[aapi] Minutes: UAI TF Meeting Tue 20 January 2015

Link: http://www.w3.org/2015/01/20-aapi-minutes.html

Plain text follows:

   [1]W3C

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

                               - DRAFT -

           Protocols and Formats Working Group Teleconference
                              20 Jan 2015

   See also: [2]IRC log

      [2] http://www.w3.org/2015/01/20-aapi-irc

Attendees

   Present
          Joseph_Scheuhammer, Joanmarie_Diggs, Bryan_Garaventa,
          Rich_Schwerdtfeger

   Regrets
          David_Bolter

   Chair
          Joseph_Scheuhammer

   Scribe
          joanie

Contents

     * [3]Topics
         1. [4]ACTION-1408/ISSUE-540: (Joseph) Email Jason Kiss
            about AAPI differences.
         2. [5]ACTION-1545: (Joseph) Editors' draft upddated:
            rowgroup exposed as "not mapped" for ATK/AT-SPI.
         3. [6]ACTION-842: (Cynthia) Create a diagram to support
            section 1.2 A11Y vs DOM tree.
         4. [7]Action 1320
         5. [8]Action 1553
         6. [9]Action 1562
         7. [10]Bryan's Email related to native checkboxes and
            mixed state
         8. [11]Menus
     * [12]Summary of Action Items
     __________________________________________________________

   <trackbot> Date: 20 January 2015

   <clown> agenda: this

   <scribe> scribe: joanie

   agenda order 5, 3, 1, 2, 4, 6, 7, 8

ACTION-1408/ISSUE-540: (Joseph) Email Jason Kiss about AAPI
differences.

   <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 2014-12-30 -- OPEN

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

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

   JS: This action is from a long time ago. Jason Kiss asked on
   one of the lists about the differences between accessibility
   APIs.
   ... Jason thinks there should be a section in the core AAM
   outlining the differences.
   ... This was assigned to me last Jan at the face to face.

   <clown>
   [14]https://lists.w3.org/Archives/Public/wai-xtech/2014Apr/0008
   .html

     [14] https://lists.w3.org/Archives/Public/wai-xtech/2014Apr/0008.html

   JS: I then asked about it here and followed up with Jason via
   email.

   <clown> isseu-540?

   <clown> issue-540?

   <trackbot> issue-540 -- Should the UAIG have a section to
   describe the API differences -- open

   <trackbot> [15]https://www.w3.org/WAI/PF/Group/track/issues/540

     [15] https://www.w3.org/WAI/PF/Group/track/issues/540

   JS: The associated issue, is issue-540.
   ... Jason there mentioned the differences between OS X mappings
   for aria-label, aria-labelledby, and aria-describedby.
   ... At the last meeting, we all agreed to do this.
   ... Then we decided to ask Jason to do so.
   ... Jason agreed to take an initial stab at it.

   (JS reads/summarizes from email)

   JS: Jason wants us all to help him with our various platforms.
   ... Ideas?

   JD: I think it's a good idea in general, but isn't that already
   addressed via the mapping tables?

   JS: Yes, but I asked Jason and he suggested more text and
   suggestions would be helpful.
   ... I think it makes sense to invite Jason to join us at some
   of these meetings.
   ... Therefore, I'd like to close Action-1408 as done, leaving
   the issue (540) open.
   ... Then I'll reply to his email suggesting he go ahead and
   start, but see if he's able to join our meetings on occasion.

   RESOLUTION: Group agrees to close action-1408 and follow-up
   with Jason. (Which Joseph will do.)

ACTION-1545: (Joseph) Editors' draft upddated: rowgroup exposed as
"not mapped" for ATK/AT-SPI.

   <clown> action-1545?

