- From: Joanmarie Diggs <jdiggs@igalia.com>
- Date: Tue, 18 Aug 2015 16:11:37 -0400
- To: "wai-xtech@w3.org" <wai-xtech@w3.org>
Link: http://www.w3.org/2015/08/18-aapi-minutes.html
Plain text follows:
[1]W3C
[1] http://www.w3.org/
Protocols and Formats Working Group Teleconference
18 Aug 2015
See also: [2]IRC log
[2] http://www.w3.org/2015/08/18-aapi-irc
Attendees
Present
Bryan_Garaventa, Joanmarie_Diggs, Joseph_Scheuhammer
Regrets
Rich_Schwerdtfeger, Cynthia_Shelly
Chair
Joseph_Scheuhammer
Scribe
joanie
Contents
* [3]Topics
1. [4]Mapping of aria-checked on UIA and UIA Express
2. [5]Any thoughts/updates in the past two weeks?
3. [6]ACTION-1526: (All) Propose practices for fallback
role of text in authoring guide [Wrong product
assignment].
4. [7]ACTION-1529: (Joseph, Joanie) Pending review:
ATK/AT-SPI mappings of aria-current.
5. [8]ACTION-1528: (Alex) Bolter to investigate the
proper IA2 mappings for aria-current.
6. [9]ACTION-1531: (Alex) Bolter to investigate ia2
mapping of role="text".
7. [10]ACTION-1533: (Joanie) Investigate and possibly
create a minimal/skeleton javascript at-spi2 listener
with the aim of automating the 1.1 tests.
8. [11]ACTION-1662: (Joseph) Edit pending: IA2 mapping
for aria-roledescription.
9. [12]ACTION-1667: (Joseph) Update the mappings for aria
role region to reflect its landmark inheritance as
defined in the aria 1.1 spec.
* [13]Summary of Action Items
__________________________________________________________
<trackbot> Date: 18 August 2015
<clown> agenda: this
<scribe> scribe: joanie
Mapping of aria-checked on UIA and UIA Express
BG: This is on radio buttons
<clown>
[14]http://rawgit.com/w3c/aria/master/core-aam/core-aam.html#ar
iaCheckedTrue
[14]
http://rawgit.com/w3c/aria/master/core-aam/core-aam.html#ariaCheckedTrue
JS: MSAA and UIA Express use the selected item pattern and set
selected on the radio buttons; not checked.
... Cynthia told me years ago that this is how it works and I
accepted it.
BG: But that's now how it works on Windows. There is no
selected state.
JS: Ok, then I'd raise an issue against the mapping spec and
possibly create an action for Cynthia to confirm this.
... At the moment I have no way to check this.
BG: I am looking at Windows 7. I don't know if Windows 8 or 10
has changed.
JS: In pure UIA it uses the toggle pattern.
... It's only MSAA and UIA Express that uses the selection
pattern.
... So you have to be sure you're looking at the right one.
JD: And this is definitely not the AT just reporting "checked"
for selected?
<clown> for the minutes, here is the start of the email thread
where bryan raised this issue:
[15]https://lists.w3.org/Archives/Public/public-pfwg/2015Aug/00
08.html
[15] https://lists.w3.org/Archives/Public/public-pfwg/2015Aug/0008.html
BG: No.
... Firefox matches what I'm seeing in Windows.
... But in IE, it's using the selected state rather than the
checked state.
... As a result, screen readers are saying it's not checked
(even though it is). And it's saying selected.
JS: I can try to duplicate this with the inspector.
... But you can either carry on the discussion on the public
mailing list with Cynthia, or raise an issue against the
mapping spec so we can discuss it here.
BG: I tried to continue the discussion, but Cynthia is on
vacation.
Conclusion: We can wait until Cynthia gets back.
Any thoughts/updates in the past two weeks?
JS: I've been away for two weeks. I don't know if you met,
looked at your action items, etc.
BG: No
JD: No
JS: I did skim the DPub stuff on aria-describedat.
... And it looks like James Craig and Rich might be leaning
towards details-summary HTML5.1 elements.
... I see it on the Thursday agenda.
ACTION-1526: (All) Propose practices for fallback role of text in
authoring guide [Wrong product assignment].
<clown> action-1526?
