See also: IRC log
<Lachy> MikeSmith, unfortunately, I can't call in, since I don't have a phone at home.
<Lachy> MikeSmith, it would be good if you could set up skype
Lachy - no can do
<smedero> Ugh, traffic - just got to the Hyatt.
<smedero> Is the meeting on the 14th floor still?
oedipus - DanC says he would rather not right now, but we can play around later at the break
I missed a few people when I did the present list
So if you are on IRC and don't see your name on the following list:
<karl> [Danc is explaining the process document with a drawing of Ian Hickson]
scribe: then please do a "Present+ YourName" to add yourself to the present list
<hober> What if you're in IRC but not at the TP?
<Hixie> anne: snakes go from the head to the tail
<anne> Hixie, oh, so I got them backwards
hober - then please do, e.g., YourName_IRC_only or something like that
<ChrisWilson> is this an internationalization difference? (In the US I think the game is marketed as chutes & ladders)
<anne> (the above ln.hixie.ch image is being discussed)
<karl> anne, there are steps you can jump over. Basically there are entrance criterias not exit criterias
<hsivonen> so the snakes went from each stage backwards
<mauro> Meeting: HTML WG F2F Meeting
<MichaelC> ARIA Overview: http://www.w3.org/WAI/intro/aria
<hsivonen> Ben arrived
Aron Leventhal speaking about ARIA ... ARIA intro
<ChrisWilson> http://www.w3.org/TR/aria-roadmap/ is a useful start too.
32 or more states in a typical accessibility API
aaronlev: MSAA (Microsoft Active Accessibility) ...
Mike Squillace arrives
<ChrisWilson> Good overview/start to MSAA is http://msdn2.microsoft.com/en-us/library/ms697707.aspx
aaronlev: HTML is missing basic things that we have on the desktop ...
... ridiculous, for example, that we don't have a tree-view [mechanism] in HTML
... widget toolkit takes care of a lot of the work
<gsnedders> The number of common widgets that need to be implemented time and time again with JS is maddening
aaronlev: because of DOM ability for mutation events, can fire an event for [accessibility needs]
aaronlev: the hard work is on the side of the JS author
<scribe> Scribenick: MikeSmith
<gsnedders> At times you'll never get as good accessibility with JS as you would with native widgets
[getting sound from Mike Squillace's machine plugged in]
<ChrisWilson> gsnedders - that's the semantic problem ARIA tries to address
Marcos is videotaping Mike's demo
<gsnedders> ChrisWilson: even with the complete ARIA, I still have my doubts about whether we can recreate everything
<gsnedders> [that could be done natively]
<ChrisWilson> by adding all the relevant controls and semantics to HTML, you mean?
Mike slows down his JAWS to 20% and mjs jokes it reminds him of TimBL slowed-down take from yesterday
[Mike is pressing controls in a webapp and getting feedback from jaws]
<gsnedders> ChrisWilson: the diversity of AT platforms and how they are implemented makes me wonder whether one thing could really support everything
Demo is "Happy Time Pizza On-line Ordering System"
<anne> (ARIA is very low-level)
<ChrisWilson> that's true, but the idea of ARIA (and similar ATs) is to allow mapping application components to a smaller, more high-level semantic set of controls.
<olivier> Mikesmith: I can see the screen, but the URI is a tad long
<ChrisWilson> sorry, not perfect decsription.
aaronlev - can you paste in the URL for the demo please
<oedipus> which widget is it -- i can probably get the uri if you tell me
<ChrisWilson> the idea is to capture the semantic roles of components; there may be multiple different components that act as a "button"
<MichaelC> Most of this demo comes from http://test.cita.uiuc.edu/aria/
[mike is now doing Dijit Tree Test demo]
<dglazkov> t.v. raman once mentioned using mobile gmail as the accessible (and quality) equivalent to the full-blown gmail
<anne> ChrisWilson, the idea is that the author implements the accessibility on top of a widget he already created
<dglazkov> perhaps that's the solution
mike: If you use Dojo toolkit now, a lot of [this] is already built into the library.
