Options:
- # Session Start: Wed Jun 27 00:00:00 2007
- # Session Ident: #html-wg
- # [00:00] <Hixie> i don't really know what test cases would mean in this context
- # [00:00] <Hixie> style="" is very well understood technically, it's just the use cases that are lacking
- # [00:00] <Hixie> nobody is proposing changing how it works
- # [00:00] <Hixie> the question is whether it's the best solution to hte problems it solves, and what problems it is solving
- # [00:01] <Hixie> it bugs me when someone sends an e-mail along the lines of "i must insist that [something or other be changed]"
- # [00:01] <DanC> well, a test case could answer "is this in the language?"
- # [00:02] <Hixie> how is that useful?
- # [00:02] <Hixie> isn't it trivial?
- # [00:02] <DanC> well, a whole pile of discussion seems to think that it's not trivial.
- # [00:02] <Hixie> re the insists thing, for everyone who insists on one change, there are two people who insist on incompatible changes. insisting helps nobody. what we need is reasoning, so that hyatt and i can find solutons that address as much as ossible.
- # [00:02] <Hixie> DanC: no i mean the test case would be trivial.
- # [00:03] <DanC> I think a lot of people would stop sending so much mail if they were assured that style="" was going to stay in the language.
- # [00:03] <Hixie> ok, here's the testcase: "<!DOCTYPE HTML><p style="">Is this conforming?"
- # [00:03] <DanC> yup
- # [00:03] <Hixie> how are we further along now?
- # [00:03] <Hixie> i don't understand how the test case helps anything
- # [00:04] <Hixie> (in this case)
- # [00:04] <DanC> well, we'll be further along with a recorded decision that yes, it's conforming.
- # [00:04] <Hixie> (i agree that for, e.g., the parser section, test cases are pretty much the only way to go)
- # [00:04] <DanC> actually, I'm not sure an empty style makes for such a good test case, now that I look at it.
- # [00:04] <Hixie> right but how does the testcase help determine the answer to that question?
- # [00:04] <gavin_> isn't the entire discussion about whether or not it should be conforming?
- # [00:04] <Hixie> ok, here's the testcase: "<!DOCTYPE HTML><p style="color:blue">Is this conforming?"
- # [00:05] * Joins: gsnedders (gsnedders@81.132.88.104)
- # [00:05] <Philip`> (Do you still need <title></title> too?)
- # [00:05] <DanC> well, yes, to some extent, the whole discussion is about whether it's conforming. but there's no black-and-white object of study to focus the discussion
- # [00:05] <Hixie> Philip`: "<!DOCTYPE HTML><title>style attribute</title><p style="color:blue">Is this conforming?"
- # [00:05] <Hixie> DanC: i don't understand how the testcase helps at all. i don't think anyone misunderstands what the question is.
- # [00:06] <DanC> I can imagine the discussion tweaking the test case until exemplifies a use case that everyone agrees merits inclusion in the language
- # [00:06] <Hixie> DanC: the problem is that we have no data to _answer_ the question
- # [00:06] <Hixie> DanC: that's use cases, not a test case
- # [00:06] <DanC> it can be both
- # [00:06] <Hixie> ok, well, feel free to try to get people to do that
- # [00:06] <Hixie> i don't really care how we get to use cases
- # [00:06] <Hixie> but my need here is use cases
- # [00:06] <Hixie> without them i can't progress on the issue
- # [00:07] <Hixie> (as i discussed in the e-mail)
- # [00:07] <DanC> I guess I haven't read enough of the discussion; I'm having a hard time imagining why the style attribute would be anything other than conforming in every sense, just like it has been as far back as I can remember.
- # [00:07] <Hixie> (use cases and other things, actually, see the e-mail for details)
- # [00:07] <Hixie> anyway, afk, bbiab
- # [00:09] * Quits: edas (edaspet@88.191.34.123) (Quit: http://eric.daspet.name/ et l'édition 2007 de http://www.paris-web.fr/ )
- # [00:13] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
- # [00:13] <DanC> ah... very nice... "Such arguments will have no effect on anything. You're not going to
- # [00:13] <DanC> convince the people who disagree with you by telling them they are wrong,
- # [00:13] <DanC> even if they _are_ wrong"
- # [00:18] * Joins: gavin (gavin@74.103.208.221)
- # [00:29] * tH resists the urge to reply "+1"
- # [00:30] <gsnedders> tH: that often gets hard to resist, especially in arguments about the use of it, doesn't it?
- # [00:30] <Hixie> (back)
- # [00:30] * Philip` wonders how many weeks it has been since the last +1
- # [00:35] * DanC thanks tH for his discipline
- # [00:47] * Quits: tH (Rob@87.102.84.66) (Quit: ChatZilla 0.9.78.1-rdmsoft [XULRunner 1.8.0.9/2006120508])
- # [00:58] * Joins: mjs (mjs@17.255.97.186)
- # [01:02] <kingryan> +1 to tH
- # [01:11] * Joins: sbuluf (ihhzoqw@200.49.140.162)
- # [01:13] <mw22> -1
- # [01:14] <gavin_> << 1
- # [01:16] <kingryan> ^1
- # [01:18] <schepers> Hixie, I think you're wrong
- # [01:19] <schepers> (sorry, had to say that)
- # [01:20] * Parts: billmason (billmason@69.30.57.156)
- # [01:22] <mjs> -1
- # [01:23] <schepers> :'(
- # [01:48] * Quits: kingryan (rking3@208.66.64.47) (Quit: kingryan)
- # [01:58] * Joins: karl (karlcow@128.30.52.30)
- # [02:05] * Joins: Lachy_ (chatzilla@203.158.59.119)
- # [02:07] * Quits: Lachy (chatzilla@203.158.59.119) (Ping timeout)
- # [02:07] * Lachy_ is now known as Lachy
- # [02:20] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
- # [02:25] * Joins: gavin (gavin@74.103.208.221)
- # [02:34] <karl> Hixie: good email on cooperative work through the wiki.
