DRAFT Minutes: January 29, 2015 WAI-PF ARIA Caucus

Draft minutes:
https://www.w3.org/2015/01/29-aria-minutes.html

[W3C]<http://www.w3.org/>
- DRAFT -
Protocols and Formats Working Group Teleconference
29 Jan 2015
See also: IRC log<http://www.w3.org/2015/01/29-aria-irc>
Attendees
Present
Regrets
Joanmarie_Diggs
Chair
SV_MEETING_CHAIR
Scribe
jon
Contents

  *   Topics<https://www.w3.org/2015/01/29-aria-minutes.html#agenda>
     *   aria-describedat<https://www.w3.org/2015/01/29-aria-minutes.html#item01>
     *   Action 1548<https://www.w3.org/2015/01/29-aria-minutes.html#item02>
  *   Summary of Action Items<https://www.w3.org/2015/01/29-aria-minutes.html#ActionSummary>

________________________________
<trackbot> Date: 29 January 2015
<clown> But only the requirement that user agents map unknown values to true.
<janina> sirc: jon
<janina> scribe: jon
<richardschwerdtfeger> yes,
<richardschwerdtfeger> coming. sorry
<clown> joanie said earlier: "but I'm hoping that y'all can do any final wordsmithing at the upcoming call and that by meeting's end there will be a comment in action-1440 telling me the specific text edits to make"
<richardschwerdtfeger> https://lists.w3.org/Archives/Public/public-pfwg/2015Jan/0138.html
aria-describedat
JS: We will start in detail next week, waiting for MC
... We are putting on the table now, so it won't be in the way of ARIA 1.1
... The next think, I propose discuss in three ways:
... Way 1: It is it imposing any requirement on the user agent UI
... MC and will have a proposal on that
... Mostly a requirement for accessibility API, expose through the visual API would be welcome though
... We need to use the word "expose", versus making the feature available to the user
... It will be helpful overal in the spec
... Way 2
... Edits ....
... Way 3
... Whether we want the feature in the spec at all
... We shouldn't be using the term "user agent" at all
<richardschwerdtfeger> https://www.w3.org/2002/09/wbs/33280/longdesc-PR/results
<clown> FWIW: http://rawgit.com/w3c/aria/master/aria/aria.html#dfn-user-agent
RS: I think it is important to read the current review of LONGDESC, please look at the list if you have not reviewed
... we need someone from George K and Benetech
... Need to find a new person from Benetech
JS: We need to know the technical issues from them and why they need it
RS: We need to have everyone hear all the issue, I am a neutral party in all this
... We are going beyond images
... Some user agents are tablets, they have different issues
... For this to useful we need to have more than what we can put in our spec, our spec just talks about exposure
MK: On that topic of UA behavior, including keyboard behavior, like landmarks has "may" language
... Even Firefox does not provide keyboard support to landmarks
... Is this a good reason to bother or not bother UA developers
... I am prejudiced that user agents would support landmarks there would be better use of landmarks
... There are a few features we may want to push on user agents, I am thinking of evolution
RS: The job of specifying that should be i the host language
... It is SVG and HTML, either of those working groups can define behaviors
MK: I think that makes sense
... MAIN is an HTML element, it is only mapped in the HTML spec, we need more in HTML spec
RS: The people working on those specs know the issue of including
<Zakim> jcraig, you wanted to say partial regrets. Only on the call for 20 minutes.
JC: I just wanted to give partial regrets, I need to give a presentation today, so will only be here for 15-20 minutes
JS: The introduction of ARIA is very good on this topic
... Currently we get a lot of good comments about it not changing behavior just exposing though the accessibility API
... ePub will be another carrot for implementation
... LONGDESC will be part of Shift-F10
FE: Is there an security issue, if someone is putting longdesc content
RS: Is that privacy?
FE: Some consider it security
Clown: I like the use of the word carrot
... If you get a cool implementation, then browser people may put it in
RS: There is an IBM customer on older browsers, so landmarks are not availble, they want skip to main links
... There is really value add for making them availble to everyone, this is not for our group though
JC: The skipTo is typically shown on focus
RS: I see your point
... Over the next week think about these things
... We need to have some of the other browser manufactures on
... I know they don't what use defining behaviors
JS: I agree and we need to shore that in our spec
RS: Anything else before we move on? We need these people next week to discuss the topic
JS: We need them for the behavior discussion
... We can have the conversation about the conformance language and the verbs
RS: Does the group want to hear from these peopel first?
MK: It doesn't matter to me
RS: The decision will not be made for weeks, so any order is OK for now
Action 1548
<clown> action-1548?
<trackbot> action-1548 -- Joanmarie Diggs to Write revised proposal for aria-current based on today's meeting for issue 587 -- due 2015-01-29 -- PENDINGREVIEW
<trackbot> https://www.w3.org/WAI/PF/Group/track/actions/1548
RS: This is for aria-current, it is out for review
... Please review
<richardschwerdtfeger> http://rawgit.com/w3c/aria/master/aria/aria.html#aria-current
