Options:
- # Session Start: Mon Mar 10 00:00:00 2008
- # Session Ident: #html-wg
- # [09:48] * Attempting to rejoin channel #html-wg
- # [09:48] * Rejoined channel #html-wg
- # [09:48] * Topic is 'HTML WG chat http://www.w3.org/html/wg/tracker (logs: http://krijnhoetmer.nl/irc-logs/ ) '
- # [09:48] * Set by DanC_lap on Mon Mar 10 03:08:44
- # [10:17] * Disconnected
- # [10:18] * Attempting to rejoin channel #html-wg
- # [10:18] * Rejoined channel #html-wg
- # [10:18] * Topic is 'HTML WG chat http://www.w3.org/html/wg/tracker (logs: http://krijnhoetmer.nl/irc-logs/ ) '
- # [10:18] * Set by DanC_lap on Mon Mar 10 03:08:44
- # [10:21] * Disconnected
- # [10:21] * Attempting to rejoin channel #html-wg
- # [10:21] * Rejoined channel #html-wg
- # [10:21] * Topic is 'HTML WG chat http://www.w3.org/html/wg/tracker (logs: http://krijnhoetmer.nl/irc-logs/ ) '
- # [10:21] * Set by DanC_lap on Mon Mar 10 03:08:44
- # [10:22] <hsivonen> as for something new like htmlns, we don't need that in order to support SVG and MathML, so we shouldn't get carried away and we shouldn't complicate the authoring of SVG and MathML by making them require more boiler plate
- # [11:28] * Disconnected
- # [11:28] * Attempting to rejoin channel #html-wg
- # [11:28] * Rejoined channel #html-wg
- # [11:28] * Topic is 'HTML WG chat http://www.w3.org/html/wg/tracker (logs: http://krijnhoetmer.nl/irc-logs/ ) '
- # [11:28] * Set by DanC_lap on Mon Mar 10 03:08:44
- # [11:34] * Quits: peepo (Jay@86.147.236.233) (Quit: later)
- # [11:39] * Joins: peepo (Jay@86.147.236.233)
- # [11:40] * Joins: tlr (tlr@128.30.52.30)
- # [11:52] <jgraham_> FWIW I think SVG in text/html has to support quoteless attributes at the parser level but I'm less concerned about whether it is conforming or not (there are good arguments in both directions)
- # [11:53] * jgraham_ can't stay to discuss right now, unfortunately
- # [11:53] <shepazu> "has to"?
- # [11:54] <jgraham_> Well given a byte stream like <svg height=50 width=50> we have to do something with it
- # [11:55] <jgraham_> (draconian errors are not a sensible option)
- # [11:55] <shepazu> spare me
- # [11:56] <jgraham_> OK :)
- # [11:56] * Quits: peepo (Jay@86.147.236.233) (Quit: later)
- # [11:56] <shepazu> this isn't legacy HTML we're talking about
- # [11:56] <jgraham_> This is legacy html we're talking about
- # [11:56] <hsivonen> shepazu: text/html has 3 key features
- # [11:57] <hsivonen> 1) the HTML part of a document is compatible with legacy UAs
- # [11:57] <hsivonen> 2) non-Draconian error handling
- # [11:57] <jgraham_> and even where the content is new it will be produced by legacy HTML CMS's which are ill suited for producing well-formed content
- # [11:58] <hsivonen> 3) can be defined to be compatible with IE without application/xhtml+xml participation from Microsoft
- # [11:58] <shepazu> jgraham_, including new kinds of content like SVG?
- # [11:58] <shepazu> hsivonen, those are design principles
- # [11:58] <hsivonen> shepazu: legacy PHP spaghetti can't ensure the well-formedness of SVG any better than XHTML
- # [11:59] <shepazu> not laws of physics
- # [11:59] <jgraham_> shepazu: Yes.
- # [11:59] <hsivonen> shepazu: I think they are statements of fact given the way the world is now
- # [11:59] <hsivonen> shepazu: calling them "key features" is an opinion
- # [11:59] <shepazu> if you ask DanC, no design decisions about the language are yet final
- # [11:59] * jgraham_ really, really has to leave now.
- # [12:00] <hsivonen> shepazu: the way shipped software *is* is final
- # [12:01] * hsivonen steps away from the computer, too
- # [12:01] * shepazu is seriously considering whether SVG in text/html is a good idea
- # [12:01] <shepazu> if such impositions are going to be put on the language, it may be best to stay away from it
- # [12:01] * Philip would like it if SVG-in-text/html could be handled in browsers like Firefox 3 by using a script that twiddles the DOM
- # [12:01] <shepazu> and only allow it in XHTML serializations
- # [12:02] <shepazu> I'm seriously concerned that ripping up the language in this way would hurt adoption
- # [12:02] <shepazu> because it would create 2 incompatible kinds of UA
- # [12:02] <anne> you're confusing language and serialization
- # [12:02] <shepazu> no, I'm not, anne
- # [12:03] <anne> i disagree
- # [12:03] <shepazu> I know the difference
- # [12:04] <shepazu> do you seriously think that using a different term for it will change the effect in the market?
- # [12:04] <hsivonen> anne: shepazu is right that enabling SVG in text/html would create something that's incompatible with the already shipped XHTML+SVG UAs
- # [12:04] <anne> i agree with that, it's a new thing
- # [12:04] <shepazu> hsivonen, no, that's not what I said
- # [12:05] <Philip> It could be nice if text/html UAs showed SVG images with a right click 'save as...' that always saves as well-formed XML
- # [12:05] <shepazu> I said that changing SVG in the way you're talking about would cause the problem
- # [12:05] <hsivonen> the justification for that incompatibility is that enabling SVG in text/html would enable wider SVG adoption in the future
- # [12:05] <shepazu> we could enforce quoted attributes just fine, you just don't wish to
- # [12:05] <hsivonen> shepazu: sorry. you said 2 incompatible kinds of UA. I thought you were contrasting SVG-in-text/html UAs with existing UAs
- # [12:05] <anne> having SVG UAs support text/html SVG is fairly trivial
- # [12:06] <anne> now text/html parsing is defined anyway
- # [12:06] <Philip> shepazu: What do you think should happen when someone uses unquoted attributes?
