/irc-logs / freenode / #whatwg / 2007-12-20 / end

Options:

# Session Start: Thu Dec 20 00:00:00 2007
# Session Ident: #whatwg
# [00:07] * Philip` wonders if the IE8 Acid2 thing makes Opera's antitrust complaint significantly less valid
# [00:20] <roc> not really
# [00:20] <roc> Acid2 is just a slice of a very large standards pie
# [00:24] * Quits: psa (n=yomode@71.93.19.66) ("All things must come to an end")
# [00:25] * Joins: psa (n=yomode@71.93.19.66)
# [00:32] <othermaciej> Microsoft themselves used to downplay the importance of Acid2 quite a bit
# [00:32] <Hixie> tehy still do
# [00:33] <Hixie> even in that post they (correctly) point out it's just one test of many
# [00:33] <Hixie> and not an especially good test from a standadrs-compliance perspective
# [00:33] <Hixie> acid3 is mildly better in that regard, as constructed at the moment
# [00:37] * Joins: heycam` (n=cam@124-168-10-57.dyn.iinet.net.au)
# [00:37] * Quits: heycam (n=cam@203-217-74-244.dyn.iinet.net.au) (Nick collision from services.)
# [00:37] * heycam` is now known as heycam
# [00:49] * Joins: jwalden_ (n=waldo@RANDOM-SEVENTY-TWO.MIT.EDU)
# [00:52] * Quits: jwalden (n=waldo@STRATTON-SIX-NINETY-EIGHT.MIT.EDU) (Read error: 110 (Connection timed out))
# [00:52] * jwalden_ is now known as jwalden
# [01:17] * Quits: Lachy (n=Lachlan@cm-84.215.41.149.getinternet.no) ("This computer has gone to sleep")
# [01:18] * Quits: billmason (n=billmaso@ip156.unival.com) (".")
# [01:24] * Quits: hasather (n=hasather@90-231-107-133-no62.tbcn.telia.com) ("leaving")
# [01:26] * Quits: othermaciej (n=mjs@17.255.106.224)
# [01:29] * Quits: tndH (i=Rob@adsl-87-102-85-140.karoo.KCOM.COM) ("ChatZilla 0.9.79-rdmsoft [XULRunner 1.8.0.9/2006120508]")
# [01:31] * Joins: Lachy (n=Lachlan@pat-tdc.opera.com)
# [01:34] * Joins: Thezilch (n=fuz007@wsip-64-147-23-102.oc.oc.cox.net)
# [01:56] * Quits: Lachy (n=Lachlan@pat-tdc.opera.com) ("This computer has gone to sleep")
# [01:59] * Quits: phsiao (n=shawn@nat/ibm/x-eb22528e2c412ab4)
# [02:13] * Joins: hdh (n=hdh@58.187.109.11)
# [02:14] * Joins: Lachy (n=Lachlan@cm-84.215.41.149.getinternet.no)
# [02:31] * Joins: weinig (n=weinig@17.203.15.140)
# [02:31] * Quits: weinig_ (n=weinig@17.203.15.140) (Read error: 104 (Connection reset by peer))
# [02:33] * Joins: othermaciej (n=mjs@dsl081-048-145.sfo1.dsl.speakeasy.net)
# [02:34] * Quits: kingryan (n=kingryan@dsl092-219-050.sfo1.dsl.speakeasy.net)
# [02:35] * Quits: dbaron (n=dbaron@corp-241.mountainview.mozilla.com) ("8403864 bytes have been tenured, next gc will be global.")
# [02:40] * Joins: dglazkov (n=dglazkov@adsl-074-229-248-021.sip.bhm.bellsouth.net)
# [02:54] * Quits: Thezilch (n=fuz007@wsip-64-147-23-102.oc.oc.cox.net) (Read error: 110 (Connection timed out))
# [02:58] * Quits: wakaba (n=w@79.163.210.220.dy.bbexcite.jp) (Read error: 110 (Connection timed out))
# [03:03] * Quits: weinig (n=weinig@17.203.15.140)
# [03:16] * Joins: phsiao (n=shawn@c-71-232-12-131.hsd1.ma.comcast.net)
# [03:53] <Hixie> hsivonen: yt?
# [03:55] <Hixie> what should the conformance criteria for <iframe>'s contents be?
# [03:55] <Hixie> especially given validation with live <script> elements, and the dynamic validity of <iframe> elements added to the dom in a browser on the fly
# [04:03] <webben> It's interesting that IEBlog commentators often call for MS to implement HTML5, including stuff like Web Forms 2.0 and VIDEO. Just wondering whether people think that would be a good thing (because there would be more implementor feedback) or a bad thing (since effectively it would prevent the spec being changed).
# [04:03] <Hixie> microsoft implementing html5 would be a good thing, especially if they actually gave us feedback
# [04:31] * Hixie prepares a 6000 line patch for the spec
# [04:32] <Hixie> -1037/+851
# [04:36] * Joins: Thezilch (n=fuz007@wsip-64-147-23-102.oc.oc.cox.net)
# [04:37] <Hixie> "block" and "inline" are a thing of the past
# [04:40] * Joins: dbaron (n=dbaron@c-71-204-145-103.hsd1.ca.comcast.net)
# [04:45] * Joins: Thezilch[FH] (n=fuz007@wsip-64-147-23-102.oc.oc.cox.net)
# [04:45] * Quits: Thezilch (n=fuz007@wsip-64-147-23-102.oc.oc.cox.net) (Read error: 104 (Connection reset by peer))
# [04:49] <takkaria> fun
# [04:50] * Quits: dglazkov (n=dglazkov@adsl-074-229-248-021.sip.bhm.bellsouth.net)
# [04:51] <Hixie> i hope someone checks that patch for mistakes
# [04:51] <Hixie> my eyes glaze over every time i look at it
# [04:51] <Hixie> maybe i should go eat
# [05:14] * Quits: othermaciej (n=mjs@dsl081-048-145.sfo1.dsl.speakeasy.net)
# [05:14] * Joins: weinig (n=weinig@c-71-198-176-23.hsd1.ca.comcast.net)
# [05:23] * Heme|Sleep is now known as Hemebond
# [05:47] * Quits: doublec (n=doublec@202.0.36.64)
# [06:00] * Quits: jruderman (n=jruderma@corp-241.mountainview.mozilla.com)
# [06:05] * Quits: weinig (n=weinig@c-71-198-176-23.hsd1.ca.comcast.net)
# [06:17] * Quits: Thezilch[FH] (n=fuz007@wsip-64-147-23-102.oc.oc.cox.net) (Read error: 104 (Connection reset by peer))
# [06:17] * Joins: Thezilch (n=fuz007@wsip-64-147-23-102.oc.oc.cox.net)
# [06:20] * Quits: roc (n=roc@202.0.36.64)
# [06:31] * Joins: jruderman (n=jruderma@c-67-180-15-227.hsd1.ca.comcast.net)
# [07:15] * Joins: kfish (n=conrad@61.194.21.25)
# [07:25] * Quits: Thezilch (n=fuz007@wsip-64-147-23-102.oc.oc.cox.net) (Read error: 110 (Connection timed out))
# [07:27] * Joins: weinig (n=weinig@c-71-198-176-23.hsd1.ca.comcast.net)
# [07:40] <Hixie> so, anyone read the changes? :-)
# [07:40] <Hemebond> Nope.
# [07:41] <Hemebond> What did you change?
# [07:42] * Joins: maikmerten (n=merten@ls5laptop14.cs.uni-dortmund.de)
# [07:42] <takkaria> Hixie: I did
# [07:42] <takkaria> seemed reasonable
# [07:42] <takkaria> but I'm an armchair spectator wrt HTML5, I don't do much web work anymore...
# [07:43] <Hixie> takkaria: hehe
# [07:43] <Hixie> Hemebond: the content models
# [07:43] <hdh> Hemebond: diff http://html5.org/tools/web-apps-tracker?from=1151&to=1152
# [07:43] <Hemebond> ""block" and "inline" are a thing of the past"
# [07:43] <Hemebond> WTF
# [07:44] <takkaria> nothing glaring popped out at me, but I think that perhaps some kind of diagram might be useful for us visual types
# [07:47] <Hixie> yeah
# [07:47] <Hixie> diagrams are hard, though, don't wanna do that until i'm sure it's ok :-)
# [07:47] <Hixie> also i'm not exactly sure what to diagram
# [07:48] * takkaria nods
# [07:50] <takkaria> on not knowing what to diagram, I have no suggestions there, either. there's certainly no rush to create them, the prose is quite readable
# [07:55] <Hixie> cool
# [07:55] <Hixie> what i really want is hsivonen's opinion
# [07:55] <Hixie> since he's implementing this stuff
# [08:01] <Hixie> so looking at the acid2 video from microsoft we see some interesting things
# [08:01] <Hixie> they implemented data: really early on in the process
# [08:02] <Hixie> also for a long time they had the eyes rendering twice
# [08:04] * Quits: dbaron (n=dbaron@c-71-204-145-103.hsd1.ca.comcast.net) ("8403864 bytes have been tenured, next gc will be global.")
