[AAPI] Minutes UAI TF Meeting, Tuesday 21 June 2016

URL: https://www.w3.org/2016/06/21-aapi-minutes.html

Plain text follows:
   [1]W3C

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

   Accessible Rich Internet Applications Working Group Teleconference

21 Jun 2016

   See also: [2]IRC log

      [2] http://www.w3.org/2016/06/21-aapi-irc

Attendees

   Present
          Joanmarie_Diggs, Joseph_Scheuhammer, Bryan_Garaventa,
          Rich_Schwerdtfeger, Cynthia_Shelly

   Regrets
   Chair
          Joseph_Scheuhammer

   Scribe
          joanie

Contents

     * [3]Topics
         1. [4]ACTION-1533 (Joanie) Periodic checkin: Webdriver
            automated accessibilty testing framework for testing
            ARIA in ATK/AT-SPI.
         2. [5]ACTION-1696/ACTION-1698 (Joanie) ATK/AT-SPI, IA2,
            and AXAPI mappings of aria-keyshortcuts.
         3. [6]ACTION-1706 (Joseph) Statement regarding
            deprecation of UIA Express.
         4. [7]ACTION-2065 (Joseph/Cynthia) Validation of
            aria-activedescendant.
         5. [8]ACTION-1739 (Joseph) Provide all the missing
            atk/at-spi2 interfaces.
         6. [9]ACTION-2022 (Rich) Followup with Alex about
            IA2_ROLE_LANDMARK (obsolete?)
         7. [10]ACTION-2056 (Rich/Joseph/Joanie) Mappings for new
            aria-haspopup values.
     * [11]Summary of Action Items
     * [12]Summary of Resolutions
     __________________________________________________________

   <scribe> agenda: this

   <joanie> scribe: joanie

ACTION-1533 (Joanie) Periodic checkin: Webdriver automated
accessibilty testing framework for testing ARIA in ATK/AT-SPI.

   <clown> action-1533

   <trackbot> action-1533 -- Joanmarie Diggs to Implement
   Webdriver automated accessibilty testing framework for testing
   ARIA in ATK/AT-SPI -- due 2016-05-24 -- OPEN

   <trackbot> [13]http://www.w3.org/WAI/ARIA/track/actions/1533

     [13] http://www.w3.org/WAI/ARIA/track/actions/1533

   JD: This is a work in progress.
   ... I have started the porting to GNU/Linux.
   ... In a nutshell, it looks awesome, but I think we will
   probably want to make some changes/tweaks.
   ... One being nitty-level stuff, like a common/shared API that
   handles differences in exposure in the various platforms (e.g.
   interfaces versus patterns).
   ... The other potentially being additions needed to test what
   we need to test in ARIA.
   ... But I won't be sure about that until I make more progress
   on the porting.
   ... But, again, what Microsoft did is pretty cool.

   JS: Change the due date or weekly reminders?

   JD: Up to you. I'm actively working on it now.

   JS: I could move it to the 5th of July -- or the 12th.
   ... If you want it added to the agenda, just say so.

   JD: Perfect.

   JS: My other question is: Are you greating a github repo for
   GNU/Linux?

   JD: My plan is ultimately for it to be in github.

   CS: I was thinking that the higher-level abstraction is called
   ARIA 2.0.

   RS: Control Patterns would fall under the branding for ARIA
   2.0.
   ... It's to look at things beyond just ARIA.

   CS: Control Patterns and Interfaces seem fairly similar.
   ... And this is why I was thinking it's part of ARIA 2.0.

   <clown> scribenick: clown

   JD: for example, we have new values for aria-haspopup
   ... in the test, the way we get those values is dependent on
   the AAPI.

   CS: It sounds like you want something that works quick-n-dirty
   in the short term.
   ... The way we did that is consider a mapping from the general
   ARIA to the platform specific values. UIA in our case.
   ... I would think you can do that platform specific stuff in
   ATK/AT-SPI.
   ... I don't want to get sucked into a full blown abstract
   generalizaion of these test cases.