- # [12:06] <shepazu> the same thing that happens in SVG
- # [12:07] <Philip> The whole containing (X)HTML document would be replaced with an error message?
- # [12:07] <shepazu> by necessity
- # [12:07] * hsivonen really steps away from the computer
- # [12:07] <shepazu> Philip, you should know by now that that's an exaggeration
- # [12:07] <anne> so you want to introduce fatal error handling in text/html?
- # [12:07] <anne> lol
- # [12:08] <Philip> shepazu: I'm not sure what else it would do, that isn't different to what happens nowadays with SVG
- # [12:08] <shepazu> what percentage of SVG content has this problem?
- # [12:09] <shepazu> well-formedness errors, that is?
- # [12:09] <anne> http://diveintomark.org/archives/2008/03/09/no-fury-like-dracon-scorned
- # [12:09] <anne> anyway, i'm done with this discussion
- # [12:10] <shepazu> approximately 0%
- # [12:10] <shepazu> because the rules are well-known and easy to understand
- # [12:11] <shepazu> anne, I agree it's a silly conversation... of course, it's one you started
- # [12:11] <shepazu> ;P
- # [12:13] <anne> really?
- # [12:14] <shepazu> yes, nobody mentioned quoteless attributes until you dropped the apple of discord in
- # [12:14] * anne shrugs
- # [12:15] <anne> it started way before that, with special casing SVG/MathML, etc.
- # [12:15] <anne> i just provided some illustration on how that would work
- # [12:16] <shepazu> in a way you know is totally unpalatable to everyone involved in SVG today
- # [12:20] <anne> for some dfn of involved you're probably right
- # [12:25] <shepazu> then I wonder why you think it's useful to bring that into the discussion?
- # [12:25] <shepazu> rather than talk about more interesting aspects we can agree on
- # [12:25] <anne> as I said, ter illustration of what special case SVG/MathML would look like
- # [12:26] <anne> and what i'd like to be able to do
- # [12:26] * Quits: tlr (tlr@128.30.52.30) (Quit: tlr)
- # [12:26] <shepazu> why would you want to do that?
- # [12:26] * Joins: tlr (tlr@128.30.52.30)
- # [12:26] <shepazu> what serious benefit is there?
- # [12:26] <anne> some SVG authors on the Web would be fine with seeing "draconian" stuff go btw
- # [12:27] <anne> I said it before, to make the HTML format (not language) internall consistent
- # [12:27] <shepazu> I'm sure that's true, but do they understand the problems that would cause for others?
- # [12:27] <anne> Sam Ruby and Jacques Distler are smart people
- # [12:27] <shepazu> internal consistency is not a very good technical argument, that's a matter of style
- # [12:28] <shepazu> yes, they are
- # [12:28] <anne> is there a technical argument for doing it differently from HTML?
- # [12:29] <shepazu> there's several pragmatic arguments, which is what it should really be about
- # [12:29] <shepazu> toolchains and UAs
- # [12:29] <anne> established HTML tokenizers, parser infrastructure is similar
- # [12:30] <shepazu> if I suggested that we changed tables in a way incompatible with older user agents, you wouldn't even entertain the idea
- # [12:30] <anne> i'm not entertaining this idea either, i'm just answering questions
- # [12:30] <shepazu> when I suggested a very pragmatic fallback for iframe contents, you dismissed it out of hand
- # [12:30] <hsivonen> shepazu: you have to weight the cost against the payoff. we are assuming that XML is holding SVG back and text/html compat would make SVG take off
- # [12:31] <anne> i forgot what you said about <iframe>, but the way it's parsed can't be changed, no
- # [12:31] <anne> (apart from maybe dropping its contents on the floor, that worked for Opera until Acid3 came along)
- # [12:31] <shepazu> anne, you seem to think that your pet use cases are somehow more relevant, but I don't think so
- # [12:31] <anne> it's not a pet use case
- # [12:32] <anne> breaking fundamental assumptions about how HTML works is just not done
- # [12:32] * Quits: tlr (tlr@128.30.52.30) (Quit: tlr)
- # [12:32] <shepazu> anne, I feel the same way about SVG
- # [12:32] <anne> changing HTML tokenizers and parsers to include XML parsing is near insane
- # [12:32] <anne> shepazu, SVG is a language, not a serialization
- # [12:32] <anne> just like we have XHTML
- # [12:33] <shepazu> hsivonen, you may be right, but it's a supposition, not one I've seen backed by evidence
- # [12:33] <shepazu> and it's a very risky one at that
- # [12:33] <anne> if it doesn't take off it's not risky
- # [12:33] <anne> we can simple take it out again :)
- # [12:34] <shepazu> you know better than that
- # [12:34] <anne> if it takes off, well, it takes off, that's good right?
- # [12:34] <hsivonen> shepazu: there is evidence, though, that XML is too hard even for smart people who actively try to do the right thing
- # [12:35] <shepazu> how about if it takes off a little, but not enough to justify changing all the older UAs, just enough to create content that makes people think SVG doesn't work the same way in different places?
- # [12:35] <anne> why is SVG so different from HTML and XHTML?
- # [12:35] <anne> why should I not be able to write SVG in three different ways just like I can do with HTML?
- # [12:35] <shepazu> hsivonen, are you talking edge cases of XML, or just normal aspects?