- # [02:37] <Hixie> now if only people listen to it :-/
- # [02:37] <Hixie> we've already had a +1 since that e-mail
- # [02:38] <karl> ;)
- # [02:38] <karl> because it is all about humans and not bots
- # [02:49] * Quits: zcorpan_ (zcorpan@84.216.41.174) (Ping timeout)
- # [02:49] <Hixie> you know, the more this naming discussion goes on, the more i wonder whether the better solution isn't just to rename xhtml2
- # [02:50] <Hixie> i've been trying to keep out of the discussion though
- # [02:51] <karl> Hixie, I proposed it to XHTML 2.0 too.
- # [02:52] <karl> s/XHTML 2.0/XHTML 2.0 WG/
- # [02:52] <Hixie> yeah
- # [02:52] <Hixie> i think danc has implied that xhtml2 might want to change name too
- # [02:52] <karl> I'm all for finding a reasonnable arangement between parties, and not that much for names ownership or "dick contest"
- # [02:53] <Hixie> yeah
- # [02:53] <karl> which until now sounds like it
- # [02:53] <Hixie> indeed
- # [02:53] <Hixie> that's why i've been trying to keep the heck away :-)
- # [02:53] * Joins: olivier (ot@128.30.52.30)
- # [03:02] * Joins: polin8 (polin8@24.184.204.6)
- # [03:16] <karl> http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2007Jun/0122
- # [03:27] <mjs> I'm not sure what any of that means
- # [03:33] * Quits: polin8 (polin8@24.184.204.6) (Quit: polin8)
- # [03:34] <karl> I think shane is saying what hixie says too and many others. That the PROSE of the spec defines the semantics. Then there are discussions on how to better integrate RDFa with future languages and old languages. And to do so in the less disruptive way.
- # [03:37] <Lachy> his argument doesn't follow. He says that a DTD only describes syntax, not semantics, and then concludes that the DOCTYPE can be used as an indicator for sementics.
- # [03:39] <karl> Lachy: yes it is what is an identifier, a label, a serial number, etc.
- # [03:39] <karl> a reference to a document
- # [03:50] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Client exited)
- # [04:27] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
- # [04:32] * Joins: gavin (gavin@74.103.208.221)
- # [04:43] <schepers> wow, Lachy, you're right, XHTML5 *is* a widely used term... about 1/10th as widely used as XHTML2
- # [04:46] <Lachy> schepers: I wasn't making a comparison of how widely each are used, it makes no difference that "XHTML2" is used more than "XHTML5"
- # [04:46] <schepers> then why did you make the argument?
- # [04:48] <Lachy> because it is a widely used term. just because A is more widely used than B, doesn't mean B isn't widely used
- # [04:49] <schepers> the only reason I see for HTML5's XML serialization to be called XHTML (as far as I can tell) is to attempt to undermine the efforts of XHTML2
- # [04:49] <schepers> why isn't HTML5/XML just as good, for example?
- # [04:50] <Lachy> so you're ignoring the fact that XHTML5 is much close to XHTML1, than XHTML2 is. If only one is to retain the name, it should be the one that is closer to XHTML1.
- # [04:50] <Lachy> however, I couldn't care less if XHTML2 retains it's name. XHTML2 is irrelevant
- # [04:50] <schepers> Lachy, I can read, and I've read that argument about 20 times now
- # [04:51] <Lachy> well, then why are you still debating?
- # [04:51] <schepers> it's clear that that's your opinion, but remember, "opinions don't matter", right?
- # [04:51] <schepers> I'm not debating, I'm pointing out that you aren't making very good arguments
- # [04:52] <schepers> I have no investment in XHTML2 at all, myself
- # [04:52] <Lachy> that XHTML2 is irrelevant is my opinion, but that XHTML5 is closer to XHTML1 than XHTML2 is, is a fact (which I already supplied evidence for)
- # [04:56] <schepers> on a syntactic level, perhaps
- # [04:56] <schepers> but not in spirit
- # [04:58] <mjs> XHTML2's effort doesn't need any undermining
- # [04:58] <mjs> they seem to be handling that fine by themselves
- # [04:58] <schepers> ok, then why are you getting in a pissing contest about the name?
- # [04:59] <mjs> I think it's fine for both to keep their current name
- # [04:59] <mjs> I'm not the one who started asking the other group to change
- # [04:59] <mjs> I think both using the XHTML1 namespace URI is a bigger problem, but it would be a problem for XHTML2 even if XHTML5 didn't exist
- # [05:02] <schepers> if XHTML2 is so doomed, why would you even want XHTML5 to be associated with the name?
- # [05:02] <schepers> or, in other news, why not just play nice and use a different name?
- # [05:04] <mjs> - the normal name so far for an XML version of HTML has been XHTML, there doesn't seem to be any deep need to change it
- # [05:04] <mjs> - why doesn't XHTML2 WG play nice and use a different name?
- # [05:04] <mjs> (not that I care if they do, but they are the ones who seem to think different names are important)
- # [05:06] <mjs> I think the XHTML2 WG passing a resolution that the HTML WG should not use a particular name is not a very friendly way to start the discussion
- # [05:08] <schepers> I believe that the discussion was started by the WHATWG appropriating the W3C name XHTML, don't you agree?
- # [05:08] <schepers> or am I mischaracterizing the situation?
