i was thinking about tim berners lee)
# [08:54] <anne> HTML5 gives you a DOM tree
# [08:54] <anne> which should be good enough
# [08:55] <anne> (XML gives you a DOM tree too)
# [08:55] <anne> it's all just serialization syntaxes... not really important
# [08:57] <sbuluf> rbut once you have a more or less well behaved dom, it would be easier to apply routines to it so that you get it progressively closer to, say, xhtml2?
# [08:57] <Lachy> sbuluf, why is it a goal to be more like XHTML2?
# [08:57] <marcos_> yeah?
# [08:57] <anne> XHTML2 is far less complete than HTML5 is
# [08:58] <sbuluf> lachy, well, as you know, i'm asking myself about tim berners lee motivations
# [08:58] <Lachy> yeah, I know
# [08:58] <anne> timbl "initiated" the HTML effort...
# [08:58] <sbuluf> more particularly, why did he ask for one html version and one xhtml version
# [08:58] <Lachy> btw, I put logs of our earlier chat in #xhtml up here http://junkyard.damowmow.com/278 and http://junkyard.damowmow.com/279
# [08:59] <marcos_> sbuluf, what are those motivations?
# [08:59] <Lachy> view source on the first one, till Hixie gets around to changing the MIME to text/plain
# [08:59] <sbuluf> marcos, dunno, i'm asking myself that
# [08:59] <sbuluf> anne, i know, but...i think he did because you would have done it anyway
# [08:59] <anne> the idea of having an XML serialization is that people who want can integrate it into other XML stuff
# [09:00] <sbuluf> anne, exactly
# [09:00] <sbuluf> anne, my point is...do we agree that timbl wanted xhtml2, and not html5?
# [09:00] <anne> and because it's trivial once you have a tree language defined
# [09:01] <marcos_> I don't think so
# [09:01] <anne> sbuluf, it seems silly to agree on that
# [09:01] <Lachy> I don't see why TimBL's opinion is relevant
# [09:01] <anne> sbuluf, i think it's a silly question, even
# [09:01] <sbuluf> because he started this group?
# [09:02] <sbuluf> lachy, not necessarily relevant...just wondering
# [09:03] <sbuluf> the way i see things, he wanted strictness, xml. he even charetered xhtml2 and was ready to allow breaking of back compat.
# [09:03] <anne> sbuluf, maybe you should just ask him
# [09:03] <anne> assuming is a very bad thing to do
# [09:03] <anne> imo
# [09:04] <sbuluf> then you pushed for this, and..it would have happened either he wanted it or not, so he accepted
# [09:04] <sbuluf> anne, right. i did try to ask him, btw. radio silence. however
# [09:05] <anne> then there's no point in discussing it
# [09:05] <sbuluf> nono, this will happen, not arguing that
# [09:05] <mjs> his personal motivations are not relevant
# [09:05] <sbuluf> i just wondered if he had more motives, or an strategy
# [09:05] <mjs> one possibility is that he sincerely changed his mind about the right approach to HTML
# [09:05] <anne> ask him...
# [09:05] <mjs> but it doesn't really matter
# [09:06] <marcos_> or read his book
# [09:06] <mjs> Tim is the inventor of the Web, not its king
# [09:06] <marcos_> I think you will find that he just wanted to get the ball rolling, not to control it's evolution
# [09:06] <marcos_> its*
# [09:06] <sbuluf> mjs..yes, that is a possibility, of course. but...between you and me...do you really believe it?
# [09:07] <anne> he's a smart guy
# [09:07] <anne> then again, say above
# [09:07] <anne> s/say/see/
# [09:07] <mjs> I can't believe cwilso keeps using Office as an example
# [09:07] <sbuluf> marcos...but he *did* give the ok for xhtml2. *and* he kept going with the whole semantic web stufff
# [09:08] <Lachy> so what?
# [09:08] <mjs> does he have any idea what people outside of Microsoft thing of OOXML?
# [09:08] <anne> if someone says OOXML I'm always confused
# [09:08] <anne> OpenOfficeXML
# [09:08] <sbuluf> lacy, so i think he might still have a plan to keep going the xhtml2 way, *and* not to loose the whole of the current web content
# [09:08] <anne> but that's clearly not it
# [09:09] <anne> sbuluf, so?
# [09:10] <sbuluf> anne, nothing, i'm wondering if he has that idea. in that case, he might see the html5 move as a sort of big "tidy" (the program)
# [09:11] <anne> sbuluf, we can't tell you. I'm not sure how I can make this more clear
# [09:11] <marcos_> sbuluf, it's not up to him... W3C is a consortium driven by business interests.. if members want to get together to standardize something web-related the w3c is a place to do it... it does not mean it will be adopted.
# [09:11] <sbuluf> something that gets him clser to move today's web content to the xhtml2 track
# [09:12] <anne> timbl doesn't seem to be remotely involved in XHTML2, but I may be wrong
# [09:12] <sbuluf> anne, fair enough. and thanks for answers, all, once more
# [09:12] <sbuluf> anne, mm, i'd say he needs strictness for the project he *is* all day into: the semantic web
# [09:12] <anne> the semantic web doesn't need strictness
# [09:13] <anne> the semantic web needs some way to convince people to actually add all that (accurate) metadata
# [09:13] <sbuluf> how about rdfa? or grddl?
# [09:13] <anne> those are just syntaxes
# [09:13] <marcos_> lol :D
# [09:14] <sbuluf> (ways of inserting and recovering rdf from webpages)
# [09:14] <anne> the semantic web is basically an idea about lots of metadata
# [09:14] <anne> that interconnects
# [09:14] <sbuluf> right, but has a link to the web we know
# [09:14] <anne> syntax is hardly relevant to its success story
# [09:14] <Lachy> the semantic web is a pipe dream
# [09:15] <anne> yes
# [09:15] <Lachy> and RDF has failed completely
# [09:15] <Zeros> provided enough of it was auto-generated it could be viable for certain things
# [09:15] <anne> it relies on "magical tools" and people adding metadata
# [09:15] <anne> "the tools will safe us" is something I don't believe in
# [09:15] <Lachy> microformats are a reasonable, practical approach to the semantic web
# [09:15] <marcos_> The pedantic web :D
# [09:15] <anne> :p
# [09:16] <sbuluf> ( i could argue fro some semweb stuff, but the discussion would get hughe. i agree that it has problems, though. some ideas might be usable still, however, imho)
# [09:16] <Zeros> Lachy, microformats will suffer similar problems.
# [09:16] <Zeros> its just more localized since the metadata isn't separate
# [09:16] <marcos_> Nothing wrong with the idea
# [09:16] <anne> microformats are based on visible metadata
# [09:16] <Lachy> Zeros, in some cases, yes
# [09:16] <anne> which is at least an improvement
# [09:16] <anne> but they are not really well documented
# [09:16] <Lachy> some microformats are getting far too complicated to see any real world use
# [09:17] <Lachy> and their specs are an issue too
# [09:17] <sbuluf> (i could argue against microformtas too, but again, i did not wish to make the discussion so big, or off topic)
# [09:17] <marcos_> you can't nail down semantics on a global scale... semantics are transient which is why tagging is successful
# [09:18] <Zeros> Lachy, a standardized way to specifying microformats would be helpful
# [09:19] <marcos_> Zeros, what's wrong with the microformats wiki approach?
# [09:19] <Lachy> Zeros, what the?
# [09:19] <Lachy> do you mean like a way to specify a formal grammar?
# [09:20] <Lachy> like a schema for microformats?
# [09:20] <anne> like a spec
# [09:20] <Zeros> yes, but human readable
# [09:20] <Zeros> we don't need BNF for them
# [09:20] <anne> atm they come down to a bunch of examples and use cases of which you need to derive how it works (in my experience anyway)
# [09:21] <marcos_> Why just use a wiki. The behavioral semantics need to be hard-coded into software anyway.
# [09:21] <mjs> marcos_: the microformats specs don't define conformance requirements
# [09:21] <mjs> marcos_: either for valid use of the microformat, or for conformant processers that try to take the microformat as input
# [09:21] <anne> it's like the HTML4 spec
# [09:22] <marcos_> mjs: has that been a big problem so far?
# [09:22] <anne> for implementors and validators, yes
# [09:22] * Quits: adele (adele@67.170.236.225) (Quit: adele)
# [09:22] <mjs> anne: unsurprising, given that Tantek things the HTML4 spec is fine as is
# [09:22] <anne> I heard about that...
# [09:22] <marcos_> anne, is Opera planning to support some microformats?
# [09:23] <mjs> marcos_: not really, but only because there are no widely deployed tools that process microformats yet
# [09:23] <anne> can't comment on that, I was just reitering some frequently cited comments
# [09:23] <mjs> (why do I keep typing "things" instead of "thinks"?)
# [09:24] <Lachy> mjs, I keep typing thinks instead of things :-)
# [09:24] <sbuluf> (no worst typer here than me, don't worry)
# [09:26] <hsivonen> what's Dmitry's PACK?
# [09:26] <marcos_> mjs: I don't disagree that things could be further specified. However, the simplicity of microformats currently makes them more usable then the alternatives... but at some point what you said would need to be specified
# [09:27] <hsivonen> marcos_: much of Microformats is based on things that you can only see if you have read microformats-discuss actively for months
# [09:28] <marcos_> hsivonen, which I don't. So I won't comment further.
# [09:28] <hsivonen> I'm not suggesting that the emperor is naked, but he sure isn't fully clothed
# [09:29] <marcos_> anyone with experience in RDF/OWL can see the big holes in microformats, of course... but anyone can appreciate the simplicity of microformats too.