- # [12:36] <hsivonen> shepazu: I'm talking about edge cases getting hit in day-to-day content management system operation
- # [12:37] <shepazu> hsivonen, give me an example of SVG that's "too hard for smart people" that is related to the serialization
- # [12:38] <shepazu> anne, because SVG UAs don't (and shouldn't have to) understand your pet serialization
- # [12:38] <anne> lol
- # [12:38] <anne> it's hard to believe you're still serious
- # [12:38] <anne> HTML a pet serialization?
- # [12:38] <shepazu> I feel exactly the same about you
- # [12:39] * krijnh hands out cookies
- # [12:40] <shepazu> anne, I mean your pet serialization of SVG, not HTML... HTML is already a mess
- # [12:40] <anne> Again, adding XML support to HTML is non-trivial. Adding HTML support to SVG UAs is trivial.
- # [12:40] * Philip warns krijnh about the dangers of accepting cookies from strangers
- # [12:40] <krijnh> :)
- # [12:41] <shepazu> anne, that's a fallacy
- # [12:41] <hsivonen> shepazu: there's so little SVG on the Web at the moment that I don't have practical examples of the SVG part itself being too hard. However, what I said was that there's evidence about XML being too hard
- # [12:41] <hsivonen> shepazu: the usual case of XML being too hard is that a trackback, comment or search string contains a forbidden Unicode code point
- # [12:42] <shepazu> anne, you're equating 2 different things: a language and a UA
- # [12:42] <hsivonen> shepazu: however, authors accidentally forgetting to close an element happens, too
- # [12:42] <hsivonen> shepazu: I have seen these issues in the XHTML part of blogs that use application/xhtml+xml in order to be able to embed SVG inline
- # [12:42] <shepazu> hsivonen, yes, and that can be easily caught and corrected
- # [12:42] <anne> shepazu, ok, whatever, adding XML support to HTML UAs is non-trivial. Adding HTML support to SVG UAs is trivial.
- # [12:43] <krijnh> http://intertwingly.net/blog/2008/03/10/x-ua-compatible-doesnt-work
- # [12:43] * Philip notes that http://www.google.com/search?q=filetype:svg says there's 783K examples of SVG on the web
- # [12:43] <Philip> (Also the first one gives me an XML parse error in Opera)
- # [12:43] <anne> consider document.write() etc.
- # [12:43] <shepazu> anne, that's not true... XML is much simpler than HTML5
- # [12:43] <anne> no
- # [12:43] <Philip> (because it uses xlink:href without an xmlns:xlink)
- # [12:43] <shepazu> for implementing, I mean
- # [12:43] <anne> I implemented both a HTML parser and an XML parser
- # [12:43] <anne> an HTML*
- # [12:43] <hsivonen> shepazu: creating a need for detect&fix firefighting is not a good idea considering the opportunity cost of the human effort involved
- # [12:44] <anne> about as complicated
- # [12:44] <anne> XML has the internal subset crazyness and all kinds of character checking, and HTML has adoption agencies...
- # [12:44] <anne> (and foster parents)
- # [12:44] <hsivonen> shepazu: I've written an HTML parser and I've non-trivially patched an XML parser from the simpler end of the spectrum
- # [12:44] <shepazu> anne, then why did you just claim the opposite, that XML was more complicated?
- # [12:44] <hsivonen> shepazu: I assure you that XML isn't simple
- # [12:44] <hsivonen> possibly not even simpler
- # [12:44] <anne> shepazu, I said adding XML to HTML is complicated
- # [12:45] <hsivonen> I'd write an XML parser if DTDs weren't a part of XML
- # [12:46] <anne> results 3 and 4 of SVG have a text/plain media type
- # [12:46] <anne> fun stuff Philip
- # [12:47] <Philip> http://www.google.com/search?q=filetype%3Asvg+xlink%3Ahref+-xmlns%3Axlink - do normal SVG UAs ignore those well-formedness errors?
- # [12:47] <anne> Adobe did
- # [12:47] <hsivonen> do they make media type barbies yet?
- # [12:47] <anne> :)
- # [12:49] <shepazu> I actually don't much care about the namespace in the root, actually
- # [12:49] <Philip> http://webdoc.kecskemet.hu/terseg_attekinto.svg doesn't use the SVG xmlns at all - do normal SVG UAs ignore that too?
- # [12:49] <shepazu> nor am I particularly rigid on the xlinkns stuff...
- # [12:49] <hsivonen> do we have a list of which UAs are "normal"?
- # [12:49] <Philip> (http://genotext.stanford.edu/figure_4a.svg too)
- # [12:50] <shepazu> though it would break links in older SVG UAs
- # [12:50] <Philip> http://artist4web.com/Documents/SVGImage.svg - oops :-(
- # [12:52] * Joins: myakura (myakura@122.29.8.215)
- # [12:52] <Philip> It seems like there's roughly as many with the xlink error as there are without
- # [12:52] <hsivonen> shepazu: ok, Philip's examples are SVG being too hard :-)
- # [12:53] <shepazu> hsivonen, that's not an issue of serialization, which is what we were talking about
- # [12:54] <shepazu> Philip, older SVGs probably won't have it, because at the time the primary UA was made, it was thought that the namespace wasn't required... newer UAs are more strict, and so new content will have it
- # [12:54] <hsivonen> shepazu: the media type thing is an issue that could be solved by embedding in text/html as it would remove the need to serve .svg with the right type
- # [12:55] <shepazu> but again, I don't think that the ns declaration is that big a deal
- # [12:55] <hsivonen> shepazu: moreover, what I've suggested would work without xmlns, although xmlns would be allowed for IE8 compat, as a talisman and as a copypaste artifact
- # [12:55] <shepazu> I could happily live without it, though I understand why it was thought to be necessary at the time
- # [12:55] <hsivonen> shepazu: the ns decl is a very big deal in conforming XML systems
- # [12:55] <anne> such as Opera is nowadays...