# [08:17] <Hemebond> IE8 passes the Acid2 test?
# [08:18] <Hemebond> That's.... good right?
# [08:22] <Hixie> yup
# [08:23] <Hemebond> Does the Acid2 test use all the spec? Or just some bits.
# [08:29] <Hixie> a tiny part
# [08:30] <Hemebond> Thought so.
# [08:53] * Joins: othermaciej (n=mjs@dsl081-048-145.sfo1.dsl.speakeasy.net)
# [08:57] * Joins: roc (n=roc@121-72-28-191.dsl.telstraclear.net)
# [09:10] <hsivonen> Hixie: I'm not familiar enough with <iframe> content usage patterns to say what should be conforming.
# [09:13] * hsivonen sees the comeback of significant inline in a light uncheckable form
# [09:24] <inimino> "When a transparent or semi-transparent element has no parent" <-- when would that happen?
# [09:26] <hsivonen> inimino: in DOM fragments
# [09:27] <inimino> ah, I see
# [09:28] <Hemebond> But... would they not still have a parent when being displayed?
# [09:28] <hsivonen> Hemebond: no, they get a parent when inserted into the document
# [09:29] <Hemebond> Yeah...
# [09:29] <hsivonen> Hixie: did you ping glazou about the content model changes?
# [09:29] <Hemebond> How do you display a DOM fragment?
# [09:29] <hsivonen> Hemebond: by inserting it into a document.
# [09:30] <Hemebond> Exactly... so it will have a parent when being displayed.
# [09:30] <hsivonen> the sentence inimino quoted has littly practical value and mostly makes readers start pondering what it is about
# [09:31] <Hemebond> Curse you Hixie!
# [09:32] <Lachy> the acid 2 test is broken. apparently a file as gone missing http://www.webstandards.org/files/acid2/test.html
# [09:32] <Lachy> this copy still works though http://www.hixie.ch/tests/evil/acid/002/
# [09:33] <inimino> I'm trying to understand the value of defining conformance for document fragments such that an element pulled from a conforming document could be a non-conforming fragment
# [09:34] <hsivonen> Lachy: have you pinged Molly about it?
# [09:37] <hsivonen> Hixie: regarding making body optional: We can't make it optional in DOMs parsed from text/html. If we do make it optional in DOMs parsed from application/xhtml+xml and in synthetic DOMs, there will be yet another special case to consider where a conforming DOM is unserializable as text/html
# [09:38] <hsivonen> Hixie: moreover, given that the main problem with <body> is CSS behavior in the XML mode, perhaps it would be better for CSS to take mitigating actions in the spec like it did for HTML
# [09:39] <hsivonen> whoa! Hixie made <title> optional!
# [09:41] <hsivonen> I'm glad I took a local copy of the spec file before Hixie landed the content model changes...
# [09:48] <Lachy> hsivonen, no
# [09:49] <hsivonen> Hixie: it might be possible to write the permitted contexts for <style scoped> in a more obvious way
# [09:50] * Joins: haavard- (n=hkmoen@pat-tdc.opera.com)
# [09:50] <hsivonen> Hixie: "Where prose content is expected before any non-<style> sibling elements and before text nodes that are not inter-element whitespace
# [09:50] <hsivonen> (the last part of the requirement is going to suck big time for RELAX NG)
# [09:51] <hsivonen> (or, rather, RELAX NG sucks big time for mixed content where the positions of the text nodes are restricted)
# [09:54] <hsivonen> Hixie: out of curiosity, do you have implementor feedback on the acceptability of <style scoped> except from Hyatt?
# [09:57] <hsivonen> http://www.w3.org/mid/c0be00e90712181835j5f8cff31mf897fba71f307998@mail.gmail.com
# [10:01] * Joins: tndH_ (i=Rob@adsl-87-102-85-140.karoo.KCOM.COM)
# [10:01] * tndH_ is now known as tndH
# [10:04] <Hixie> hsivonen: ok. if you find out anything about iframes, do let me know, i basically have no idea how to define it right now.
# [10:05] <Hixie> hsivonen: the comeback of "significant inline" actualy happened at the same time as it was removed (assuming you mean that you "should" not have empty elements)
# [10:05] <Hixie> hsivonen: haven't spoken to glazou yet, wanted to get your take on it first
# [10:05] <Hixie> hsivonen: agreed on <body>. I'll note that.
# [10:06] <Hixie> hsivonen: I made <title> optional? oops
# [10:06] <Hixie> hsivonen: re <style scoped>, no, not yet.
# [10:08] <Hixie> hsivonen: ok, i have a fix for <title> pending
# [10:08] <hsivonen> Hixie: ok
# [10:09] * hsivonen continues reviewing the changes
# [10:09] <Hixie> and i've removed the note i have about body (originally a proposal from glazou)
# [10:09] <Hixie> http://www.whatwg.org/specs/web-apps/current-work/index-diff should be a diff of the big change, btw
# [10:10] <hsivonen> Hixie: that diff has green-on-green styling for inserted text
# [10:11] <Hixie> fixed
# [10:11] <hsivonen> Hixie: thanks
# [10:12] <Hixie> i recommend taking a copy of that page, it'll get nuked when i next regen the spec
# [10:13] <hsivonen> Hixie: copy taken
# [10:14] <Hixie> i also have a fix for <style scoped> pending
# [10:15] <hsivonen> thanks
# [10:15] <Hixie> i think that's everything you mentioned
# [10:15] <hsivonen> yes
# [10:16] * hsivonen needs to learn to think that prose means %Flow
# [10:16] <Hixie> yeah the new terms take some getting used to.
# [10:16] <Hixie> i think the win from not clashing with css is good, though
# [10:20] <hsivonen> hmm. now that bimorphic is gone, <del> becomes less complex unless there are new hidden complexities
# [10:22] <hsivonen> does Heading content participate in any parent-child content models?
# [10:22] <hsivonen> or only exclusions?
# [10:24] <hsivonen> <address> has probably one of the worst usefulness/complexity ratios
# [10:26] * Quits: Lachy (n=Lachlan@cm-84.215.41.149.getinternet.no) ("This computer has gone to sleep")
# [10:28] * hdh recommends tracking whatwg with git-svn or bzr-svn or something like such
# [10:29] <hsivonen> hdh: what do git-svn and bzr-svn do and what's the benefit over using svn directly?
# [10:29] <hdh> can't get an overview like that index-diff though
# [10:29] <Hixie> hsivonen: i think <del> is simple. but i'm not 100% sure.
# [10:30] <Hixie> hsivonen: Heading is involved in the (increasingly badly explained and to be rewritten) outline algorithms
# [10:30] <Hixie> hsivonen: <address> could be simplified by making it phrasing-only as in html4
# [10:30] <hdh> it copies the whole svn history to my machine, so I can get the diffs offline
# [10:31] <hdh> I checked only the recent revs, my internet access is rather costy
# [10:31] <hdh> costly even
# [10:32] <hsivonen> Hixie: The must not about <p><br></p> is going to be violated all the time
# [10:33] <Philip`> svnsync can copy the whole SVN history too, without involving any other source-control system, assuming the server is new enough
# [10:33] <Philip`> (It's a standard part of SVN 1.4, I think)
# [10:34] <Hixie> hsivonen: sure. so is the requirement not to use <table> for layout, etc. those violations are real violations, though.