# [09:30] <hsivonen> marcos_: the major holes I see are the lack of content conformance requirements and the lack of a specified processing model for consuming apps
# [09:30] <anne> lol
# [09:30] <anne> the #xhtml minutes are funny
# [09:31] * sbuluf glup
# [09:32] <sbuluf> does html4 allow for insertion of arbitrary datastructures? does html5? if not, shouldn't it?
# [09:32] <Lachy> sbuluf, what kind of data structures?
# [09:33] <mjs> marcos_: I think the concept of microformats is good, but the specs are not very good in spec terms
# [09:33] <sbuluf> (where datastructure = bunch of [ordered] fields)
# [09:33] <mjs> anne: pointer?
# [09:33] <marcos_> mjs, agreed.
# [09:33] <Lachy> sbuluf, for what purpose?
# [09:33] <Lachy> sbuluf, use case?
# [09:33] <Lachy> sbuluf, example?
# [09:34] <Lachy> mjs, see the links I posted above
# [09:34] <anne> "<ShaneM> yes. I think the w3c proces will cause some change that hickson wont like and the whole thing will go to shit."
# [09:34] <Lachy> http://junkyard.damowmow.com/278 and http://junkyard.damowmow.com/279
# [09:35] <anne> :D
# [09:35] * Joins: htmlr (htmlr@203.206.237.84)
# [09:35] <sbuluf> lachy, you define say data for an employee. then you could insert that data into html, without having to fit it into, say a <ul>, or a <ol>, or a table, which are the allowed datastrutures in html
# [09:36] <Lachy> so you want some kind of arbitrary markup? Like EmployeeXML or something?
# [09:36] * anne rather has a use case than a feature request
# [09:36] <sbuluf> yes, some sort of way to include arbitrary datastructres into html, without needing to fit it into already existing ones
# [09:37] <Lachy> sbuluf, we need use cases, not arbitrary solutions
# [09:37] <mjs> Lachy: thanks
# [09:38] <mjs> "solve real problems"
# [09:38] <Zeros> sbuluf, why not just use XSL to transform your EmployeXML into HTML?
# [09:38] <mjs> Lachy: I think the 278 one is missing a whitespace: pre or something
# [09:39] <mjs> Lachy: was it supposed to be plaintext? Firefox and Safari both seem to be rendering it as HTML
# [09:39] <anne> mjs, it has the wrong MIME type
# [09:39] <anne> mjs, view source
# [09:39] <Lachy> mjs, when I uploaded it, I forgot to change the MIME to text/plain
# [09:39] <Lachy> Hixie said he can fix it, but he hasn't done it yet. He'll do it when he gets back
# [09:39] <sbuluf> zeros, because you would need to plug your employee data, or your car data, into just the few datastrutures html allows (like ul, ol, table). isn't that a bit like looking for ways to plug square blocks into tringle holes?
# [09:39] <anne> sbuluf, what's the use case?
# [09:40] <mjs> ok, thanks
# [09:40] <sbuluf> anne, being able to display arbitrary data straight from any database in html? is that what you mean?
# [09:40] <anne> did Shane suggest timbl is not that clever? :)
# [09:40] <anne> or cwilso?
# [09:41] * anne can't really believe it for either
# [09:41] <sbuluf> timbl, i think
# [09:41] <Lachy> anne, quote?
# [09:41] <anne> "
# [09:41] <anne> Apr 13 14:47:18 <ShaneM> I think you give him too much credit. he's not that clever. really he's not. oh - you mean the "XML serialization" of the new HTML?"
# [09:41] <sbuluf> timbl
# [09:41] <anne> yeah, sees to be about timbl... ouch
# [09:42] <anne> sbuluf, that's not a use case
# [09:42] <Zeros> sbuluf, the arbitrary semantics attached to your CarML or EmployeeML will be useless to browsers and screen readers.
# [09:42] <Zeros> google won't know what to do with it
# [09:42] <anne> sbuluf, and if it is, that's already possible
# [09:42] <anne> just export the thing to some <table>
# [09:45] <sbuluf> mm, that's true, a tree of fields/values could always be rendered as a table
# [09:45] <Zeros> or lists
# [09:45] <Lachy> sbuluf, look at the way existing databases export data to HTML. Things like PHPMyAdmin for MySQL use a table
# [09:47] * Quits: mjs (mjs@64.81.48.145) (Client exited)
# [09:47] * Joins: mjs (mjs@64.81.48.145)
# [09:48] <sbuluf> zeroa, regarding meaning, what if you defined the structure in some schema language (hanged somewhere on the web), where you could specify meanings as well?
# [09:49] <anne> not a use case
# [09:53] <mjs> Lachy: I am reading those logs with wide eyes and mouth agape
# [09:54] <Lachy> mjs, aha! I was typing my responses the same way.
# [09:54] <Lachy> krijnh, yt?
# [09:55] <sbuluf> anne, i will think some more, re-read microformats, perhaps be back at it.
# [09:55] <anne> sure
# [09:55] * anne feels fricking tired
# [09:56] <sbuluf> (thanks everyone for answers and comments, once more)
# [10:00] <Zeros> sbuluf, There's no way to encode abstract meaning like that
# [10:00] <Zeros> DC, microformats, HTML; they all have predetermined meaning
# [10:02] * Quits: myakura (myakura@60.239.122.32) (Quit: Leaving...)
# [10:02] <sbuluf> zeros, i need to refresh microformats.
# [10:03] <Zeros> :)
# [10:03] * Joins: ROBOd (robod@86.34.246.154)
# [10:03] <sbuluf> last time i saw them, once, more, my impression was poeple trying to plug square pegs into just tringle holes...cause thy could not have holes of any shape they wanted.
# [10:05] <marcos_> later y'all
# [10:05] <sbuluf> later, marcos
# [10:05] <Lachy> bye marcos_
# [10:05] <Zeros> night marcos_
# [10:10] <sbuluf> zeros, a side note: there seem to be at least two kinds of "meanings". one is document-meanings, and the other is real meanings. for example: <p>Bush is a cretin.</p> the "document-meaning" is "this is a paragraph", wile the real meaning, is "my opinion of bush". that's why i feel uneasy each time i read about "semantic html"
# [10:14] * Joins: mikeday (mikeday@144.136.3.123)
# [10:15] <sbuluf> html provides only for document-meanings, and not for real-meanings. so i tend to think html is not really semantic.
# [10:17] <Lachy> semantic HTML is about describing the "document-meaning", not describing how to interpret the natural language used within it
# [10:17] <sbuluf> right
# [10:18] <mikeday> Has the HTML table layout model ever been formally specified?
# [10:19] <anne> nope
# [10:20] <anne> dbaron did some work on reverse engineering IE
# [10:20] <anne> the IE team doesn't get their own code
# [10:20] <mikeday> doesn't get?
# [10:20] <anne> Hixie stopped working on it
# [10:20] <anne> s/get/understand/
# [10:20] <mikeday> did he produce anything publically available before stopping?
# [10:20] <anne> (As I understand it it's quite a complicated beast.)
# [10:21] <anne> Hixie released some demo/tests and dbaron wrote a document on intrinsic widths
# [10:21] <mikeday> we are currently attempting to reverse engineer Mozilla and Opera
# [10:21] <mikeday> but it seems suboptimal for each group to reverse engineer their predecessors every time
# [10:21] <anne> Mozilla changed their model for Firefox 3
# [10:22] <anne> at least I believe dbaron made changes to it per his document
# [10:22] <anne> I think you best reverse engineer IE
# [10:22] <krijnh> Lachy: Yes
# [10:22] <Lachy> krijnh, can you join #xhtml and keep logs of it on your site?
# [10:22] <krijnh> On this server?
# [10:23] <Lachy> yes
# [10:23] <krijnh> Sure
# [10:23] <krijnh> One moment plz, k10x
# [10:23] <Lachy> that's the XHTML2WG channel, it should be entertaining for us :-)
# [10:24] <Lachy> I put some logs up on damowmow (see links above) that you can import into your logs if you like
# [10:24] <krijnh> I should get Google Ads with al this bandwith usage just for entertainment ;)
# [10:24] <krijnh> bandwidth*
# [10:24] <mikeday> anne: would you have a link to dbaron's work?
# [10:26] * Joins: hasather (hasather@81.235.209.174)
# [10:29] <Lachy> krijnh, is it possible for you to set up links like /whatwg/latest, /html-wg/latest and /xhtml/latest that redirect to the most recent day, so I can bookmark those
# [10:29] <krijnh> Sure
# [10:47] <krijnh> Lachy: http://krijnhoetmer.nl/irc-logs/html-wg/latest/
# [10:47] * Joins: krijnh2 (krijnhoetm@213.84.148.98)
# [10:47] * Parts: krijnh2 (krijnhoetm@213.84.148.98)
# [10:53] <anne> mikeday, http://dbaron.org/css/test/intrinsic/ http://dbaron.org/css/intrinsic/
# [10:53] <mikeday> Thanks!
# [11:07] <anne> Btw, I feel there shouldn't be such a thing as HTML table layout model by the way. That should be the way the CSS table layout model works.
# [11:08] <anne> And the (X)HTML table elements should just be implemented in terms of the CSS table layout model (and not break stuff!)
# [11:09] <Zeros> Wasn't that the whole point of adding the CSS table display values in the first place?
# [11:11] * Quits: ROBOd (robod@86.34.246.154) (Ping timeout)
# [11:13] * Joins: ROBOd (robod@86.34.246.154)
# [11:17] * Joins: heycam (cam@203.214.79.176)
# [11:22] * Joins: myakura (myakura@60.239.122.32)
# [11:27] <sbuluf> http://developers.slashdot.org/developers/07/04/12/152245.shtml <--i assume everyone knows html5 has hit slashdot, right?
# [11:31] <mikeday> The trouble is, CSS table layout does not cover the full generality of HTML tables.
# [11:31] <mikeday> Hence the need for an extended specification, wherever it ends up.
# [11:34] <mjs> sbuluf: I found out when my manager's manager walked by my couch and said, "Hey Maciej, you know you made slashdot today, right?"