- # [12:55] <hsivonen> shepazu: should we cherry-pick XML?
- # [12:56] <shepazu> hsivonen, are you trying to be productive or sarcastic?
- # [12:57] <anne> man, try acting in good faith for a change
- # [12:57] <shepazu> anne, that's rich from you
- # [12:57] <hsivonen> shepazu: productive by pointing out that saying that ns decl doesn't matter but quotes do is not a very consistent positition
- # [12:57] <shepazu> hsivonen, I think there are some things that are pragmatically and technically important... I think quoted attributes are one of them
- # [12:58] <shepazu> pragamatically, because all older UAs (ASV, for example) need them
- # [12:58] <hsivonen> if it isn't XML exactly warts and all, then surely various parts are up for discussion
- # [12:59] <shepazu> while ASV doesn't need ns dclarations, and 'svg' is unique enough an element name that it could be usefully identified without the ns decl
- # [13:00] <hsivonen> (that's the basis of my tree builder scope-based proposal)
- # [13:00] <shepazu> history plays a part in this story too, as you know from HTML
- # [13:00] <zcorpan> shepazu: but the Namespaces in XML spec doesn't allow UAs to guess the namespace based on element names
- # [13:01] <shepazu> zcorpan, NSX isn't XML
- # [13:01] <shepazu> and I'm not arguing from a purist perspective
- # [13:01] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Client exited)
- # [13:01] <zcorpan> shepazu: no namespace declaration means no namespace in XML per spec, and unbound prefixes is a namespace-wellformedness error
- # [13:02] <zcorpan> shepazu: top browsers have namespace-aware XML parsers
- # [13:02] <gsnedders> shepazu: you mean ASV supports <svg> inline in HTML? or…?
- # [13:02] <hsivonen> ASV is a zombie
- # [13:02] <shepazu> gsnedders, well... after a fashion
- # [13:03] <gsnedders> I've never heard anything about that, ever.
- # [13:03] <shepazu> gsnedders, you can make a file that uses SVG inline in (X)HTML that will work in both modern browsers and in IE+ASV, using IE's data islands
- # [13:04] <gsnedders> ah.
- # [13:04] <zcorpan> gsnedders: it works in IE7 with ActiveX using either the <object> trigger or application/xhtml+xml as trigger
- # [13:04] <shepazu> in fact, I did so on my home page many years ago
- # [13:04] <shepazu> ASV is indeed a zombie
- # [13:05] <shepazu> but since IE doesn't support SVG...
- # [13:05] <zcorpan> with IE8 it would work in text/html without the <object> to trigger, AIUI
- # [13:05] <hsivonen> is there a non-zombie SVG ActiveX control ready to hook into the new IE8 mechanism?
- # [13:06] <shepazu> hsivonen, I doubt it... but I do know of potential replacements
- # [13:06] <shepazu> sadly, nothing I can talk about publicly
- # [13:06] <hsivonen> :-(
- # [13:06] <shepazu> or even privately, outside W3C Team
- # [13:07] * Quits: ROBOd (robod@89.122.216.38) (Quit: http://www.robodesign.ro )
- # [13:09] <shepazu> the real shame of it is, ASV is still one of the best SVG UAs out there, even if it's not as complete as I'd like it
- # [13:09] <shepazu> it performs better than any of the native implementations in most cases
- # [13:10] <shepazu> (there are exceptions, of course... some things are faster in Opera or Safari)
- # [13:10] <Philip> (Is nothing faster in Firefox? :-) )
- # [13:12] <shepazu> Philip, not that I know of... though FF does have support for a couple of things I think weren't implemented in ASV
- # [13:13] <Philip> Could ASV do things like foreignObject?
- # [13:13] <shepazu> I love FF, and it's the most widespread modern SVG UA on the desktop, but it's not the most performant
- # [13:13] <shepazu> Philip, that's a great example where ASV fell down
- # [13:14] <Philip> I'd guess that's kind of hard to do as a plugin
- # [13:15] <shepazu> well, I heard rumors they actually had XHTML working in internal builds, but I don't think that made it out the door
- # [13:15] <shepazu> but it handled it itself, it didn't hand it off to the browser, IIRC
- # [13:16] <shepazu> so, probably not fun to script for compound documents
- # [13:16] <Philip> Like it was implementing an XHTML UA itself? That does sound kind of hard
- # [13:17] <shepazu> not from other sources I know who've done the same... they said that most of XHTML was pretty trivial
- # [13:18] <shepazu> I don't know if they implemented everything, of course
- # [13:20] * shepazu shakes his head at all the great code that never sees the light of day :(
- # [13:21] <Philip> Most of my code seems much less great once it has seen the light of day and everyone has started finding bugs in it :-)
- # [13:26] <gsnedders> who needs good code anyway? :P
- # [13:27] <Philip> Space shuttle pilots?
- # [13:27] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
- # [13:27] <gsnedders> There aren't many.
- # [13:28] <shepazu> yeah, that's an edge case, Philip
- # [13:28] <shepazu> those aren't real-world users
- # [13:28] <shepazu> by definition
- # [13:28] <gsnedders> they're out of this world users.
- # [13:29] <anne> oh man
- # [13:29] * Philip wonders how many rockets have been lost due to software issues
- # [13:29] <anne> are we now going to get "Please think of the space pilots" pleas whenever someone mentions "real-world"? :d
- # [13:30] <gsnedders> anne: well, we just respond with they're out of this world.
- # [13:30] <Philip> (The only one I remember is the first Ariane 5, but I'd assume there must be others)
- # [13:30] <gsnedders> Philip: yeah, plenty others
- # [13:34] <hsivonen> Ariane didn't have an on-broard human pilot, did it?