# [10:34] <Hixie> (and always have been)
# [10:34] <hsivonen> Hixie: evidently, these MUSTs are not held to the same standard as the Ogg SHOULD :-)
# [10:35] <Hixie> hsivonen: sure, one's an author requirement and the other a UA requirement
# [10:35] <Hixie> the bar for UA requirements is much higher since there are so many fewer of them
# [10:35] <Hixie> uas, that is
# [10:35] <Hixie> as opposed to authors
# [10:36] <hdh> Philip`: didn't know that, thanks
# [10:37] <hsivonen> Is there a spec-blessed word for the elements that are prose but not phrasing?
# [10:38] <Hixie> no
# [10:38] <Hixie> i ended up not really needing one
# [10:38] <Hixie> if you need one i can add one
# [10:38] <Hixie> you get the honour of deciding what it should be
# [10:38] <hsivonen> I'm going to need a word, but then bimorphic wasn't in the spec, either
# [10:39] <Hixie> (and exactly what it should cover)
# [10:39] <hsivonen> Hixie: ok. I'll think about it when I start implementing this
# [10:39] <Hixie> k
# [10:39] <hsivonen> (not going to happen before Christmas)
# [10:39] <Hixie> does the general idea look sound?
# [10:39] <Hixie> (and is it an improvement?)
# [10:40] <hsivonen> Hixie: it isn't an improvement considering a markup-aesthetic gut feeling, but it is a big improvement considering actual experience with applying the previous model to actual pages
# [10:41] <hsivonen> so I guess the practical needs should override the aesthetic gut feeling
# [10:41] <Hixie> cool, that's what i was aiming for
# [10:41] <Hixie> i tried to address the aesthetic aspect with the new definition of paragraph
# [10:42] <hsivonen> I congratulate myself for postponing improving bimorphic violation error messages
# [10:43] * Joins: doublec (n=doublec@203-211-95-159.ue.woosh.co.nz)
# [10:43] <Hixie> hehe
# [10:43] <Hixie> now i want to know what lachy's gonna think :-)
# [10:43] <Hixie> i'm sure he'll hate it
# [10:43] <Hixie> i guess i should post to a list
# [10:44] * Joins: Lachy (n=Lachlan@pat-tdc.opera.com)
# [10:44] <hsivonen> there he is
# [10:44] <Hixie> wow, good timing
# [10:44] <Hixie> Lachy: so the spec changed
# [10:44] <Hixie> Lachy: i'm sure you're going to hate it
# [10:45] <Hixie> Lachy: but i want to know your opinion anyway :-)
# [10:45] <hsivonen> Hixie: did you consider allowing <span> to contain more stuff the the extent interoperably parsed by current browsers?
# [10:45] <hsivonen> s/the the/to the/
# [10:45] <Hixie> hsivonen: not while i was doing this change, no
# [10:45] <Hixie> hsivonen: it's been proposed before though, and i'm sure i'll have to respond to that feedback at some point
# [10:45] <hsivonen> Hixie: ok
# [10:45] <Hixie> hsivonen: i'm not sure it's worth it, though, if it's just a migration thing
# [10:46] <hsivonen> Migration is very important considering out design principles :-)
# [10:46] <hsivonen> in fact, migration considerations are a big thing going for HTML5 compared to competing specs :-)
# [10:47] <hsivonen> s/out/our/
# [10:47] <hsivonen> typo++
# [10:47] <Hixie> i guess
# [10:47] <Hixie> i dunno
# [10:47] * Lachy looks
# [10:47] <Hixie> it seems like a big hack to me
# [10:47] <Hixie> why not use <div>?
# [10:48] <hsivonen> http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cp%3Efoo%3Cspan%3E%3Col%3E%3Cli%3Eitem%3C%2Fli%3E%3Cli%3Eitem%3C%2Fli%3E%3C%2Fol%3E%3C%2Fspan%3Ebar%3C%2Fp%3E
# [10:49] <hsivonen> that works in Firefox
# [10:49] <hsivonen> this doesn't: http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cp%3Efoo%3Cdiv%3E%3Col%3E%3Cli%3Eitem%3C%2Fli%3E%3Cli%3Eitem%3C%2Fli%3E%3C%2Fol%3E%3C%2Fdiv%3Ebar%3C%2Fp%3E
# [10:50] <Hixie> oh, for putting lists inside paragraphs
# [10:50] <Hixie> i thought you meant for <section>s and company
# [10:50] <Hixie> i basically concluded we should give up with the whole <ol>s-in-paragraphs thing
# [10:50] <Hixie> the benefits aren't compelling
# [10:52] <hsivonen> Hixie: there's demand and allowing the span hack doesn't seem to have a serious downside
# [10:52] <Hixie> is there really a demand?
# [10:52] <Hixie> the pain is pretty big
# [10:52] <hsivonen> Hixie: the pain?
# [10:54] <hsivonen> Hixie: well, even if J. Random Author doesn't care, the issue has come up again and again among markup enthusiasts and experts
# [10:54] <Hixie> e.g. worrying about ending up with nested <p>s
# [10:54] <Hixie> so has <acronym>, our solution to that was to just chop it off altogether
# [10:55] * Joins: ROBOd (n=robod@89.122.216.38)
# [10:55] <hsivonen> Hixie: http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cp%3Efoo%3Cspan%3E%3Col%3E%3Cli%3E%3Cp%3Eitem%3C%2Fp%3E%3C%2Fli%3E%3Cli%3E%3Cp%3Eitem%3C%2Fp%3E%3C%2Fli%3E%3C%2Fol%3E%3C%2Fspan%3Ebar%3C%2Fp%3E works for me in Firefox
# [10:57] <Hixie> yeah
# [10:57] <Hixie> that's a bad thing
# [10:57] <Hixie> what does it mean to have a paragraph containing a paragraph?
# [10:57] <Hixie> that's not typographically sane
# [10:57] <hsivonen> well, I already have the schema bits around for allowing lists whose items are phrasing only
# [11:00] <Hixie> yeah but then we're back to people being confused and... it just doesn't seem worth the effort
# [11:00] <Lachy> Hixie, you're right about one thing: I hate it!
# [11:00] <Hixie> hsivonen: i mean, what do we gain?
# [11:00] <Hixie> Lachy: :-)
# [11:01] <Hixie> Lachy: i tried to mitigate the damage with the new implied paragraph concept
# [11:01] <hsivonen> Hixie: we fix a long-standing spec bug for people doing conversions from LaTeX and DocBook
# [11:01] <Philip`> http://janus.state.me.us/legis/statutes/17-a/title17-Asec1105-A.html has subparagraphs in lists in paragraphs in lists in paragraphs
# [11:02] * Joins: Camaban (n=adrianle@host217-41-27-233.in-addr.btopenworld.com)
# [11:02] <Hixie> hsivonen: we fix a longstanding bug for roundtripping from and back to LaTeX and DocBook, but... do we care?
# [11:03] <hsivonen> I kinda care, yes
# [11:03] <Hixie> Philip`: actually, i only see one paragraph there
# [11:03] <Hixie> hsivonen: so what are the elements you would allow there?
# [11:03] <Hixie> hsivonen: just <ol> and <ul>?
# [11:03] <hsivonen> Hixie: yes
# [11:03] <hdh> dl?
# [11:04] <Hixie> hsivonen: and limit all decendants to phrasing-only?
# [11:04] <hsivonen> Hixie: I'd be OK with phrasing-only <li>s in that case
# [11:04] <Hixie> (and <li>)
# [11:05] <Lachy> Hixie, for h1-h6, shouldn't the expected context say something like "Where prose content is expected, except where heading content is forbidden"
# [11:06] <hdh> Any content model that expects prose content also expects phrasing content ← should the second expects be accepts?