# [11:35] <mjs> I was worried that this would lead to a dressing down and/or involuntary change of employment
# [11:35] <mjs> what really surprises me is that it did not make digg or reddit
# [11:40] * Quits: Zeros (Zeros-Elip@67.154.87.254) (Quit: Leaving)
# [11:41] <beowulf> does it worry you getting a lot of publicity at this stage in the process?
# [11:43] <mjs> I have mixed feelings about the publicity
# [11:43] * Joins: zcorpan (zcorpan@84.216.42.156)
# [11:43] <mjs> I'd rather see publicity when we accomplish something
# [11:44] <mjs> the proposal being formally adopted might be a newsworthy item
# [11:44] <mjs> or even something like the Design Principles doc being published as a Note
# [11:45] <beowulf> it worries me the effect of not delivering
# [11:46] <mjs> however, I think the proposal to adopt HTML5 as a starting point has been probably the biggest kickstart the group has seen
# [11:46] <beowulf> for sure, but if that doesn't get accepted or gets bogged down... eek
# [11:47] <mjs> then we can start determining if the group has failed
# [11:47] <anne> wasn't it microsoft intention to do what developers wanted?
# [11:48] <mjs> sentiment seems strongly in favor of html5, so it's just a matter of the chairs deciding when it is reasonable to hold a vote
# [11:48] <mjs> design principles upgrade to more formal status might carry and chairs might agree on a vote soon
# [11:48] <mjs> since many of them are becoming widely cited
# [11:50] <hsivonen> people have such a fixation on languages being somehow less proper until there is a schema :-(
# [11:50] * anne just wrote the exact same thing as hsivonen
# [11:50] <anne> only yours reached the list a little earlier :)
# [11:50] <mjs> there can totally be a schema
# [11:50] <mjs> but the schema language will have to be Python or Perl or something
# [11:50] <beowulf> i'd also like to see more discussion on the gulf between the majority of html writers and the sort of person who writes to the spec
# [11:51] <anne> the spec is designed with inperfection in mind
# [11:52] <beowulf> which is sort of the last point of the charter success criteria
# [11:52] <beowulf> i don't really mean perfection, just knowing that div is not the only fruit
# [11:54] <sbuluf> <mjs> sentiment seems strongly in favor of html5 <-- here maybe. not in slashdot, it seems, altought i assume that does not count
# [11:56] <sbuluf> as for this WG...the html5 proposal was the big kickstarting, yes. the chris wilson reply was the big stopper, though. current exchanges are most interesting, imho.
# [11:56] <beowulf> i think the whole thing is fascinating :)
# [11:56] <sbuluf> are there chances of microsft walking away if they do not get what they want?
# [11:57] <sbuluf> would a vote really matter if they do not agree?
# [12:00] <anne> does it help speculating?
# [12:00] <gsnedders> Chris is far from an idiot – I expect he'll find some way of having a swtich.
# [12:01] <anne> hsivonen, I think part of the problem is that we make people make a lot of mind-shifts
# [12:02] <anne> hsivonen, no SGML, no schema, defined error handling, HTML recommended over XML, etc.
# [12:02] <anne> hsivonen, no DOCTYPE
# [12:02] <anne> It's probably hard to grasp that all you ever thought was right -- is wrong
# [12:03] <anne> bluntly put
# [12:03] <beowulf> that's well put
# [12:06] <mjs> sbuluf: I'm just counting sentiments expressed in the working group
# [12:06] <sbuluf> right
# [12:06] <mjs> Chris's essay has created a sideshow but isn't directly related to the adtopt / not adopt question
# [12:06] <anne> chris second essay will address that
# [12:07] <mjs> if Microsoft is willing to talk away over losing one vote on a minor issue, then we will be sad but we can live without them
# [12:08] <gsnedders> I want to see versioning, but I have no idea as to how we should do it.
# [12:08] <hsivonen> anne: yeah, HTML5 is about taking the red pill
# [12:08] <gsnedders> I don't want it in the DOCTYPE, but it can't be on the <html> element as it can be ommitted
# [12:09] <hsivonen> gsnedders: omitting the start tag is a non-issue of version information can be omitted as well :-)
# [12:10] <gsnedders> hsivonen: that's true, but I want it somewhere where it can always be in any document
# [12:11] * Quits: Yudai (Yudai@59.147.29.149) (Quit: SIGTERM received; exit)
# [12:11] * Joins: Yudai (Yudai@59.147.29.149)
# [12:12] <anne> hsivonen, thanks, used as slug :)
# [12:16] <hsivonen> gsnedders: every document has the html element even if you don't see it in the syntax
# [12:17] <gsnedders> hsivonen: I'm aware, but how can we add an attribute to it if it isn't in the markup?
# [12:17] <hsivonen> gsnedders: you add it to the markup when you add the attribute
# [12:17] <hsivonen> gsnedders: assuming we end up defining an attribute, which I doubt
# [12:18] <hsivonen> gsnedders: processing models will need to deal with the absence of the attribute anyway
# [12:18] <gsnedders> and that model should be the latest version of HTML.
# [12:21] <mikeday> I thought HTML5 had a RELAX NG schema, even if it's only informative?
# [12:21] * Joins: foca_ (foca@190.64.4.141)
# [12:25] <hsivonen> mikeday: it isn't even an informative part of the spec
# [12:26] <mikeday> an uninformative part? :)
# [12:26] <hsivonen> mikeday: not any kind of part of the spec
# [12:26] <mikeday> What is the purpose of avoiding a schema?
# [12:27] <hsivonen> mikeday: 1) Avoiding casting a part of implementation in concrete and 2) avoiding making people think that a schema is sufficient
# [12:28] <hsivonen> mikeday: you wouldn't put normative lines of C++ in the spec, either
# [12:28] <mikeday> 1) won't the spec cast the content model in concrete? 2) sufficient for what?
# [12:28] <gsnedders> mikeday: 2) the requirements of the specification expressable in prose but not in any schema
# [12:29] <hsivonen> 1) yes 2) conformance checking
# [12:29] <gsnedders> mikeday: there are thousands of HTML documents that people think are valid because the validator says so, yet they aren't, because the DTD can't express what the spec says
# [12:29] <mikeday> If the spec needs to describe the content model, why not describe in a form that is widely understood?
# [12:29] <hsivonen> mikeday: English is :-)
# [12:29] <mikeday> Hmm, that is a limitation of the validator, not the DTD.
# [12:29] <gsnedders> mikeday: because using a schema puts limits on what we can describe
# [12:29] <gsnedders> mikeday: no, a limitation of the schema
# [12:30] <gsnedders> mikeday: there's a limit to how complex rules can be put in a schema
# [12:30] <mikeday> People generally don't validate their HTML against a DTD, they pass the HTML to a validator;
# [12:30] <mikeday> the validator does not need to be restricted to just a DTD/schema.
# [12:30] <gsnedders> mikeday: the validator checks it again a DTD
# [12:30] <hsivonen> mikeday: well, there you go. a schema isn't sufficient
# [12:30] <gsnedders> mikeday: why have a DTD at all then?
# [12:30] <gsnedders> mikeday: if the schema can't express everything, why mislead people into thinking it explains the whole standard?
# [12:31] <mikeday> No standard is wholly explained by a schema, but many standards have schemas nonetheless, as they can be useful
# [12:32] <mikeday> The "schema is not sufficient to express everything" argument is in indictment of using schemas at all, for anything.
# [12:32] <hsivonen> mikeday: well, let's look at it this way: in theory, the formal power of RELAX NG could deal with exclusions. However, to use that power, an insane number of grammar productions would be needed. I made the call to leave big holes in RELAX NG and plug those in Schematron
# [12:32] * mikeday nods
# [12:32] <hsivonen> mikeday: if the normative schema made a different call, my implementation would be seen as illegitimae
# [12:32] <hsivonen> illegitimate
# [12:33] <mikeday> Hmm, your schema is either useful or non-useful
# [12:34] <mikeday> for example, someone might want to take the basic schema and restrict it further, to create a subset of HTML,
# [12:34] <mikeday> perhaps to use in a content management system
# [12:34] <mikeday> that doesn't make the schema illegitimate, just useful for a different purpose.
# [12:35] <hsivonen> my point is that if someone were to create a RELAX NG *only* schema that is as useful as possible, the schema would be different from the RELAX NG part of a more useful RELAX NG + Schematron solution
# [12:35] <mikeday> Go with the more expressive RELAX NG + Schematron approach, then
# [12:35] <mikeday> sounds like it would produce a more readable and comprehensive schema
# [12:36] <mikeday> there's a big difference between including Schematron assertions and C++ code in an appendix to the spec
# [12:36] <hsivonen> mikeday: if all schemas are just non-normative implementations, fine. but if one schema is elevated as the One Normative Schema, other implementation approaches need more explaining to people who just see that someone made a homegrown schema
# [12:37] <hsivonen> I already went with RELAX NG + Schematron
# [12:37] <hsivonen> which makes the result less useful for e.g. the editor autocompletion use case
# [12:38] * mikeday ponders
# [12:38] <mikeday> including two schemas as informative appendices to the specification sounds like the way to go, then :)
# [12:39] <mikeday> it would give people something to work with,
# [12:39] <hsivonen> mikeday: even if the schema written by fantasai and I was elevated as the One True Schema, I'd still like to be able to fix bugs without users flaming me if the schema and the prose disagree
# [12:39] <mikeday> and demonstrate concretely that there is no One True Schema.
# [12:40] <hsivonen> mikeday: as precedent, the schema in the appendix of the Atom spec sucks compared to the Feed Validator
# [12:40] <mikeday> ie. "Schema for (Partial) Document Validation", "Schema for Editor Autocompletion", ...
# [12:41] <mikeday> hmm, does the feed validator do much stuff that can't be expressed in RELAX NG + Schematron?
# [12:41] <hsivonen> mikeday: yes
# [12:41] <zcorpan> why have a schema in the spec at all? people who want a schema would find one regardless of whether there is one in the spec or not
# [12:42] <mikeday> personally I think that expressing things formally can assist in the development of the spec,
# [12:42] <mikeday> as it is easier to hide ambiguities in prose
# [12:43] * Quits: sbuluf (kzxdkz@200.49.140.77) (Ping timeout)
# [12:43] <hsivonen> mikeday: check out the XForms schema for some ambiguities :-)
# [12:43] <mikeday> consider the XML grammar, which is quite rigorous and easy to follow, however,
# [12:43] <mikeday> parameter entities are only described in the prose, and are often implemented incorrectly by parsers
# [12:43] <hsivonen> mikeday: developing a conformance checker in parallel with the spec is a good way to find bugs in prose