   <trackbot> action-1545 -- Joseph Scheuhammer to Make sure the
   rowgroup mapping for atk/atspi is accurate -- due 2015-01-20 --
   CLOSED

   <trackbot>
   [16]https://www.w3.org/WAI/PF/Group/track/actions/1545

     [16] https://www.w3.org/WAI/PF/Group/track/actions/1545

   JS: This is just an update from last week.
   ... We had agreed that the ATK/AT-SPI mapping for rowgroup was
   "not mapped."
   ... The text is....

   <clown>
   [17]http://rawgit.com/w3c/aria/master/core-aam/core-aam.html#ro
   le-map-rowgroup

     [17]
http://rawgit.com/w3c/aria/master/core-aam/core-aam.html#role-map-rowgroup

   JS: Now both ATK/AT-SPI and AX API share the "not mapped"
   mapping. This is not a change for AX API.

ACTION-842: (Cynthia) Create a diagram to support section 1.2 A11Y vs
DOM tree.

   JS: Cynthia is not here.

Action 1320

   action-1320

   <trackbot> action-1320 -- Joseph Scheuhammer to Consider
   mapping the "offscreen" api properties in the situation of
   aria-hidden="false" on non-rendered elements. -- due 2015-01-20
   -- OPEN

   <trackbot>
   [18]https://www.w3.org/WAI/PF/Group/track/actions/1320

     [18] https://www.w3.org/WAI/PF/Group/track/actions/1320

   JS: I think I left it there because there was discussion last
   week regarding whether or not Rich wanted an action to talk to
   Freedom Scientific about this.
   ... I'm going to move this back a week and bring it back in
   case richardschwerdtfeger is still here.

Action 1553

   action-1553

   <trackbot> action-1553 -- Joseph Scheuhammer to Provide
   ATK/AT-SPI2 mapping for role="searchbox" -- due 2015-01-21 --
   OPEN

   <trackbot>
   [19]https://www.w3.org/WAI/PF/Group/track/actions/1553

     [19] https://www.w3.org/WAI/PF/Group/track/actions/1553

   JS: Ok, this makes perfect sense to me.
   ... I'll try to have this done for next meeting.

Action 1562

   action-1562

   <trackbot> action-1562 -- Joseph Scheuhammer to Provide
   ATK/AT-SPI2 mapping for role="switch" -- due 2015-01-22 -- OPEN

   <trackbot>
   [20]https://www.w3.org/WAI/PF/Group/track/actions/1562

     [20] https://www.w3.org/WAI/PF/Group/track/actions/1562

   JS: There was a discussion about the name/value on switches
   versus toggle buttons.

   JS and BG: When this first came up we had questions about why
   we need this role instead of switch.

   JS: I forget the exact justification James Craig provided but I
   think he convinced me. :)
   ... I'd also like to see role toggle button in ARIA.
   ... I'd try to do this action too.

Bryan's Email related to native checkboxes and mixed state

   <clown>
   [21]https://lists.w3.org/Archives/Public/public-pfwg/2014Dec/01
   62.html

     [21] https://lists.w3.org/Archives/Public/public-pfwg/2014Dec/0162.html

   (Group reads email above)

   JS: Native host semantics don't trump ARIA on states and
   properties.

   <clown>
   [22]http://rawgit.com/w3c/aria/master/core-aam/core-aam.html#ma
   pping_conflicts

     [22]
http://rawgit.com/w3c/aria/master/core-aam/core-aam.html#mapping_conflicts

   JS: (Reads from above)
   ... So in this particular case (Bryan), if the HTML spec says
   you can't do that, then you can't do that.

   BG: Does it say you can't do it?

   <clown> [23]http://www.w3.org/TR/html/

     [23] http://www.w3.org/TR/html/

   JS: (Looks at input from above spec)

   <clown>
   [24]http://www.w3.org/TR/html/forms.html#checkbox-state-%28type
   =checkbox%29

     [24]
http://www.w3.org/TR/html/forms.html#checkbox-state-%28type=checkbox%29

   Allowed ARIA state and property attributes:

   Global aria-* attributes

   Any aria-* attributes applicable to the allowed roles.

