Protocols and Formats Working Group Teleconference -- 06 Oct 2014

<trackbot> Date: 06 October 2014

<richardschwerdtfeger> meeting: W3C WAI-PF ARIA Caucus

<richardschwerdtfeger> http://lists.w3.org/Archives/Public/public-pfwg/2014Oct/0008.html

TPAC

ready

Janina: We almost have a calendar, thursday is booked
... Starting with WebTV, best time
... 11:15 HTML working group on two topics
... Accessible graphics and event handler enumeration

JC: Recommend the PF group only raise this issue again if a technical proposal is ready to go with it, we need at least one implementor on board in order to raise this again

Janina: We need a separate time to discuss this recommendation
... We are concerned about DRM stuff getting in the way

RS: Comcast is building an accessible set top box

Janina: I have also seen it at CSUN

RS: Is there something specific?
... What are they looking to standardize?

Janina: Say something that long, so there are some opportunities that would have captions and multiple languages, can be used for testing, talking to mark about this

RS: Isnt there a mandate

Janina: Not all in a mandate, especially in audio descriptions

RS: You have to have something concrete

JC: The same thing probably applies, event enumeration has be raised and declined many times

Janina: We have a chain of e-mails on the list

JC: The original reasons were not on the list

Janina: Shane stated security was one of the main reasons

JN: Tools can already do this, they just need to write plateform specific test tools

JC: I agree

RS: We are going to have UI events, and design patterns, I am more focused on where we are going, than where we have been

JN: It is the same issue on security

RS: Query event handlers?

JC: Everything we need is already there, this group has already rejected this

Janina: We need to put hat together on the compelling reason

JN: We should drop it

JC: One of my directors sometimes says, "I only have a few silver bullets… I don't want to waste of them one on this issue."

<Zakim> Joseph_Scheuhammer, you wanted to discuss longdesc

RS: Please cannot discuss longdesc?

Janina: Can we go back to the agenda
... We are still working on nailing the meeting done, 66% done
... Digital publishing in the after lunch

<richardschwerdtfeger> +1

Janina: We have another 30 minutes with the HTML group
... We then have time with Mark H

RS: The big things are the taxonomy for graphics, and the validation issues
... the whole think with keyboard, the biggest advances are CSS and web apps
... It would be nice to have a modal dialog box, but we don't see that much on mobile

JC: Another issue is data providers for tables, everyone has there own
... It is a huge issue for accessibility and performance
... I am not sure I have time to prepare a summary of the topic

Janina: So not something they should work on in the near future

JC: I think we can ask them too
... Let me think about it more, there are major benefits for accessibility and some on performance, it would take me some time to think about it

Janina: We want to think about the future and take advantage of meeting time with other groups

JC: The way most table views, similar to set-size and ste-position, maybe only a couple 100 view elements, some unordered view
... only the element in the current view are available, the data source will provide more when needed, this is all managed by the author now

MK: How does pos inset, press the down arrow, it requests a row element and gets the next row

JC: You can easily get ahead of what the data is available, so there are gaps in the user view of information
... Let me search for something in the table, but not all the data is there in these rendered views
... milliseconds versus double-digit seconds
... Scroll wheel scrolling has similar effects, an API to get to these data sources is much more effecient

MK: If there is something not rendered, I would not expect a screen reader to search it

JC: There maybe data in the memory, but the data is not rendered
... if you had something like safari native find it could search these APIs

MK: How you want to do that search is different between applications

JC: All native platforms render table views and list views this way, but the Web doesn't have a solution for table and collection view data providers.

MK: I don't see how solving it for accessibility is only for accessibility

JC: It is not specific for accesibility, but it will help

RS: Anything else on this one?

Jania: Not for me

RS: The DOM thing would be nice, but the other ones are critical

MK: There was a discussion about parametrized queries?

RS: That wil move to 2.x

Name computation (issue 541)

<clown> issue-541

<trackbot> issue-541 -- Text Alt Comp too specific to ARIA attributes (aria-valuetext and aria-valuenow) -- raised

<trackbot> https://www.w3.org/WAI/PF/Group/track/issues/541

<richardschwerdtfeger> https://www.w3.org/WAI/PF/Group/track/issues/541

<clown> http://rawgit.com/w3c/aria/master/accessiblename/accName.html#mapping_additional_nd_te

JC: I have proposed text there

<jcraig> Current text: If the embedded control is a range (e.g., a spinbutton or slider), use the value of the aria-valuetext attribute if available, or otherwise the value of the aria-valuenow attribute.

<jcraig> Proposed: If the embedded control is a range (e.g., a spinbutton or slider), use the text value if defined (e.g. aria-valuetext), or otherwise the value of the control (e.g., aria-valuenow, or host language value attribute).

JC: This is editroial

RS: I agrre
... Can we assign this to the implementation guide

JS: We are close tot being done

RS: Put it on the 1.1 implementation guide

MC: We can do it on the implementation guide or something new

RS: JS are you all set?