<richardschwerdtfeger> http://rawgit.com/w3c/aria/master/aria/aria.html#aria-current
clown: Review the phrase "could be used"
... I take these bullets as ....
RS: Probably an authoring practices needs to look at uses
clown: Maybe they are not definitions
RS: What if we said this "A page token ... for example..."
... A page token used to indicate a .....
... What action item?
... Are other people OK with that woding? I will edit the action item
MK: At one time I had a more lengthy wording of select, ....
... So image you are reading ARIA 1.1 for the first time, will there will be confusion between selection and aria-current
... In the note we include a link to the APG
... In the spec, I did find links to the PAG and we are working on the APG
... Is it bad to link to the APG, since it is a note
FE: You mean circular references?
MK: From a W3C view?
SJ: I missed the question
MK: We are talking about aria-current and have a note that links to the APG, and the APG is a non-normative note, is that a W3C practice
SJ: MC would know more, but it doesn't seem to be a problem for me
MK: Alex was concerned
clown: If it matters all notes are considere non-normative, that should be OK
RS: As long as the URL doesn't break
SJ: Thats why we think carefully about URLs
RS: We need an action to add the note
MK: Give me 2 weeks
yes
<richardschwerdtfeger> ACTION: Matt Create a proposed alternative to the aria-current note. [recorded in http://www.w3.org/2015/01/29-aria-minutes.html#action01]
<trackbot> Error finding 'Matt'. You can review and register nicknames at <https://www.w3.org/WAI/PF/Group/track/users>.
<richardschwerdtfeger> ACTION: mattking Create a proposed alternative to the aria-current note. [recorded in http://www.w3.org/2015/01/29-aria-minutes.html#action02]
<trackbot> Created ACTION-1573 - Create a proposed alternative to the aria-current note. [on Matthew King - due 2015-02-05].
RS: Reviewing 1573 date, 2 weeks?
MK: Yes
RS: I am also got a note to change the text
... In addition to rewording the examples for aria-current, fix the spelling of flow chart
<clown> "Any value not included in the list of allowed values must be treated by user agents as if the value true had been provided."
clown: One other compliant from Alex
... he wants any invalid values should be considered a value of true
MK: Are we going to allow people to pass any string
RS: what does the browser do with that, do you know what I mean
clown: Is there another type of container, like "street", should that mapped to "true"?
... True is like th catch all
... He has to go through each of the legal values and if not there set to "true"
... It might be premature, since we do not have an API mapping yet
RS: It passes it a value
clown: ATK could do the work of assign a value to a undefined values
MK: HOw does that work for other mappings
RS: Change it to call matt
MK: Everything will be rendered as false
clown: There is a similar issue as aria-invalid, undefined values are mapped to "true"
... the application author should set to true, it actually does not make any requirements on the user agent
MK: Thats informing versus exposing
clown: For future expansion, .... as if the value "ture" was provided
MK: I am wondering what mozilla doing today
what's in the test hareness???
RS: I think that should have been there
MK: No aria-invalid for "foobar"?
clown: test case 587
RS: What was the results?
<clown> https://www.w3.org/WAI/PF/testharness/testresults?testsuite_id=1&testcase_id=587
RS and MK: JOseph is amazing with these links
clown: there were some passed and some failed, no tests on safari
MK: There was only 1 pass?
... discussion of history of this test ....
<clown> https://lists.w3.org/Archives/Public/public-pfwg/2015Jan/0148.html
RS: Let me get the post from Alex
clown: here it is
RS: I am respnding from my notes
clown: I would caution about the tests are from 2012
CS: The expected test results have been updated by me
... Is that an actually an API value
... I passed it for the invalid property is true, it seems to be an invalid test case
... Is that suppose to be an object value
clown: It is a state in ATK
... it is still passing
CS: the expected results is badly worded
... Checking IE....
clown: I am not sure who is mapping, mozilla or ATK
... I will look in the implementation guide
... expose invalid true as a text value, that is strange for a checkbox
... It just says what comes out of the accessibility API, not who does it
... I am not sure what FF is doing, either way the mapping come out correct
... It might be doing it, I would need to bring up a debugger
SJ: OK
clown: the other way is to find the source code in FF
SJ: Thats a needle in the hay stack
MK: Are there any other issues raised with current?
... I think we are done
RS: Let me get to the next one
Action 1540 Issue for handling Generic roles (Related to issue 638)
<trackbot> Error finding '1540'. You can review and register nicknames at <https://www.w3.org/WAI/PF/Group/track/users>.
(Rich)
<clown> action-1540
<trackbot> action-1540 -- Cynthia Shelly to Investigate and propose mappings for aria-modal property for uia. -- due 2014-12-02 -- PENDINGREVIEW
<trackbot> https://www.w3.org/WAI/PF/Group/track/actions/1540
<mattking> action-15540?
<trackbot> Sorry, but action-15540 does not exist.
<mattking> action-1540?
<trackbot> action-1540 -- Cynthia Shelly to Investigate and propose mappings for aria-modal property for uia. -- due 2014-12-02 -- PENDINGREVIEW
