XHTML2 WG Weekly Teleconference -- 28 Nov 2007

Calls over the holiday period

Roland: I'm not here the 26th

Steven: I'm away for December

Roland: Call OK for next week

<oedipus> GJR: no problem with call on december 19

Roland: 19th OK?
... Ok we assume the 19th is OK
... no call on the 26th

<oedipus> boxing day the only skipped date?

Roland: we resume on the 2nd?
... so we miss just one week, the 26th of December

Roland: I will do agendas for this month

Steven: I will be around in case of need (for instance transition calls)

Announcements

Roland: We produced a new draft of CURIEs last week
... and we have a request outstanding for FPWD of Access

Roland: And ITS review?

Shane: We are happy with it

<scribe> ACTION: Steven to send ITS review (=thankyou, it's fine) [recorded in http://www.w3.org/2007/11/28-xhtml-minutes.html#action01]

Testing of XHTML Basic 1.1

Roland: Yam, are you able to report on this now?

Yam: At least my colleague is testing inputmode, so it is on the way

Roland: When might it be done?

Yam: Next week I think

Steven: That's good news

Yam: I think we are on schedule with the implementation
... the team spotted two errors, and are fixing that
... all three test cases are tested in our technical team

Title issue (RDFa vs Role)

<oedipus> GJR: RDFa versus Role to be discussed at today's telecon

Roland: There was going to be a followup meeting on this

Steven: Rich, what is your take on the current opinions in the PF about the use of role for title?

Rich: Well, clearly title isn't a unique thing
... since it may apply to many sections
... it hasn't been proven to me why you need a title element for sections and the whole document
... maybe because of iframe

Steven: So we have a way to do that already, the property attribute equal to dc:title
... <h1 property="dc:title">My life</h1>

Rich: We also have labelledby

<ShaneM> Note that the RDFa spec does not say that something with a property of dc:title is the same as / overrides the title element.

<oedipus> GJR: labelledby for today, RDFa for next round

Steven: I agree that role is the wrong way to do it
... but the use case is still covered

Roland: So the short term goal is role, and the long term goal is RDFa

<oedipus> GJR: short term goal is "labelledby" (ARIA syntax); long term goal is to use RDFa and explore means of communicating RDFa to Assistive Technologies

Steven: You can also say <h2 about="#section3" property="dc:title">Future tasks</h2>
... which says what the title is for some subsection

<oedipus> original example: <h1>Record #35793: <span property="dc:title">Title of Document or Sub-Document or portion of document</span></h1>

<oedipus> thread starts here: http://lists.w3.org/Archives/Member/w3c-wai-pf/2007OctDec/0191.html

Roland: If we are happy, let's move on

<Rich> short term solution: <div role="secondary" aria-labelledby="x26">

<Rich> <div role="header" id="x26">Last week's News Stories</div>

<oedipus> aria- is what has been implemented so far -- Jon Gunderson still using aria: in some examples

<oedipus> aria- being pushed by implementors/developers -- not a patient lot -- don't care about standardization, but product delivery

<Rich> aria:aria-labelledby

<oedipus> +1 to Tina's observation about using DIV to substitute for semantically meaningful markup (theory, i suppose, is that ARIA makes it semantically meaningful, but that is a strawman, in my opinion

<oedipus> extensibility for all but those who don't want it (HTML5 devs) -- SVG integration into HTML5 may be made redundant if CANVAS carries

Steven: I don't think that the aria attributes should be chameleon
... you want them to work everywhere, without changing the languages; and I think that the XML languages should use aria:labelledby, and let HTML5 us the hyphen for its namespace separator if it is so allergic to colons.

Shane: Well, why is role chameleon then?

Roland: So that language designers can decide for themselves which they use.

TAG role comment

Roland: This is ongoing
... I still have to do the response to them

Events2 Comment (Purpose)

Rich: We need a description of the handler

<oedipus> DanC keeps reiterating that "NOTHING" has been decided; there has been, however a "formal survey" on immediate role graphics and CANVAS

(which will become the first decision)

<oedipus> right

Rich: And <purpose> was there for that

<oedipus> Steven: handler specs, purpose element -- strictly speaking with so many types of handlers, good to have solution that addresses that as well; idea of purpose description is very good one -- if can find way to do it; Mark suggested role, i can envision using RDFa to pass description of handler (metadata in a way)

<oedipus> RS: place to put text, associate text with handler via RDFa -- mechanism to put actual text in handler?

<oedipus> Rich: liked Roland's proposal -- similar to solution for XForms

<oedipus> Roland: why not provide information - this is a label, this is what it does, hint can be obtained on user request (tell me more about this)

<oedipus> Steven: label in relation to handler?

<oedipus> Rich: don't see how label fits for handler

<oedipus> Roland: if here is a trigger and this is the handler for it, could get label from handler

<oedipus> Steven: have to think about that

<oedipus> Roland: have as many as want in UI - don't have to write twice; and can get supporting materials

<oedipus> Rich: XML Events -- how to do?

<oedipus> Roland: can still have element as child of handler

<Roland> http://lists.w3.org/Archives/Public/www-html-editor/2007OctDec/0027.html

<Roland> <hint>performing this action will delight you.</hint>.

<Roland> doactivate(event);

Shane: Any of this is fine
... only as long as we don't overengineer it

Roland: No one says we have to do all three
... but in XForms we do
... a label for the UI layer

Shane: What is the UA requirement for a label?

Roland: That it is presented to the user

Shane: If I have a span
... and associate an access key with it
... and a handler that is invoked by that key
... if the handler has a label, how does that get manifested?

Roland: I don't think it does

Shane: OK, good

Rich: So do we need all 3 in XML Events2?

Roland: Well, I was suggesting synergy with stuff we have already
... you don't HAVE to use them

Rich: I see, fine
... I see a way of mapping it to the accessibility API

<oedipus> GJR: can use "prolog" in ARIA concept to provide richer, fuller human-understandable information about specifics of choice (if choice available)

Gregory: See my comment about 'prolog' in aria
... above

<ShaneM> So I think what I need to do is steal the label, hint, and help elements from XForms and put them into XML Events 2?

Rich: If we had a declarative model, like XForms data model, we could do this all a lot easier

Roland: So I'm saying XForms, and therefore XHTML2 has this already, so why not reuse it rather than reinventing stuff

Shane: Do you mean add it to handlers?

Roland: Yes

<oedipus> +1 add to handlers and reuse in XForms

<alessio> +1

Rich: So we'll be using XML Events2 in XHTML2?

Steven: That is the plan, but there is an overlap with XForms,which still uses XML Events 1
... and we will use XForms 1.1

Rich: XHTML2 is not fully supported in the browser, and there is HTML, we are looking at a dual timeline in the WAI area

CURIEs

Shane: I would like to go to last call in 2 weeks on CURIEs
... just to push the fight
... we can deal with the comments

Steven: Length?

Roland: Till late January
... for the ftf

Shane: 6 or 7 weeks

Steven: Any objections?

Shane: I will prepare a draft for next week's call

Steven: Will put on the agenda

<scribe> ACTION: Roland put last call of CURIES on next week's agenda [recorded in http://www.w3.org/2007/11/28-xhtml-minutes.html#action02]

<scribe> ACTION: Shane prepare a new draft of CURIES for next week [recorded in http://www.w3.org/2007/11/28-xhtml-minutes.html#action03]

- DRAFT -

XHTML2 WG Weekly Teleconference

28 Nov 2007

Attendees

Contents