- # [05:08] <mjs> WHATWG used the name "Web Apps 1.0"
- # [05:08] <mjs> although unofficially people referred to it as HTML5 and XHTML5
- # [05:09] <mjs> now it is a W3C spec maintained by the HTML WG
- # [05:09] <mjs> which has officially resolved to name it HTML5, but has not picked an official name for the XHTML serialization
- # [05:09] <mjs> although popular and sensible proposals include XHTML 5 and XHTML 1.5
- # [05:10] <schepers> that wasn't my question, mjs
- # [05:10] <mjs> I think it's a mischaracterization to say "the discussion was started by the WHATWG appropriating the W3C name XHTML"
- # [05:10] <mjs> which was in fact your question
- # [05:11] <schepers> how is it a mischaracterization?
- # [05:11] <schepers> Lachy stated that the term XHTML5 has been used for 3 years.... either it has been or it hasn't
- # [05:12] <schepers> and, in fact, it has
- # [05:12] <mjs> WHATWG didn't name anything XHTML5, random people called it that, now the WG is considering making it official
- # [05:12] <schepers> you should run for office :)
- # [05:13] <mjs> no one told XHTML2 WG what they should or shouldn't name their spec
- # [05:13] <mjs> both groups will be defining languages evolved from XHTML 1
- # [05:13] <mjs> I am sure arguments could be made why one side has more right to the name than the other
- # [05:13] <mjs> I think sharing is just fine
- # [05:13] <schepers> I think they were chartered to work on the XHTML spec, right?
- # [05:15] <schepers> I think it's not a very friendly way to start the discussion by appropriating the name of a language that they were chartered to define
- # [05:15] <mjs> they are chartered to "continue the XHTML2 work"
- # [05:15] <mjs> I didn't know they were chartered to define XHTML5
- # [05:15] <mjs> I don't think it says that in their charter
- # [05:16] <mjs> I don't see a problem with both languages being named XHTML N for different values of N
- # [05:16] <schepers> you can spin it howver you like, maciej, but I'll say again, the only reason I can see for the HTML WG (and the WHATWG) to use the term XHTML5 is to sow confusion and undermine the XHTML2 WG
- # [05:17] <schepers> and I suspect you think the same thing
- # [05:17] <mjs> I'd be fine with XHTML 1.5 as a name if it made people happy (makes it more clear that it is related to the XHTML 1.x family of languages, not XHTML2)
- # [05:17] <mjs> but I don't think it would make people happy
- # [05:17] <schepers> you know it wouldn't, and you know why
- # [05:18] <mjs> I personally don't care what the XHTML2 WG does and I wish they would leave us alone
- # [05:18] <mjs> we don't go to their mailing list to start turf wars or pissing matches
- # [05:18] <schepers> there's an easy way to do that: stop stepping on their toes... there are plenty of perfectly good acronyms out there
- # [05:19] <mjs> it could certainly reduce confusion if they didn't use "H", "T", "M" and "L" in that order in their name
- # [05:19] <schepers> but like I said, I'm just a spectator here
- # [05:19] <mjs> since it may lead people to think their language is suitable for use as an on-the-wire format, or would work in browsers
- # [05:19] <Lachy> there has been no real evidence for such confusion presented yet
- # [05:19] <mjs> but I don't think it's a big problem for them to keep the XHTML2 name
- # [05:20] <schepers> ah, yes, "evidence"
- # [05:20] <schepers> that's starting to sound like a dirty word
- # [05:21] <mjs> I actually don't know why XHTML 1.5 would be unacceptable either - it lets XHTML2 keep the privilege of the larger number and the appearance of the more radically forward-looking language
- # [05:21] <schepers> define evidence, please... and do avoid using the word "evident" because it's clear that's not what you mean
- # [05:22] <schepers> mjs, because there would be a name clash through iterations either way
- # [05:22] <mjs> I mean, I guess it is ok for the XHTML2 WG to make a polite request to change the name of the XML serialization of our language, but I also think it is ok for us to say "no" to that
- # [05:22] <Lachy> evidence for confusion would be things like forum posts, emails, blog posts, etc. that ask questions like what's the difference between XHTML2 and XHTML5, and saying I don't understand something, where such statements arise from the similarity in names
- # [05:22] <schepers> (assuming either language survives more than a couple decades, which I personally doubt)
- # [05:23] <mjs> arguments could be made both ways why one has more or less right to claim their language is a successor to XHTML
- # [05:23] <mjs> but arguing such things is a waste of time
- # [05:23] <mjs> let's just be nice and share
- # [05:23] <schepers> lol
- # [05:24] <schepers> I honestly think you're a good guy, mjs, but I don't think you're being straightforward here
- # [05:25] <schepers> Lachy, I've been at a meeting where there was, in fact, great confusion about the relationship of XHTML2 and XHTML5
- # [05:25] <schepers> it was a large meeting of mostly technical people
- # [05:25] <mjs> sample argument: XHTML5 will likely get served as application/xhtml+xml on the public web as a final format, XHTML2 likely won't (unless carefully restricted to an XHTML5 or XHTML1-compatible subset, or using trickery like client-side XSLT to transform to an actual end delivery format)
- # [05:26] <mjs> seems like the thing that will be served with the XHTML MIME type should be named XHTML
- # [05:26] <mjs> (this is not to disparage the potential value of XHTML2 as a back-end authoring format)
- # [05:27] <mjs> you can see how arguments along these lines could easily turn into a mean-spirited dick size flamewar
- # [06:26] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
- # [06:35] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
- # [06:40] * Joins: gavin (gavin@74.103.208.221)
- # [08:43] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
- # [08:47] * Joins: gavin (gavin@74.103.208.221)
- # [08:54] <anne> hmm, the namespace we use also has "xhtml" in it
- # [08:56] <Lachy> I suppose we should document all these arguments in the wiki, so we can point people there instead of debating it again and again
- # [09:08] <anne> karl, all user agents support <image>
- # [09:09] * anne is not going to send another e-mail debating <image> suppor
- # [09:09] * anne ... support*
- # [09:09] <anne> I'm not aware of any table, I'm aware of tests that have been done
- # [09:10] <karl> anne: by user agents, do you mean desktop browsers?