<trackbot> action-1526 -- Matthew King to Propose practices for
fallback role of text in authoring guide -- due 2014-11-17 --
OPEN
<trackbot>
[16]https://www.w3.org/WAI/PF/Group/track/actions/1526
[16] https://www.w3.org/WAI/PF/Group/track/actions/1526
JS: For this meeting, I went through the actions list looking
for things that was overdue and came across the above.
... Why is this (action-1526) in the mapping spec? It's best
practices.
<clown> issue-435?
<trackbot> issue-435 -- Consider role="text" to expose elements
(and contents) as static text node -- open
<trackbot> [17]https://www.w3.org/WAI/PF/Group/track/issues/435
[17] https://www.w3.org/WAI/PF/Group/track/issues/435
JS: Oh, it's the associated issue (issue-435).
... The product for that is implementation spec, but it
shouldn't be. Right?
... I'm going to force action-1526 to be part of the authoring
guide. Any objections?
BG: I haven't heard from Matt recently.
JS: I've changed the product for action-1526 to the authoring
guide.
ACTION-1529: (Joseph, Joanie) Pending review: ATK/AT-SPI mappings of
aria-current.
<clown> action-1529?
<trackbot> action-1529 -- Joseph Scheuhammer to Investigate the
proper atk/at-spi mappings for aria-current -- due 2014-11-18
-- PENDINGREVIEW
<trackbot>
[18]https://www.w3.org/WAI/PF/Group/track/actions/1529
[18] https://www.w3.org/WAI/PF/Group/track/actions/1529
JS: I did the edits for aria-current and I just want to confirm
that it's right.
<clown>
[19]https://github.com/w3c/aria/commit/de39ed3cdf37d6d7d3c701af
c089cddf842cfd05
[19]
https://github.com/w3c/aria/commit/de39ed3cdf37d6d7d3c701afc089cddf842cfd05
JD: Mapping it to ATK/AT-SPI2_STATE_ACTIVE looks good. I'm not
sure about the object attribute name.
... Alex might also object to the object attribute name.
JS: And in IA2, STATE_ACTIVE means something else I think.
... Like active descendant.
JD: Were I you, I'd ask on the ia2 mailing list as Alex, and
James Teh, and I, and others might have some things to discuss.
JS: If you scroll down to the bottom of the commit, you'll see
I've also added the event for ATK.
<clown>
[20]http://rawgit.com/w3c/aria/master/core-aam/core-aam.html#ma
pping_events_state-change#event-aria-current
[20]
http://rawgit.com/w3c/aria/master/core-aam/core-aam.html#mapping_events_state-change
JD: In ATK/AT-SPI2, it's object:state-changed:active; not
object:state-changed:current. We don't have a "current" state.
... Just out of curiosity, is it really ToggleStateProperty for
UIA event?
JS: No. I'll correct that.
... It was a copy&paste error.
ACTION-1528: (Alex) Bolter to investigate the proper IA2 mappings for
aria-current.
JS: I'm going to write Alex and David on the IA2 mailing list
to get their opinion.
ACTION-1531: (Alex) Bolter to investigate ia2 mapping of role="text".
JS: I was going to email them this morning. I'll put it off for
another week and email Alex and David about the mapping for
role="text".
... Though I think he already has something implemented.
BG: One use case for the text role: Some people will have image
links which have embedded images. What happens in cases like
that is the AT (like JAWS) reports multiple links rather than
one.
... So using the text role would flatten it out.
<clown> [21]http://w3c.github.io/aria/aria/aria.html#text
[21] http://w3c.github.io/aria/aria/aria.html#text
<clown> "An element whose entire subtree should be exposed to
accessibility APIs as plain text."
<clown> IA2_ROLE_TEXT_FRAME
JD: Were it me, I'd ask on list before assuming the mapping is
IA2_ROLE_TEXT_FRAME. My concern is that James Teh might be
surprised if an image or something else which is not text shows
up as a "text frame".
<clown> "Collection of objects that constitute a logical text
entity. "
JD: Regarding the above "text entity", note that the text
entity might be a non-caret-navigable image with no string.
... So just as a sanity check, I'd be sure James Teh is aware
of this and has no issues.
... We dealt with this in ATK by mapping to the new "static"
role.
JS: If it's not text frame, then what is it? They don't seem to
have anything like role static in IA2.
JD: I don't know. But we should ask James Teh.
ACTION-1533: (Joanie) Investigate and possibly create a
minimal/skeleton javascript at-spi2 listener with the aim of
automating the 1.1 tests.
<clown> action-1533?