- # [13:34] <Philip> No
- # [13:35] <Philip> I don't remember any manned flights being lost due to software
- # [13:37] <hsivonen> must be fun when Vista revokes display drivers on a laptop on a space station
- # [13:37] <Philip> I think they don't use laptops as mission-critical infrastructure :-p
- # [13:37] * shepazu wonders if DVD players even work in space... they are definitely out of their region
- # [13:38] <MikeSmith> by the hoary hosts of Hoggoth
- # [13:38] <hsivonen> wasn't there just a story of a cosmonaut having to spend very expensive time with a software license manager on a space station?
- # [13:38] <MikeSmith> load on my mail server at 27+
- # [13:38] <MikeSmith> what the hell
- # [13:39] <shepazu> unless the boundaries of a country extend indefinitely upward...
- # [13:39] <MikeSmith> tim to hit the air brakes
- # [13:46] * Joins: Julian (chatzilla@217.91.35.233)
- # [13:52] * Quits: zcorpan (zcorpan@88.131.66.80) (Ping timeout)
- # [14:22] * Joins: aaronlev (chatzilla@24.61.74.31)
- # [14:37] <Dashiva> shepazu: You could always bring a DVD player from the surface, instead of trying to create a new one :P
- # [15:12] * Joins: matt (matt@128.30.52.30)
- # [15:13] * Joins: ROBOd (robod@89.122.216.38)
- # [15:23] * Quits: Lachy (Lachlan@213.236.208.22) (Quit: Leaving)
- # [15:23] * Joins: billmason (billmason@69.30.57.25)
- # [15:33] <anne> 01:00:00 UTC
- # [15:33] <anne> there goes Europe
- # [15:34] <hsivonen> ?
- # [15:35] <anne> http://www.w3.org/mid/47D54508.5080805@w3.org
- # [15:35] <hsivonen> ah.
- # [15:36] <anne> i've probably done my 3 hours of test making already
- # [15:36] * Joins: Lachy (Lachlan@213.236.208.22)
- # [15:40] <Dashiva> Nobody important lives in Europe anyway
- # [15:41] <krijnh> -_-
- # [15:42] <gsnedders> :|
- # [15:43] <anne> MikeSmith, heh, you didn't have to do that :)
- # [15:43] <anne> we want you to make tests!
- # [15:48] <MikeSmith> I need an personal intern
- # [15:49] <MikeSmith> hey intern, write a few hundred test cases please
- # [15:51] <MikeSmith> anne - seriously, I think it's actually more important to choose a time that allows, say, Philip, jgraham_, zcorpan, you to be involved in test-case discussion
- # [15:51] <MikeSmith> I said as much here earlier today
- # [15:53] <MikeSmith> issue-37?
- # [15:53] * trackbot-ng getting information on ISSUE-37
- # [15:53] <trackbot-ng> ISSUE-37 -- Integration of SVG and MathML into text/html -- RAISED
- # [15:53] <trackbot-ng> http://www.w3.org/html/wg/tracker/issues/37
- # [15:53] <anne> fair enough
- # [15:53] * Joins: aroben (aroben@71.58.69.205)
- # [15:53] <MikeSmith> jgraham_, great right-up
- # [15:54] <MikeSmith> "write-up", make that
- # [15:54] <anne> so even native speakers mix that up
- # [15:54] <anne> good to know
- # [15:55] <MikeSmith> mentally challenged native speakers like me at least
- # [15:55] <anne> hehe
- # [15:56] <MikeSmith> I guess I should say differently abled native speakers like me
- # [15:56] <MikeSmith> not sure what my different abilities are, though
- # [15:58] <MikeSmith> btw, for those who care about relative mailing-list volume:
- # [15:59] <MikeSmith> with some help from Gerald Oskoboiny today, I looked at what the highest-traffic W3C mailing lists are
- # [15:59] <MikeSmith> actually, not really with help
- # [15:59] <MikeSmith> Gerald did the work
- # [16:00] <MikeSmith> anyway, not surprisingly, public-html is in the top 3
- # [16:00] * Quits: DanC_lap (connolly@128.30.52.30) (Ping timeout)
- # [16:00] <Dashiva> What are the other two?
- # [16:00] <MikeSmith> highest is www-style
- # [16:00] <MikeSmith> no big surprise there
- # [16:01] <anne> what timeframe did you use?
- # [16:01] * anne wonders what the last one is
- # [16:01] <MikeSmith> just looking through the archives that shoe numbers for each month
- # [16:01] <anne> so all months are included?
- # [16:01] <MikeSmith> http://lists.w3.org/Archives/Public/public-owl-wg/
- # [16:02] <anne> because public-html is on a declining line
- # [16:02] <MikeSmith> Gerald looked at just the last 30 days
- # [16:02] <MikeSmith> OWL WG list had more messages than public-html in January
- # [16:02] <MikeSmith> and 600+ messages in November
- # [16:03] <anne> we had 1773 in April last year :p
- # [16:03] <MikeSmith> I like the look of that number
- # [16:03] <MikeSmith> let's aim for 1776 for this April
- # [16:03] <anne> or 1337
- # [16:03] <MikeSmith> we shall call it "Tribute to America"
- # [16:03] <MikeSmith> oh yeah, 1337 way better
- # [16:03] <anne> heh
- # [16:04] <gsnedders> oh, and FWIW, I should make it to the TP this year
- # [16:04] <anne> we were close in May 2007
- # [16:04] <anne> 1307
- # [16:05] <MikeSmith> anyway, it definitely won't do to have a Semantic Web mailing list outdoing us in terms of volume
- # [16:05] <MikeSmith> so we need to do something to correct this
- # [16:05] <MikeSmith> I recommend using a variation of "Takahashi" presentation style for replies
- # [16:07] <MikeSmith> break replies into multiple messages, send them out in serial fashion
- # [16:07] <MikeSmith> heightens the suspense
- # [16:07] <MikeSmith> that's how Charles Dickens became famous
- # [16:07] <anne> did you take into account blog traffic on html5 and the WHATWG mailing list?
