[aapi] Minutes: UAI TF Meeting Tue 24 March 2015

Link: http://www.w3.org/2015/03/24-aapi-minutes.html

Plain text follows:

   [1]W3C

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

                               - DRAFT -

           Protocols and Formats Working Group Teleconference
                              24 Mar 2015

   See also: [2]IRC log

      [2] http://www.w3.org/2015/03/24-aapi-irc

Attendees

   Present
          Bryan_Garaventa, Cynthia_Shelly, Joanmarie_Diggs,
          Joseph_Scheuhammer, Rich_Schwerdtfeger

   Regrets
          David_Bolter

   Chair
          Joseph_Scheuhammer

   Scribe
          joanie

Contents

     * [3]Topics
         1. [4]ISSUE-700/ACTION-1544/ACTION-1546: (Rich, Cynthia,
            David, Alex) Mapping of role rowgroup.
         2. [5]ACTION-1373/ISSUE-441: (Cynthia) Test case for when
            aria-posinset and aria-setsize are provided explicitly
            on some but not all elements.
         3. [6]ACTION-1590: (Joseph/James) Confirm that the AXAPI
            mapping of aria-owns is AXOwns array.
         4. [7]Placeholder discussion
     * [8]Summary of Action Items
     __________________________________________________________

   <trackbot> Date: 24 March 2015

   <clown> agenda: this

   <scribe> scribe: joanie

ISSUE-700/ACTION-1544/ACTION-1546: (Rich, Cynthia, David, Alex)
Mapping of role rowgroup.

   <clown> issue-700?

   <trackbot> issue-700 -- Discuss rowgroup, tbody, and thead in
   aapis -- open

   <trackbot> [9]https://www.w3.org/WAI/PF/Group/track/issues/700

      [9] https://www.w3.org/WAI/PF/Group/track/issues/700

   JS: We started this last week, but I don't recall finishing it.

   RS: What has Alex actually said?

   JS: You (Rich) sent him a question and I think he answered it.

   (Looks for the mailing list thread)

   <clown>
   [10]https://lists.w3.org/Archives/Public/public-pfwg/2015Feb/00
   60.html

     [10] https://lists.w3.org/Archives/Public/public-pfwg/2015Feb/0060.html

   JS: The above is Alex's reply.

   1. I only see one use case for role="rowgroup" is to break rows
   into

   groups, each groups is a separate set and position inside it
   and set size

   are calculated independently.

   RS: If I read this correctly, he wants the actual number of
   rows within the group to be reflected; not just the entire
   table. Correct?
   ... If you were going through a number of rows, and all of a
   sudden the row count goes from 3 to a row count of 500, in the
   same table, you'd probably go crazy if you didn't know those
   rows were part of a group.

   JS: Sample markup?

   <richardschwerdtfeger> <table>

   <richardschwerdtfeger> <rowgroup>

   <richardschwerdtfeger> <row aria-rowindex=“1” aria-rowcount=2”>

   <richardschwerdtfeger> <row aria-rowindex=“2” aria-rowcount=2”>

   <richardschwerdtfeger> </rowgroup>

   <richardschwerdtfeger> <row aria-rowindex=“5”
   aria-rowcount=500”>

   <richardschwerdtfeger> …

   <richardschwerdtfeger> </table>

   RS: So the rowcount is in the context of the rowgroup (in the
   example above).

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

     [11]
http://rawgit.com/w3c/aria/master/core-aam/core-aam.html#role-map-rowgroup

   JS: The above is the current mapping.
   ... Three are exposing something.

   <clown> scribenick: clown

   JD: Ideally, I don't want rowgroup exposed.
   ... I don't see why it would have a whole separate rowcount
   ... It's confusing.

   RS: We didn't have rowcounts before.
   ... We can make this part of our response.
   ... We want rowcount to apply to the whole table, not just
   groups.

   <scribe> scribenick: Joanie

   JS: He then says they don't expose groups for thead, tfoot, or
   tbody.
   ... Unless they make it focusable.
   ... I do think he has a point with focusable. In terms of the
   user agent, you can Tab to that row group. And that row group
   has DOM focus.

   RS: Do you want it to be a panel or a group?
   ... Joanie's going to be drafting a panel role. Have you done
   that yet?

   JD: I am still working on table-related things.