- # [09:10] <karl> just to know
- # [09:10] <anne> Opera Mini, Opera Mobile, S60, etc. too at least
- # [09:10] <anne> I've no idea how far other mobile browsers care about the web versus walled garden content
- # [09:11] <karl> ok. thanks. because I try to search a bit and didn't find anything about it.
- # [09:11] <karl> I was curious
- # [09:11] * Joins: frippz (frippz@193.15.86.40)
- # [09:11] <anne> test it yourself...
- # [09:12] <anne> here's a testcase you can use: http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C%21DOCTYPE%20html%3E%3Cimage%20src%3Dimage%3E
- # [09:12] <karl> If I do not find anything is what I will do :))) indeed
- # [09:12] * Joins: tH (Rob@87.102.84.66)
- # [09:12] <karl> hmmm the element name is even converted in the DOM, I see
- # [09:12] <anne> that's the whole point
- # [09:13] <anne> it's a parsing quirk
- # [09:13] <anne> the language stays unaffected
- # [09:13] <anne> I don't think anyone on the mailing list gets that though, but that's ok
- # [09:13] <karl> yirkkkk cat pictures :p
- # [09:14] <anne> however, I'd recommend against testing parsing quirk for parsing quirk
- # [09:14] <anne> that seems like a waste of time
- # [09:18] <anne> "There is no reason not to publish as a working draft. Although it's clearly incomplete and there are many open issues, it closely reflects the current state of HTML5, which is good enough for a WD."
- # [09:18] <anne> Lachy, what open issues?
- # [09:18] * anne also wonders what is missing
- # [09:19] <Lachy> open issues in HTML5 in general, such as the missing headers="", summary="", etc.
- # [09:19] <anne> oh
- # [09:21] <karl> anne: do you need me to send a new version of the status for the document?
- # [09:21] <karl> or will you include my version with the comments which have been made?
- # [09:22] <anne> could you give me a HTML version?
- # [09:22] <anne> I modified the abstract, fwiw
- # [09:22] <karl> yes :) I will
- # [09:26] <karl> oh btw mjs: I have been playing a bit with DashCode, nice piece of software
- # [09:26] <mjs> karl: glad to hear you like it
- # [09:29] * Joins: edas (edaspet@88.191.34.123)
- # [09:37] * Quits: olivier (ot@128.30.52.30) (Quit: This computer has gone to sleep)
- # [09:37] * Quits: karl (karlcow@128.30.52.30) (Quit: Where dwelt Ymir, or wherein did he find sustenance?)
- # [10:23] * Joins: zcorpan_ (zcorpan@84.216.41.153)
- # [10:26] * Joins: Jero (Jero@213.46.207.230)
- # [10:49] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
- # [10:54] * Joins: gavin (gavin@74.103.208.221)
- # [10:59] * Quits: frippz (frippz@193.15.86.40) (Quit: frippz)
- # [10:59] * Joins: polin8 (polin8@24.184.204.6)
- # [11:09] <hsivonen> Hixie: in case you are not reading every message on public-html at the moment, I'd like to call your attention to Aaron Leventhal's contributions to the "market hasn't spoken" thread
- # [11:15] * Joins: ROBOd (robod@86.34.246.154)
- # [11:42] <mjs> I think an interim accessibility solution is useful but it doesn't need to be conforming to be (best practice would be use the appropriate element) and I kinda hate the baggage of the role attribute
- # [11:42] <mjs> the accessibility-only version sort of makes sense but it has become a magnet for semantics freaks
- # [11:42] <mjs> an interim solution based on class could be conforming w/o any changes to the HTML spec, and could be done as a separate spec
- # [11:43] <mjs> in fact there is a spec already for how to embed role information in class (not necessarily a very good one mind you)
- # [12:00] <hsivonen> mjs: not making it conforming leads to silliness like the recent IBM-authored Note
- # [12:01] <mjs> that being the one that says how to use class for accessibility info?
