Minutes: [aapi] UAI TF Meeting Tue 1 April 2014

http://www.w3.org/2014/04/01-aapi-minutes.html

Plain text follows:

   [1]W3C

      [1] http://www.w3.org/

                               - DRAFT -

           Protocols and Formats Working Group Teleconference
                              01 Apr 2014

   See also: [2]IRC log

      [2] http://www.w3.org/2014/04/01-aapi-irc

Attendees

   Present
          Joseph_Scheuhammer, Bryan_Garaventa, Joanie_Diggs,
          +1.416.848.aaaa, David_Bolter

   Regrets

          Cynthia Shelly

   Chair
          Joseph_Scheuhammer

   Scribe
          joanie, clown

Contents

     * [3]Topics
         1. [4](All) Update re: relevant aria-* attributes using
            selectors for referencing other nodes:
            http://lists.w3.org/Archives/Public/public-pfwg/2014Ma
            r/0091.html
         2. [5]ACTION-1369 (David): Investigate FF implementation
            of aria-live removal events (see also ISSUE-481).
         3. [6]ACTION-1408 (Joseph): Email Jason Kiss re: AAPI
            that describes API differences.
         4. [7]ISSUE-583 (Joseph): Update regarding status.
         5. [8]ACTION-1409 (Joanie): File a bug against ATK re:
            adding position, level, and other API.
         6. [9]ISSUE-645 (All): Remove 'checkable' object
            attribute from aria-pressed mappings.
     * [10]Summary of Action Items
     __________________________________________________________

   <trackbot> Date: 01 April 2014

   <clown> zakim GVoice is Joseph_Scheuhammer

   <clown> agenda: this

   <joanie> ScribeNick: joanie

   regrets Cynthia Shelly

(All) Update re: relevant aria-* attributes using selectors for
referencing other nodes:
[11]http://lists.w3.org/Archives/Public/public-pfwg/2014Mar/0091.html

     [11] http://lists.w3.org/Archives/Public/public-pfwg/2014Mar/0091.html

   JS: Wanted to bring everyone up to date who wasn't on the call
   yesterday
   ... There was a request from Dominic from Google, not
   specifically to do with web components
   ... but how to provide access to shadow DOMs because you cannot
   use an ID
   ... because the shadow DOM is sandboxed
   ... document.querySelector
   ... Function from DOM4 (not sure, might be DOM3)
   ... There is a discussion to let aria attributes with an ID ref
   to work with active-descendant
   ... Discussion between javascript versus selection syntax run
   in the browser
   ... just like CSS selectors

   DB: Did Dominic provide examples?

   JS: My recollection is no, but he's beginning to put that out
   on the lists now.
   ... Today I saw something from an actual project he's working
   on

   <clown> [12]http://www.polymer-project.org/

     [12] http://www.polymer-project.org/

   JS: Here's the URL (above)
   ... (quotes from the email Dominic sent about polymer and these
   sorts of examples)
   ... And Rich brought up an example in which the
   relation/reference could go both ways

   DB: I am happy that Dominic is putting energy toward solving
   this

   BG: I like the selector method better than the javascript
   method
   ... because it is more clear

   DB: My gut reaction is to agree with you Bryan

   JS: And Rich wants to get this for 1.1, I think

   DB: When is 1.1's target?

   JS: April 2016

   DB: April 1st! :)

   JS: Realize we started the final ARIA 1.0 push in late October
   2013
   ... That's almost half a year, just to push the document
   through the whole process
   ... In order to release it by April 2016, it has to be done by
   November 2015. That's not too bad.

   BG: They were unclear about selectors referencing one or
   multiple IDs.
   ... It might be important to distinguish

   DB: It would be important for us all to do the same thing

   JS: What's the status of shadow DOM in Firefox?

   DB: I'm not sure

   JS: Your example was great
   ... (describes example of video player where all the controls
   are in the shadow DOM)
   ... So if you actually look at the document, you don't see any
   buttons to push

   BG: You don't see that a checkbox is checked

   JS: The difference is that it looks totally different depending
   on which browser you use

ACTION-1369 (David): Investigate FF implementation of aria-live
removal events (see also ISSUE-481).

   <clown> action-1369?