   RS: (Writing response). Rowcount and rowindex reflect the
   entire table; not the rowgroup.
   ... Due to this, rowgroup has no impact on these values
   (aria-rowcount and aria-rowindex)

   <richardschwerdtfeger> 1. The group has determined that the
   aria-rowcount needs to be indicative of the total number of
   rows in the entire table or grid.

   <richardschwerdtfeger> 2. The aria-rowindex is indicative of
   the index of the row within the entire set of rows for the
   entire grid or table

   <richardschwerdtfeger> 3. Due to 1 and 2 rowgroups have no
   impact on the values for aria-rowcount and aria-rowindex

   <richardschwerdtfeger> 4. So, a rowgroup will have no impact on
   context withing row count.

   <richardschwerdtfeger> 1. The group has determined that the
   aria-rowcount needs to be indicative of the total number of
   rows in the entire table or grid.

   <richardschwerdtfeger> 2. The aria-rowindex is indicative of
   the index of the row within the entire set of rows for the
   entire grid or table

   <richardschwerdtfeger> 3. Due to 1 and 2 rowgroups have no
   impact on the values for aria-rowcount and aria-rowindex

   <richardschwerdtfeger> 4. So, a rowgroup will have no impact on
   context with respect to row count.

   (Discussion about appropriate role for rowgroup)

   CS: Group and panel are not the same in MSAA or UIA.
   ... There's pane and group. The pane would be like the thing on
   the right hand side in Office before the Ribbon.

   <clown>
   [12]https://developer.mozilla.org/en-US/docs/Web/Accessibility/
   AT-APIs/MSAA/Roles

     [12]
https://developer.mozilla.org/en-US/docs/Web/Accessibility/AT-APIs/MSAA/Roles

   CS: Typical applications you can use a function key to navigate
   amongst panes in Windows.

   RS: So is the group of any value in a rowgroup?

   JS: There has to be an accessible if it gets focus.

   CS: For UIA and MSAA it will need to be a group.

   JS: Joanie wants it to be a panel for ATK.

   RS: Does a div map to a group also?

   CS: Yes. Fieldset also maps to group.

   <clown>
   [13]https://developer.gnome.org/atk/unstable/AtkObject.html#ATK
   -ROLE-PANEL:CAPS

     [13]
https://developer.gnome.org/atk/unstable/AtkObject.html#ATK-ROLE-PANEL:CAPS

   JD: An ATK_ROLE_PANEL is a collection of "stuff" that is not a
   block of text.

   RS: What does Alex have it mapped to now?

   JS: He has it mapped to group.

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

     [14]
http://rawgit.com/w3c/aria/master/core-aam/core-aam.html#role-map-rowgroup

   <richardschwerdtfeger> 1. The group has determined that the
   aria-rowcount needs to be indicative of the total number of
   rows in the entire table or grid.

   <richardschwerdtfeger> 2. The aria-rowindex is indicative of
   the index of the row within the entire set of rows for the
   entire grid or table

   <richardschwerdtfeger> 3. Due to 1 and 2 rowgroups have no
   impact on the values for aria-rowcount and aria-rowindex

   <richardschwerdtfeger> 4. So, a rowgroup will have no impact on
   context with respect to row count. So, if it has focus we will
   give a role of "group" for MSAA based platforms, ATK/ATSPI will
   make it a ATK_ROLE_PANEL,

   RS: So I think the action is to ask James if it's ok to map
   this to AXGroup.

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

     [15]
http://rawgit.com/w3c/aria/master/core-aam/core-aam.html#role-map-rowgroup

   JS: The current mappings are on the assumption that it's not
   focused. Do you want to expose anything if it isn't focused and
   cannot be focused?

   CS: Actually no.
   ... We'd follow the same rules as div. For instance, if it had
   a name, it would be exposed.

   JD: You need to be consistent: If it's focusable, and we must
   expose it at all, then it needs to be in the tree.

   CS: That's what we do with everything. If it has tabindex, then
   it goes into the tree.

   RS: I'll modify our response to reflect that.

   CS: Don't we have rules about this (when to expose objects)?

   JS: Yes.

   <richardschwerdtfeger> 1. The group has determined that the
   aria-rowcount needs to be indicative of the total number of
   rows in the entire table or grid.