- # [12:01] <hsivonen> mjs: http://www.w3.org/WAI/PF/adaptable/HTML4/embedding-20061212.html
- # [12:01] <mjs> the only thing I find silly about it is that I don't get why they didn't do it that way in the first place
- # [12:02] <Lachy> hsivonen: Aaron's message is certainly a refreshingly objective post about accessibility
- # [12:02] <hsivonen> it's silly to hide stuff in another place to trick the validator and them move it with scripting instead of just making it ok to put it in the right place from the start
- # [12:02] <hsivonen> mjs: "that way" as in class or as in moving stuff with a script
- # [12:03] <mjs> as in using classes
- # [12:03] <mjs> instead of making up new global attributes
- # [12:05] <mjs> notice there are four different attributes in their simple XML example
- # [12:05] <anne> they have a lot more iirc
- # [12:05] <anne> it's quite complex stuff
- # [12:06] <mjs> so anyway I think a simple design based on class or other general-purpose extensibility hooks might be worth it, but as it is, it's a very purpose-specific and quite complex system, for an interim solution
- # [12:07] <mjs> better to build interim solutions on top of general mechanisms - then we know the general mechanisms are more likely to be ready for the next time someone needs an interim solution
- # [12:08] <mjs> http://www.w3.org/WAI/PF/GUI/roleTaxonomy-20060221.html
- # [12:08] <mjs> they have some quite silly roles in there
- # [12:08] <mjs> like "separator"
- # [12:09] * Quits: heycam (cam@203.214.72.248) (Ping timeout)
- # [12:09] <mjs> styling an <hr> already works in every browser afaik
- # [12:09] <hsivonen> as much as Hixie hates namespace prefixes, they could be the extensibility mechanims for special edge case stuff like this
- # [12:09] <beowulf> mm, styling an hr is a nightmare, i just don't use them
- # [12:10] * Joins: heycam (cam@203.214.72.248)
- # [12:10] * Quits: Lachy (chatzilla@203.158.59.119) (Ping timeout)
- # [12:10] * Joins: Lachy_ (chatzilla@203.158.59.119)
- # [12:10] * Lachy_ is now known as Lachy
- # [12:10] <mjs> maybe
- # [12:12] <mjs> wow, the WAI states are even sillier than the roles
- # [12:13] <mjs> anyway I guess namespaced attributes are a pretty obvious way to allow key/value data to be attached to an element for later use by scripting
- # [12:13] <mjs> just not sure if there's a way of using them that is backwards compatible and also XHTML - HTML cross-compatible
- # [12:20] <hsivonen> mjs: adding namespaces to HTML would make it XHTML-HTML cross-compatible. compat with shipped Firefox would need a script similar to what IBM is already advocating for HTML in the Note
- # [12:25] <zcorpan_> do we need to define role processing?
- # [12:26] <zcorpan_> isn't it better to try to get <progress> &c implemented?
- # [12:27] <hsivonen> zcorpan_: in the long term, yes. in the short term, making the role stuff (what's actually implemented) non-conforming is probably more harm than good
- # [12:28] <hsivonen> zcorpan_: I can see the point in what Aaron has implemented in Gecko. I don't see the point of the RDF complication
- # [12:28] <hsivonen> the new entity stuff is hard to get right when bending over backwards to avoid repetitive object allocation in an inner loop
- # [12:29] <mjs> hsivonen: what would you use to access namespaced attributes from script?
- # [12:29] <zcorpan_> hsivonen: we can make role conforming without defining processing
- # [12:29] <mjs> getAttributeNS would not do the right thing in old HTML versions
- # [12:29] <hsivonen> mjs: getAttributeNS() in HTML5 UAs and getAttribute() in legacy
- # [12:30] <zcorpan_> hsivonen: much like we can make RDF conforming without defining processing
- # [12:30] <hsivonen> zcorpan_: also true
- # [12:30] <mjs> hsivonen: how do you tell which one you are in? UA switching?
- # [12:30] <hsivonen> zcorpan_: although leaving it unspecified doesn't help other UA interoperate with Gecko
- # [12:31] <zcorpan_> hsivonen: indeed. perhaps they will opt implementing <progress> instead. :)
- # [12:31] <hsivonen> mjs: if the first returns null (in DOM5 return value semantics), try the other
- # [12:32] <zcorpan_> what i'm saying is that i see little point in reverse engineering and speccing role as implemented in firefox/jaws
- # [12:32] <mjs> hsivonen: and how would you set such attributes from script? also try both?
- # [12:33] <zcorpan_> because i think other vendors should implement the new stuff instead, like WF2
- # [12:34] <zcorpan_> if we spec how role is to be implemented, and other vendors implement that, it takes even longer time before <progress> and WF2 gets implemented
- # [12:34] <mjs> I don't think <progress> and WF2 are direct substitutes
- # [12:35] <zcorpan_> i think they are more important to implement than role
- # [12:35] <mjs> the direct substitute is being able to write <input type="checkbox"> but give it custom appearance and behavior
- # [12:35] <zcorpan_> i.e. xbl
- # [12:35] <mjs> so XBL2 would be one technology that can substitute
- # [12:35] <zcorpan_> yeah
- # [12:35] <mjs> (or for simpler cases CSS styling of form controls)
- # [12:35] <zcorpan_> indeed
- # [12:35] <hsivonen> mjs: I guess you'd have to try both
- # [12:36] <hsivonen> mjs: oops no
- # [12:36] <hsivonen> mjs: you'd have to try what currently works in Gecko, which AFAIK uses setAttrNS in text/html
- # [12:36] <hsivonen> mjs: I'm not an expert in the detail
- # [12:36] <hsivonen> s
- # [12:37] <hsivonen> mjs, zcorpan_: feel free to point out holes on the mailing list to get them on the record
- # [12:38] <mjs> hsivonen: I am not sure you can make a reasonable set of namespaced attribute operations that works across existing browsers in both HTML and XHTML mode
- # [12:39] <mjs> (without undesirable random side effects)
- # [12:39] <hsivonen> mjs: they need to work in existing browsers only to the extent the IBM script works today