- # [16:07] <MikeSmith> despite his poor writing skills
- # [16:08] <anne> and all our other W3C mailing lists?
- # [16:08] <anne> that our approach is decentralized (with a centralized editor, mind you) should not put us at a disadvantage :)
- # [16:09] <gsnedders> why can we decentralize our editor?
- # [16:09] <MikeSmith> no, seriously, I was just looking at in the context of people saying that the traffic on public-html is excessively high
- # [16:09] <gsnedders> *can't
- # [16:09] <MikeSmith> and data does not seems to support it being excessive nor impossible to keep up with
- # [16:10] <MikeSmith> not relative to lists for other high-interest W3C work
- # [16:10] <MikeSmith> at least
- # [16:11] <MikeSmith> ..
- # [16:12] <MikeSmith> on other subjects of importance, it looks like we will be able to have the next WAF f2f in Dublin for sure
- # [16:13] <MikeSmith> which means less travel for me, yay-hay
- # [16:13] * anne won't be there
- # [16:13] <MikeSmith> anne - ?
- # [16:13] <MikeSmith> damn
- # [16:13] <MikeSmith> I thought you'd be at XTech
- # [16:14] <MikeSmith> you have a schedule conflict?
- # [16:14] <anne> I was planning to go there, yes :)
- # [16:14] <MikeSmith> hmm, if Turin in June would work better for you, maybe we should do Turin in June instead
- # [16:14] * Joins: aroben_ (aroben@71.58.69.205)
- # [16:16] * Quits: aroben (aroben@71.58.69.205) (Ping timeout)
- # [16:18] <anne> MikeSmith, it's not that, it's that the meeting is for Widgets and I'm no longer doing that
- # [16:20] * Quits: Zeros (Zeros-Elip@69.140.40.140) (Quit: This computer has gone to sleep)
- # [16:20] <MikeSmith> ah yeah, I had forgotten it was for Widgets only
- # [16:27] * Joins: DanC_lap (connolly@128.30.52.30)
- # [16:41] * Quits: Lionheart (robin@66.57.69.65) (Connection reset by peer)
- # [16:42] * Joins: Lionheart (robin@66.57.69.65)
- # [16:42] * Joins: Sander_ (svl@86.87.68.167)
- # [16:43] * Quits: Sander (svl@86.87.68.167) (Ping timeout)
- # [16:43] * Sander_ is now known as Sander
- # [16:44] * Quits: DanC_lap (connolly@128.30.52.30) (Ping timeout)
- # [16:47] * Quits: myakura (myakura@122.29.8.215) (Quit: Leaving...)
- # [16:49] * Quits: Lachy (Lachlan@213.236.208.22) (Quit: This computer has gone to sleep)
- # [17:06] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Client exited)
- # [17:26] * Joins: dbaron (dbaron@67.160.251.228)
- # [17:45] * Joins: adele (adele@17.203.14.235)
- # [17:45] * Joins: tlr (tlr@128.30.52.30)
- # [17:52] * Joins: Lachy (Lachlan@84.215.54.100)
- # [18:13] * Quits: gsnedders (gsnedders@86.138.199.53) (Quit: Partying in teh intarwebs)
- # [18:14] * Quits: hsivonen (hsivonen@130.233.41.50) (Ping timeout)
- # [18:15] * Joins: gsnedders (gsnedders@86.138.199.53)
- # [18:18] * Joins: hsivonen (hsivonen@130.233.41.50)
- # [18:33] * Joins: DanC_lap (connolly@128.30.52.30)
- # [18:38] * Quits: tlr (tlr@128.30.52.30) (Quit: tlr)
- # [18:43] * Joins: aroben__ (aroben@71.58.69.205)
- # [18:44] * Joins: smedero (smedero@192.223.6.251)
- # [18:46] * Quits: aroben_ (aroben@71.58.69.205) (Ping timeout)
- # [18:46] * aroben__ is now known as aroben|meeting
- # [18:52] * Quits: DanC_lap (connolly@128.30.52.30) (Ping timeout)
- # [18:56] * Quits: aaronlev (chatzilla@24.61.74.31) (Client exited)
- # [18:58] * Joins: aaronlev (chatzilla@24.61.74.31)
- # [19:33] * Quits: dbaron (dbaron@67.160.251.228) (Quit: 8403864 bytes have been tenured, next gc will be global.)
- # [19:45] * Quits: billmason (billmason@69.30.57.25) (Connection reset by peer)
- # [19:48] * Joins: billmason (billmason@69.30.57.25)
- # [19:50] * Joins: brianlandau (brianlanda@24.106.184.150)
- # [19:50] * Joins: dbaron (dbaron@63.245.220.241)
- # [19:53] * Quits: aaronlev (chatzilla@24.61.74.31) (Ping timeout)
- # [20:03] * Quits: Lionheart (robin@66.57.69.65) (Ping timeout)
- # [20:14] * Quits: Julian (chatzilla@217.91.35.233) (Quit: ChatZilla 0.9.81 [Firefox 2.0.0.12/2008020121])
- # [20:20] * Joins: aaronlev (chatzilla@65.78.0.149)
- # [20:26] <aaronlev> anne: i need someone to poke holes in my draft for ARIA user agent best practices
- # [20:27] <aaronlev> http://developer.mozilla.org/en/docs/ARIA_UA_Best_Practices
- # [20:27] <anne> I'd suggest Simon
- # [20:27] * anne looks
- # [20:28] <anne> hmm large :)
- # [20:28] * hsivonen looks
- # [20:31] <aaronlev> anne: simon can't get to it for a while
- # [20:31] <hsivonen> editorial: some list items are broken so that the first line is in the list item and the rest is in a paragraph that breaks the list
- # [20:32] <aaronlev> i see one place
- # [20:32] * Quits: Lachy (Lachlan@84.215.54.100) (Ping timeout)
- # [20:32] <hsivonen> two items under 11.3.1.1 HTML 5 Tabindex
- # [20:35] * Quits: dbaron (dbaron@63.245.220.241) (Quit: 8403864 bytes have been tenured, next gc will be global.)