<dglazkov> instead of taking a complex Web app and attaching AT hooks to it
aaronlev: Mike is using the standard keys for a tree view
<dglazkov> build your app with simple blocks, available in HTML
<dglazkov> and use JS to enhance the UX
<gsnedders> if I create a slider in JS, is there any way it could behave identically to a native one on OS X using ARIA?
mike: just as if I were in Windows Explorer ... just like on a desktop
<ChrisWilson> anne - umm, sort of. the idea is that the author describes the semantic role of their components, so that accessibility tools can make use of it
<oedipus> the dijit tree test works on my FF3 install (last release) with JAWS 8.0.2173
<dglazkov> while providing the opportunity to turn off the enhancement
olivier: Is there a specific intonation for indicating [this metadata]
<anne> ChrisWilson, is it used for anything else?
<gsnedders> (even if we ignore the visible part of the UI for now)
<ChrisWilson> anne, the semantic role?
<anne> dglazkov, that's the idea of CSS + XBL basically, if I understand you correctly (but that's more a long term solution :-(
tex: wouldn't you want to say, give me the list of all elements at this level? ...
<dglazkov> yes.. I guess I just don't like the ARIA proposal, because it's ultimately a dead-end.
tex: why not just map this to a completely different model?
<oedipus> dglazkov: why do you say it is a dead-end?
<dglazkov> well, oedipus, I may be to harsh in this characterization
aaronlev: there probably is a more efficient model from presenting the data for this context ... but it would be something completely foreign to the average desktop user
... JS widgets will always be a reality ...
<dglazkov> but I think that we shouldn't be taking complex UIs and making them accessible
<Lachy> because it's way too complex for average authors to even think about using and you're relying on the distribution and use of pre-written toolkits for this to be even remotely successful
<mjs> it's a dead end only if all control types covered by AX apis end up available in HTML, and we get enough power to fully customize look and feel of all widgets
aaronlev: sometimes the best UI is just one that people [already] know
<oedipus> ARIA is the first step and suffices for generalized markup languages, but the intermediary where specialized knowledge domain markup is used -- consult: http://www.linux-foundation.org/en/Accessibility/Handlers
<anne> mjs, yeah, CSS is the problem...
<gsnedders> Lachy: most authors just use text/links, though, in fairness
<mjs> anne: I dunno if it's just CSS - HTML doesn't have built-in support for many OS-standard widget types, even in the current HTML5 draft
<jgraham_> Lachy: you rely on prewritten toolkits in desktop apps too; to be fair
<oedipus> mjg, by AX do you mean accessibility (a11y)?
<mjs> oedipus, yes, AX is the abbreviation used by the Mac OS X accessibility APIs
<oedipus> thanks, i didn't know
kevin: Think of something like CoverFlow
Hixie: people are still making up their own widgets
<oedipus> mjs: the last time i used a mac, i could still see... it's hard to get agencies for the blind to train on anything other than windows
<mjs> CoverFlow is just a list view with a bonus magical display
<mjs> oedipus, that's too bad - VoiceOver is supposedly pretty good
[one more demo from mike now]
<Lachy> I used VoiceOver briefly to test it out, and it was remarkebly simple even for me
[mike is showing a demo of navigating a table]
mike: I hate navigating spreadsheets
... the last thing I want to do is go navigating cell-by-cell through a spreadsheet
[mike moving through column headers]
[mike showing how to edit contents of cells]
[mike hitting Ctrl+Alt to pull up menus]
hsivonen - are the table semantics from HTML only or with ARIA markup?
aaronlev - some ARIA to say that the row headers are read-only
[demo finished ... applause ... DanC stands and stretches]
[next up is demo from Ben]
<karl> oedipus, we are collecting questions
[waiting from Ben to get set up]
<MichaelC> Regarding customization of presentation: ARIA provides the semantics to allow Assistive Technology (AT) to customize presentation - AT could compete on how they optimize the presentation of particular widgets based on these semantics, but we can't expect authors to provide multiple parallel versions
[we got musical accompaniment from DanC on the guitar]
[Ben Millard introducing topic of data tables]
<anne> He explains: http://sitesurgeon.co.uk/tables/
ben: found out there are many tables on the Web that weren't in HTML ... some ASCII tables, some Excel, etc.