# [11:07] <hsivonen> all my elaborate <del> magic went to waste :-)
# [11:09] <Lachy> it's not all bad I guess. Prose content seems to be rougly equivalent to %flow; in the HTML4 DTD
# [11:09] <hsivonen> (I bet that some people who went Strict to be cool will feel betrayed.)
# [11:10] <Hixie> Lachy: yeah basically prose = flow and phrasing = inline
# [11:10] <Hixie> hsivonen: i know i do...
# [11:10] <Hixie> Lachy: well, everything is allowed where it's allowed except where it's forbidden
# [11:12] <Hixie> hsivonen: so... this is going to require parser changes too, i take it
# [11:13] <hsivonen> Hixie: what's "this"?
# [11:13] <Hixie> the whole span-ol thing
# [11:13] * Quits: Lachy (n=Lachlan@pat-tdc.opera.com) (calvino.freenode.net irc.freenode.net)
# [11:13] <hsivonen> oh :-( the whole point was not requiring parsing changes
# [11:14] <Hixie> seems the spec isn't what safari and mozilla do for <p><span><ol>
# [11:15] <Hixie> in fact it's not what they do for any formatting element
# [11:15] <Hixie> even other blocks handle being put inside an inline.
# [11:15] <Hixie> huh.
# [11:16] * Joins: Lachy (n=Lachlan@pat-tdc.opera.com)
# [11:16] <Hixie> i wonder what i need to do to make the spec work
# [11:16] <Hixie> the parser work rather
# [11:17] <hsivonen> Does http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cp%3Efoo%3Cspan%3E%3Col%3E%3Cli%3E%3Cp%3Eitem%3C%2Fp%3E%3C%2Fli%3E%3Cli%3E%3Cp%3Eitem%3C%2Fp%3E%3C%2Fli%3E%3C%2Fol%3E%3C%2Fspan%3Ebar%3C%2Fp%3E work in IE like it works in Firefox and Safari?
# [11:17] <Hixie> i don't have IE right here
# [11:17] <hsivonen> in Opera, "bar" is misplaced compared to Firefox and Safari
# [11:18] <hsivonen> IIRC, Gecko parsing was changed to allow blocks in <span>, because IE did and there were sites depending on it
# [11:18] <hsivonen> If Firefox, Safari and IE agree here, I think the parsing spec should be changed to match, although I would have hoped that parsing was done and a closed issue
# [11:19] <jgraham_> hsivonen: That was a bit optimistic, no?
# [11:19] <hsivonen> jgraham_: yeah
# [11:21] <Hixie> man, the parsing spec doesn't match browsers at all when it comes to handling of block-level start tags inside inlines
# [11:21] <Philip`> hsivonen: In IE6, the OL is a child of both the SPAN and then of BODY; and "bar" comes after that (as a child of BODY)
# [11:22] <hsivonen> Philip`: I don't understand what you mean
# [11:22] <hdh> body > span > ol?
# [11:24] <Philip`> data:text/plain,%23comment%3A%20CTYPE%20ht%0D%0AHTML%0D%0A%20%20%20%20HEAD%0D%0A%20%20%20%20%20%20%20%20TITLE%0D%0A%20%20%20%20BODY%0D%0A%20%20%20%20%20%20%20%20P%0D%0A%20%20%20%20%20%20%20%20%20%20%20%20%23text%3A%20foo%0D%0A%20%20%20%20%20%20%20%20%20%20%20%20SPAN%0D%0A%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20OL%0D%0A%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20LI%0D%0A%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20
# [11:24] <Philip`> Uh, that's too long
# [11:24] <Hixie> http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%3Cp%3E%3Ci%3E%3Cp%3E%3C%2Fp%3E%3C%2Fi%3E%3C%2Fp%3E is crazy
# [11:25] <Hixie> luckily, safari agrees with the spec on that one
# [11:25] <Philip`> http://tinyurl.com/2bslye
# [11:25] <Hixie> http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%3Cp%3E%3Ci%3E%3Cdiv%3E%3Cp%3E%3C%2Fp%3E%3C%2Fdiv%3E%3C%2Fi%3E%3C%2Fp%3E is crazy even in safari
# [11:26] <Hixie> i'm not sure exactly what to do here
# [11:35] <hsivonen> Hixie: why is div's content model transparent instead of prose?
# [11:36] <hsivonen> yay, <font> is phrasing
# [11:37] <hsivonen> "XXX publish a "Valid HTML5!" button with a kitten on it. Made by an artist. (Doodle?)"
# [11:37] <hsivonen> didn't zcorpan already take care of that action item?
# [11:38] <hsivonen> Hixie: the big content model landing looks good
# [11:38] <Philip`> I'm not sure he got the "Made by an artist" bit :-)
# [11:38] <Hixie> <div> being transparent _should_ make no difference to being prose, but transparent better conveys the intent. i might change it back to prose though.
# [11:38] <Hixie> <font> is phrasing but note that it is transparent too
# [11:39] <hsivonen> Hixie: the main practical problem for me will involve banning text siblings before <style scoped>, <param> and <source>
# [11:39] <Hixie> so it allows blocks inside it
# [11:39] <hsivonen> Hixie: not a big deal in Java, though
# [11:39] <hsivonen> Hixie: making <font> transparent sucks
# [11:39] <Hixie> yeah well <font> in general sucks and is an open issue
# [11:39] <hsivonen> Hixie: I'd prefer changing div's content model to prose
# [11:40] <Hixie> the idea is to jsut allow <font> <div> <p> ... bla
# [11:40] <Hixie> since everyone who's gonna use <font> is gonna do that anyway
# [11:40] <Hixie> (and it seems no more harmful than allowing font in the firtst place)
# [11:40] <hsivonen> Hixie: every time you say something out of the ordinary, everyone has to stop and think what the spec *really* means
# [11:40] <Hixie> hsivonen: yeah, that was somewhat the intent. but i'll change it to prose.
# [11:47] * Quits: doublec (n=doublec@203-211-95-159.ue.woosh.co.nz)
# [11:55] <hsivonen> :-( :-( Creative Commons deploys now qNames in content: http://wiki.creativecommons.org/CCPlus
# [11:55] <hsivonen> s/now/new/
# [12:04] <Hemebond> Hixie: That's actually something that kind of puts me off HTML5... legitimising crap.
# [12:05] <Hixie> well, telling people they're wrong without really a good reason hasn't worked so far
# [12:05] <Hemebond> What are you talking about? The web is finally starting to clean itself up thanks to the standards movement.
# [12:06] <Hemebond> There are already specs that allow FONT and that rubbish.
# [12:06] * Quits: Lachy (n=Lachlan@pat-tdc.opera.com) (calvino.freenode.net irc.freenode.net)
# [12:06] <Hixie> oh i agree that we shouldn't allow <font>
# [12:06] <Hemebond> Let the (insert derogatory term) use that spec.
# [12:07] <Hixie> (though if we don't allow it, it's not clear what to do with wysiwyg software, since the state of the art doesn't yet have a solution for semantic editor ui)
# [12:07] <Hemebond> yet
# [12:08] <Hemebond> I think there are WYSIWYG editors out there, web-based, that create fairly clean markup.
# [12:08] <Hemebond> I guess I'm always going to prefer the direction of XHTML2.
# [12:08] <Hemebond> Also...
# [12:09] <Hemebond> I remember someone saying a day or so ago about how crap XUL was...
# [12:09] <Hemebond> Was that you?
# [12:09] <Hemebond> Or how it was a bad idea to use it for web apps.
# [12:09] <Hixie> i don't recall talking about xul recently
# [12:09] <Hixie> you can join the xhtml2 working group, btw, they're as open as the html5 working group
# [12:10] * Quits: heycam (n=cam@124-168-10-57.dyn.iinet.net.au) ("on holidays, back in a few days")
# [12:10] <Hixie> and while there are wysiwyg editors that generate clean markup, i haven't seen any that generate correct non-presentational semantic markup that are usable by your average user
# [12:10] <akaroa> Hixie: that's not actually the case (xhtml2)
# [12:10] <Hixie> what isn't the case?
# [12:11] <akaroa> I tried to join and got the : " no, members only" deal
# [12:12] <Hemebond> Hixie: Even if I could join XHTML2 WG, which is pointless cause I don't know shit, I'm really just trying to understand the reason for the revolt.
# [12:12] <Philip`> akaroa: Do you mean joining just the mailing list, or joining the group as an invited expert?
# [12:13] <Philip`> *public invited expert, or whatever they're called
# [12:13] <Hixie> akaroa: recently?
# [12:13] <akaroa> Then SP emailed back and said that I'd have to submit a paragraph saying how I'd be of use to the group and that this would be sent to the managment for consideration
# [12:13] <Hixie> weird
# [12:13] * Joins: Lachy (n=Lachlan@pat-tdc.opera.com)
# [12:13] <Hixie> they have the same charter as the html working group as far as i can tell
# [12:15] <Hixie> did you follow the instructions on this page?: http://www.w3.org/2004/01/pp-impl/32107/instructions
# [12:15] <Philip`> As far as I can see, the charter doesn't specify the policy for accepting invited expert requests
# [12:15] <akaroa> Hixie: Yes
# [12:16] <Hixie> akaroa: did you submit a paragraph?
# [12:16] <akaroa> no
# [12:16] <hsivonen> does the XHTML2 WG have (public) Invited Experts at all or only W3C Invited Experts?
# [12:16] <Philip`> and http://www.w3.org/2007/04/html-ie-faq#sameproc says it is the same process as normal, which requires approval from various people and isn't just automatic
# [12:16] <akaroa> I gave up because I believe the future of xhtml is with xhtml5
# [12:16] <Hixie> akaroa: at which step in those steps did they ask you for a paragraph?
# [12:17] <akaroa> At first I got the no we don't have invited experts, then SP emailed and said that I'd need to write and ask etc...