   JS: So I think your example with mixed is valid. I'd have to
   check it.

   (Discussion about styling)

   BG: It's a boolean property, so you can't do mixed/between
   them.

   JS: Someone (Paciello Group?) has a document outlining
   suggestions about when it's best to use ARIA.
   ... And when the form is submitted, the ARIA value wouldn't be
   included.

   BG: It's probably easier just to style it.

   JS: In terms of the spec, it seems to be allowed.

   BG: If that's true, it doesn't seem to be done correctly by any
   of the APIs.

   (Group laments lack of tri-state checkboxes)

   JS: (Reads from Alex Surkov's response on-list)

Menus

   BG: On the authoring guide, we're talking about menus best
   practices.
   ... James and Matt believe persistent state menus should be
   allowed. I'm less certain.
   ... If you look at the UAIG, it says there are certain
   properties that only exist only when the menu is rendered.
   ... Does that sound accurate?

   <clown>
   [25]http://rawgit.com/w3c/aria/master/core-aam/core-aam.html#ma
   pping_events_menus

     [25]
http://rawgit.com/w3c/aria/master/core-aam/core-aam.html#mapping_events_menus

   JS: I don't know what the phrase "persistent menu" means.

   BG: When you open the page, the properties are already there.
   ... You can get to them via the keyboard.
   ... But they're not dynamically rendered.

   JS: So it doesn't behave anything like a menu.

   BG: Correct.

   JS: But it's declared as a menu.

   BG: Correct.

   JS: So your question is, is it going to fire these events?
   ... There's no menu-start event because it's always on -- or
   maybe it's there upon load.
   ... I'm using the MSAA event names btw.
   ... As you navigate the menu, it's supposed to fire focus
   events.

   BG: I think that would fire correctly as you navigate into it.
   ... But there's no platform equivalent of automatically-opening
   menus.
   ... The other question they were asking was about the
   difference between a menu bar and menu.
   ... Because from this list it doesn't look like there's any
   differences in events triggered.

   JS: Menu bars trigger the start; menus don't.
   ... If focus shifts to the menubar, that should trigger the
   menu start event.
   ... When you leave the menubar, the menu end event occurs.

   BG: They were saying that functionally there wasn't any
   difference between the two.

   JS: Do you have an example?

   BG: I have a hypothetical one.
   ... James was talking about if you have a complex header
   navigation structure with nested lists, you could have
   role="menu" for each nested list.
   ... And you'd have role=menuitem for the links.
   ... The problem is that if you put role=menu on the ul, the
   structure is lost.
   ... Menus themselves are not identified like regions.
   ... So if you're in applications mode, you're fine. Otherwise
   you cannot tell where one menu ends and another begins.
   ... A header structure that's five levels deep with hundreds of
   items, and all you want is "login," that's really hard.
   ... What kind of API issues might be involved in doing that?

   JS: I find it odd to call these things menus, myself.
   ... The point of ARIA was: You have some markup which was
   styled to look like menus and menu bars and it reacted to input
   events like it was a menu and menu bar.
   ... But there was nothing to tell assistive technologies that
   was what it was.

   BG: Precisely

   JS: But in this example, it doesn't behave like a menu to begin
   with.
   ... So why do you want to do this?
   ... James wants to move past them quickly?

   BG: Yes.

   JS: Is it a tree then?

   BG: It's one flat series of links I think.

   JS: Conclusion, I need to see an example.
   ... There's a lot to investigate.

   BG: I think Matt built a small demo and it wasn't working for
   him either.

   JS: Java Swing allowed you, if you had a menu popped open, to
   jump quickly by typing the name.
   ... Native Mac menus do this too.

   <bgaraventa1979> sorry got booted

Summary of Action Items

   [End of minutes]

Received on Tuesday, 20 January 2015 21:09:00 UTC