- # [12:39] <mjs> at least I can't immediately think of such a thing
- # [12:39] <hsivonen> whew. I now pass test cases in the new entity tokenization world
- # [12:39] <hsivonen> that sure took time in the debugger
- # [12:40] <mjs> hsivonen: I am assuming namespaced attributes in HTML5 would be held to a higher standard than in that Note
- # [12:40] <mjs> i.e., suitable for general use and with a sane backward compatibility story
- # [12:41] <hsivonen> mjs: I think the best backwards compat story we could get on the serialization level is compat with whatever weird stuff IE does today with prefixes
- # [12:41] <zcorpan_> which is insane
- # [12:42] <mjs> I can't tell how well that script library works since the Note doesn't actually include or reference it
- # [12:42] <mjs> just describes sort of how it could work
- # [12:44] <hsivonen> I expect markp could have an opinion or two on what makes sense here
- # [12:48] <mjs> plus there's the fact that namespaces suck
- # [12:49] <mjs> so since the literal form of XML namespaced attributes doesn't have a good backwards compat story, maybe there is a different mechanism for extensions adding key/value pairs that would work reasonably
- # [12:50] * Quits: sbuluf (ihhzoqw@200.49.140.162) (Ping timeout)
- # [12:51] * Quits: heycam (cam@203.214.72.248) (Ping timeout)
- # [12:52] * Joins: Lachy_ (chatzilla@203.158.59.119)
- # [12:52] * Joins: heycam (cam@203.214.72.248)
- # [12:52] * Quits: Lachy (chatzilla@203.158.59.119) (Ping timeout)
- # [12:52] * Lachy_ is now known as Lachy
- # [12:52] <anne> maybe have a simple state= attribute that would take "foo=bar x=z" as values that would be exposed in some .state interface
- # [12:53] <anne> the problem with this, of course, is that you want semantics to be expressed in the element and not in some weird attribute only scripts deal with
- # [12:53] <mjs> yeah, and nesting key/value pairs in a key/value mechanism is lame
- # [12:53] <mjs> you could use a non-XML prefixing mechanism
- # [12:54] <anne> author-foo
- # [12:54] <mjs> dojo_extendedAttr="some-value"
- # [12:55] <mjs> any punctuation character that's allowed for attributes but not used in standard attributes would do, really
- # [12:55] <mjs> as a separator
- # [12:55] <mjs> it would be easier to make this work sanely than XML-style namespaces
- # [12:55] <mjs> (w/ backwards compat and all)
- # [12:56] <mjs> another possibility would be a child element that allows attributes of any name, but that doesn't work for void elements
- # [12:56] <mjs> but it does give up the idea that the true namespace identifier is a URI
- # [12:57] <hsivonen> mjs: I'd still try to go for XML NS compat by using the colon as the separator and having a long list of hardwired prefixes
- # [12:58] <anne> aah
- # [12:58] <mjs> the colon is what introduces the compat issues though
- # [12:58] <anne> yeah, colon == bad
- # [12:59] <anne> if we need extensibility lets at least have a sane way of doing it
- # [12:59] <anne> copy & paste compatible and all that
- # [13:01] <hsivonen> anne: what do you tell those who want an algorithmic mapping from any XHTML+custom attributes document to HTML5?
- # [13:01] <hsivonen> or elements even
- # [13:02] <mjs> that's quite a different problem than the desire for an extension mechanism
- # [13:03] <mjs> it may be you can solve both with the same solution, but the XML-mapping path is at odds with the goal for something with maximum backwards and cross-serialization compatibility
- # [13:06] * Joins: Patapouffe (Quetz16@84.100.205.108)
- # [13:06] * Parts: Patapouffe (Quetz16@84.100.205.108)
- # [13:36] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
- # [13:37] * Quits: mjs (mjs@17.255.97.186) (Ping timeout)
- # [13:41] * Joins: gavin (gavin@74.103.208.221)
- # [14:00] <anne> hsivonen, tell them to use HTML5 syntax in XHTML
- # [14:01] <hsivonen> anne: that would be incompatible with whatever existing code there is written for a particular XML vocabulary
- # [14:09] <anne> like there's much of that...
- # [14:09] <anne> don't they have these cool XML tools?
- # [14:12] * Quits: polin8 (polin8@24.184.204.6) (Quit: polin8)
- # [14:18] <hsivonen> anne: exactly
- # [14:19] <hsivonen> the problem is we don't know if it is a real problem because that part of code is hidden on the server side before content hits the Web
- # [14:24] * Joins: mjs (mjs@64.81.48.145)
- # [15:30] * Joins: polin8 (polin8@209.176.7.3)
- # [15:41] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Quit: Less talk, more pimp walk.)
- # [15:42] * Quits: Lachy (chatzilla@203.158.59.119) (Ping timeout)
- # [15:44] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
- # [15:44] * Joins: myakura (myakura@58.88.37.26)
- # [15:49] * Joins: gavin (gavin@74.103.208.221)
- # [16:20] <zcorpan_> Simon Pieters (good standing) Opera Software
- # [16:22] <anne> welcome to the dark side
- # [16:22] <anne> euh, wait
- # [16:22] <zcorpan_> :)
- # [16:26] <zcorpan_> am i subscribed to the list with my gmail.com or opera.com address now?
- # [16:26] * Quits: nickshanks (nicholas@195.137.85.17) (Quit: nickshanks)
- # [16:26] <anne> opera prolly
- # [16:26] <hsivonen> we now have one "No, disagree"
- # [16:27] * Joins: billmason (billmason@69.30.57.156)
- # [16:31] <zcorpan_> well, seems the mails are still coming in to my gmail
- # [16:31] <anne> and several unaddressed comments
- # [16:31] <anne> zcorpan_, you haven't left the WG as IE yet I think
- # [16:31] <hsivonen> anne: the W3C system allows a person change affiliation, right?
- # [16:32] * Joins: briansuda (briansuda@85.220.95.76)
- # [16:32] <hsivonen> anne: the system just changed zcorpan_'s affiliation record without changing the mail record
- # [16:32] <hsivonen> it seems
- # [16:33] <anne> zcorpan_, oh right, you can change your e-mail address if you want
- # [16:34] <zcorpan_> anne: where?