   JD: I hear you, and we can do the quick-n-dirty approach.
   ... But, I'm not yet at a point to say exactly how to do that.
   ... My plan is to contact your colleague and submit patches.

   CS: do you have examples of what you are thinking of?

   JD: I haven't gotten that far yet.
   ... My plan is to figure out how to do this, and submit
   patches.

   CS: What you say sounds interesting. I would like to see it
   more concretely.

   <joanie> scribe: joanie

   JS: I'll bring this up again July 12th.
   ... Sooner if you let me know.

ACTION-1696/ACTION-1698 (Joanie) ATK/AT-SPI, IA2, and AXAPI mappings
of aria-keyshortcuts.

   <clown> action-1696

   <trackbot> action-1696 -- Joanmarie Diggs to Propose ATK/AT-SPI
   mapping of aria-keyshortcuts -- due 2016-06-14 -- OPEN

   <trackbot> [14]http://www.w3.org/WAI/ARIA/track/actions/1696

     [14] http://www.w3.org/WAI/ARIA/track/actions/1696

   JS: There has been some discussion with Alex Surkove about
   this.
   ... This is on Joanie to come up with a way to expose
   aria-keyshortcuts.
   ... Joanie's proposal recently was to use the AtkAction
   interface.
   ... AtkAction has name, description, is connected to an
   accessible, and also has a do_action method.
   ... The user can interact with the widget via this interface.
   ... One of Joanie's proposals was to use this as a means to
   expose the keyshortcuts.
   ... But the problem with this is that the action cannot
   actually be performed.

   <clown> scribenick: clown

   JD: Originally, I emailed IA2 list for synchronicity between
   IA2 and ATK.
   ... Jamie said that IA2 already have accKeyShortcut property.
   ... So, I said ATK coudl either use an object attribute or
   AtkAction.
   ... But object attributes are not performant. AtkAction would
   be better.
   ... I asked Alex what he would want to do.
   ... And, he said AtkAction is fine, and the browser will guess
   what action to perform, or send a click action

   CS: We analyzed a whole bunch of stuff, and mapped it to a
   click.
   ... But we did not do this for aria-keyshortcuts.

   JD: That is how my platform works.
   ... I don't think we can use an action for ATK for
   aria-keyshortcuts.
   ... I'm landing on just using object attributes.

   CS: We are not doing anything except putting the string in a
   property.
   ... Actually, real widgets' shortcuts work the same way — it's
   descriptive. It doesn't do anything.

   JD: I can imagine authors scripting multiple keyboard shortcuts
   without there being any widget present at all.
   ... I'm going to email Alex, and just say object attribute.

   <joanie> scribe: joanie

   JS: I replied to Alex that there might not even be a click
   handler.
   ... So doing a click might result in nothing.

   CS: There are a lot of cases where there is a click handler.
   ... So that's why we went with that.
   ... It usually will do what the author has already done.

   JS: There's even a section in the Core AAM about setting click
   actions as the default action.

   <clown>
   [15]http://localhost/~clown/GitHub/aria/core-aam/core-aam.html#
   mapping_actions

     [15]
http://localhost/~clown/GitHub/aria/core-aam/core-aam.html#mapping_actions

   <clown>
   [16]https://rawgit.com/w3c/aria/master/core-aam/core-aam.html#m
   apping_actions

     [16]
https://rawgit.com/w3c/aria/master/core-aam/core-aam.html#mapping_actions

   CS: That was always a fallback that we intended to improve.

   JS: But what's here is a SHOULD.

   CS: In the short term, having it as a descriptive string sounds
   like the right thing to do.

   JS: I'll leave it to Joanie to follow up with Alex.
   ... And I will leave the due date as-is.

   JD: I think this is ready for you Joseph.
   ... It's expose as an object attribute
   ... The exact name is, I think, keyshortcuts (i.e. chop off
   "aria-")
   ... You should probably verify this.

