- From: Joanmarie Diggs <jdiggs@igalia.com>
- Date: Tue, 18 Feb 2014 15:58:59 -0500
- To: "wai-xtech@w3.org" <wai-xtech@w3.org>
Link: http://www.w3.org/2014/02/18-aapi-minutes.html
Plain text follows:
[1]W3C
[1] http://www.w3.org/
- DRAFT -
Protocols and Formats Working Group Teleconference
18 Feb 2014
See also: [2]IRC log
[2] http://www.w3.org/2014/02/18-aapi-irc
Attendees
Present
Joseph_Scheuhammer, David_Bolter, Bryan_Geraventa,
Joanmarie_Diggs
Regrets
Cynthia_Shelly
Chair
Joseph_Scheuhammer
Scribe
joanie, clown
Contents
* [3]Topics
1. [4]ACTION-1383: (All) Implications for UAIG re: role
'none'.
2. [5]ISSUE-481/ACTION-1368/ACTION-1369: (David, Cynthia)
aria-live removal events should/must happen before
object removal.
3. [6]ISSUE-612/ACTION-1370: (David) Add text back for
implicit aria-level computation via aria-owns
hierarchy.
4. [7]ISSUE-441/ACTION-1372: (Cynthia) Test case for
incomplete use of aria-posinset/aria-setsize.
5. [8]ISSUE-618/ACTION-1371: (Cynthia) Investigate how IE
implements item de-selection in single selection
container.
6. [9]ISSUE-635/ACTION-1372: (Cynthia) Create test case
for rowgroup role.
7. [10]ISSUE-644: (All) Update ATK/AT-SPI mapping for
aria-hidden="true" to reflect elements should not be
exposed/mapped.
8. [11]1.1 ISSUES and ACTIONS: (All) Start with
ISSUE-540,
https://www.w3.org/WAI/PF/Group/track/products/23
9. [12]be done.
* [13]Summary of Action Items
__________________________________________________________
<trackbot> Date: 18 February 2014
<clown> agenda: this
<joanie> scribenick: joanie
ACTION-1383: (All) Implications for UAIG re: role 'none'.
<clown> action-1383?
<trackbot> action-1383 -- James Craig to Add role none to spec
-- due 2014-02-17 -- OPEN
<trackbot>
[14]https://www.w3.org/WAI/PF/Group/track/actions/1383
[14] https://www.w3.org/WAI/PF/Group/track/actions/1383
JS: Wants to touch base. If aria-role="none", opinion on what
we should do vis-a-vis UAIG?
DB: I would just add it to the UAIG
<clown>
[15]http://www.w3.org/TR/wai-aria-implementation/#exclude_eleme
nts2
[15] http://www.w3.org/TR/wai-aria-implementation/#exclude_elements2
JS: Points out above as one mention
DB: There's a bugzilla bug for Firefox
<davidb> joanie, the bugzilla comment is here
[16]https://bugzilla.mozilla.org/show_bug.cgi?id=971212#c1
[16] https://bugzilla.mozilla.org/show_bug.cgi?id=971212#c1
<davidb> from Bryan "In sharing the new role proposal with some
devs, many incorrectly associated role=none with display:none
because they share the same value."
JS: Rolemapping table probably says something about it
<clown>
[17]http://www.w3.org/TR/wai-aria-implementation/#mapping_role_
table
[17] http://www.w3.org/TR/wai-aria-implementation/#mapping_role_table
JS: Reads text. It's going to link to the exclusion section.
<clown> If the object is in the accessibility tree, expose as
ROLE_SYSTEM_PANE
JS: There is (the above) under IA2
<clown> If the object is in the accessibility tree, expose as
ROLE_PANEL
JS: And under ATK the above
... It's exposed as a panel
... If it has role="presentation" and is included in the tree
... I would assume, mapping for none is the same
DB: yeah
... As a side note, if it is included in the tree refers to
there being other markup requiring it to be in the tree
JS: If it's a button, in a presentation role, why would it have
role panel?
... I would have thought if something is interactive, ignore
the role
DB: Is that a real example?