<trackbot> https://www.w3.org/WAI/PF/Group/track/actions/1540
<clown> issue-638?
<trackbot> issue-638 -- Generic container roles for things like div/span... -- open
<trackbot> https://www.w3.org/WAI/PF/Group/track/issues/638
<richardschwerdtfeger> https://lists.w3.org/Archives/Public/public-pfwg/2015Jan/0104.html
RS: I had a post to the list, this is a big problem with the mapping guide, a DIV is not mapped consistenty, some map to group others panel
clown: I replied to this
MK: Yes you did, and discussed what a panel is...
TS: OSX maps to a group....
RS: A div is really just a container of stuff, panels are not groups and can have a text area
clown: That is not what it is in ATK
RS: There is inconsistency across platforms, and we need to clean this up
... definition in Java for panes.....
... Group is a ....
... discussion of various mapping problems for HTML elements ....
... This is bigger than HTML, there are issues with SVG
... I beleive we are overloading group semantics
... In IBM DIV elements need roles if used ....
<richardschwerdtfeger> One proposal:
<richardschwerdtfeger> 1. Map <div> to a role of section on all platforms or a single generic role
<richardschwerdtfeger> on all platforms that says this is just an container and nothing more
RS: We need to harmonize with HTML5 section element
<richardschwerdtfeger> 2. When applying role="presentation" map the semantic HTML structural
<richardschwerdtfeger> alements to the role determined in 1
<richardschwerdtfeger> 3. In HTML5 map <section> to the generic role in 1. and map it to region
<richardschwerdtfeger> when a label is applied to it.
<richardschwerdtfeger> 4. Regarding 1. for a Mac either make the generic container a new section
<richardschwerdtfeger> role and leave axGroup for role="group" or make role="group" map to
<richardschwerdtfeger> something different than axGroup and leave axGroup generic.
RS: This way you preserve your structure
MK: I have a question related to OSX and IA2 and FF
... If you have a container element...., it was a text node or something...
clown: In IA2 it comes out as a section
NK: There is a role section
<clown> IA2_ROLE_SECTION
MK: RS OSX doesn't have the equivalent?
RS: right
... Only JC can answer this, how is role group used across all applications
... I did some testing with voice over on the OSX, and it said "group", "proup"....
... I don't know what to do with this
... This is a fundamental problem, what do end users on OSX expect
... What the strategy is for OSX
... If we mark a table as presentational, does everything become this generic section, we should do this
MK: Even thoiugh DIV and SPAN are generic containers, from an AT perspective they are very different
<Zakim> Joseph_Scheuhammer, you wanted to ask joanie's question: if the group with VO on OSX was with Safari
MK: If you put in 3 spans, they get treated as a single text run, ...., for example a sentence, but DIV would break it up
JN: It is probably something to do with the CSS, I do not encourage people to do that
... Changes in CSS can change the element semantics
MK: We want the semantics to be in the role and not in the CSS
JN: We want it to do what the author intended to do
MK: If CSS can override the element semantics then CSS ....
... In your proposal, do you have a DIV and SPAN semantics
RS: I did not address a SPAN type semantics
... SPAN is used to style tyhings
MK: Not always, we use SPAN to put in ARIA roles
... You can put IDs on SPANs
RS: An role example?
FE: You can change the flow
<clown> "a <span> with just a style change and no semantics may not get its own accessible object, but the style change will be exposed by other means."
MK: Stephan raised some issue that I thought were not good practice
RS: The big issue is global attributes
... SPANs are not used for regions
<jamesn> you could put aria-live on a span....
RS: Do we want to really mark that ....
MK: We have a que
<Zakim> joanie, you wanted to give examples when span needs to be exposed
Joanie: They ca have event handlers
<jamesn> .... or if it has a child that is....
Joaine: What is you have a title attribute
Joanie: In ATK we have a dedicated role, SPAN should not be mapped as DIV
... We have a new role, "STATIC", it is similar to ROLE=TEXT
... I am done I think
MK: I think we really need to talk about the difference between, DIV label is from author, and SPAN label is form content
<clown> "a <span> with just a style change and no semantics may not get its own accessible object, but the style change will be exposed by other means."
clown: I found this is the SPAN is .....
... Otherwise it gets sucked into its parent
RS: OK
... So what happens if a SPAN has a ROLE
<joanie> ATK_ROLE_STATIC
clown: It would have a role
MK: It you want to create a separate role under section, we need something like STATIC, the span may not contain text
joanie: FF does ....
FE: One of the things I have seen is SPAN is used to represent data
... A SPAN can represent a whole piece of data
Summary of Action Items
[NEW] ACTION: Matt Create a proposed alternative to the aria-current note. [recorded in http://www.w3.org/2015/01/29-aria-minutes.html#action01]
[NEW] ACTION: mattking Create a proposed alternative to the aria-current note. [recorded in http://www.w3.org/2015/01/29-aria-minutes.html#action02]