   JS: I will try to have this done next week.

ACTION-1706 (Joseph) Statement regarding deprecation of UIA Express.

   <clown> action-1706

   <trackbot> action-1706 -- Joseph Scheuhammer to Compose a
   statement regarding the "deprecation" of uia express mappings
   in the core-aam, referencing the aria 1.0 mapping spec. -- due
   2016-06-07 -- PENDINGREVIEW

   <trackbot> [17]http://www.w3.org/WAI/ARIA/track/actions/1706

     [17] http://www.w3.org/WAI/ARIA/track/actions/1706

   JS: I got the statement from Cynthia a while back.

   <clown>
   [18]https://github.com/w3c/aria/commit/dd087eaa68b9936bc8173ae9
   2d3ec47b570860d6

     [18]
https://github.com/w3c/aria/commit/dd087eaa68b9936bc8173ae92d3ec47b570860d6

   JS: I added that statement to the spec (URL above).
   ... I did that last week.
   ... You (Cynthia) can delete your branch after you verify it.

   CS: Looks good to me.

   <clown>
   [19]https://rawgit.com/w3c/aria/Action-1706/core-aam/core-aam.h
   tml#intro_aapi

     [19]
https://rawgit.com/w3c/aria/Action-1706/core-aam/core-aam.html#intro_aapi

   JS: I manually squashed and merged it.

   CS: I'm also going to close the pull request.

   JS: And I'm going to close the action.

ACTION-2065 (Joseph/Cynthia) Validation of aria-activedescendant.

   <clown> action-2065?

   <trackbot> action-2065 -- Joseph Scheuhammer to Add this
   statement to the core-aam: “user agents are not expected to
   validate that the active descendant is a descendant of the
   focused container. “ -- due 2016-05-19 -- PENDINGREVIEW

   <trackbot> [20]http://www.w3.org/WAI/ARIA/track/actions/2065

     [20] http://www.w3.org/WAI/ARIA/track/actions/2065

   JS: This is your pull request for the statement regarding user
   agents not needing to validate activedescendant.
   ... I've merged in your text and added the link we discussed
   last week.

   <clown>
   [21]https://github.com/w3c/aria/commit/b51868eca0acc75bfe696010
   12437f0cef404001

     [21]
https://github.com/w3c/aria/commit/b51868eca0acc75bfe69601012437f0cef404001

   <clown>
   [22]https://rawgit.com/w3c/aria/master/core-aam/core-aam.html#k
   eyboard-focus_aria-activedescendant

     [22]
https://rawgit.com/w3c/aria/master/core-aam/core-aam.html#keyboard-focus_aria-activedescendant

   JS: It's at the URL above. In the first paragraph, second
   sentence.
   ... (Reads outloud)
   ... Any objections?

   CS: Looks good to me.

   JS: You can close your branch (Cynthia).
   ... I will close the action.

ACTION-1739 (Joseph) Provide all the missing atk/at-spi2 interfaces.

   <clown> action-1739

   <trackbot> action-1739 -- Joseph Scheuhammer to Provide all the
   missing atk/at-spi2 interfaces for joseph -- due 2016-07-05 --
   OPEN

   <trackbot> [23]http://www.w3.org/WAI/ARIA/track/actions/1739

     [23] http://www.w3.org/WAI/ARIA/track/actions/1739

   JS: I'm supposed to add, based on Joanie's advice, all the
   relevant interfaces.
   ... I've started, but not yet finished.
   ... So nothing to review yet.
   ... We'll come back to it next week.

ACTION-2022 (Rich) Followup with Alex about IA2_ROLE_LANDMARK
(obsolete?)

   <clown> action-2022?

   <trackbot> action-2022 -- Richard Schwerdtfeger to Followup
   with alex about ia2 discussion re: adding a landmark role,
   similar to atk -- due 2016-02-23 -- OPEN

   <trackbot> [24]http://www.w3.org/WAI/ARIA/track/actions/2022

     [24] http://www.w3.org/WAI/ARIA/track/actions/2022

   JS: I came across this action, which I think is obsolete.