# [12:44] <mikeday> schema = conformance checker implemented in declarative programming language :)
# [12:45] <Hixie> feel free to write more schemas :-)
# [12:45] <Hixie> however, imho at least, a schema has the same status as any other implementation of the standard
# [12:45] <hsivonen> I'd bet that en-GB-x-Hixie is more readable than Prolog to most people
# [12:46] <Hixie> i think the spec mostly uses en-US now :-(
# [12:46] <mikeday> heh
# [12:46] <gsnedders> Hixie: yeah, I've been noticing more and more en-US
# [12:46] <mikeday> Hixie: were you working on describing HTML table rendering at one point?
# [12:46] <Hixie> mikeday: no, dbaron did though
# [12:47] <Hixie> anyone see any interesting or fun comments in the slashdot threads?
# [12:47] <Hixie> i don't feel like reading them all
# [12:47] <zcorpan> perhaps i should start to advocate Transitional doctypes. they have the handling of content i want css to say is the correct
# [12:47] <zcorpan> Hixie: i looked though them quickly yesterday, but didn't find anything interesting
# [12:48] <zcorpan> Hixie: except one guy liked the elegance in being able to omit tags
# [12:48] <Hixie> heh
# [12:49] * Quits: karl (karlcow@128.30.52.30) (Quit: Where dwelt Ymir, or wherein did he find sustenance?)
# [12:49] <Hixie> nn
# [12:51] <mikeday> Is there any chance on getting HTML table rendering specified in HTML5? (ie. if someone else does the work and submits it to the group)
# [12:52] <mikeday> as it seems odd to add new features when the old features are not specified yet.
# [12:52] <gsnedders> that's probably more relevant defining display: table* in CSS, and then the HTML references that
# [12:53] <Lachy> I'm fairly sure table rendering will be defined by the CSS WG in due course
# [12:54] <mikeday> question still stands for Hixie then, given that CSS 2.1 punted on the issue
# [12:55] <mikeday> (is punted an Americanism meaning "kicked it away, choosing not to run with the ball"?)
# [12:55] <Lachy> CSS 2.1 is feature complete, table rendering should be defined in CSS3
# [12:55] <mikeday> Shame that so many years of work were not able to define what browsers have been doing all this time, ah well.
# [12:56] <hsivonen> in November, the slot allocation of cells in CSS 2.1 was totally unrealistic
# [12:56] <Dashiva> Well, how many years now since html was shut down? That's a lot of years spent on XHTML
# [12:56] * hsivonen checks
# [12:57] <Lachy> hsivonen, do you mean for overlapping cells in XHTML docs?
# [12:58] <hsivonen> Lachy: yes
# [12:58] <hsivonen> and yes, the latest CSS 2.1 WD is still unrealistic on this point
# [12:58] <mikeday> unrealistic = useless for user agent implementors and authors
# [12:58] <Lachy> that difference, along with the other XHTML differences, should be taken out
# [12:58] <mikeday> particularly user agent implementors, as authors can always just use trial and error :)
# [12:58] <Lachy> it won't make it out of CR with it, since UAs don't implement it
# [13:02] <mikeday> I was thinking more of width calculation than cell allocation
# [13:03] <mikeday> as the width calculation in CSS 2.1 seems divorced from browser reality
# [13:03] * Parts: foca_ (foca@190.64.4.141)
# [13:29] <mikeday> somewhat off topic, new version of Prince is out with improved HTML table support: princexml.com/alpha/2007-04-13/
# [13:30] <mikeday> proudly supporting the table "border" attribute since 2007. ahem.
# [13:34] <Dashiva> Has there been discussion about changing <html> to something different? A desperate attempt to be able to separate old pages from new pages that don't want IE5 quirks mode just because they leave out a doctype
# [13:42] <hsivonen> Dashiva: not really. the old parsers would infer <html> anyway which would interfere with changing the name of the root element
# [13:45] <zcorpan> mikeday: does <table border> (or <table border="">) result in a 1px border (equivalent to <table border="1">)?
# [13:47] <Dashiva> hsivonen: But old parsers can be replaced, old pages are forever
# [13:51] <zcorpan> Dashiva: how would changing the root element help?
# [13:52] <Dashiva> It's the same idea as <!DOCTYPE html> to say 'Standard mode here please', but making it an integral part of the document instead of a doctype that can be forgotten or left out
# [13:53] <zcorpan> i don't understand the difference
# [13:53] <zcorpan> <html> can be forgotten or left out
# [13:53] <zcorpan> doctype is an integral part of the document
# [13:53] <Dashiva> Hardly
# [13:54] <Dashiva> The doctype is an anti-quirks mode talisman
# [13:54] <zcorpan> fair enough, but your proposal would make the root element tag the same
# [13:54] <zcorpan> no?
# [13:54] <Dashiva> Correct
# [13:54] <zcorpan> and wouldn't help today's browsers
# [13:55] <zcorpan> nor would it solve the problem of getting interop on today's content
# [13:55] <zcorpan> so i don't get the point
# [13:55] <Dashiva> It's not attempting to solve those problems. It's to avoid the situation where new pages are quirks mode by default, even if they're created in 2007
# [13:55] * Joins: marcos__ (chatzilla@203.206.31.102)
# [13:56] <zcorpan> well, you'd have to make authors use the new tag, which is just as hard as making them use <!doctype html>
# [13:56] <zcorpan> <!doctype html> already works so why not use that?
# [13:57] <Dashiva> As I said, doctypes can be left out. I see it all the time when doing JS help
# [13:57] <zcorpan> your magic tag can also be left out
# [13:58] <zcorpan> i see it more often than the doctype ;)
# [13:58] <Dashiva> Replaced, not left out
# [13:58] <Dashiva> The difference is that the pros leave out <html>, the amateurs leave out doctype
# [13:58] <zcorpan> all pages today don't have your magic tag, so it can obviously be left out
# [13:59] <zcorpan> that's because doctypes are hard to remember
# [13:59] <Dashiva> Not left out, replaced by <html>
# [13:59] <zcorpan> <!doctype html> isn't hard to remember
# [13:59] <Dashiva> It's a two-stage process. You have to remember you need a doctype first, then you can remember which one
# [14:00] <zcorpan> i don't see the difference from your magic tag. you need to remember you need one, and you need to remember which one
# [14:00] <zcorpan> no?
# [14:01] <Dashiva> How often do you see pages without <html>?
# [14:01] <zcorpan> not often, but i haven't seen any page with your magic tag
# [14:01] <Dashiva> Obviously, it's just a proposal
# [14:01] <zcorpan> getting people to replace <html> with your magic tag is not simler than making them use <!doctype html>
# [14:02] <mikeday> zcorpan: no, have to add that to the next release; thanks for pointing it out!
# [14:02] <zcorpan> mikeday: np
# [14:03] <Dashiva> Getting people to use <html5> (pure example) for HTML5 could even be made into a natural thing with PR. Doctypes remain an arbitrary attachment
# [14:03] <Dashiva> (the 5 should not be taken as an argument for versioning ;)
# [14:04] <zcorpan> how would you style it? in today's browsers it would be a bogus element child of the <body>
# [14:04] <zcorpan> changing the html parser is not trivial
# [14:05] <Dashiva> This is not a change for today, it's an idea to eventually get out of the situation where we're locked in a quirks default
# [14:05] * Quits: htmlr (htmlr@203.206.237.84) (Quit: htmlr)
# [14:06] <zcorpan> ok. don't think it would work though
# [14:07] <mikeday> idea: add an unobtrusive marker in the browser chrome when page is in quirks mode
# [14:07] <Dashiva> That may be, we wouldn't be in today's situation if there was an easy way out
# [14:07] <mikeday> just like the popup notification in Firefox, or the way the URL bar is highlighted for SSL
# [14:07] * Quits: loic (loic@90.41.1.61) (Ping timeout)
# [14:07] <mikeday> over time, make the quirks mode chrome more and more obvious and unappealing
# [14:08] <mikeday> sites that trigger quirks mode will start to look ugly, dangerous, outdated, unappealing
# [14:08] <Dashiva> That suffers the same problem as XML error handling
# [14:08] <mikeday> then in say, 2046, drop support for quirks mode :)
# [14:08] <Dashiva> The browser without it will be more attractive
# [14:08] <zcorpan> mikeday: dream on ;) no browser would do that
# [14:08] <zcorpan> that would result in less market share
# [14:09] <mikeday> or more practically, drop the layout parts of quirks mode, so that sites still "work" but don't look as good
# [14:09] <Dashiva> Another way (but still highly improbable) would be to default pages to standards mode and include a "Is this page broken? Try old page compatability mode, click here"
# [14:09] <zcorpan> "my browser complains about 9 pages out of 10. i better switch back to my old browser or to that other browser that doesn't complain"
# [14:09] <Dashiva> The problem remains that you're pushing inconvenience on the users
# [14:10] <mikeday> perhaps find a better way to trigger standards mode than <!DOCTYPE
# [14:10] <mikeday> then 9 pages out of 10 wouldn't trigger quirks mode... hmm.
# [14:10] <zcorpan> "9 out of 10 pages look broken in my browser. i better switch back to my old browser or to that other browser that works"
# [14:10] <zcorpan> ok, let's say 5 out of ten
# [14:10] <zcorpan> doesn't matter
# [14:11] <Dashiva> Even if it's just 1 out of 1000, it's probably a bank site
# [14:11] * mikeday grins
# [14:11] * Quits: myakura (myakura@60.239.122.32) (Ping timeout)
# [14:12] <mikeday> to be honest I don't think people switch browsers because pages look broken
# [14:12] <zcorpan> they do
# [14:12] <mikeday> most pages look broken by design, anyway
# [14:12] <Dashiva> Users switch browsers for the silliest reasons
# [14:12] <Dashiva> Well, typically switch back to the previous one, rather
# [14:12] <zcorpan> yes
# [14:12] <mikeday> seems like the primary reason for switching browsers is upgrading from Windows 98 to XP :)
# [14:13] <zcorpan> let's say firefox dropped quirks mode. the majority of pages wouldn't render correct. the users would either switch back to and older version of firefox or to internet explorer
# [14:14] <mikeday> right. but why drop the whole thing overnight?
# [14:14] <mikeday> what if Firefox dropped some of the more obscure corners of quirks mode
# [14:14] <zcorpan> why drop it at all?
# [14:14] <mikeday> most pages wouldn't look any different
# [14:15] <mikeday> "why drop it at all?" Question: would design of HTML5 be easier or harder if quirks mode didn't exist? :)
# [14:15] <zcorpan> there are billions of pages that rely on quirks mode. you can't drop quirks mode, even over a long period of time, because in 100 years from now you still want to render those pages that rely on the quirks
# [14:15] <zcorpan> we're stuck with it
# [14:15] <mikeday> "render those pages" => "render those pages pixel perfect with todays rendering" ?
# [14:15] * Joins: myakura (myakura@60.239.122.32)
# [14:16] <mikeday> given that fonts change, displays become higher resolution (or smaller, in the case of phones)
# [14:16] <mikeday> browser chrome changes, form controls change
# [14:16] <mikeday> browser default style sheets and user style sheets...
# [14:16] <zcorpan> as i said, anything that makes the new browser render the web worse than the old browser won't be accepted
# [14:16] <mikeday> "worse"? :)
# [14:16] <Dashiva> We need to give the guys at Web Repair Initiative root access to all servers with pages without a doctype!