... found some instances of tables lacking any header rows at all
<ChrisWilson> oedipus, maybe we should make that requirement for all discussion.
[Ben is walking us through the "How Authors Indicate Headers" part of page at URL above]
<oedipus> ChrisW: sounds good to me -- should i go down into the basement to mix up the medicine?
[DanC preparing to do some music]
<Hixie> IH: I'm encouraged by these numbers
<Hixie> (they indicate higher quality tables are more common than i expected)
[getting ready for demo from Maciej]
[Maciej beginning his demo]
<ChrisWilson> [mjs demoing arbitrary affine transforms thru CSS]
<karl> so anne can talk
... has really good fallback behavior
... if you view same content in a browser that doesn't support the animation properties
... no special hardware being used
Lachy: because it's cool, like us
<ChrisWilson> karl, I was hearing the Ginsu knife TV ad...
<karl> might be a cultural thing
[maciej is showing the markup source and CSS]
<anne> dglazkov, someone makes a movie, so maybe later
maciej: this demo maybe shows a need for extensions to form controls
... <button type=toggle> or <button type=radio>
<smedero> I'm on it
<anne> I think that in that case it doesn't match to radio really well
<anne> Which typically has a single one selected
"but wait, there's more" was invented by RonCo
<anne> (no toggling behavior)
ChrisWilson: you gotta put "q+ to ..."
<ChrisWilson> MikeSmith, thanks, forgot to use "to" rather than for"
<Zakim> hsivonen, you wanted to pitch conformance checking test case discussion
<Zakim> ChrisWilson, you wanted to give versioning pitch
<Zakim> Ben, you wanted to pitch for more info on tables
<Zakim> jgraham_, you wanted to give test case organisation
<anne> Henri wants to talk about systematic validator testing
<anne> Chris wants to talk about versioning, backwards compat, etc.
jgraham_ points out that we have accumulated a small pile of disorganized test cases ...
<anne> Ben wants to talk about data tables (header association, etc.)
scribe: what I want is to have a discussion about what test cases we need, how to best organizes the ones we already got
<anne> James wants to talk about organizing existing testcases, licensing tests, etc.
<Zakim> Hixie, you wanted to pitch for status thingy and to pitch for test thingy
Hixie: spec has various sections with varying levels of stability ...
<anne> Ian wants to talk about stability annotation within the specification (status thingy)
Hixie: would like to have a way to annotate stability of different parts of spec
<anne> Ian wants to teach people to write tests
Hixie: tutorial session on how to write test cases
... You will write test cases and I will tell you what you are doing wrong [laughs]
mjs: session or sweatshop?
I like the word niutes
<Zakim> danc, you wanted to talk about "testing is hard"
DanC: Storytelling and test cases ...
... karl has some great slides on this
<Zakim> mjs, you wanted to pitch for offline support and to pitch for media elements
mjs: offline support ... we have done some implementation work, I know Mozilla has also
<Zakim> karl, you wanted to pitch for HTML 5 for author
karl: with HTML5 we have a big clash with [previous expections] of the community toward W3C specs ...
... [need to consider how to address the expectations of the authoring community]
<Hixie> there are 12 items
<Hixie> at the moment
song for Håkon: 6ish people raise hands
song for Håkon, 6ish people raise hands
<smedero> mjs: do you have a public URL for your CSS animation demo? (I have the URL for the WebKit blog entry on it)
<mjs> smedero: I will post it publicly soon
<mjs> smedero: I only finished this version right before presenting
<smedero> alright, no rush... just compiling info for the issue. :)
<mjs> smedero: last version used content that I couldn't redistribute
[we are effectively breaking until 4pm ... ]
<molly> wondering if it's worth coming over for the remainder of the day or having a little me time and catching up tomorrow. Advice?