- # [20:35] <hsivonen> aaronlev: ID matching should be defined. I hope the matching is equivalent to getElementById
- # [20:35] <aaronlev> hsivonen: i use nsIContent::GetID()
- # [20:35] <aaronlev> which in HTML i think it just getElementById
- # [20:36] <aaronlev> but i don't know what other complexities are in there
- # [20:36] <aaronlev> i guess you can have a different id attribute technically
- # [20:36] <aaronlev> ?
- # [20:37] <hsivonen> in XML, yes
- # [20:37] <aaronlev> right
- # [20:37] <aaronlev> so in the future of ARIA can be applied to XML
- # [20:37] <aaronlev> then you'd want to do that
- # [20:37] <aaronlev> should we word it so it's future proofed
- # [20:38] * Joins: jgraham (james@81.86.215.67)
- # [20:38] <aaronlev> if you want to edit the wiki feel free
- # [20:38] <aaronlev> anne: we still need info on how to expose this stuff via universal access
- # [20:38] * Joins: Lachy (Lachlan@88.91.109.123)
- # [20:38] <hsivonen> is Universal Access the OS X API?
- # [20:38] <aaronlev> hsivonen: yes
- # [20:39] <aaronlev> anne: and opera is taking the lead on that -- do you guys have any docs on how you're exposing ARIA?
- # [20:39] <hsivonen> aaronlev: it doesn't say how aria-foo=bar maps to a name-value pair
- # [20:39] <hsivonen> is the name foo or aria-foo or something else?
- # [20:40] <aaronlev> that should be under states an object properties
- # [20:40] <hsivonen> ah. ok
- # [20:40] <anne> aaronlev, no docs that I know of
- # [20:41] <aaronlev> anne: well then we can't yell at other user agent mfrs when they do it differently from you guys
- # [20:41] <anne> I'm afraid that other than "I'd like the syntax to be like this" I can't be of much help
- # [20:41] <aaronlev> anne: unless you want people to have to reverse engineer Opera's ARIA support
- # [20:41] <anne> It'd be best to ask Simon
- # [20:41] <aaronlev> k
- # [20:41] * Quits: Lachy (Lachlan@88.91.109.123) (Ping timeout)
- # [20:41] <anne> for MS though it would be best to run their test cases in both Opera and MS
- # [20:42] <anne> that's what we do when we add new functionality anyway
- # [20:42] * Quits: ROBOd (robod@89.122.216.38) (Quit: http://www.robodesign.ro )
- # [20:43] <anne> aaronlev, the <iframe> stuff, is that also applicable to <embed> and <object>?
- # [20:44] * Joins: DanC_lap (connolly@128.30.52.30)
- # [20:44] <hsivonen> aaronlev: "collect the name from the content subtrees pointed to by aria-labelledby" is not precise enough
- # [20:45] <hsivonen> does it mean the document-order concatenation of text (and CDATA) node children with runs of one or more whitespace characters replaced with a single space?
- # [20:45] <hsivonen> (oh it does say depth first)
- # [20:46] <hsivonen> <title> typo
- # [20:47] <aaronlev> hsivonen: it's described in the next list below that one
- # [20:47] * Quits: billmason (billmason@69.30.57.25) (Quit: .)
- # [20:47] * Joins: Lachy (Lachlan@84.215.54.100)
- # [20:48] <hsivonen> aaronlev: ah. ok. (meta comment: the doc doesn't que the reader for subsequent definitions)
- # [20:48] <aaronlev> i just fixed it in that one case
- # [20:48] <aaronlev> are there others?
- # [20:49] <hsivonen> aaronlev: the name-value thing
- # [20:50] <aaronlev> ok
- # [20:50] <aaronlev> hsivonen: can you suggest some text for the ID matching thing?
- # [20:50] * Joins: billmason (billmason@69.30.57.25)
- # [20:51] <hsivonen> aaronlev: something along the lines that an ID matches the element that is returned by getElementById with the ID as an argument
- # [20:51] <hsivonen> assuming that the API you are using indeed matches gEBI
- # [20:51] <aaronlev> should we allow ARIA to be used with general XML?
- # [20:52] <aaronlev> and goes getElementById use the ID attribute defined by the XML?
- # [20:52] <anne> getElementById does ID based matching
- # [20:52] <anne> so yes
- # [20:52] <anne> I'm not sure if ARIA should be used with general XML...
- # [20:52] <aaronlev> what's the argument against it
- # [20:53] <aaronlev> why bother restricting it
- # [20:55] <anne> proprietary markup shouldn't be send over the wire
- # [20:55] <hsivonen> I wouldn't restrict it
- # [20:55] <anne> http://ln.hixie.ch/?start=1064828134&count=1
- # [20:56] <hsivonen> aaronlev: doesn't aria-templateid couple a site and an AT?
- # [20:56] <hsivonen> is that a good idea?
- # [20:57] <aaronlev> hsivonen: i'm asking pfwg for clarification, it seems bad to me so far
- # [20:57] * Quits: DanC_lap (connolly@128.30.52.30) (Ping timeout)
- # [20:57] <aaronlev> apparently it happened when i wasn't there
- # [20:57] <aaronlev> i don't see any implementation for it
- # [20:57] <aaronlev> but al is saying that it's not AT dependent
- # [20:57] <Dashiva> Why is there a picture of bin laden next to that post... I wonder what keyword pick that out
- # [20:57] <aaronlev> it's more of an ID than a pointer to a script
- # [20:57] <aaronlev> hsivonen: which is weird since the def says it's a URI
- # [20:58] <hsivonen> URIs-as-identifiers strikes again
- # [20:58] <aaronlev> anne: but if XBL gets impl'd in browsers proprietary market would be fine right?