   <richardschwerdtfeger> 2. The aria-rowindex is indicative of
   the index of the row within the entire set of rows for the
   entire grid or table

   <richardschwerdtfeger> 3. Due to 1 and 2 rowgroups have no
   impact on the values for aria-rowcount and aria-rowindex

   <richardschwerdtfeger> 4. So, a rowgroup will have no impact on
   context with respect to row count. So, if it is focusable (has
   a tabindex attribute) or it has an ARIA global attribute
   applied we will give a role of "group" for MSAA based
   platforms, ATK/ATSPI will make it a ROLE_PANEL otherwise it
   will not be mapped.

   <clown>
   [16]http://rawgit.com/w3c/aria/master/core-aam/core-aam.html#in
   clude_elements

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

   <cyns> add UIA control type is also Group

   JS: The above part of the spec has this.
   ... (Reads from spec)

   <richardschwerdtfeger> This is in response to your first bullet
   on rowgroup after today's #aapi discussion:

   <richardschwerdtfeger> 1. The group has determined that the
   aria-rowcount needs to be indicative of the total number of
   rows in the entire table or grid.

   <richardschwerdtfeger> 2. The aria-rowindex is indicative of
   the index of the row within the entire set of rows for the
   entire grid or table

   <richardschwerdtfeger> 3. Due to 1 and 2 rowgroups have no
   impact on the values for aria-rowcount and aria-rowindex

   <richardschwerdtfeger> 4. So, a rowgroup will have no impact on
   context with respect to row count. So, if it is focusable (has
   a tabindex attribute) or it has an ARIA global attribute
   applied we will give a role of "group" for MSAA based platforms
   or UIA, ATK/ATSPI will make it a ROLE_PANEL otherwise it will
   not be mapped.

   JS: The mapping table entry for rowgroup currently treats it as
   if it's a rowgroup with nothing else on it. I need to add in an
   "if" clause.
   ... This is about the core mapping spec.
   ... I will take what you've written and use it as a guide for
   the text I will write for the mapping spec entry for rowgroup

   RS: I'll send this text to Alex.

   (Group agrees)

ACTION-1373/ISSUE-441: (Cynthia) Test case for when aria-posinset and
aria-setsize are provided explicitly on some but not all elements.

   <clown> action-1373?

   <trackbot> action-1373 -- Cynthia Shelly to Create a test case
   for when aria-posinset and aria-setsize are provided explicitly
   on some but not all elements. -- due 2015-03-24 -- OPEN

   <trackbot>
   [17]https://www.w3.org/WAI/PF/Group/track/actions/1373

     [17] https://www.w3.org/WAI/PF/Group/track/actions/1373

   JS: This is Cynthia's action. I think we decided to come back
   to it this week.
   ... It's to create a test case/file.

   CS: Leave it overdue.

   JS: This might be relevant to the rowindex, etc.

ACTION-1590: (Joseph/James) Confirm that the AXAPI mapping of
aria-owns is AXOwns array.

   <clown> action-1590?

   <trackbot> action-1590 -- James Craig to Confirm that the AXAPI
   mapping of aria-owns is AXOwns array -- due 2015-02-27 -- OPEN

   <trackbot>
   [18]https://www.w3.org/WAI/PF/Group/track/actions/1590

     [18] https://www.w3.org/WAI/PF/Group/track/actions/1590

   JS: I sent email to James.
   ... This is about if you use aria-owns, the AX API mapping is
   the AXLinkedUIElements array.
   ... I happened to be testing something else and discovered
   there's an AXOwns array.
   ... So I've emailed James to get confirmation.

   <clown> AXLinkedUIElements

Placeholder discussion

   BG: Have you followed the discussion on the list.

   RS: I haven't had time yet.

   JS: I've seen it, but I haven't made up my mind yet.

   RS: WCAG says you should not use it as a label.

   BG: What's messing up my testing is that placeholder is in the
   name calculation before title.
   ... So if you have both title and placeholder, the placeholder
   is used; not the title.

   JS: I'm guessing Steve based this on observing what happens in
   the practice. (?)

   BG: But in the accessibility tree, even when the placeholder
   text is gone, the label remains the placeholder.

   <joanie_>
   [19]https://lists.w3.org/Archives/Public/public-pfwg/2015Mar/01
   32.html

     [19] https://lists.w3.org/Archives/Public/public-pfwg/2015Mar/0132.html

Summary of Action Items

   [End of minutes]

Received on Tuesday, 24 March 2015 20:20:18 UTC