# [12:18] <Hixie> well that's weird
# [12:18] <hsivonen> Hemebond: if a user selects text in a WYSIWYG editor and chooses a color for the text, what markup should be generated?
# [12:18] <Hixie> i didn't realise they were screening members
# [12:18] <Hixie> oh well
# [12:18] <Hixie> yet another reason 5>2
# [12:19] <akaroa> Hixie: yes
# [12:19] <Philip`> I expect they look at the HTML WG and don't exactly want to copy that type of community
# [12:24] * Joins: arnath01 (n=arnath@d54C1C929.access.telenet.be)
# [12:28] <Dashiva> Did anyone else get double content in the latest commit-watchers mail?
# [12:28] <Hemebond> hsivonen: Well, it wouldn't work the way other WYSIWYG editors.
# [12:28] <Hemebond> You would need to change the way they construct documents.
# [12:28] <Hixie> indeed. the problem is that nobody has yet shown how to do that.
# [12:29] <Philip`> Dashiva: You mean other than the way it normally sends both 'index' and 'source' diffs?
# [12:29] <hsivonen> Hemebond: do you have a plan for marketing the paradigm shift?
# [12:29] <Dashiva> This time I got index+source, massive whitespace, index+source again, massive whitespace
# [12:29] <hsivonen> back in 2002, I thought that problems could be solved if only editors had a different UI paradigm
# [12:29] <Hemebond> Yes, I have a 5-point plan I prepared earlier.
# [12:29] * Hemebond pulls the proposal out of the oven.
# [12:30] * Philip` didn't read that 300KB commit message at all
# [12:30] <Hemebond> *sniff* ahh, absolute rubbish.
# [12:31] <Philip`> Dashiva: Hmm, I seem to have just one set of content in that message
# [12:31] <akaroa> Hixie: 5>2, yes. You can't develop open standards for the open web in closed working groups. Open standards need to be developed in open groups
# [12:31] <Dashiva> Philip`: Okay, probably a bug in my client then
# [12:31] <Hixie> anyone know of any bugs in browsers in the DOM relative to what the DOM2 specs say?
# [12:32] <hsivonen> Hixie: getAttribute return value when attribute not set. (but one could argue it is a spec bug now)
# [12:32] <hsivonen> Hixie: namespace stuff in XHR-loaded DOMs
# [12:32] <Dashiva> hrefs being resolved vs not is also a bit vague
# [12:33] <Hixie> XHR isn't (yet) a spec, sadly
# [12:34] <Hixie> these are things for acid3, i'm trying to limit myself to things i can prove are bugs using just specs published in 2004 or earlier.
# [12:34] <Hixie> dom bugs, specifically
# [12:34] <hsivonen> Hixie: actually, I think Acid3 should test that getAttribute violates the spec in the interoperable way
# [12:35] <Hixie> acid3 can't test anything that the 2004-or-earlier specs don't explicitly require
# [12:35] <Philip`> Opera supports getElementById('id\0garbage'), but that's not really a DOM problem because it affects the entire browser API but I like complaining about it anyway
# [12:35] <Hixie> (though i can avoid testing things i know are wrong)
# [12:35] <Hixie> Philip`: that counts
# [12:36] <Hixie> you have test 38
# [12:37] <Hixie> load http://www.hixie.ch/tests/evil/acid/003/NOT_READY_PLEASE_DO_NOT_USE.html in opera, wait for the animation to stop, and click the "Acid3" text
# [12:37] <Hixie> see if the alert mentioned test 38 fails
# [12:37] * Philip` will cherish it
# [12:38] <Philip`> Fails in 9.2 and 9.5 and works in FF2, which sounds right
# [12:39] <Philip`> Maybe you should keep search engines away from http://www.hixie.ch/tests/evil/acid/003/reference.html so it isn't the top result for "acid3" and makes people think their browser scored 100%
# [12:39] * Quits: bakarat (n=arnath@d54C1C929.access.telenet.be) (Read error: 110 (Connection timed out))
# [12:40] <Hixie> it'll fix itself when acid3 actually exists
# [12:40] * Quits: roc (n=roc@121-72-28-191.dsl.telstraclear.net)
# [12:41] <hsivonen> Hixie: judging the smoothness of the animation is problematic
# [12:42] <Hixie> hsivonen: really?
# [12:42] <Hemebond> You could run it without a tic and time how long it takes to run.
# [12:42] <Hemebond> Liek a timedemo.
# [12:42] <hsivonen> Hixie: in Firefox 2 on Mac, I'm not sure if the animation counts as smooth
# [12:43] <Hixie> hsivonen: then it probably doesn't
# [12:46] * Parts: akaroa (n=opera@121-72-47-224.dsl.telstraclear.net)
# [12:46] * Quits: kfish (n=conrad@61.194.21.25) ("Pike!")
# [12:47] <hsivonen> Hixie: standards-minded people might feel less betrayed if you allowed Content-Style-Type text/css as a conforming talisman
# [12:47] <Hixie> with what UA requirements if it's not text/css?
# [12:48] <hsivonen> Hixie: assuming that UAs today ignore it, requiring it to be ignored
# [12:48] <Hixie> ugh
# [12:50] * Hemebond is now known as Heme|Sleep
# [12:50] <hsivonen> I'm not a fan of talismans, but Validome telling people to put it in and then Validator.nu telling people to take it out seems unproductive
# [12:51] <Hixie> i see one possible fix...
# [12:51] <Hixie> :-)
# [12:53] <hsivonen> perhaps it is good to point out the bogosity of some talismans. hard call
# [12:54] * Hixie is watching microsoft's acid2 video
# [12:54] <Hixie> man, microsoft sure are nice about this test now that they pass it
# [12:54] <Hixie> they even said it was well-constructed :-D
# [13:13] <othermaciej> heh
# [13:14] <othermaciej> it's funny given how much they complained in the past that the test was silly, that competitions to pass tests like these are meaningless, and that browsers shouldn't compete on standards compliance
# [13:22] * Joins: MikeSmith (n=MikeSmit@71-218-59-92.hlrn.qwest.net)
# [13:27] <Camaban> well, they shouldn't compete on standards, but they will while some (mostly) support them and some make a real hash of them
# [13:29] * Joins: dedridge (n=opera@121-72-16-18.dsl.telstraclear.net)
# [13:33] <Hixie> othermaciej: if you have any ideas for acid3, do let me know
# [13:33] <othermaciej> Hixie: I need to look over what it currently tests at some point
# [13:34] <othermaciej> maybe now is a good time, if I can't manage to get to sleep soon
# [13:34] <Hixie> http://www.hixie.ch/tests/evil/acid/003/NOT_READY_PLEASE_DO_NOT_USE.html
# [13:34] <Hixie> it's much more modular than acid2
# [13:34] <othermaciej> my least favorite IE quirk is the direct JS property <--> DOM attribute mapping
# [13:35] <Hixie> and i'm only testing things that i can pretty strongly argue are required in specs that were in CR or better in 2004