- # [16:36] <anne> zcorpan_, I think http://www.w3.org/Systems/db/memUser
- # [16:36] <hsivonen> ooh. no year in URI!
- # [16:38] * Joins: Schnitz (Miranda@84.153.87.189)
- # [16:38] <Schnitz> hi
- # [16:38] <hsivonen> hi
- # [16:38] <zcorpan_> anne: thanks
- # [16:53] * Quits: myakura (myakura@58.88.37.26) (Quit: Leaving...)
- # [16:54] <zcorpan_> anne: perhaps s/dropped/not included/ in html4-differences
- # [16:56] <zcorpan_> although that might be a larger edit than what is allowed
- # [16:56] <anne> DanC?
- # [16:56] <Schnitz> bye...
- # [16:56] <zcorpan_> Schnitz: cya
- # [16:57] * Quits: Schnitz (Miranda@84.153.87.189) (Quit: Schnitz)
- # [16:57] <DanC> umm... hi anne; I've got a telcon in 3min. what's up?
- # [16:58] <anne> DanC, see the suggestion from zcorpan_ above about s/dropped/not included/
- # [16:58] <DanC> that's definitely in the grey zone.
- # [16:58] <DanC> but yes, if others concur (zcorpan) then let's make that change. carefully.
- # [16:59] <anne> I don't feel strongly either way
- # [16:59] <DanC> I'm not sure I do either.
- # [16:59] * DanC must dash...
- # [16:59] <anne> people on the web are in favor of the document already; and this version is an improved version of the one "they" reviewed
- # [17:00] <zcorpan_> the edit wouldn't change what the document was intended to say, it would just make it clearer
- # [17:01] * Quits: polin8 (polin8@209.176.7.3) (Ping timeout)
- # [17:18] * Joins: icaaq (icaaaq@217.13.228.226)
- # [17:30] * Parts: icaaq (icaaaq@217.13.228.226)
- # [17:34] * Quits: edas (edaspet@88.191.34.123) (Ping timeout)
- # [17:45] * Joins: Lachy (chatzilla@203.158.61.14)
- # [17:51] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
- # [17:54] * Joins: polin8 (polin8@209.176.7.3)
- # [17:54] * Joins: Sander (svl@71.57.109.108)
- # [17:56] * Joins: gavin (gavin@74.103.208.221)
- # [18:31] * Joins: edas (edaspet@88.191.34.123)
- # [18:38] * Quits: jmb (jmb@81.86.70.47) (Ping timeout)
- # [18:45] * Joins: jmb (jmb@81.86.70.47)
- # [18:49] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
- # [18:52] * Parts: anne (annevk@213.236.208.22)
- # [18:52] * Joins: anne (annevk@213.236.208.22)
- # [18:53] <anne> indeed, it's not an argument... it's a fact
- # [18:53] * anne ...
- # [19:02] <anne> bloating style= in such a way doesn't make much sense to me
- # [19:03] <anne> also considering all the API changes that would need to be made
- # [19:05] * Joins: hasather (hasather@80.203.71.22)
- # [19:29] * Parts: hasather (hasather@80.203.71.22)
- # [19:30] * Joins: hasather (hasather@80.203.71.22)
- # [19:48] * Joins: Lionheart (robin@198.86.248.1)
- # [19:58] * Quits: Lionheart (robin@198.86.248.1) (Quit: Leaving.)
- # [19:58] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
- # [20:03] * Joins: gavin (gavin@74.103.208.221)
- # [20:19] * Joins: Roger (roger@213.64.74.230)
- # [20:27] * Quits: polin8 (polin8@209.176.7.3) (Ping timeout)
- # [20:30] * Quits: Roger (roger@213.64.74.230) (Quit: Roger)
- # [20:31] * Joins: polin8 (polin8@209.176.7.3)
- # [20:33] * Quits: Jero (Jero@213.46.207.230) (Connection reset by peer)
- # [20:49] <anne> yay, formal objection
- # [20:50] <billmason> Love the analogy to the Patriot Act.
- # [20:51] <tH> can't say i'm surprised :(
- # [20:51] <anne> I suppose he'll formally object until we're done
- # [20:52] <anne> Not entirely sure why people say that it has already been published by the WHATWG...
- # [20:52] <anne> It was a WHATWG Wiki page at some point...
- # [21:02] * Quits: polin8 (polin8@209.176.7.3) (Quit: polin8)
- # [21:05] * Joins: polin8 (polin8@209.176.7.3)
- # [21:06] <DanC> the Patriot Act is so bar beyond the bounds of civility...
- # [21:09] <anne> it seems the FO completely missed which document is being published
- # [21:09] <anne> "i am EXTREMELY uncomfortable in being forced to choose a draft, in toto, as our bsic working draft slash foundational document"
- # [21:17] <tH> on the plus side, i now know more latin than i did when i woke up this morning
- # [21:24] * Joins: Jero (Jero@213.46.207.230)
- # [21:40] <beowulf> as i understand it this is to satisfy the "heartbeat" requirement, and will be continually republished over time?