JS: When I reviewed all the HTML references it was language similar to JC wording, I like JC wording

<richardschwerdtfeger> https://www.w3.org/WAI/PF/Group/track/issues/542

Issue 542

RS: Proper fall back

JC: This has slipped between the cracks in the CSS spec
... If its using comma separated lists for alternative content, webkit supports, I doubt there is other implementation
... There is an ALT property in the CSS specs
... The test tools need to look at the live DOM and this information is the DOM

JS: Is that specific to HTML or can it also be used in SVG

JC: The data-* attributes are specific to HTML, but the attr() content is not.

JS: It can be attribute X

JC: ALT doesn't mean anything on divs
... It is for localization via the stylesheet

JS: We have COR, HTML specific and SVG specific, where does this go?

JC: Let me change the example, you can't render content directly on the element

RS: If this is going into CSS, we need to look at it

<jcraig> div::after {

<jcraig> content: url(icon.png), attr(data-fallback);"

<jcraig> }

JC: This has been in since CSS 2

RS: Do you sit in on all this stuff

JC: I don't go to the calls, but I follow the specs

RS: I have not seen people use it

JC: It works in one browser
... Chicken and egg problem

Janina: If we talk about this friday morning, who should we invite from CSS

RS: We need to look at this alot more

<Birkir_> Birkir Gunnarsson on phone too, long time absent. 919 number.

JC: This has only been added recently
... fall back text content ....

RS: Are there more things we need to look at

JC: Potentially....

RS: The separation of content and presentation is getting blurry, so this is a concern

JN: Which CSS spec is this in?

JC: CSS 4 selectors...

JN: I will take a look

RS: Is WCAG looking at this? This is a big can of worms, test tools and authoring practices...

ISSUE 542

<MichaelC> issue-542?

<trackbot> issue-542 -- Text Alt Comp section on CSS generated content does not allow for proper CSS content fallback -- raised

<trackbot> https://www.w3.org/WAI/PF/Group/track/issues/542

MC: What are you asking me?

RS: The issue here, you have the ability to add ALT content form CSS instead of the HTML source
... It has to go into the name commputation, authoring practices, we need to have a closer look at this CSS feature, we need to pay ...

<Birkir_> assistive technologies, at least screen readers, do not search for alt content of CSS either (at lesat speaking definitivel for Jaws/NVDA)

JC: We discussed this, but didn't get proposed to the CSS working group
... It was this group's inaction that made me propose to the CSS working group separately, because it was delaying accessibility of several Apple products.

RS: Other that JC, we are not looking at this holistically

Janina: I think WCAG will pay attention to it, we also need to pay more attention to it

RS: Can we create an action for the WCAG group look at this

MC: There are several people in this group that are in the WCAG group, we can do it as a chair to chair thing, or as a technique group issue
... We still have to get someone to write a technique, and then we need to make sure there is accessibility suppoort
... We need to wai for implementation and create test cases

RS: Can we get access to this through the DOM in webkit?

<MichaelC> qq-

JC: In webkit nightly we make label available in the DOM and it will include CSS information
... 5 or 6 years in webkit

RS: Current tools are using the DOM, is there a way to get at it in the CSS ....

JS: It might not be able to get the computer label, so you want a way to get the computed label

<jcraig> CSS "alt" discussion on www-style: http://lists.w3.org/Archives/Public/www-style/2012Nov/0319.html

MC: You may mean test the WCAG techniques, but the group recommends testing for more ....

<Zakim> MichaelC, you wanted to clarify ¨testing to WCAG compliance¨

MC: We need to have accessibility support, even if it can in theory be accessible
... There is a CSS object model spec, but I don't know alot about it

<MichaelC> CSS Object Model

RS: CSS is becoming more popular that Javascript

Janina: Can we get some time at TPAC on the topic

MC: I recollect that there was some work, but it may have been dropped

<jamesn> jcraig, I saw the mail. I still can't find it in a spec. I'm concerned about how this works. Calling it fallback seems incorrect and I just want to see how it is specified

RS: RS: You mean an accessibility API mapping?
... We had talked about a accessible name mapping

JC: That was proposed for the DOM.....
... get computed role and get computed ... I discussed in January

RS: One way or another we need to get this information, either the final name and or through the CSS

JN: This can be done in a browser specific environment, through extensions
... Simiar to even model stuff

RS: Is there a way to do this across browser

JN: If you are using Firefox and Webkit you have access to this information through extensions

RS: It depends on how much they are doing with CSS

JG: We do this with OpenAjax tools

RS: I am concerned that CSS is changing features on how we find content in the browser
... Only in the last week we have been looking at this from a tool perspective

JN: I know about generated content; I wrote the tool we use as a FF extension

JC: I wrote one for webkit

<clown> https://www.w3.org/WAI/PF/testharness/testresults?testsuite_id=1&testcase_id=553

MK: We used generated content in place of sprits

JS: There is a test case in the test suite test for this

RS: We don't have it for this one, for generating ALT text
... We need to look at this in much more detail

Janina: We will look at this all Friday morning

RS: JN if someone puts in an image VIA CSS

JN: FF doesn''t support it yet, but when it does it will
... It is required to have a text alternative

RS: The group meeting I was in required to have ALT attribute

JN: There was a big discussion about this in WCAG, there are in the techniques now, we did not get TITLE

RS: We beat that to death, ....