# [13:35] <othermaciej> for elements
# [13:35] <othermaciej> but I have to figure out what aspects of that are non-compliant
# [13:35] <othermaciej> I'm pretty sure getAttribute() returning a non-string is non-compliant
# [13:35] <othermaciej> getElementById finding things by name also seems like a clear violation
# [13:38] * Quits: jgraham (n=jgraham@81-86-217-3.dsl.pipex.com) (Read error: 110 (Connection timed out))
# [13:39] * Quits: jgraham_ (n=james@81-86-217-3.dsl.pipex.com) (Read error: 110 (Connection timed out))
# [13:39] <Hixie> what attribute returns a non-string?
# [13:40] <othermaciej> any event listener attribute would return a function object if the attribute is set
# [13:40] <Hixie> ooo ok
# [13:41] <hsivonen> doesn't any getAttribute in IE return whatever the corresponding DOM property would return?
# [13:41] <othermaciej> yes
# [13:41] <othermaciej> a bunch of those happen to be strings
# [13:41] <hsivonen> also, there's the getAttribute('className') thing
# [13:41] <othermaciej> event handler attributes are a case where it won't even convert to string to look right
# [13:42] <othermaciej> there was some blog post about IE DOM bugs recently that covered this
# [13:42] <othermaciej> but I can't remember where I saw it
# [13:43] <hsivonen> might have been sitepoint
# [13:43] <othermaciej> ah: http://www.sitepoint.com/blogs/2007/11/30/internet-explorer-doesnt-just-suck-it-also-blows/
# [13:44] <Hixie> ok cool, i'll look through those tomorrow
# [13:44] <Hixie> thanks for the help
# [13:44] <Hixie> send me further ideas at ian@hixie.ch
# [13:45] <Hixie> going to bed now
# [13:45] <othermaciej> roger that
# [13:45] <othermaciej> good night
# [13:45] * Joins: colione (n=colione@17.247.241.83.in-addr.dgcsystems.net)
# [14:17] * Joins: zcorpan (n=zcorpan@c-cb21e353.1451-1-64736c12.cust.bredbandsbolaget.se)
# [14:21] * Joins: lucas- (n=lucas@ip20-185-173-82.adsl2.versatel.nl)
# [14:53] <Lachy> wow, this video is a huge download. http://channel9.msdn.com/showpost.aspx?postid=367207
# [14:54] <Lachy> for 584MB, it had better be a good quality video
# [14:57] <Camaban> for a 32min vid? it had better be good quality...
# [15:05] <Lachy> oh no. <p> can no longer contain <ol> or <ul>, even for XHTML. :-(
# [15:06] <webben> I wonder if we could have some sort of p-continuation element that would provide the same semantic in HTML and XHTML
# [15:06] <webben> because that's what it actually is ... it's when a paragraph continues after a block
# [15:14] * Quits: zcorpan (n=zcorpan@c-cb21e353.1451-1-64736c12.cust.bredbandsbolaget.se) (Read error: 110 (Connection timed out))
# [15:14] <webben> it would also provide a nice styling hook : continuation {text-indent:0;margin-top:1ex;}
# [15:22] <othermaciej> if a looser content model is really desirable it would make more sense to add a new element for paragraphs in addition to <p>
# [15:22] <othermaciej> but it's probably not worth it
# [15:23] * Joins: gsnedders (n=gsnedder@host86-135-224-200.range86-135.btcentralplus.com)
# [15:23] <Camaban> webben: something like that would be cool for lists too :) quite how to do it in a 'correct' way I'm not sure on though
# [15:50] * Joins: hasather (n=hasather@90-231-107-133-no62.tbcn.telia.com)
# [15:58] * Quits: MikeSmith (n=MikeSmit@71-218-59-92.hlrn.qwest.net) ("Less talk, more pimp walk.")
# [16:02] * Joins: colione_ (n=colione@17.247.241.83.in-addr.dgcsystems.net)
# [16:02] * Quits: colione (n=colione@17.247.241.83.in-addr.dgcsystems.net) (Read error: 104 (Connection reset by peer))
# [16:03] * Joins: _Ivo (n=ivo@89-180-57-187.net.novis.pt)
# [16:10] <webben> othermaciej: A looser content model isn't the aim. Being able to express the media independent semantic of and simply style them is the aim.
# [16:11] <webben> *of paragraph continuations
# [16:11] <webben> one could do it with a <paragraph> or <para> element, but then you have two elements that mean the same thing
# [16:11] <webben> if one were designing a lang from scratch, doubtless a different content model would be the way to go
# [16:13] <webben> i suppose another possibility would be <p type="continuation">
# [16:13] <othermaciej> <para>some text <ol>...</ol> more text</para> reads better to me than <p>some text</p><ol>...</ol><continuation>more text</continuation>
# [16:13] * Quits: maikmerten (n=merten@ls5laptop14.cs.uni-dortmund.de) ("Verlassend")
# [16:13] <othermaciej> and certainly expresses the semantics better
# [16:14] <webben> othermaciej: The semantic is the same in either case. It's a nicer way to do it, but having 2 elements do the same thing would probably be unpopular.
# [16:14] <webben> (I wouldn't mind writing para ... not sure others would feel the same way)
# [16:16] <hdh> the type=cont way would need reparenting elements somehow, so I'd go for allowing <ol> inside <p>
# [16:17] <Camaban> I'd kind of prefer 'para', it maybe gives more a mental clue as to the purpose of the tag. I'm not sure how many people at the moment really associate <p> with 'a paragraph of text', though I guess that could be a remnant of using whatever code makes it look right
# [16:18] * Joins: annevk (n=annevk@c529c1b12.cable.wanadoo.nl)
# [16:19] * Quits: Lachy (n=Lachlan@pat-tdc.opera.com) ("This computer has gone to sleep")
# [16:22] <webben> hdh: You /can't/ allow OL inside P in text/html since it breaks backwards compat.
# [16:22] <webben> (browsers close P when they hit an OL)
# [16:23] <webben> the main problem with type="continuation" is that it would turn into <p type="continuation" class="continuation"> for styling in IE6 (which doesn't support attribute selectors)
# [16:24] <webben> which is a lot more typing than <continuation>
# [16:24] <hdh> webben: ah sorry, I remembered it to be the p > span > ol example
# [16:25] * Joins: dolphinling_ (n=chatzill@rbpool6-36.shoreham.net)
# [16:30] * Joins: billmason (n=billmaso@ip156.unival.com)
# [16:32] * Quits: dedridge (n=opera@121-72-16-18.dsl.telstraclear.net) (Read error: 104 (Connection reset by peer))
# [16:34] * Joins: Lachy (n=Lachlan@cm-84.215.41.149.getinternet.no)
# [16:36] <gsnedders> http://tom-frank.blogspot.com/ — what can I say?
# [16:38] * Quits: dolphinling (n=chatzill@rbpool6-76.shoreham.net) (Read error: 110 (Connection timed out))
# [16:39] * Quits: annevk (n=annevk@c529c1b12.cable.wanadoo.nl) (Read error: 110 (Connection timed out))
# [16:45] * Quits: csarven (n=nevrasc@81-5-133-33.static.nfwebsolutions.com) (Read error: 104 (Connection reset by peer))
# [16:52] * Joins: dolphinling__ (n=chatzill@rbpool2-34.shoreham.net)
# [16:52] * dolphinling__ is now known as dolphinling
# [16:56] * Quits: billmason (n=billmaso@ip156.unival.com) (".")
# [16:59] * Joins: billmason (n=billmaso@ip156.unival.com)
# [17:01] * Quits: dolphinling_ (n=chatzill@rbpool6-36.shoreham.net) (Read error: 110 (Connection timed out))
# [17:15] <krijn> Are the IRC logs working for anybody?
# [17:20] <gsnedders> krijn: yeah
# [17:20] <krijn> :/
# [17:32] * Quits: jwalden (n=waldo@RANDOM-SEVENTY-TWO.MIT.EDU) (Remote closed the connection)
# [17:46] <webben> Hixie: Just a thought that it might be good if someone familiar with the issues wrote a WHATWG FAQ item about why VIDEO is a good idea as opposed to defining a video API for object.