   RS: We discussed this with Alex.
   ... We got agreement, but I don't know if they did it yet.

   JS: The mappings are already using it in the Core AAM.
   ... We'll find out when we do the tests.
   ... I suggest we close action-2022.
   ... Objections?

   (None)

ACTION-2056 (Rich/Joseph/Joanie) Mappings for new aria-haspopup
values.

   <clown> action-2056?

   <trackbot> action-2056 -- Joanmarie Diggs to Coordinate the
   mappings for the various AAPIs of the enumerated aria-haspopup
   values -- due 2016-06-07 -- OPEN

   <trackbot> [25]http://www.w3.org/WAI/ARIA/track/actions/2056

     [25] http://www.w3.org/WAI/ARIA/track/actions/2056

   JS: Before Rich went on vacation, he gave me some text.
   ... I did that but would like your (Rich's) review.

   <clown>
   [26]https://rawgit.com/w3c/aria/master/core-aam/core-aam.html#a
   riaHaspopupDialog

     [26]
https://rawgit.com/w3c/aria/master/core-aam/core-aam.html#ariaHaspopupDialog

   JS: The first one is at the above URL.

   RS: I had put stuff in the spec for those things.

   JS: Right, and I've merged them in.
   ... I think we have everyone's but Apple's.

   JD: There's an email on the list for that.

   RS: The reason we have this in here is because of combo box.
   ... We want to avoid people getting confused by having people
   wind up in, say, a grid after activating a combobox.

   JS: I think James is going to come back with you can find out
   via examining the object associated with
   aria-controls/aria-owns.
   ... So an AT can already figure this out.

   RS: So have them map it to true.

   JS: I think I just copied the "true" values for AXAPI. Or you
   did, Rich.

   RS: I did.
   ... As long as VoiceOver can do the extra work, then there
   should be no problem.

   JS: Should I add a statement to this effect?

   RS + JD: No. That's what VoiceOver does. It's not a mapping.

   JS: We cannot close this because James has an objection.

   <clown> Of note is that this change is not backwards-compatible
   with ARIA 1.1 because authors changing aria-haspopup="true" to
   one of the new enumerated types will find that this breaks all
   current implementations. It will currently be treated as
   aria-haspopup="false" (the default value for ARIA 1.0).

   JS: He also said the above.
   ... I don't think that's right.

   RS: I don't either.

   <bgaraventa1979> sorry guys getting booted from the room

   <clown> "When exposing as a platform API Boolean
   [27]https://www.w3.org/TR/wai-aria-implementation/#def_state,
   treat "", "undefined" or no
   [28]https://www.w3.org/TR/wai-aria-implementation/#def_attribut
   e present as false. Treat any other value as true."

     [27] https://www.w3.org/TR/wai-aria-implementation/#def_state,
     [28] https://www.w3.org/TR/wai-aria-implementation/#def_attribute

   JD: I think James is correct.

   <clown> "When exposing as a platform API Boolean state, treat
   "", "undefined" or no attribute present as false. Treat any
   other value as true."

   <clown>
   [29]https://www.w3.org/TR/wai-aria-implementation/#document-han
   dling_author-errors

     [29]
https://www.w3.org/TR/wai-aria-implementation/#document-handling_author-errors

   CS: For our platform, we plan on making everything that isn't
   "false" map to "true"

   JS: I found the above quote from the 1.0 UAIG.
   ... Is there any conflict between the ARIA spec and mapping
   spec regarding error handling for booleans?

   RS: I think we're fine.

   JS: I'm tempted to put together some markup to test this with
   Safari.
   ... And see how it gets mapped currently.
   ... But I don't know if I'll have time to do this.

   RS: I think James' issue is addressed.
   ... And there's no API change.

Summary of Action Items

Summary of Resolutions

   [End of minutes]

Received on Tuesday, 21 June 2016 20:10:01 UTC