<trackbot> action-1533 -- Joanmarie Diggs to Investigate and
possibly create a minimal/skeleton javascript at-spi2 listener
with the aim of automating the 1.1 tests. -- due 2015-07-23 --
OPEN
<trackbot>
[22]https://www.w3.org/WAI/PF/Group/track/actions/1533
[22] https://www.w3.org/WAI/PF/Group/track/actions/1533
JD: Let's put this back another couple of months.
... This might be obsoleted by something Rich and Cynthia had
talked about.
<clown>
[23]https://wiki.mozilla.org/Accessibility/WebAccessibilityAPI#
Node_interface_extension
[23]
https://wiki.mozilla.org/Accessibility/WebAccessibilityAPI#Node_interface_extension
JS: Alex is also working on something in this area. See the URL
above.
... Ok, I'm setting this action to October 23rd.
ACTION-1662: (Joseph) Edit pending: IA2 mapping for
aria-roledescription.
<clown> action-1662?
<trackbot> action-1662 -- Joseph Scheuhammer to Propose IA2
mapping for aria-roledescription -- due 2015-06-30 -- OPEN
<trackbot>
[24]https://www.w3.org/WAI/PF/Group/track/actions/1662
[24] https://www.w3.org/WAI/PF/Group/track/actions/1662
JS: I just did this today.
<clown> issue-636?
<trackbot> issue-636 -- Continue discussion of localized role
"name" (or custom role "descriptions" on known role types) --
closed
<trackbot> [25]https://www.w3.org/WAI/PF/Group/track/issues/636
[25] https://www.w3.org/WAI/PF/Group/track/issues/636
JS: I found issue-636. If you look at the fourth note, James
Craig back in 2014 already knew the mappings.
<clown>
[26]http://rawgit.com/w3c/aria/master/core-aam/core-aam.html#ar
iaRoleDescription
[26]
http://rawgit.com/w3c/aria/master/core-aam/core-aam.html#ariaRoleDescription
JS: So I took discussions from Alex, stuff from Joanie, and
what James said in that note. I put them all together, and the
result is at the above URL.
... Everything is done but UIA Express.
... Probably UIA Express will use what UIA has.
... (Reads resulting mappings)
... I'm inclined to close action-1662.
... Also action-1660 (the ATK/AT-SPI mapping), but leave
action-1661 open.
... Objections?
(None)
ACTION-1667: (Joseph) Update the mappings for aria role region to
reflect its landmark inheritance as defined in the aria 1.1 spec.
<clown> action-1667?
<trackbot> action-1667 -- Joseph Scheuhammer to Update the
mapping spec for aria role region to reflect its landmark
inheritance as defined in the aria 1.1 spec. -- due 2015-07-06
-- OPEN
<trackbot>
[27]https://www.w3.org/WAI/PF/Group/track/actions/1667
[27] https://www.w3.org/WAI/PF/Group/track/actions/1667
JS: This came up as due. I read it and had a question.
<clown>
[28]https://rawgit.com/w3c/aria/master/aria/aria.html#region
[28] https://rawgit.com/w3c/aria/master/aria/aria.html#region
JS: But I'm not quite sure what I'm supposed to do.
... I do remember that we, in the ARIA spec, rewrote the region
role so that it inherits from landmark. So it must have a
label.
<clown>
[29]http://rawgit.com/w3c/aria/master/core-aam/core-aam.html#ro
le-map-region
[29]
http://rawgit.com/w3c/aria/master/core-aam/core-aam.html#role-map-region
JS: If we look at the region role mapping (above), all the
mappings are pane/panel related.
JD: In the case of ATK/AT-SPI2, if a region is a landmark, then
I would think map it to ROLE_LANDMARK.
... And then I guess expose "region" as the xml-roles value.
JS: There is a comment in the action, what happens if there is
no label?
<clown>
[30]https://rawgit.com/w3c/aria/master/aria/aria.html#region
[30] https://rawgit.com/w3c/aria/master/aria/aria.html#region
JD: If memory serves me, I had forgotten that region had to
have a label (i.e. it was not optional).
JS: Confirmed.
JD + JS: So it will never be mapped to ROLE_PANEL; a region is
always a ROLE_LANDMARK now.
JD: This might be yet another thing to ask on the IA2 mailing
list.
... Because IA2 doesn't have a landmark role.
Summary of Action Items
[End of minutes]
Received on Tuesday, 18 August 2015 20:12:15 UTC