# [18:09] * Quits: othermaciej (n=mjs@dsl081-048-145.sfo1.dsl.speakeasy.net)
# [18:16] <krijn> Is IE3 still used today? :|
# [18:18] <Camaban> webben: heh
# [18:20] <Camaban> webben: I agree though, documenting the why behind some of this stuff would be good
# [18:25] * Joins: webben_ (n=benh@82.152.16.177)
# [18:25] * Quits: webben_ (n=benh@82.152.16.177) (Read error: 104 (Connection reset by peer))
# [18:30] <hdh> what does 1H08 mean in the IE8 post?
# [18:30] <hdh> oh, first half
# [18:32] <hdh> that page use <font>...
# [18:32] * Joins: webben_ (n=benh@82.152.16.177)
# [18:33] <hdh> generated by MS Office
# [18:33] <Philip`> http://blogs.msdn.com/ie/archive/2007/12/19/internet-explorer-8-and-acid2-a-milestone.aspx ? I don't see any "<font" on there
# [18:34] <hdh> <p class="MsoNormal" style="margin: 0in 0in 0pt;"><font face="Calibri">
# [18:34] <hdh> /html/body/form/div[2]/div/div[3]/div[2]/div/div/div/p[3]/font
# [18:34] <hdh> wtf is that form doing there?
# [18:35] <hdh> argh, I mean http://myitforum.com/cs2/blogs/rtrent/archive/2007/12/19/ie8-news-on-standards-and-acid2-test-announcement.aspx
# [18:36] <hdh> sorry MS
# [18:38] <Philip`> "I watched the Channel 9 video and the UI of IE8 looks awfully similar to IE7" - hmm, I only noticed the UI looking totally different to IE7 though quite like IE3 - it just had a single toolbar with a few buttons, presumably since it was for debugging the engine and not for browsing the web
# [18:38] <hdh> qt's demo browser even has a search box
# [18:38] * Philip` wonders what sites Chris Wilson was measuring when he said that half use 'standards' mode
# [18:39] <hdh> defaulting to Qt doc searching to boot
# [18:42] * Quits: webben (n=benh@dip5-fw.corp.ukl.yahoo.com) (Connection timed out)
# [18:43] <hdh> 1997.webhistory.org ← first time I see such name
# [18:48] <hdh> why didn't they link to http://channel9.msdn.com/showpost.aspx?postid=367207 but channel9's / ?
# [18:56] * Parts: Camaban (n=adrianle@host217-41-27-233.in-addr.btopenworld.com)
# [18:58] <webben_> hdh: webhistory is awesome ... they've got archives (searchable of yahoo-groups) from the earliest days of HTML.
# [18:58] <webben_> (They're not complete, because there was a failure on the original mailserver.)
# [18:58] <hdh> kinda wayback machines for news:// ?
# [18:58] <webben_> hdh: Sorry ... no just for the mailing list for the original HTML specs
# [18:59] <hdh> sorry, I missed the last phrase xD
# [18:59] <webben_> hdh: http://1997.webhistory.org/lists/lists.html
# [19:01] <hdh> If the material can be digitally uploaded, you will be able to submit it to the archives via the Internet once this capability is in place. Until then, please contact the Webmaster at webmaster@webhistory.org. ← that IE8 video qualifies here?
# [19:02] <webben_> hdh: Arguably... however, I'm not sure that the project is active.
# [19:02] <webben_> It doesn't seem to have changed since 2003 or something.
# [19:02] <webben_> You could always try mailing em and see what happens though.
# [19:02] <hdh> well, at least the hosting is still up
# [19:09] * hdh googled 'make dl look like table' and found nothing, has anyone here did that?
# [19:13] <hdh> okay, adding css to the query helps; ftr, someone did http://www.macrankin.co.uk/web_projects/tables/table_forge_3.html via http://www.killersites.com/mvnforum/mvnforum/viewthread?thread=4416
# [19:19] * Joins: othermaciej (n=mjs@17.203.15.179)
# [19:30] * Joins: hober (n=ted@unaffiliated/hober)
# [19:30] * Quits: ROBOd (n=robod@89.122.216.38) (Read error: 104 (Connection reset by peer))
# [19:33] * Joins: webben (n=benh@82.152.16.177)
# [19:33] * Quits: webben (n=benh@82.152.16.177) (Read error: 104 (Connection reset by peer))
# [19:40] * Joins: zcorpan (n=zcorpan@c-cb21e353.1451-1-64736c12.cust.bredbandsbolaget.se)
# [19:41] * Joins: jgraham (n=jgraham@81-86-208-38.dsl.pipex.com)
# [19:41] * Joins: jgraham_ (n=james@81-86-208-38.dsl.pipex.com)
# [19:41] <zcorpan> Hixie: shouldn't the content model for <del> be just "transparent"?
# [19:43] <zcorpan> Hixie: <figure> still talks about embedded content
# [19:47] <zcorpan> Hixie: it seems to me that <noscript> has similar content model requirements as <iframe> would have
# [19:47] * Joins: maikmerten (n=maikmert@Lae9a.l.pppool.de)
# [19:50] <zcorpan> Hixie: furthermore, iirc, opera has a setting to disable frames, where <noframes> and <iframe> is parsed as pcdata (like <noscript> when scripting is disabled)
# [19:50] <zcorpan> Hixie: same with plugins and <noembed>
# [19:50] * Quits: tndH (i=Rob@adsl-87-102-85-140.karoo.KCOM.COM) (Read error: 104 (Connection reset by peer))
# [19:51] * Joins: tndH_ (i=Rob@adsl-87-102-44-124.karoo.KCOM.COM)
# [19:51] <zcorpan> (though, having three toggles in the html parser sucks)
# [19:51] * tndH_ is now known as tndH
# [19:51] * Quits: webben_ (n=benh@82.152.16.177) (Read error: 110 (Connection timed out))
# [19:54] * Joins: kingryan (n=kingryan@dsl092-002-056.sfo1.dsl.speakeasy.net)
# [19:57] * Quits: kingryan (n=kingryan@dsl092-002-056.sfo1.dsl.speakeasy.net) (Client Quit)
# [19:58] * Joins: kingryan (n=kingryan@dsl092-002-056.sfo1.dsl.speakeasy.net)
# [19:58] * Joins: csarven (n=nevrasc@81-5-133-33.static.nfwebsolutions.com)
# [20:00] * Joins: webben (n=benh@82.152.16.177)
# [20:03] <zcorpan> Hixie: why does one need to log in to view test cases and demos that the spec annotation system links to?
# [20:05] * Quits: Heme|Sleep (n=Hemebond@ip-118-90-125-176.xdsl.xnet.co.nz) ("Leaving.")
# [20:08] <zcorpan> Hixie: while we're on content models, i think <table> should allow <col> as children
# [20:09] <zcorpan> in fact, i'm not sure the parser shouldn't imply <colgroup> around <col>s (only ie does that)
# [20:14] <zcorpan> Hixie: why allow prose content in <th>?
# [20:14] <zcorpan> (and not in <h1>, <dt>, etc)
# [20:16] * Quits: weinig (n=weinig@c-71-198-176-23.hsd1.ca.comcast.net)
# [20:20] * Quits: csarven (n=nevrasc@81-5-133-33.static.nfwebsolutions.com) (Read error: 104 (Connection reset by peer))
# [20:20] * Quits: csarven- (n=nevrasc@81-5-133-33.static.nfwebsolutions.com) (Read error: 104 (Connection reset by peer))
# [20:22] * Joins: csarven (n=nevrasc@81-5-133-33.static.nfwebsolutions.com)
# [20:28] * Joins: virtuelv (n=virtuelv@233.80-203-100.nextgentel.com)
# [20:34] * Joins: roc (n=roc@121-72-28-191.dsl.telstraclear.net)
# [20:47] * Quits: virtuelv (n=virtuelv@233.80-203-100.nextgentel.com) ("Leaving")
# [20:48] * Quits: phsiao (n=shawn@c-71-232-12-131.hsd1.ma.comcast.net)
# [20:48] * Joins: phsiao (n=shawn@c-71-232-12-131.hsd1.ma.comcast.net)
# [20:59] * Parts: Lachy (n=Lachlan@cm-84.215.41.149.getinternet.no) ("Leaving")
# [21:04] * Joins: csarven- (n=nevrasc@81-5-133-33.static.nfwebsolutions.com)
# [21:09] * Joins: othermaciej_ (n=mjs@17.255.106.224)
# [21:09] * Joins: weinig (n=weinig@17.203.15.140)
# [21:15] * Joins: Lachy (n=Lachlan@cm-84.215.41.149.getinternet.no)
# [21:15] * Quits: zcorpan (n=zcorpan@c-cb21e353.1451-1-64736c12.cust.bredbandsbolaget.se) (Read error: 110 (Connection timed out))
# [21:19] * Quits: roc (n=roc@121-72-28-191.dsl.telstraclear.net)
# [21:24] * Quits: othermaciej (n=mjs@17.203.15.179) (Read error: 110 (Connection timed out))
# [21:29] * Quits: maikmerten (n=maikmert@Lae9a.l.pppool.de) ("Leaving")
# [21:45] * Joins: zcorpan (n=zcorpan@c-cb21e353.1451-1-64736c12.cust.bredbandsbolaget.se)
# [21:45] <zcorpan> Hixie: nested <menu>s are now banned
# [21:46] <Hixie> i have a fix pending for <del> transparent; figure's mention of embedding content will be gone
# [21:47] <Hixie> noscript and iframe are similar and i nearly used the exact same text but it doesn't work because iframe could have <script> elements inside it, which could detect the validator testing the page for correctness
# [21:47] <Hixie> the log in problem is an oversight
# [21:47] <Hixie> i don't see why <table> should allow <col>
# [21:47] <zcorpan> <table><col> is allowed in html4 and xhtml1
# [21:47] <Hixie> i wasn't going to allow prose in <th> but i'm assuming we need to for back compat, i'm willing to try a change if we want
# [21:48] <Hixie> and i'll look into <menu> nesting
# [21:49] <zcorpan> Hixie: i think the common case for prose in <th> is <th><p> or <th><div>, which on the one hand seem pretty harmless but otoh pretty bogus