   <trackbot> action-1369 -- David Bolter to Investigate
   implementation of aria-live region events in FF -- due
   2014-02-25 -- OPEN

   <trackbot>
   [13]https://www.w3.org/WAI/PF/Group/track/actions/1369

     [13] https://www.w3.org/WAI/PF/Group/track/actions/1369

   DB: I haven't done it yet.
   ... We also had agreed that it wasn't high priority.

   (Group reads over issue above)

   DB: Let's go with April 15th.

   JS: (changes the ticket)

   DB: I might have questions.

   <clown> issue-481?

   <trackbot> issue-481 -- aria-live removal events should or must
   happen before object removal -- open

   <trackbot> [14]https://www.w3.org/WAI/PF/Group/track/issues/481

     [14] https://www.w3.org/WAI/PF/Group/track/issues/481

   JS: Yeah, it's changing the SHOULD to a MUST.
   ... What possible reason would one have to not let an AT know
   something is going away?
   ... All your action is: What does Firefox do now?

   DB: I'm 99% sure we made the change to make this work.
   ... It actually was non-trivial.

   <davidb> but code churn may have happened since

   JS: I will do something when I get some answers from somebody.

ACTION-1408 (Joseph): Email Jason Kiss re: AAPI that describes API
differences.

   JS: I was going to do it as soon as I did the agenda.

   <clown> action-1408?

   <trackbot> action-1408 -- Joseph Scheuhammer to Send email to
   jason kiss to ask him if he could write an aapi section that
   describes api differences. -- due 2014-04-01 -- OPEN

   <trackbot>
   [15]https://www.w3.org/WAI/PF/Group/track/actions/1408

     [15] https://www.w3.org/WAI/PF/Group/track/actions/1408

   JS: It's due today. I will change it to a week from today.

   DB: I wonder if we should add a role description to ATK and
   IA2.

ISSUE-583 (Joseph): Update regarding status.

   <clown> issue-583?

   <trackbot> issue-583 -- Elements that are descendants of an
   element having aria-activedescendant should not all be
   focusable -- open

   <trackbot> [16]https://www.w3.org/WAI/PF/Group/track/issues/583

     [16] https://www.w3.org/WAI/PF/Group/track/issues/583

   JS: When we left this off a week ago, I was trying to find the
   test case.
   ... We concluded that this is finished and we could just close
   this issue.
   ... There's enough info in the UAIG

   <clown>
   [17]https://www.w3.org/WAI/PF/testharness/testcases/edit?testsu
   ite_id=1&testcase_id=784

     [17]
https://www.w3.org/WAI/PF/testharness/testcases/edit?testsuite_id=1&testcase_id=784

   JS: But I cannot find the test case
   ... The URL in the issue (above) doesn't seem to exist
   ... I ended up seeing email to Rich and Michael asking, "Where
   is this test case?"
   ... It's going to stay open until I find this test case.
   ... I think it is going to work though

   DB: I'm trying to read the issue

   JS: I'll try to summarize the issue
   ... Rich brings up that in an active-descendant case there
   should be some restrictions on what should be focusable
   ... Because there can be all sorts of descendants which may be,
   for example, descriptive
   ... But the UAIG says it has to have an ID and a role

   <clown>
   [18]http://www.w3.org/WAI/PF/aria-implementation-1.1/#keyboard-
   focus_aria-activedescendant

     [18]
http://www.w3.org/WAI/PF/aria-implementation-1.1/#keyboard-focus_aria-activedescendant

   JS: (to David) You and I asked Alex about this and he said,
   "yes."
   ... If you go down to step 4 (in the above)
   ... 4a in particular

   DB: So this is about the focusable state?

   JS: Yes

   DB: So it's a heuristic to determine which objects should have
   state focusable

   <clown> [19]https://www.w3.org/WAI/PF/Group/track/actions/1222

     [19] https://www.w3.org/WAI/PF/Group/track/actions/1222

   JS: (to David) Look at the above issue. And read the note.
   ... I want to make sure the test case actually works (i.e.
   before I close it). Because Rich cites a test case.

   i maybe gone

   i hear nothing

   <davidb> i don't hear you either

   <davidb> but 3:30! :)

   ha

   calling back in

   <clown> likewise, I don't hear joanie, but I do hear davidb.

   sorry

   <davidb> np

   <scribe> ScribeNick: Joanie

ACTION-1409 (Joanie): File a bug against ATK re: adding position,
level, and other API.