# [14:17] <mikeday> eg. Safari has Aqua style form controls, looks different to MacIE
# [14:17] <zcorpan> this discussion is not productive. i'll go do something else
# [14:17] <mikeday> Firefox on Linux seems to change its default font settings with every distro upgrade
# [14:17] * mikeday apologises
# [14:18] <mikeday> quirks mode can stay, then :)
# [14:18] <zcorpan> :)
# [14:18] <mikeday> I'm just cranky because I'll probably have to implement the damn thing.
# [14:18] <zcorpan> i'll go do something else anyway :)
# [14:18] <zcorpan> mikeday: yes, but it would be simpler to implement if it was specced. then you wouldn't have to reverse engineer ie6
# [14:19] <mikeday> Indeed!
# [14:19] <zcorpan> so we agree then? :)
# [14:19] <mikeday> Now if only one or two of the 1000+ people who just joined the HTML WG was interested in specifying quirks mode! :)
# [14:19] * zcorpan is
# [14:19] <zcorpan> and i think Hixie is too
# [14:19] <zcorpan> probably annevk as well
# [14:20] <zcorpan> so there you go
# [14:20] <mikeday> awesome
# [14:20] <zcorpan> bbl
# [14:20] * mikeday waves
# [14:22] * Joins: loic (loic@90.27.88.46)
# [14:22] * Joins: marcos___ (chatzilla@203.206.31.102)
# [14:22] * marcos___ is now known as marcos
# [14:24] * Quits: mikeday (mikeday@144.136.3.123) (Quit: -)
# [15:01] * Quits: gsnedders (gsnedders@86.139.123.225) (Quit: gsnedders)
# [15:14] * Quits: marcos__ (chatzilla@203.206.31.102) (Ping timeout)
# [15:17] * Joins: gsnedders (gsnedders@86.139.123.225)
# [15:20] <anne> it's indeed unfortunate that the CSS WG still hasn't solved the table problem
# [16:06] * Joins: marcos__ (chatzilla@203.206.31.102)
# [16:16] * Quits: anne (annevk@83.82.206.111) (Ping timeout)
# [16:16] * Quits: marcos__ (chatzilla@203.206.31.102) (Ping timeout)
# [16:18] * Joins: alexf (alejandro@85.152.42.1)
# [16:46] * Quits: edas (edaspet@88.191.34.123) (Quit: http://eric.daspet.name/ et l'édition 2007 de http://www.paris-web.fr/ )
# [16:48] <Lachy> 727 emails this month on public-html - no wonder I'm having difficulty keeping up with it all!
# [16:53] * Lachy makes that 728. Sorry
# [16:55] <Lachy> aargh! I had to reply to someone with a silly white-list e-mail address autoresponder system :-(
# [17:00] <alexf> lachy: i usually spend almost my working time reading the emails from the list
# [17:02] <Lachy> I still have 16 left to read and and 218 on whatwg
# [17:02] <Lachy> and a few on other lists I just won't read
# [17:03] <krijnh> The versioning thread is interesting
# [17:03] <alexf> fortunately I'm not suscribed to whatwg (please don't tell to anyone) ;-)
# [17:06] <gsnedders> anyone: alexf isn't subscribed to whatwg! ;)
# [17:07] <krijnh> :O
# [17:08] * gsnedders runs
# [17:08] <alexf> ups, now my secret is not a secret, but one month ago I did't knew what's the meaning of whatwg, heheheh
# [17:11] * Quits: myakura (myakura@60.239.122.32) (Ping timeout)
# [17:11] <zcorpan> it's a synonym of whattf, which is an abbreviation of what the f...
# [17:12] <alexf> zcorpan: thanks for bringing light to me, hehe
# [17:13] <zcorpan> np
# [17:16] * Joins: myakura (myakura@60.239.122.32)
# [17:17] * Lachy takes a deep break, re-reads the rant he just wrote before subjecting it to the subscribers on the list...
# [17:17] * gsnedders braces himself to read the rant
# [17:17] <gsnedders> after… WHAT!? no new posts to the mailing list!
# [17:17] <gsnedders> I'm sorry, but that's just rare.
# [17:18] <Lachy> it's only short, but I'm just checking it can't be taken as an attack, or at least not too offensive)
# [17:19] <krijnh> At least nobody can fire you for it Lachy
# [17:19] * gsnedders should do something about the homework that he's had his arms leaning on while typing for the past twenty minutes
# [17:25] <alexf> Lachy: do you mean offensive for suscribers or just for a group of them?
# [17:27] <alexf> sorry, a group of us...
# [17:27] <Lachy> no, I mean offensive to Chris Wilson
# [17:30] <alexf> ok, maybe sometimes is not easy to represent ms, even with the ms hat off...
# [17:40] * Lachy made some changes and gets ready to send
# [17:43] <beowulf> the suspense is killing me
# [17:45] <Lachy> it's sent, you should have received it by now
# [17:45] <beowulf> my mail app is going for a dramtic pause
# [17:46] <beowulf> when they make the movie of the story of this working group, who do you see playing yourself?
# [17:50] <Lachy> ha! :-)
# [17:50] <Lachy> it would have to be someone tall
# [17:50] <alexf> maybe tom hanks with a fee of 50 million dollars
# [17:51] <Lachy> ... and sexy ;-)
# [17:51] <krijnh> *cough*
# [17:51] <Lachy> yeah, maybe Tom Hanks
# [17:51] <Mallory> <- Woody Allen
# [17:51] <gsnedders> I haven't got it either :P
# [17:52] <Lachy> it'd have to be someone with an aussie accent
# [17:54] <beowulf> russell crowe?
# [17:54] <alexf> gladiatorwg
# [17:54] <Lachy> yeah, maybe
# [17:55] <Lachy> anyone read my rant yet?
# [17:56] <krijnh> I did
# [17:56] <krijnh> I agree with you
# [17:56] <Lachy> I know you did, you got a sneak peak ;-)
# [17:56] <krijnh> But I don't get why Chris would overlook that
# [17:57] <krijnh> Any idea on that? :))
# [17:57] <Lachy> overlook what?
# [17:57] <krijnh> The fact that Opera, Safari and Fx can implement the spec without breaking the web
# [17:58] <krijnh> Or at least don't want versioning
# [17:58] <Lachy> I don't know, that's why I said it's illogical
# [18:02] * Parts: alexf (alejandro@85.152.42.1)
# [18:04] * Joins: anne (annevk@81.68.67.12)
# [18:05] * Joins: hober (ted@69.45.6.105)
# [18:05] * Joins: edas (edaspet@88.191.34.123)
# [18:09] * Joins: dbaron (dbaron@71.198.189.81)
# [18:09] * Quits: Yudai (Yudai@59.147.29.149) (Quit: SIGTERM received; exit)
# [18:09] * Joins: Yudai (Yudai@59.147.29.149)
# [18:22] <anne> Lachy, some good stuff in your interview
# [18:24] <Lachy> thanks
# [18:25] * anne had the same interview when figuring out how XHTML worked was hip
# [18:25] <Lachy> yeah, I know, I read that when preparing for my own
# [18:29] <anne> whoa, my todo list is amazingly small
# [18:29] <anne> guess i was productive today...
# [18:29] <anne> hmm
# [18:30] <anne> it's prolly more so that i didn't add new stuff, such as preparing my presentation
# [18:37] * Quits: myakura (myakura@60.239.122.32) (Ping timeout)
# [18:38] * Joins: myakura (myakura@60.239.122.32)
# [18:42] <krijnh> Best Viewed on IE 4 or Netscape 4.0 and above at 800x600
# [18:42] <krijnh> Wow, that still exists :|
# [18:43] <Lachy> krijnh, does the site still work in IE7?
# [18:43] <anne> http://www.google.com/search?q=%22best+viewed+with+internet+explorer%22
# [18:43] <krijnh> Lachy: yes
# [18:43] <anne> http://www.google.com/search?q=%22best+viewed+with+netscape%22 (half a million hits)
# [18:44] <krijnh> And in Opera, and in Firefox
# [18:44] <krijnh> Tight markup is overrated :)
# [18:44] <anne> valid markup is
# [18:44] <krijnh> Valid markup can still be very untight
# [18:53] <anne> jgraham, there's a point to be made though, as it kind of presumes you won't switch editors
# [18:55] <Lachy> jgraham, are there any editors that remember the cursor position each time you load a file? That seems unrealistic