<Lachy> MikeSmith, now that the break is on, is there someone who can organise a skypecast? Just need someone with skype and a decent mic to pick up the speakers
<oedipus> if there is my skypename is oedipusnj -- marcos captured the first 2 demos (at least) on his webcam
<Lachy> oedipus, are you participating via IRC only?
<oedipus> the first demo is (temporarily at): http://22.214.171.124/aria.avi - note it takes quite some time to load, but eventually will
<Lachy> oedipus, has that server got a slow upload speed? It's not going very fast at all
<oedipus> lachy, it took forever and a day, and then suddenly started playing -- for some reason, won't play before the whole thing is buffered, so when it started, i saved it locally and then listened to it
<Lachy> yeah, I'm trying to save it locally
<Lachy> what codec is it using?
<oedipus> it is an AVI file
<Lachy> Could it be converted to ogg theora or something with a bit better compression?
<Lachy> AVI is just a container format. It could contain almost any codec
<oedipus> i have a codec conversion utility -- i'll "see" what i can do
<Lachy> run it through ffmpeg2theora
<Lachy> sending the file by skype would be much slower, skype isn't good for transferring large files
<molly> still loading over here
<Lachy> 40 to 50 minutes remaining at this speed :-(
<Lachy> well, you could try it if you have a decent upload speed. Otherwise, I can get marcos to skype it to me when he gets back to his computer
<Lachy> my skype name is lachlanhunt
<oedipus> thanks, lachy -- let me try to convert it first
<Lachy> you can get ffmpeg2theora here. It's a command line utility for windows. http://v2v.cc/~j/ffmpeg2theora/download.html
<molly> is there any other female involved in this wg discussion?
<oedipus> we don't know whether RRSAgent is male, female or indifferent...
<molly> oedipus I think we can probably agree that the agent is genderless :D
<gsnedders> anne… be silly? oh wow.
<smedero> molly, there were two females present in the conference room earlier... I don't enough of these folks to tell you whos who.
<molly> smedero thanks! I am just curious. I'll be back in the room tomorrow
<molly> i'm not bringing it up to call foul or anything, it's just a point of curiosity for me
<smedero> molly, no I understand.
<smedero> particularly in light of the A List Apart survey recently...
<jane> oedipus: har :\
<molly> smedero yes! that was an amazing survey. Frightening in some ways
<gsnedders> it had issues for people not in full-time jobs though, IIRC
<dbaron> mjs: "I'm trying to schedule based on facts, not moral imperatives."
<molly> Leslie's here, now I know there's some additional girl power in the room ;)
<Marcos> no probs... next time I'll bring a proper web cam
<Marcos> If I can, I'm going to put the Aria video on youtube....
<shepazu> sorry about that
<kingryan> Lachy: could I join the skype call? my skype username is ryansking
<Lachy> I'll see if we can set up a skypecast. Hold on.
<ChrisWilson> [media elements discussion begins]
<scribe> Scribenick: MikeSmith
mjs: Does anybody have problems with the current parts of the HTML5 spec for audio and video?
DanC: yeah, I do. I think we should take it out.
<Lachy> skypecast wouldn't work
Håkon: first thing that happens when people turn on the OLPC machine is, they see themselves ...
scribe: because it has a built-in camera ...
scribe: also Tom-Tom jamming application ...
[Håkon demo'ing video stuff in special Opera build]
<jgraham_> aria people are squatting in rthe empty xsl-fo room
[Håkon attempting to show Green-Winged Macaw page from Wikipedia, but having problems]
ChrisWilson: As I understand it, the goal is to make it easier and more declarative to support videos
mjs: there's not reason why it needs to be more complicated than <img> for the simple case
DanC: I think the Design Principles document can help grow a set of requirements
DaveSinger: The justification might help us figure out when we have completed the design.
<Hixie> The agenda is here: http://esw.w3.org/topic/HTML/F2F
ChrisWilson: I think some of the goals that mjs has brought up require the codec discussion.