   <clown> action-1409?

   <trackbot> action-1409 -- Joanmarie Diggs to File a bug against
   atk in gnome's bugzilla requesting new position, level, and
   other api be added. -- due 2014-04-01 -- OPEN

   <trackbot>
   [20]https://www.w3.org/WAI/PF/Group/track/actions/1409

     [20] https://www.w3.org/WAI/PF/Group/track/actions/1409

   JD: I did my action item. There is a bug filed in GNOME's
   bugzilla

   <clown> [21]https://bugzilla.gnome.org/show_bug.cgi?id=727453

     [21] https://bugzilla.gnome.org/show_bug.cgi?id=727453

   JD: I am not anticipating much discussion, but there may be
   some.
   ... I want to ping the Gtk+ accessibility developer because I
   think this will solve some issues he pointed out to me.

   JS: So do we close this issue out?

   JD: No, I'd like to keep it open in case there is subsequent
   discussion that requires our input.

   <clown> action-1409?

   <trackbot> action-1409 -- Joanmarie Diggs to Monitor progress
   of bug against atk in gnome's bugzilla requesting new position,
   level, and other api be added
   ([22]https://bugzilla.gnome.org/show_bug.cgi?id=727453) -- due
   2014-04-15 -- OPEN

     [22] https://bugzilla.gnome.org/show_bug.cgi?id=727453)

   <trackbot>
   [23]https://www.w3.org/WAI/PF/Group/track/actions/1409

     [23] https://www.w3.org/WAI/PF/Group/track/actions/1409

   davidb: it's to get rid of 30 crazy things for implementors :)

   <davidb> joanie: you don't like object attributes eh? :)

   <davidb> we've crammed too much

   <clown> davidb, you are correct, sir.

   JD: The whole point of this new API, as was discussed last
   week, is not only to simplify things for implementors

   <davidb> up oh those ":"'s are gonna look like speakers

   JD: But also because ATs don't give a _____ where it came from,
   they just want the answer consistently.

ISSUE-645 (All): Remove 'checkable' object attribute from
aria-pressed mappings.

   <clown> issue-645?

   <trackbot> issue-645 -- remove 'checkable' object attribute
   requirement from aria-pressed mapping requirements -- open

   <trackbot> [24]https://www.w3.org/WAI/PF/Group/track/issues/645

     [24] https://www.w3.org/WAI/PF/Group/track/issues/645

   JS: I wish more people were here now. :)
   ... If you look at the issue, there is an aria-pressed
   attribute
   ... And the current UAIG says that if it has true or false, the
   checkable object attribute must be present

   <clown>
   [25]http://www.w3.org/WAI/PF/aria-implementation-1.1/#mapping_s
   tate-property_table

     [25]
http://www.w3.org/WAI/PF/aria-implementation-1.1/#mapping_state-property_table

   JS: I can confirm that aria-pressed only applies to toggle
   buttons
   ... So I agree with Alex that checkable is redundant
   ... I am willing to take it out of the UAIG for all but Apple
   (because it's not there for Apple)

   BG: iOS used to say "checked" for toggle buttons

   JS: There is no issue with UIA
   ... So the action is to remove the "checkable" from ATK and IA2

   JD: I agree with this decision because it is not only
   redundant, but it should eliminate extraneous state-changed
   events that currently have to be filtered out by the AT.

   JS: I want David's opinion, but he'll probably agree because
   Alex has already removed this from Firefox (if I read this
   correctly)
   ... Oh, it's not closed yet.
   ... There's been activity.

   <clown> [26]https://bugzilla.mozilla.org/show_bug.cgi?id=825114

     [26] https://bugzilla.mozilla.org/show_bug.cgi?id=825114

   <clown>
   [27]https://bugzilla.mozilla.org/show_bug.cgi?id=825114#c11

     [27] https://bugzilla.mozilla.org/show_bug.cgi?id=825114#c11

   (Discussion about the issue and differences in ATK and IA2
   w.r.t toggle buttons)

   <clown> [28]https://bugzilla.mozilla.org/show_bug.cgi?id=989958

     [28] https://bugzilla.mozilla.org/show_bug.cgi?id=989958

   JS: Marco is "stealing" the former bug (825114) to fix it in
   the newer but (989958)

   <bgaraventa1979> I like getting rid of checkable within toggle
   buttons

   JS: There's no due date on issues, so let's wait until this is
   definitely closed.