# [18:57] * Joins: mw22 (chatzilla@63.245.220.228)
# [18:58] <edas> Lachy, I think vim remember the position in the file
# [19:00] <hober> heh. that's precisely why I make sure 'vi' invokes nvi, not vim, on my machines
# [19:18] * Joins: adele (adele@17.255.100.139)
# [19:28] * Quits: edas (edaspet@88.191.34.123) (Ping timeout)
# [19:29] * Joins: kingryan (rking3@66.92.187.33)
# [19:36] <jgraham> Lachy: Well MS Word does for Word files.
# [19:37] * Joins: edas (edaspet@88.191.34.123)
# [19:47] <anne> http://wiki.whatwg.org/wiki/Layout_tables
# [19:54] * Quits: myakura (myakura@60.239.122.32) (Quit: Leaving...)
# [19:55] * Quits: mw22 (chatzilla@63.245.220.228) (Ping timeout)
# [19:58] * Quits: dbaron (dbaron@71.198.189.81) (Quit: 8403864 bytes have been tenured, next gc will be global.)
# [20:00] <edas> anne, are we trying to correct a CSS lack with new html presentational markup ?
# [20:00] <edas> shouldn't we promote a correction in css ? (may be in another group)
# [20:00] <krijnh> anne: April 1 is over :)
# [20:01] <anne> I'm in complete agreement with you both :)
# [20:01] <anne> http://www.flickr.com/photos/tomascaspers/404571262/in/pool-htmljokes/ is also funny
# [20:03] <krijnh> Hehe
# [20:05] * Joins: mw22 (chatzilla@63.245.220.228)
# [20:22] * Joins: dbaron (dbaron@63.245.220.242)
# [20:33] * Quits: claudio (claudioc@89.97.35.74) (Quit: Leaving)
# [20:38] * Joins: Zeros (Zeros-Elip@67.154.87.254)
# [20:49] * Joins: DanC (connolly@128.30.52.30)
# [20:50] <DanC> I just updated the issues list a bit http://www.w3.org/html/wg/il16 and the WG homepage http://www.w3.org/html/wg/#current
# [20:51] <anne> Item 7 seems weird
# [20:51] <anne> "principle of modularity"
# [20:52] <DanC> it's perhaps a little wierd; I wasn't as careful when I added the 1st few...
# [20:53] <DanC> but modularity is a pretty well-known principle, no?
# [20:53] <anne> i don't particularly care for it
# [20:53] <DanC> http://en.wikipedia.org/wiki/Modularity_%28programming%29
# [20:53] <anne> in a spec, that is
# [20:54] <DanC> well, as chair, modularity is pretty darned valuable. otherwise I have to keep everything in my head all at once.
# [20:54] <anne> specs that consists of modules are mostly confusing and don't really work for the web as everyone has to support everything anyway
# [20:54] <DanC> everyone? I don't think so. technorati's search engines don't mess with javascript, for example.
# [20:55] <anne> that doesn't make it a separate module
# [20:55] <Zeros> I think modularizing the conformance requirements would be nice. ie. "If you are going to implement X you must implement Y"
# [20:55] <anne> that just makes it a different switch
# [20:55] <DanC> I sure hope we can specify parsing without reference to scripting. i'm still studying whether that's feasible/useful.
# [20:55] <anne> document.write() injects directly into the tokenizer
# [20:56] <Zeros> We know people won't implement the whole thing in one go, but if we group what parts should be implemented together...
# [20:56] <anne> innerHTML is intertwined with the parsing algorithm
# [20:56] <anne> (and needs to be)
# [20:56] <anne> Zeros, browsers vendors are capable of determining that themselves
# [20:56] <DanC> I wonder whether it really needs to be. technorati, google, etc. don't run scripts when they index the web, do they?
# [20:57] <anne> DanC, large part of HTML is rendering it
# [20:57] <Zeros> anne, clearly they aren't
# [20:57] <Zeros> anne, Why else would different portions of CSS2 be missing from different browsers
# [20:57] <anne> DanC, if scripts are disabled those parts of the parser are simply not actived (and don't have to be implemented)
# [20:57] <DanC> yes, there's lots of interesting stuff to do with HTML besides rendering it.
# [20:57] <Zeros> half the table model implemented in some
# [20:57] <Zeros> (css model)
# [20:57] <Zeros> partial print CSS support
# [20:57] <anne> Zeros, the CSS table model is not done yet
# [20:57] <anne> Zeros, CSS print is also quite vague I'd say
# [20:58] <Zeros> anne, the whole thing is fragmented. If you look at what gets added to Gecko and Webkit there's no order. Its not "lets get this and all related parts done" its all over the place
# [20:58] <anne> it's whatever their customers and developers feel like, yes
# [20:58] <DanC> if there's a spec for how to parse HTML with scripts disabled, that means the modules are separable.
# [20:58] <anne> you won't be able to change that
# [20:58] <anne> DanC, by writing the parsing spec twice?!
# [20:59] <anne> DanC, seems too much extra work for almost no benefit
# [20:59] <anne> no benefit even
# [20:59] <DanC> how you package the writing is up to the editor.
# [20:59] <anne> then i'm not sure what your point is
# [20:59] <Zeros> Why not just express the parser with abstract bindings for innerHTML and write() like functionality
# [20:59] <Zeros> you don't have to mention those directly
# [21:00] <DanC> replay: I sure hope we can specify parsing without reference to scripting. i'm still studying whether that's feasible/useful.
# [21:00] <anne> if you can rewrite the parsing section to do that and have it still be clear...
# [21:01] <DanC> if it's too much work to actually separate the specs, other stuff can still be separated; e.g. testing work
# [21:01] <DanC> the html5lib python parsing codes doesn't include a javascript interpreter, does it?
# [21:01] <DanC> s/codes/code/
# [21:01] <anne> nope, but we wrote a simple library, not a browser
# [21:02] <anne> part of parsing is the script execution model and events, etc.
# [21:02] <DanC> I hear claims that it implements the HTML5 tree generation algorithm. if that's true, then the tree generation algorithm is separate (or at least separable) from document.write() and the ECMAscript spec.
# [21:02] <Zeros> You could provide a binding for SpiderMonkey or KJS
# [21:03] <Zeros> Then you'd have both
# [21:03] <Zeros> Does Opera's JS runtime have a separate name outside "Presto"?
# [21:04] <anne> DanC, we implement the innerHTML feature too
# [21:04] <anne> DanC, sure, you don't have to support scripting to become a conforming parser
# [21:04] <anne> DanC, scripting could be disabled in browsers as well
# [21:05] <DanC> (btw... http://en.wikipedia.org/wiki/Modularity_%28programming%29 looks pretty bad.)
# [21:05] * Quits: adele (adele@17.255.100.139) (Quit: adele)
# [21:05] <anne> Zeros, don't think so
# [21:05] <Zeros> ah
# [21:06] <anne> re: the name of our script engine
# [21:06] <Zeros> yeah
# [21:06] <mjs> DanC: there is a no-script mode for parsing, but when you have scripting support it has to interact with parsing in a deep way
# [21:07] <DanC> maybe some enterprising person will be willing to extract the no-script mode as an informative note.
# [21:08] * DanC is pretty depressed that HTML as become turing-complete
# [21:08] <anne> Wouldn't you want to have innerHTML in there for serialization and parsing of subtrees?
# [21:08] <mjs> I think the highest value things to potentially separate would be things that other specs may want refer to in a non-HTML context
# [21:08] <anne> oh, the document.write() stuff
# [21:09] <anne> just the serialization that is turing complete
# [21:10] <anne> which in turn, is just a string, just like xml
# [21:10] <anne> you prolly can't call the language turing-complete
# [21:10] <mjs> for instance, if there is some API that you can imagine also being useful with SVG or MathML somehow, then that might be a candidate to split out
# [21:11] <mjs> the parsing alone might be useful for specialized html implementations, but it's probably not useful for anything non-HTML
# [21:11] <zcorpan> even if some things are usable by other languages, they can just use it and reference the html spec
# [21:11] <anne> parsing and writing it just applies to text/html
# [21:12] <anne> the rest of the language applies to every DOM thingie
# [21:14] <DanC> oh my... a "Formal definition of HTML5" thread. Let's discuss everything all at once! oh well.
# [21:14] <anne> still catching up?
# [21:15] <krijnh> "I'd like to enable web developers spend as near as zero time as possible fixing browser interoperability problems." - how ironic
# [21:15] <anne> whoa, 739 messages already
# [21:15] <DanC> catching up? I read in fairly random order.
# [21:16] <anne> krijnh, did cwilso say that?
# [21:16] <krijnh> anne: yeah
# [21:16] <anne> heh
# [21:17] <mjs> I was gonna make a request to start the ball rolling on formally adopting the Design Principles, but I didn't want to add too much noise to the list
# [21:17] <krijnh> He's the only one sending mails atm :)
# [21:17] <DanC> you think it's time to put the question on design principles, mjs? I guess I better finish reviewing them soon
# [21:17] <krijnh> How many +1's do 'we' need before the WHATWG drafts are adopted?
# [21:18] <anne> prolly a +1 from MS and the W3C maybe
# [21:18] <krijnh> K, so mine isn't missed :)
# [21:18] <zcorpan> why does he resist on the idea that ie would have to remove classid and activex just because html5 doesn't define them?
# [21:18] <DanC> I'd like to hear from Nokia and Boeing on HTML5.
# [21:18] <anne> DanC, Nokia already +1'd it
# [21:19] <DanC> oh? I missed that; pointer, anne ?
# [21:19] <DanC> ah... found it
# [21:19] <anne> twice
# [21:19] <anne> http://lists.w3.org/Archives/Public/public-html/2007Apr/0719.html
# [21:19] <anne> http://lists.w3.org/Archives/Public/public-html/2007Apr/0720.html
# [21:19] <zcorpan> haven't we pointed out enough that ie can support classid and activex without the spec saying anything about it?