- # [20:58] <aaronlev> hsivonen: i guess that's what it is
- # [20:58] <aaronlev> i don't see the need
- # [20:59] <aaronlev> i think the need should be proven
- # [20:59] <anne> aaronlev, I don't think so...
- # [20:59] <hsivonen> does ATK reuse MSAA interface names for AccessibleValue etc?
- # [20:59] <anne> XBL is an optional language
- # [20:59] * Quits: Lachy (Lachlan@84.215.54.100) (Quit: Leaving)
- # [20:59] * Joins: Lachy (Lachlan@84.215.54.100)
- # [21:00] <aaronlev> it is now but because not all the major browsers support it
- # [21:00] <aaronlev> that could change
- # [21:00] <hsivonen> CSS is optional only in theory
- # [21:00] <aaronlev> my opinion is that the web is barely managing right now, doing everything in HTML and JS
- # [21:00] <aaronlev> and that some better way of doing custom widget elements will be needed at some pt
- # [21:01] <aaronlev> i just don't see the need to restrict ARIA
- # [21:01] <aaronlev> if people don't think it's a good idea to send XML fine, but why stop ARIA from being used if they do
- # [21:01] <aaronlev> mozilla doesn't check the namespace of the element anymore, when ARIA is used
- # [21:02] <anne> that sounds like a violation of ARIA
- # [21:02] <anne> but i don't feel strongly about that
- # [21:03] <aaronlev> is you show me where
- # [21:03] <aaronlev> s/is/if
- # [21:03] <aaronlev> that will help
- # [21:03] <hsivonen> I think it's reasonable to implement it without checking the namespace
- # [21:03] <hsivonen> anne: it's good for forward compat
- # [21:03] <anne> as i said, i don't care
- # [21:03] <aaronlev> ok
- # [21:03] <hsivonen> anne: if FooML magically becomes part of the Web platform, ARIA works
- # [21:03] <anne> simon told me they added a note to the spec that said it was not allowed or something
- # [21:04] <aaronlev> anne: i didn't see that note
- # [21:04] <anne> we should've done that for id=
- # [21:04] <aaronlev> i think simon means his spec
- # [21:04] <anne> and class=
- # [21:04] <aaronlev> not the w3.org spec
- # [21:04] <anne> no, I don't think so
- # [21:05] <hsivonen> aaronlev: I don't have further comments on the document. It looks good, but I don't know the APIs well enough to know for sure.
- # [21:05] <aaronlev> hsivonen: that's fine great thanks
- # [21:05] <aaronlev> hsivonen: btw another forward compat feature i'm working on is to expose general aria- attributes
- # [21:05] <aaronlev> just to pass them through if we don't know anything about them
- # [21:06] <anne> from the ARIA spec
- # [21:06] <anne> "User agents MUST NOT implement ARIA support in such a manner that they recognize the ARIA attributes in any namespace"
- # [21:06] <aaronlev> that will allow some freedom to experiment outside of fire vox and DOM-based screen readers
- # [21:06] <aaronlev> ok
- # [21:06] <anne> not sure what that statement means though
- # [21:06] <aaronlev> i think it means what you said
- # [21:06] <anne> i guess so
- # [21:07] <anne> that's what simon said anyway
- # [21:17] * Quits: aroben|meeting (aroben@71.58.69.205) (Quit: Leaving)
- # [21:17] * Joins: aroben (aroben@71.58.69.205)
- # [21:23] * Quits: brianlandau (brianlanda@24.106.184.150) (Quit: brianlandau)
- # [21:34] * Joins: DanC_lap (connolly@128.30.52.30)
- # [22:01] * Quits: trackbot-ng (trackbot-n@128.30.52.30) (Client exited)
- # [22:06] * Joins: trackbot-ng (trackbot-n@128.30.52.30)
- # [22:06] * trackbot-ng HTML Issue Tracking http://www.w3.org/html/wg/tracker/
- # [22:06] * Quits: jgraham (james@81.86.215.67) (Ping timeout)
- # [22:06] * Quits: jgraham_ (jgraham@81.86.215.67) (Ping timeout)
- # [22:06] * Joins: jgraham_ (jgraham@81.86.216.20)
- # [22:07] * Quits: trackbot-ng (trackbot-n@128.30.52.30) (Client exited)
- # [22:07] * Joins: jgraham (james@81.86.216.20)
- # [22:09] * Joins: dbaron (dbaron@63.245.220.241)
- # [22:10] * Joins: trackbot-ng (trackbot-n@128.30.52.30)
- # [22:10] * trackbot-ng HTML Issue Tracking http://www.w3.org/html/wg/tracker/
- # [22:58] * Quits: aroben (aroben@71.58.69.205) (Quit: aroben)
- # [22:58] * Quits: jgraham (james@81.86.216.20) (Quit: I get eaten by the worms)
- # [22:59] * Joins: aroben (aroben@71.58.69.205)
- # [23:00] * Quits: DanC_lap (connolly@128.30.52.30) (Ping timeout)
- # [23:00] * Joins: jgraham (james@81.86.216.20)
- # [23:21] * Quits: jgraham (james@81.86.216.20) (Quit: I get eaten by the worms)
- # [23:23] * Quits: matt (matt@128.30.52.30) (Quit: matt)
- # [23:27] * Joins: jgraham (james@81.86.216.20)
- # [23:30] * Quits: jgraham (james@81.86.216.20) (Quit: I get eaten by the worms)
- # [23:48] * Joins: thorn (David_Lind@24.147.187.203)
- # Session Close: Tue Mar 11 00:00:00 2008
The end :)