   BG: (agrees with getting rid of it)

   JS: Joseph will get rid of it

   <clown> ACTION: Joseph to modify IA2 and ATK/AT-SPI mappings
   for aria-presssed to remove "expose object attribute
   checkable:true". [recorded in
   [29]http://www.w3.org/2014/04/01-aapi-minutes.html#action01]

   <trackbot> Created ACTION-1418 - Modify ia2 and atk/at-spi
   mappings for aria-presssed to remove "expose object attribute
   checkable:true". [on Joseph Scheuhammer - due 2014-04-08].

   <clown> action-1418?

   <trackbot> action-1418 -- Joseph Scheuhammer to Modify ia2 and
   atk/at-spi mappings for aria-presssed to remove "expose object
   attribute checkable:true". -- due 2014-04-08 -- OPEN

   <trackbot>
   [30]https://www.w3.org/WAI/PF/Group/track/actions/1418

     [30] https://www.w3.org/WAI/PF/Group/track/actions/1418

   <clown>
   [31]http://www.w3.org/WAI/PF/aria-implementation-1.1/#mapping_e
   vents_state-change

     [31]
http://www.w3.org/WAI/PF/aria-implementation-1.1/#mapping_events_state-change

   JS: The problem was that "checkable:true" was present for
   toggle buttons
   ... But if the state changed on that toggle button, the
   state-changed event was for the "pressed" state and not the
   "checked" state.
   ... At the moment, aria-pressed is only used on toggle buttons
   ... But I can imagine it being applicable to other widgets.
   ... I think the worry was that if you put aria-pressed on
   something that didn't have a clear role
   ... You needed a way to know it could be pressed/toggled.
   ... Visually... Toggle buttons, you cannot necessarily see that
   they are toggle-able until they are toggled.

   BG: JAWS and NVDA have role toggle button.

   <clown> scribenick: clown

   JD: Actually I think that mixed state on a toggle button is
   wrong.
   ... Toggle buttons are either on or off, not mixed.

   JS: That's an issue with the spec, not the UAIG.

   <bgaraventa1979> I agree, it is confusing

   <joanie> ACTION: Joanie to investigate the validity of
   aria-pressed=mixed [recorded in
   [32]http://www.w3.org/2014/04/01-aapi-minutes.html#action02]

   <trackbot> Created ACTION-1419 - Investigate the validity of
   aria-pressed=mixed [on Joanmarie Diggs - due 2014-04-08].

   <joanie> JS: There's definitely aria-checked=mixed because I
   wrote that section.

   <joanie> JS: Reads the docs about the meaning of "mixed"

   <joanie> JS: I do not see any examples of mixed-state toggle
   buttons; I do see them for check boxes

   <joanie> JS: But definitely the spec provides for the existence
   of tri-state toggle buttons

   <joanie> JS: So that's where you need to go if you decide to
   complain

   <joanie> JD: I will complain

   <joanie> JS: Meeting over

Summary of Action Items

   [NEW] ACTION: Joanie to investigate the validity of
   aria-pressed=mixed [recorded in
   [33]http://www.w3.org/2014/04/01-aapi-minutes.html#action02]
   [NEW] ACTION: Joseph to modify IA2 and ATK/AT-SPI mappings for
   aria-presssed to remove "expose object attribute
   checkable:true". [recorded in
   [34]http://www.w3.org/2014/04/01-aapi-minutes.html#action01]

   [End of minutes]
     __________________________________________________________


    Minutes formatted by David Booth's [35]scribe.perl version
    1.138 ([36]CVS log)
    $Date: 2014-04-01 20:10:42 $
     __________________________________________________________

     [35] http://dev.w3.org/cvsweb/~checkout~/2002/scribe/scribedoc.htm
     [36] http://dev.w3.org/cvsweb/2002/scribe/

Received on Tuesday, 1 April 2014 20:55:34 UTC