# [21:19] <DanC> yes, there was some trouble connecting that mailbox to the list.
# [21:20] <DanC> mjs, have you seen any ammendments to the HTML5 proposal that you're inclined to adopt?
# [21:20] <anne> (the guy who wrote that was / is heavily involved in XForms)
# [21:20] <mjs> DanC: there isn't much objection these days, although I do have one email objecting to the idea of having any Design Principles in general
# [21:20] <mjs> DanC: let me get caught up on email before I answer that
# [21:21] <DanC> I'd like to have more than one editor
# [21:22] <mjs> are there any other nominees for editor?
# [21:22] <DanC> I have written and deleted several "we are the working group" draft messages... people seem to think this is some sort of customer service mailing list. "One <indent> tag, please"
# [21:22] <mjs> I think we'd need to evaluate specific proposed editor teams, not just the idea of having more than one in the abstract
# [21:23] <DanC> I have talked to one or two editor nominees, but this eye injury got in the way this week.
# [21:24] <mjs> personally, given Ian's stated preference for working alone and demonstrated ability to be effective doing so, I'd rather have just him as editor, and nominate other people for other roles, like managing tracking of incoming issues, editing the test suite, developing an agenda to guide mailing list discussion
# [21:24] <anne> are there some good reasons for having multiple editors?
# [21:24] <zcorpan> mjs: agreed
# [21:24] <krijnh> I also agree with mjs
# [21:25] <krijnh> Although Hixie shouldn't break a leg or something :)
# [21:25] <mjs> but if anyone wants to publicly nominate a different editor or editing team, I would consider it
# [21:25] <DanC> yes, the "if he gets hit by a bus" risk is one thing, though that's fairly remote...
# [21:25] * zcorpan thought he was immortal :D
# [21:25] <DanC> and having other people in supporting roles is likely to work
# [21:25] <krijnh> He has 9 lives, no?
# [21:26] <mjs> I think some of the other roles above require enough expertise in the spec to do well, that people doing them could make good successors as editor if needed
# [21:27] <DanC> his frank style has caused a number of W3C collaborators to stop sending comments on specs that he edits. "Why send comments? they'll just get declined/ignored" is what I hear.
# [21:27] <anne> it's hard to beat someone who does it as his fulltime job and has the kind of QA expertise he has
# [21:27] <anne> DanC, for XBL for instance?
# [21:27] <DanC> I think so, anne
# [21:28] * anne thinks Hixie does a far better job than say, SVG...
# [21:28] <anne> He declined several requests, but all for good reasons in which the WAF WG supported him
# [21:29] <anne> The number of edits he made in response to comments (and satisfied comments) far outnumbered that though
# [21:29] <anne> ( http://dev.w3.org/cvsweb/~checkout~/2006/xbl2/disposition-of-comments has the history of the comments)
# [21:30] <mjs> DanC: I've never felt ignored, even when he disagreed with me, unlike with, say, SVG or CDF
# [21:30] <DanC> I have no reason to object to him as the only editor. but I'd prefer to have some redundancy.
# [21:30] <mjs> but that could be partly due to my commenting style
# [21:31] * DanC is still looking into some SVG process issues.
# [21:31] <anne> have fun :)
# [21:31] <mjs> I tend to enumerate the details for why something doesn't work, then propose at least one viable alternative that I think does suffice
# [21:31] * Joins: hasather_ (hasather@81.235.209.174)
# [21:32] <mjs> so unless that either leads to a change or is replied to with clear reasoning, it can feel like a waste of time
# [21:32] <mjs> on the other hand, sometimes I see other people ask for a change without giving clear justification
# [21:32] <DanC> mjs, please train all the HTML WG members to "propose at least one viable alternative" when they argue against something. That should just take a couple minutes. ;-)
# [21:33] <DanC> and now for something almost completely different: XHTML 1.1 Basic. it's been in last call for a while. I'd like the HTML WG to review it.
# [21:33] <mjs> DanC: leading by example is not very effective in creating such a norm, alas
# [21:33] <DanC> there are worse ways than leading by example
# [21:34] <anne> so why does the XHTML2 WG still own XHTML1 specs?
# [21:34] <anne> and how does that work with the HTML namespace?
# [21:34] <krijnh> Shall we continue in #xhtml ?
# [21:34] <mjs> I think my point of review would be "XHTML profiles are harmful, please don't publish this"
# [21:34] <anne> i'd rather not
# [21:35] <DanC> the XHTML2 WG still own XHTML1 specs as a result of the rather messy chartering process that I would like to now leave behind us.
# [21:36] <anne> if we want an XML serialization of HTML5 without silly objections from that group we should own http://www.w3.org/1999/xhtml by now
# [21:36] <DanC> anne, the current status is that both groups are subject to objections from the other.
# [21:37] <mjs> I don't think they will complain, unless they want to reuse that namespace for XHTML2
# [21:37] <anne> they do
# [21:37] <DanC> silly objections are a waste of everybody's time, but thoughtful objections will get reviewed by The Director
# [21:37] <mjs> so someone should talk them off the ledge on that
# [21:37] * mjs does not volunteer
# [21:37] <anne> they also have some hairy DOCTYPE based versioning strategy for all their XHTML 1.x sillyness
# [21:38] <DanC> anne, I don't understand why they're still using DOCTYPE stuff, but the mobile world is pretty interested in XHTML Basic 1.1, and I'm trying to learn why.
# [21:38] * Quits: ROBOd (robod@86.34.246.154) (Quit: http://www.robodesign.ro )
# [21:39] <zcorpan> DanC: i'd like to know as well
# [21:39] <krijnh> Me too
# [21:39] <DanC> with more mobile browsers than desktops (as of 2006, I gather), I'd like to learn more.
# [21:39] <DanC> that's why I'm excited that Nokia has joined. I'm recruiting other mobile folks.
# [21:40] <DanC> so as politically uncomfortable as it may be, I'd like this WG to take a serious look at XHTLM Basic.
# [21:40] <anne> Openwave joined
# [21:40] <anne> we joined
# [21:40] <anne> would be nice if Acess did I suppose
# [21:40] <Zeros> DanC, assuming the content is valid XML the parser can be quite a bit simpler, and there's already plenty of XML parsers available for Brew, J2ME, etc.
# [21:40] <anne> Access, even
# [21:40] <Zeros> Are they backing Basic 1.1 with intent to support soup though?
# [21:40] <DanC> yes, if you have contacts at Access, do nudge them, please.
# [21:41] <anne> i don't
# [21:41] <mjs> I think it's becoming increasingly clear that mobile devices can support real browsing
# [21:41] <anne> Zeros, all XHTML on phones is tag soup
# [21:41] <mjs> and even for those that can't, I am not sure a profile helps interoperability
# [21:41] <anne> for those that can't: Opera Mini
# [21:41] <DanC> I don't know, Zeros; I hear both "the mobile world is XML-clean" and "the mobile world follows MS IE bug-for-bug". I'm trying to figure out which is true. Maybe both are true, to some extent.
# [21:42] <anne> XML-clean is not true, at least not according to the OMA
# [21:42] <Zeros> DanC, possibly. I don't know enough about mobile browsers to make an educated comment on the state of that market.
# [21:42] <anne> I don't think anything on the web is "XML" clean except perhaps some WS stuff which suffers from so many interop problems that properly testing XML support prolly hasn't been on the agenda
# [21:43] * DanC is late getting some lunch
# [21:45] <hober> re: xhtml-on-phones, http://simon.html5.org/articles/mobile-results (in case someone hadn't seen this already)
# [21:51] * Quits: mjs (mjs@64.81.48.145) (Quit: mjs)
# [21:58] <anne> So Microsoft versioning is like:
# [21:58] <anne> 1. We implement HTML5.
# [21:58] <anne> 2. We ship with <!doctype html5> which triggers HTML5-mode.
# [21:58] <anne> 3. We'll fix bugs.
# [21:58] <anne> 4. If more than .5% of the content uses <!doctype html5> we stop fixing bugs.
# [21:58] <anne> Is that correct?
# [21:59] <anne> i'll ask on the list
# [22:01] <edas> could someone give me some advice about http://eric.daspet.name/htmlwg/opt-in.txt ?
# [22:01] <edas> anne, I've undertood that <!doctype html> trigger standard mode in today IE but <!doctype html5> don't. We surely want to trigger the today standard mode
# [22:02] <zcorpan> edas: <!doctype html5> triggers standards mode in ie but not in firefox or safari
# [22:03] <anne> edas, that's not what this question is about
# [22:03] <edas> but appart from that, I've understand the same about IE, with the add of : 5. we will find another opt-in mecanism to correct the bugs after point 4, and redo point 3 to 5"
# [22:04] <edas> zcorpan, thanks for the clarification
# [22:05] <anne> added
# [22:05] <anne> "5. Back to step one incrementing the HTML version number by one."
# [22:09] <zcorpan> the entire versioning concept makes it orders of magnitude harder for new vendors to enter the market, or to write a browser from scratch in 500 years in order to read today's content
# [22:09] <anne> no news to me
# [22:10] <DanC> that argument appeals to me, but I'm still studying it.
# [22:11] * anne wants to improve HTML, not add another 50 variants
# [22:11] <anne> undocumented Microsoft specific variants, even
# [22:12] <edas> zcorpan, I agree, but if we trash the versionning concept, we may be pragmatics about the "don't break the web principle"
# [22:13] <anne> i think that's included in the principle
# [22:14] <anne> in the used definition, that is
# [22:16] * anne wonders who oedipus is
# [22:17] <Zeros> from greek mythology?
# [22:17] <kingryan> DanC: about "technorati's search engines don't mess with javascript, for example."
# [22:18] <anne> from the mailing list
# [22:18] <kingryan> DanC: I'd actually like to implement document.write() in our parser and may do so based on the current WHAT-WG spec
# [22:19] <DanC> do tell. why?
# [22:19] <anne> to support the web?
# [22:19] <anne> document.write() contains content :(
# [22:19] <anne> (also ads, but also content)
# [22:21] <DanC> I'd rather not reward those who hide content inside scripts. And I'd like to think there's little enough of it that technorati wouldn't be bothered to index it.