<Hixie> Bert: there is -- <img>, <video>, <embed>, <audio>, <iframe>
<mjs> MikeSmith: "codec", not "kodak"
<anne> I don't think the ARIA session is being minuted
<smedero> Video codec starter issue: http://www.w3.org/html/wg/tracker/issues/7
<smedero> (starting to collect emails and such)
<oedipus> there are people in the pf channel, but no notes
ChrisWilson: [re discussion with Dave Singer], you just introduced some "risk of editorship"
<dsinger_> we need video to get consistent attributes, dom, accessibility; to expose its time-based behavior; to make it a 1st-class citizen that is properly integrated; ...
ChrisWilson: there is an open question about how do we do synchronization
[mjs prepares to do demo of video in Webkit]
mjs: this in in the Webkit nightlies
[demoing example of opacity/compositing]
mjs: we will be adding the controls attribute soon ...
... and fully supporting soon the complete HTML5 spec for <video>
<myakura> WebKit Nightlies http://nightly.webkit.org/
<karl> <video id="carvideo" src="car.rm" region="videoregion" title="Car video"
<karl> alt="Illustration of relativistic time dilation and length
<karl> longdesc="carvideodesc.html" readIndex="3"/>
<karl> <audio id="caraudio" src="caraudio.rm" region="videoregion"
<karl> title="Car presentation voiceover" begin="bar.begin"/>
<karl> <animation id="cardiagram" src="car.svg" region="animregion"
<karl> title="Diagram of the car" readIndex="2"/>
<karl> <img id="scvad" src="scv.png" region="videoregion"
<karl> title="Advertisement for Sugar Coated Vegetables"
<smedero> doh. dupe issues. :-/
smedero - no problem ... I'll just delete the one I opened
<smedero> Ahh, thanks
<smedero> useful page: http://people.opera.com/howcome/2007/video/
<smedero> Ahh, thanks Bert - I'm just about to open up an issue on that
<karl> where did you put it Bert?
<anne> object=textstream, object=animation ?
<Bert> Nowhere, Karl, except here.
<smedero> Karl, I'm creating an issue
<smedero> for <video> and SMIL issues
<karl> Bert I think that would be better to have a full comparison not only names but attributes/features too
Tex: we need a statement of scope
ChrisWilson: I would like to see us define how autoplay works, and how synchronization works
<dsinger_> I think how autoplay works is defined
Hixie: current spec is that each <video> element is independent
Håkon: We should err on the side of simplicity
scribe: what's currently in there is a little too much more than most people would need
<smedero> ChrisWilson: should there be a separate issue for <video>'s autoplay implementation?
<Lachy> autoplay is well defined here http://www.whatwg.org/specs/web-apps/current-work/#loading
RESOLUTION: Add AudioVideoScope page to the Wiki
<karl> Lachy, you have the wrong uri :p - http://www.w3.org/html/wg/html5/#loading
mjs: YouTube, MySpace, Apple's QuickTime trailer site, Wikipedia are good sites for video examples
<Lachy> karl, no I don't. The specs are identical
<Hixie> feel free to stick stuff on there
mjs: If every major browser could support one common video codec, then authors could post just one video format on their site (instead of multiple ones).
... Ogg, patent risk is more unknown
... MPEG4, open standard, technically pretty good, used by Apple, haven't been sued yet, but patented and the patent licenses are not cheap ...
... many mobile devices have built-in support for H264 (MPEG4) ...
... third candidate is Windows Media ...
<Lachy> Dirac will be another good alternative when it's finished
<Lachy> it's being developed by the BBC, and since they're big, they're taking the risk of being sued for it
dsinger_: H263 ... which is also widely supported in mobiles
<Lachy> or animated GIF! ;-P
mjs: tradeoff of IP issues vs. codec quality
<Philip> Lachy, animated PNG would allow higher quality videos :-)
<Lachy> but we know GIF patents have expired already :-)
ChrisWilson: Our legal people do not believe that Ogg Theora has much of any chance of being free from patent issues.