[End of minutes]
________________________________
Minutes formatted by David Booth's scribe.perl<http://dev.w3.org/cvsweb/%7Echeckout%7E/2002/scribe/scribedoc.htm> version 1.140 (CVS log<http://dev.w3.org/cvsweb/2002/scribe/>)
$Date: 2015-01-29 19:01:31 $
________________________________
Scribe.perl diagnostic output
[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.140  of Date: 2014-11-06 18:16:30
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/<http://dev.w3.org/cvsweb/%7Echeckout%7E/2002/scribe/>

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/fals/false/
No ScribeNick specified.  Guessing ScribeNick: jongund
Found Scribe: jon

WARNING: No "Present: ... " found!
Possibly Present: Apple CS Clown FE FWIW Fred_Esch GVoice JC JN JS James_Nurthen Joaine Joanie Joanmarie_Diggs Jon_Gunderson Joseph_Scheuhammer MK Matt_King Microsoft NK P3 RS Rich_Schwerdtfeger SJ TS aria cyns fesch https jamesn janina jcraig joined mattking richardschwerdtfeger sirc trackbot
You can indicate people for the Present list like this:
        <dbooth> Present: dbooth jonathan mary
        <dbooth> Present+ amy

Regrets: Joanmarie_Diggs

WARNING: No meeting chair found!
You should specify the meeting chair like this:
<dbooth> Chair: dbooth

Found Date: 29 Jan 2015
Guessing minutes URL: http://www.w3.org/2015/01/29-aria-minutes.html
People with action items: matt mattking

[End of scribe.perl<http://dev.w3.org/cvsweb/%7Echeckout%7E/2002/scribe/scribedoc.htm> diagnostic output]