# [21:49] <Hixie> yeah
# [21:49] <Hixie> not sure what to do about that
# [21:49] <zcorpan> i've also seen <th><h3> though, from authors who want to be overly explicit
# [21:49] <zcorpan> which will probably just screw up things
# [21:50] <Hixie> i can allow heading content too
# [21:51] <Hixie> i could say phrasing content or heading content but not both
# [21:52] <Hixie> i dunno
# [21:52] <Hixie> hsivonen: any opinion?
# [21:52] <zcorpan> i'm not sure why heading in <th> would be useful
# [21:52] <Hixie> yeah
# [21:52] <zcorpan> wouldn't it screw up the outline?
# [21:52] <Hixie> i've changed it to phrasing content
# [21:52] <Hixie> we'll see if we get complaints
# [21:52] <zcorpan> ok
# [21:53] <Hixie> <table><col> i recall had some issue with parsing which is why i didn't allow it
# [21:53] <zcorpan> likely :)
# [21:53] <Hixie> i think
# [21:53] <zcorpan> ie implies <colgroup>
# [21:53] <zcorpan> but it also implies <tbody>, which is not required :)
# [21:55] <Hixie> fixed requiring login for showing tests and demos in a separate tab
# [21:56] <Hixie> yeah
# [21:59] <gsnedders> does http://pastebin.ca/825838 seem overly harsh?
# [22:00] <Philip`> I hope that's not intended for posting to public-html :-p
# [22:00] <gsnedders> No
# [22:01] <gsnedders> :P
# [22:01] <hober> gsnedders: looks pretty good to me. -public-html +www-archive that baby
# [22:01] <gsnedders> I'm not that dumb
# [22:01] <gsnedders> hober: yeah, and cc'd to chairs, I was planning
# [22:01] <gsnedders> s/dumb/dumb normally/
# [22:01] <gsnedders> :)
# [22:01] * Parts: arnath01 (n=arnath@d54C1C929.access.telenet.be) ("Leaving")
# [22:02] <Philip`> Why CC to chairs?
# [22:03] * Joins: roc (n=roc@121-72-28-191.dsl.telstraclear.net)
# [22:03] <gsnedders> Philip`: issue with conduct on mailing lists
# [22:04] * othermaciej_ is now known as othermaciej
# [22:05] <gsnedders> http://lists.w3.org/Archives/Public/www-archive/2007Dec/0087.html
# [22:07] <kingryan> gsnedders: thanks for that email. I had given up on that thread being productive.
# [22:08] <gsnedders> kingryan: Oh, I scrapped various drafts of emails like that to various people over the months.
# [22:14] * Quits: zcorpan (n=zcorpan@c-cb21e353.1451-1-64736c12.cust.bredbandsbolaget.se) (Read error: 110 (Connection timed out))
# [22:16] * Philip` doesn't like it when the spec says "you must only do x where y", because he doesn't find it immediately clear whether that's "you must do x, and must not do it without y" (i.e. "you must do x and y") or "you may do x, and must not do it without y" (i.e. "you must not do x unless y")
# [22:17] <Philip`> (e.g. "Authors must only use elements in the HTML namespace in the contexts where they are allowed, as defined for each element")
# [22:18] * Quits: roc (n=roc@121-72-28-191.dsl.telstraclear.net)
# [22:20] <inimino> that sentence doesn't seem to match the "only x where y" form
# [22:22] * Quits: _Ivo (n=ivo@89-180-57-187.net.novis.pt) (Connection timed out)
# [22:25] <Philip`> It's not a syntactic match, but there's the same ambiguity of "authors must not use elements in the HTML namespace except in the contexts where they are allowed" vs "authors must not use elements which aren't in the HTML namespace, in contexts where HTML elements are allowed"
# [22:25] <Philip`> Actually, maybe that's still not the same
# [22:25] * Parts: hdh (n=hdh@58.187.109.11)
# [22:26] <Philip`> But it's still the "z must only w" form, so it's not too far off :-)
# [22:28] <Hixie> Philip`: how would you phrase it?
# [22:29] <Philip`> Hixie: As "Authors must not use elements in the HTML namespace except in the contexts where they are allowed" or "Authors must not use elements which aren't in the HTML namespace, in contexts where HTML elements are allowed"
# [22:29] <Hixie> that disallows wrong use, but doesn't allow correct use
# [22:32] <Philip`> The spec doesn't explicitly allow me to wear socks, but that doesn't make me a non-conforming HTML5 UA if I do - the default state for unspecified things is that they are allowed (I think)
# [22:33] * Joins: zcorpan (n=zcorpan@c-cb21e353.1451-1-64736c12.cust.bredbandsbolaget.se)
# [22:33] <Hixie> i guess
# [22:33] <Hixie> can you send mail with those suggestions to the whatwg list?
# [22:33] <Hixie> i'll fix it at some point
# [22:33] <Hixie> (i want to think about exactly how best to phrase it)
# [22:34] <Philip`> "must only" seems an abuse of the RFC2119 definition of "must", since it relies on parsing English to understand what are the requirements and allowances in the sentence
# [22:35] * Philip` will mail
# [22:36] <Philip`> Incidentally, s/preceeded/preceded/ in the concept of "preceeded or followed"
# [22:36] <Hixie> yeah i agree that the must only sentence is poor
# [22:37] <Hixie> i guess it's time for me to regen the spec with all the fixes
# [22:38] * Joins: doublec (n=doublec@202.0.36.64)
# [22:43] <Hixie> right, spec regenned and checked in
# [22:46] <Lachy> woot! hasLayout is gone from IE8! :-) http://lists.w3.org/Archives/Public/www-style/2007Dec/0151.html
# [22:47] <webben> lots of good news coming out of the IE team ... finally
# [22:48] * gsnedders blames middle-management of MS
# [22:50] * Quits: zcorpan (n=zcorpan@c-cb21e353.1451-1-64736c12.cust.bredbandsbolaget.se) (Read error: 104 (Connection reset by peer))
# [22:51] * Joins: zcorpan (n=zcorpan@c-cb21e353.1451-1-64736c12.cust.bredbandsbolaget.se)
# [22:55] <Philip`> Is http://lists.whatwg.org dead?
# [23:04] * Joins: jwalden (n=waldo@STRATTON-FOUR-FIFTY-EIGHT.MIT.EDU)
# [23:16] * Quits: zcorpan (n=zcorpan@c-cb21e353.1451-1-64736c12.cust.bredbandsbolaget.se) (Read error: 110 (Connection timed out))
# [23:42] <Lachy> Philip`, yes, it appears to be
# [23:52] * Joins: Thezilch (n=fuz007@wsip-64-147-23-102.oc.oc.cox.net)
# [23:54] <Hixie> Philip`: back up
# [23:58] * Joins: MikeSmith (n=MikeSmit@71-218-59-131.hlrn.qwest.net)
# [23:58] * Quits: MikeSmith (n=MikeSmit@71-218-59-131.hlrn.qwest.net) (Read error: 104 (Connection reset by peer))
# [23:58] * Joins: MikeSmith (n=MikeSmit@71-218-59-131.hlrn.qwest.net)
# Session Close: Fri Dec 21 00:00:00 2007

The end :)