<scribe> ACTION: Dave Singer to draft a summary of this discussion for the mailing list. [recorded in http://www.w3.org/2007/11/08-html-wg-minutes.html#action01]
<trackbot-ng> Sorry, couldn't find user - Dave
<trackbot-ng> Reloading Tracker config
<trackbot-ng> Tracking ISSUEs and ACTIONs from http://www.w3.org/html/wg/tracker/
<scribe> ACTION: David to draft a summary of this discussion for the mailing list. [recorded in http://www.w3.org/2007/11/08-html-wg-minutes.html#action02]
<trackbot-ng> Sorry, amibiguous username (more than one match) - David
<trackbot-ng> Try using a different identifier, such as family name or username (eg. dsinger2, hyatt)
<scribe> ACTION: dsinger2 to draft a summary of this discussion for the mailing list. [recorded in http://www.w3.org/2007/11/08-html-wg-minutes.html#action03]
<trackbot-ng> Created ACTION-4 - Draft a summary of this discussion for the mailing list. [on David Singer - due 2007-11-15].
<kingryan> spec section reference: http://www.whatwg.org/specs/web-apps/current-work/#browser
<smedero> kingryan, yes. or http://www.whatwg.org/specs/web-apps/current-work/multipage/section-browser.html#browser for folks who loathe the single page version.
<Lachy> simple video test cases: http://lachy.id.au/dev/markup/tests/html5/video/001
<Lachy> if someone wants to write some more video test cases, that would be aweome. You can use the videos here http://lachy.id.au/dev/markup/tests/html5/support/
<smedero> Lachy: the FAIL videos you produced are indirectly hilarious.
<Lachy> how so?
<smedero> in the playing off internet lingo sense
<oedipus> lachy: what UA are you using? the only thing i get is "You should see a video below" and then "FAIL"
<Lachy> oedipus, you need a recent build of Opera with Video support
<Lachy> there was one released on labs.opera.com yesterday
<Philip> (Ooh, 3d canvas too?)
<Lachy> here's one http://hixie.ch/tests/html40/test41-1b.html - note that pass condition contains red!
<ChrisWilson> video] http://dig.csail.mit.edu/breadcrumbs/node/127
<kingryan> +1 to mercurial
<gsnedders> +1 to Mercurial
<kingryan> it separates the acts of creating a changeset and publishing a changeset
<kingryan> every working copy is a a full repository
<gsnedders> (I'd say any distributed SCM over any centralised one, though I have a preference for Mercurial)
<Philip> (Opera says: cannot find d3d9.dll :-( )
<kingryan> welcome tantek to the skype conf call
<tantek> greetings. who on the call is actually at the meeting? i suggest everyone else on the call click the mute (mic with slash) button.
<Lachy> tantek, Marcos is at the meeting
<anne> http://tc.labs.opera.com/apis/focus/ has my three tests
<Lachy> anne, 003 needs to give instructions to click the text
<anne> Lachy, no
<anne> Lachy, maybe to not click the test
<Lachy> oh, right.
<smedero> tantek: you did
<kingryan> i think someone just made tantek's day
<tantek> is that a test page janet?
<kingryan> tantek: I think it's an agenda item :)
<ChrisWilson> Yes, Janet, that's an agenda item.
<ChrisWilson> tantek - you're correct. Maciej found a bug in MacIE. :) I paid him back by finding a couple of bugs in Safari.
<tantek> nicely done ChrisWilson
<Lachy> I wrote another test case http://lachy.id.au/dev/markup/tests/html5/video/004
<Lachy> that's for whoever it was Hixie just told to look them up
Lachy - thanks. That was Tex Texin asking about that
kingryan - yeah
make it more fun
<smedero> for the guys discussing text selection range support: http://www.quirksmode.org/dom/range_intro.html
<kingryan> this might also be of interest: http://www.w3.org/Style/CSS/Test/testsuitedocumentation.html
<anne> maybe e-mail them to email@example.com
<Lachy> yeah, that would be good
anne - yeah
<ChrisWilson> anne - agree, I'll mail mine.
<ChrisWilson> okay, time to head back to the hotel and make dinner plans.
<ChrisWilson> see you all tomorrow