- # [21:40] <beowulf> until it reaches Recommendation
- # [21:40] <anne> in theory a WG has to publish something every three months or indicate in some other way what it's up to to the public
- # [21:41] <beowulf> yes, that's what i though
- # [21:41] <beowulf> thought
- # [21:41] <anne> there are no real restrictions on what's published though
- # [21:41] <zcorpan_> we can publish tutorials... :)
- # [21:41] <beowulf> in that case i think this is perfect, nothing gives you a better idea of whats what than this (better than design principals imo)
- # [21:44] * Quits: polin8 (polin8@209.176.7.3) (Quit: polin8)
- # [21:44] <beowulf> ah wait, i still haven't worked out my member affiliations
- # [21:45] * beowulf doesn't hit submit
- # [21:51] * Parts: hasather (hasather@80.203.71.22)
- # [21:51] * Joins: hasather (hasather@80.203.71.22)
- # [21:53] * Quits: briansuda (briansuda@85.220.95.76) (Quit: briansuda)
- # [21:55] * Joins: briansuda (briansuda@85.220.95.76)
- # [21:55] * Quits: briansuda (briansuda@85.220.95.76) (Client exited)
- # [22:02] <zcorpan_> i'm positively surprised that firefox doesn't pay attention to namespaces for role
- # [22:03] <anne> oh, I could have told you that
- # [22:03] <zcorpan_> -_-
- # [22:03] <zcorpan_> also for aaa:foo="" ?
- # [22:04] <anne> I knew they made it all work in HTML
- # [22:05] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
- # [22:05] <zcorpan_> is that documented?
- # [22:05] * Joins: polin8 (polin8@209.176.7.3)
- # [22:05] <anne> Don't know about that. Got most of my information from meetings and maybe some bug reports...
- # [22:07] <anne> http://developer.mozilla.org/en/docs/Accessible_DHTML seems to suggest they are namespaced though
- # [22:08] <zcorpan_> that's the thing -- i don't trust specs or documentation
- # [22:09] <anne> yeah, they're mostly wrong
- # [22:09] <DanC> test cases. test cases. test cases.
- # [22:09] <anne> problem with stuff like role= is that reverse engineering is even more annoying
- # [22:10] <zcorpan_> i don't plan to reverse engineer role. unless someone wants me to (i.e., if opera intends to implement it)
- # [22:11] * Joins: gavin (gavin@74.103.208.221)
- # [22:11] <zcorpan_> DanC: yeah. test cases == reverse engineering :)
- # [22:14] * anne made some tests today
- # [22:17] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Quit: Less talk, more pimp walk.)
- # [22:18] * Quits: heycam (cam@203.214.72.248) (Ping timeout)
- # [22:19] * Joins: Lionheart (robin@66.57.69.65)
- # [22:29] * Joins: Bob_le_Pointu (mallory@80.248.208.232)
- # [22:29] <Bob_le_Pointu> Hi there.
- # [22:31] <anne> hi
- # [22:32] <Bob_le_Pointu> Do you think it would be a good idea to make this proposal : using a checksum attribute in some <a> tags to make the UA check itself the integrity of a downloadable file ?
- # [22:32] <gavin_> have you been following the proposal made by Gerv in the Mozilla newsgroups?
- # [22:33] <gavin_> he calls it "Link Fingerprints", and it's exactly that
- # [22:33] <Bob_le_Pointu> No, but if you can send me an URL, I'll see that.
- # [22:33] <gavin_> http://www.gerv.net/security/link-fingerprints/
- # [22:33] <Bob_le_Pointu> Thanks:)
- # [22:34] <gavin_> see also the various threads in mozilla.dev.apps.firefox
- # [22:34] <gavin_> (on google groups or nntp://news.mozilla.org)
- # [22:35] * Joins: heycam (cam@124.168.130.154)
- # [22:35] <Bob_le_Pointu> It's even better than my proposal.
- # [22:36] <Bob_le_Pointu> And it was proposed a long time ago :/
- # [22:37] <schepers> nevertheless, a good idea
- # [22:38] <Bob_le_Pointu> Was it submitted on the list ?
- # [22:38] <gavin_> not on the whatwg/html-wg lists, that I know of
- # [22:39] <gavin_> maybe you could ask Gerv about that, he's a member of this WG
- # [22:39] <gavin_> oh, maybe it was on the whatwg list
- # [22:39] <gavin_> http://wiki.whatwg.org/wiki/Link_Hashes
- # [22:40] <gavin_> http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2006-November/thread.html#7825
- # [22:43] <mjs> we have something like that in Safari but it depends on downloading a checksum file from a trusted site followed by the content from a potentially untrusted site
- # [22:49] <gavin_> the Mozilla extension update system does the same
- # [22:49] <Bob_le_Pointu> So, it's on the way.
- # [22:49] <gavin_> (hash from trusted site, extension from potentially untrusted)
- # [22:49] <gavin_> no, I don't think that has really anything to do with Gerv's proposal
- # [22:49] <gavin_> different things, really
- # [22:54] <Bob_le_Pointu> So, should I ask Gerv to make a proposal on html-wg ?
- # [23:00] * Quits: Jero (Jero@213.46.207.230) (Quit: ChatZilla 0.9.78.1 [Firefox 2.0.0.4/2007051502])
- # [23:00] * Quits: edas (edaspet@88.191.34.123) (Ping timeout)
- # [23:01] * Quits: ROBOd (robod@86.34.246.154) (Client exited)
- # [23:05] * Joins: myakura (myakura@58.88.37.26)
- # [23:17] * Joins: kingryan (rking3@71.202.121.218)
- # [23:36] * Quits: polin8 (polin8@209.176.7.3) (Quit: polin8)
- # [23:38] * Joins: polin8 (polin8@209.176.7.3)
- # [23:40] * Quits: mjs (mjs@64.81.48.145) (Ping timeout)
- # [23:48] * Quits: polin8 (polin8@209.176.7.3) (Quit: polin8)
- # [23:52] * Joins: polin8 (polin8@209.176.7.3)
- # [23:54] * Parts: mw22 (chatzilla@84.41.169.151)
- # [23:55] * Quits: polin8 (polin8@209.176.7.3) (Quit: polin8)
- # Session Close: Thu Jun 28 00:00:00 2007
The end :)