JN: There is not technique for TITLE

<Zakim> jcraig, you wanted to talk about CSS WG process

JC: We need more people in other working groups with accessibility interests...

<MichaelC> WCAG-ACTION-282?

<trackbot> WCAG-ACTION-282 -- Michael Cooper to Raise issue of css technique for generated content fallbacks https://www.w3.org/wai/pf/group/track/issues/542 -- due 2014-10-13 -- OPEN

<trackbot> http://www.w3.org/WAI/GL/track/actions/282

Janina: It has to be a two way street
... We did that in HTML reasonably well

JC: Rich is in SVG

RS: We need someone in CSS

JC: We need more people in CSS with accessibility concerns

RS: we need to do that
... like the canvas thing, we need to have someone to work in that area

<richardschwerdtfeger> https://www.w3.org/WAI/PF/Group/track/issues/609

RS: Any proposed text?

JC: There is proposed text, so assign to implementation guide

RS: Signed over

ISSUE-642

<trackbot> ISSUE-642 -- alt text computation should distinguish between nameFrom contents during root node calculation versus recursive calculation. -- raised

<trackbot> https://www.w3.org/WAI/PF/Group/track/issues/642

<richardschwerdtfeger> https://www.w3.org/WAI/PF/Group/track/issues/642

RS: We don't do that

JC: JS just commented, so this is a larger name that needs to happen

RS: JS can you take this on?

JS: Sure

JC: This was a constant point of confusion in the psuedo code

RS: A section on name and description

Janina: Do we want to talk to the HTML time?

JC: We should invite SF to the discussion
... Jason Kiss too

<clown> action-1073?

<trackbot> action-1073 -- Matthew King to Update aria-selected to reflect that it communicates selectability and clarify responsibility for ensuring aria-selected=false is on selectable elements -- due 2014-09-29 -- OPEN

<trackbot> https://www.w3.org/WAI/PF/Group/track/actions/1073

<richardschwerdtfeger> https://www.w3.org/WAI/PF/Group/track/actions/1073

MK: Some how I got this, I did not ask for
... I am still willing to work on it
... Lets put it out three weeks

RS: After TPAC, how about November 18th

ACTION-1335

<trackbot> ACTION-1335 -- Joanmarie Diggs to Propose edit for issue-493 to explicitly disallow strings matching "" or "undefined" in this sentence: content authors must provide values for required states and properties. -- due 2014-09-30 -- OPEN

<trackbot> https://www.w3.org/WAI/PF/Group/track/actions/1335

<clown> action-1335

<trackbot> action-1335 -- Joanmarie Diggs to Propose edit for issue-493 to explicitly disallow strings matching "" or "undefined" in this sentence: content authors must provide values for required states and properties. -- due 2014-09-30 -- OPEN

<trackbot> https://www.w3.org/WAI/PF/Group/track/actions/1335

Joanie: I am learning how to edit the spec tomorrow, can we push it back a week

RS: You want to put a "author must"
... We will have to test

<clown> http://www.w3.org/TR/wai-aria/complete#requiredState

JC: We just say does this content pass it ...

RS: We do have to cover it with the director

JS: it is already in the spec...

RS: Just a clarification, are you happy with the text

JC: I said my language is clunky, but it is accurate

RS: A week

JC: You are already familiar with git?

<jcraig> Content authors must provide a non-empty value for required attributes; this value may not match the format "undefined".

Joanie: Yes

JC: ... discussing spec editing and who should do what and outcomes ....

RS: Joanie do you want any notes

Joanie: I just want the clarification of what JC

JC: The phase that author must ... it must be part of the requirement
... Let me type it into IRC

<jcraig> RFC-2119 reqs should clearly define who is required to do what e.g. Authors MUST or User Agents MUST, etc.

JC: I would phrase this to be more like (pause)

RS: So you put in a semi colon

JC: I am pointing out that is the wrong way to do it

<jcraig> Content authors MUST provide a non-empty value for required attributes and Authors MUST NOT use the value "undefined" for required attrs.

<clown> +1

<clown> "required"

JC: This is more explicit, that we require that "undefined" must not be used

RS: Feel free to improve the text

- DRAFT -

Protocols and Formats Working Group Teleconference
06 Oct 2014

Attendees

Contents

TPAC

Name computation (issue 541)

Issue 542

Summary of Action Items

Scribe.perl diagnostic output

- DRAFT -

Protocols and Formats Working Group Teleconference 06 Oct 2014

Attendees

Contents

TPAC

Name computation (issue 541)

Issue 542

Summary of Action Items

Scribe.perl diagnostic output

Protocols and Formats Working Group Teleconference
06 Oct 2014