# [22:21] <DanC> but I'm prepared to be further educated/depreseed.
# [22:22] <kingryan> DanC: I'd rather not "reward those who hide content inside scripts" either, which is what we currently do by *not* handling it
# [22:23] <kingryan> I want to support a subset of scripting for 2 reasons–
# [22:23] <DanC> oh joy.
# [22:23] <kingryan> 1. to make our spider's view more consistent with what people see in their browsers
# [22:23] <kingryan> 2. to fight spam :D
# [22:24] <DanC> I wonder if you could elaborate, maybe in a blog artcile, about this sort of spam
# [22:24] <kingryan> unfortunately, I can't elaborate too much
# [22:24] <kingryan> though there is some good research in AIRWeb
# [22:25] <DanC> "AIRWeb"? new to me. googling...
# [22:25] <DanC> http://airweb.cse.lehigh.edu/
# [22:26] <DanC> that one?
# [22:26] <kingryan> http://www2007.org/workshop-W1.php
# [22:26] <kingryan> yeah, same thing
# [22:27] <anne> spam pages hide their content from search engines but not from users or something?
# [22:27] <kingryan> right
# [22:27] <anne> and so by supporting script you can see the spam too and ban them?
# [22:27] <kingryan> yup
# [22:27] <anne> interesting
# [22:27] <anne> these things used to work the other way around :)
# [22:28] <DanC> depressing. but unstoppable.
# [22:28] <kingryan> well, they work the other way, too. some site will send different content to spiders than they do to browsers
# [22:28] <kingryan> sites*
# [22:28] <anne> yeah, bit of an overstatement from me
# [22:28] <DanC> I wonder how much greed the internet can survive.
# [22:28] <Zeros> I'm pretty sure most search engines ban if you report it
# [22:29] <kingryan> Zeros: yes, but that requires human review, which is much more expensive than CPU cycles
# [22:29] <anne> DanC, there's lots of openness between all the spam
# [22:30] <Zeros> kingryan, does google automate any of that?
# [22:30] <kingryan> no idea, but it wouldn't surprise me
# [22:30] <DanC> openness? I'm not sure what you mean.
# [22:30] <anne> open source projects, wikipedia, etc.
# [22:33] <DanC> yes, the internet is great when people want to work together... but I wonder how much bad-actor stuff it can survive before tragedy-of-the-commons hits
# [22:34] <kingryan> DanC: I think the internet can survive a lot more. it's value is so great that "bad actors" have a lot of work to do to ruin it
# [22:34] * DanC adds security review to http://esw.w3.org/topic/HtmlTaskBrainstorm
# [22:34] <anne> security isn't done yet
# [22:34] <DanC> I wonder... what's the state of the art in free email? last time i created a yahoo inbox, it was totally overrun with drek in a few weeks.
# [22:35] <anne> gmail?
# [22:35] <kingryan> gmail is quite good at spam filtering
# [22:35] <DanC> I now use gmail for my personal mailbox; it was overrun with spam even though it was unpublished
# [22:35] <Zeros> My gmail box is much better at filtering than yahoo
# [22:36] <DanC> it took me a while to figure out google-apps-for-your-domain, but it works.
# [22:36] <Zeros> lots of stuff seems to slip through yahoo's filters
# [22:36] <anne> DanC, overrun with spam means spam in your inbox right (not filtered away but still lots)?
# [22:36] <DanC> right; in the inbox
# [22:37] <anne> maybe it takes a while to learn what you consider spam...
# [22:37] * anne almost never has it
# [22:37] <anne> and my e-mail address is on my weblog
# [22:37] <DanC> W3C spends zillions of CPU/person hours fighting spam every day.
# [22:38] <anne> well, i've 1700 spam in the filter
# [22:38] <anne> and about 40000 e-mails indexed
# [22:38] <DanC> the fundamental promise of SMTP is long gone, i.e. that your mail will either get delivered or you'll get a bounce.
# [22:39] <zcorpan> my email address is publicly available, without being escaped or anything. i almost never get spam in my inbox
# [22:39] <edas> lucky you
# [22:39] <anne> oh, mine is escaped
# [22:40] <anne> using traditional hexadecimal character references
# [22:40] <krijnh> Opera M2 is also pretty good at filtering
# [22:40] <anne> yeah
# [22:40] * anne uses M2 for his work account
# [22:40] <zcorpan> i've got 25 spam in my spambox in 2 days
# [22:40] <krijnh> Even my mother likes it
# [22:40] <anne> hehe
# [22:40] <krijnh> Who used Outlook before
# [22:41] <kingryan> DanC: nasty (probably NSFW) javascript spam example http://www.instantboards.com/?mforum=girlsflashingin (though I don't know the point of this spam)
# [22:42] <krijnh> Note to self; add rel="nofollow" to all links on /irc-logs/ :+
# [22:42] * Joins: adele (adele@17.255.100.139)
# [22:42] <kingryan> krijnh: good idea :D
# [22:42] <kingryan> DanC: to figure out what the javascript does on that page, I'd need to execute it
# [22:43] <krijnh> Done
# [22:44] <krijnh> No more taking advantage of my huge PR
# [22:44] <krijnh> Hrhr
# [22:49] <krijnh> http://annevankesteren.nl/2005/07/html5-doctype#comment-4391 - "There's no versioning problem."
# [22:50] <zcorpan> my t-shirt has been sent
# [22:50] <anne> you got it already?
# [22:50] <zcorpan> no
# [22:50] <zcorpan> got an email saying that they have sent it
# [22:51] <Zeros> kingryan, couldn't you just create an instance of Gecko, render the buffer without a window and grab the document.documentElement.innerHTML ?
# [22:51] <anne> oh, duh
# [22:56] * Joins: Roger (roger@213.64.74.230)
# [22:57] <hasather> hey Roger
# [22:57] <Roger> hey
# [22:57] * Joins: foca (foca@190.64.4.27)
# [22:58] <Roger> anybody else around?
# [22:58] <krijnh> o/
# [22:58] <kingryan> Zeros: in theory, yes, though that might not be the best solution for us
# [22:58] <anne> good evening
# [22:58] <kingryan> DanC: more good js spam examples: http://06-gay-pride-day-in-atlanta-blog.blogspot.com/ , http://13trudysfantasynovel.blogspot.com/
# [22:59] <kingryan> they use js to render a captcha over the content (presumably for distributed captcha breaking)
# [23:00] <anne> you could just *.blogspot.com -> /dev/null
# [23:00] <anne> 80% accurate
# [23:00] <krijnh> Hehe
# [23:01] <Roger> i do that ;-)
# [23:03] * Quits: zcorpan (zcorpan@84.216.42.156) (Ping timeout)
# [23:04] <Roger> so... I have been trying to read and understand the huge amount of messages on the mailing list...
# [23:05] <Roger> but the discussions seem beyond my capacity to understand
# [23:06] <anne> any particular discussion?
# [23:06] <Roger> "don't break the web" in particular
# [23:06] <Roger> i don't understand how html 5 could break current tag soup sites
# [23:07] <Roger> unless all html documents automagically are treated as html 5 once the spec is released
# [23:07] <anne> the catch is the case
# [23:08] <anne> browsers treat all HTML already identically
# [23:08] <anne> DOCTYPEs, for one, are ignored
# [23:08] <anne> (except for triggering some compat mode)
# [23:08] <anne> (known as standards, almost standards and quirks mode)
# [23:08] <Roger> right
# [23:09] <krijnh> I think Mr Johansson knows that already?
# [23:09] <Roger> hehe. yes
# [23:09] <anne> so if we define parsing for HTML it should not break compat
# [23:09] <anne> for instance
# [23:09] <anne> <b> <i> </b> should not get a different tree than it gets now
# [23:09] <anne> the problem is that UAs treat it differently now, etc.
# [23:10] <anne> and that authors do UA sniffing which cuases another bunch of breaking the web cases
# [23:10] <Roger> hmm.
# [23:11] <krijnh> Authors sniffing UAs shouldn't be using the HTML5 doctype then
# [23:11] <anne> that doesn't help
# [23:11] <anne> you have content A and B and UA C and D
# [23:11] <Roger> but won't browsers continue to treat tag soup as tag soup just the way they do now?
# [23:11] <anne> content A is compatible with C and incompatible with HTML5
# [23:12] <anne> content B is compatible with D and with HTML5
# [23:12] <anne> browser C implements HTML5
# [23:12] <anne> Roger, we want to standardize tag soup parsing
# [23:12] <Roger> right
# [23:13] <anne> Roger, to ensure that the information we accumulate can be read centuries later
# [23:13] <Roger> so once a browser that supports html 5 is released, it has to treat all content as html 5?
# [23:13] <anne> yes
# [23:13] <Roger> ah
# [23:13] * Joins: claudio (claudioc@89.97.35.74)
# [23:13] <anne> that's the goal anyway
# [23:13] <Roger> and... why is that?
# [23:14] <Roger> because that's what would break stuff, right?
# [23:14] <anne> see above
# [23:14] <anne> to ensure we can read the information later
# [23:14] <anne> to ensure that it doesn't get out of hand, etc.
# [23:14] <anne> if we introduce yet another switch there's even more to fix
# [23:14] <Roger> but how will that prevent people from continuing to produce crap html?
# [23:15] <krijnh> Crap html will we specced
# [23:15] <krijnh> At least the parsing bit
# [23:16] <anne> as we basically would have to document every switch UAs make
# [23:16] <anne> not documenting one means information loss
# [23:16] <beowulf> it won't, i don't think
# [23:19] <krijnh> Hmm, gf beats html discussions, good night people
# [23:20] * Disconnected
# [23:20] * Attempting to rejoin channel #html-wg
# [23:25] * Attempting to rejoin channel #html-wg
# [23:25] * Rejoined channel #html-wg
# [23:25] * Topic is 'W3C HTML WG http://www.w3.org/html/wg/ - http://krijnhoetmer.nl/irc-logs/ (logged) - http://esw.w3.org/topic/HTML/ProposedDesignPrinciples'
# [23:25] * Set by anne on Tue Mar 27 12:28:46
# [23:26] <zcorpan> Roger: it probably doesn't. how could it? should it?
# [23:26] * Joins: anne (annevk@81.68.67.12)
# [23:26] * Joins: Hixie (ianh@129.241.93.37)
# [23:26] <zcorpan> (is this channel unstable right now or what?)
# [23:26] <krijnh> It is
# [23:26] <krijnh> So much for logging :)
# [23:26] <zcorpan> indeed
# [23:27] <krijnh> Anyway, nn
# [23:27] <zcorpan> nn
# [23:27] <Roger> i don't think it could... but I also don't see how inculding bad practices in the spec helps
# [23:27] * Joins: deltab (deltab@82.46.154.93)
# [23:27] <zcorpan> i don't understand, what are the bad practices that are being included in the spec?
# [23:29] <anne> and adding another switch increases code size complexity, etc.
# Session Close: Sat Apr 14 00:00:00 2007

The end :)