JS: I don't know
... Currently the mapping table says in that case (his reading
of the mapping table), expose it as a panel
DB: It might be that they want to use a button element to do
something not button-like
JS: That's author error (quite likely)
DB: Sounds like there is more discussion to be had
JS: I think if aria-role="none" goes in, change to UIAG would
be to mention/add it to the docs for presentation
DB: agrees
... If this is if it is accepted
JS: Would not like two presentation roles
<davidb> (and deprecate presentation)
JS: If the purpose to switch to none is get rid of the
confusion / it's just a name change
ISSUE-481/ACTION-1368/ACTION-1369: (David, Cynthia) aria-live removal
events should/must happen before object removal.
<clown> action-1369?
<trackbot> action-1369 -- David Bolter to Investigate
implementation of aria-live region events in FF -- due
2014-02-11 -- OPEN
<trackbot>
[18]https://www.w3.org/WAI/PF/Group/track/actions/1369
[18] https://www.w3.org/WAI/PF/Group/track/actions/1369
DB: For this one, I didn't do my homework. But not sure he
understands it.
<clown> issue-481?
<trackbot> issue-481 -- aria-live removal events should or must
happen before object removal -- open
<trackbot> [19]https://www.w3.org/WAI/PF/Group/track/issues/481
[19] https://www.w3.org/WAI/PF/Group/track/issues/481
JS: It relates back to the above issue
... It says....
JS quotes the description
JS: He does say you must notify first before removing it
... If you notify it after, it goes to grab what was removed
and by then it's gone
DB: We've done some work on this
... As a meta-comment, why aren't these linked?
JS: Points out they are linked
DB: I have more homework to do here.
... I have the feeling this might be a "should" for now.
JS: He *thinks* Cynthia was going to look into it first to see
what IE does.
<clown> scribnick: clown
<clown> JD: <reinforces what Joseph said: impossible to present
material that is already gone.>
<clown> scribenick: clown
JD: Dead accessibles are difficult to deal with.
DB: I thought we did a whole bunch of work to deal with live
regions, and removal of text.
... We had a text cache, outisde of content and layout.
... Even had a "diff" algorithm to do this.
... I think we are close. I just have to dig.
... The UAIG section speaks about the deletion of a DOM node.
... This might be about ATs that look at the DOM.
<joanie> scribenick: joanie
JS: Do we want to change due date?
... Was due a week ago.
DB: I can make it next week.
<davidb> famous last words
ISSUE-612/ACTION-1370: (David) Add text back for implicit aria-level
computation via aria-owns hierarchy.
<clown> action-1370?
<trackbot> action-1370 -- Joseph Scheuhammer to Add text back
into uaig regarding computing implicit level information when a
tree is built out of aria-owns relationships. -- due 2014-02-11
-- OPEN
<trackbot>
[20]https://www.w3.org/WAI/PF/Group/track/actions/1370
[20] https://www.w3.org/WAI/PF/Group/track/actions/1370
<davidb> this is possibly
[21]https://bugzilla.mozilla.org/show_bug.cgi?id=499917
[21] https://bugzilla.mozilla.org/show_bug.cgi?id=499917
JS: This is about aria-owns on a flat DOM, but it makes a tree.
And the test case doesn't pass.
DB: Whether we have the spec from the UAIG, this is going to be
a "known issue" in Gecko for a little while
... due to priorities.
JS: From Monday call, this was a 1.1
... Michael said, this is what the spec says
... Rich and James pointed out it was a special case and likely
would not be seen in the wild.
<davidb> it smells bad
JS: We decided there was not enough time to get all the browser
people together, so let's just punt it for now
... I added the URL in a note.
... Do we want to change the date?
DB: What do we want to do for it?
JS: The action is on me. I removed the MUST statement from the
UAIG. But the proviso was to put it back in 1.1.
... I've been wanting to create a 1.1 draft.
... But Michael said Richard wants to split it in pieces
... Core, HTML, SVG are the three pieces.
... Core tells you 80%+ of what you need to know.
... Then there are a few things in HTML and SVG that override
core.
... Example, making title attribute the tooltip text.
... That wouldn't appear in the core, but in the HTML piece.
... I want to "tag and branch" to start on "trunk" (1.1)
<davidb> do it
ISSUE-441/ACTION-1372: (Cynthia) Test case for incomplete use of
aria-posinset/aria-setsize.
JS: pass on that because she's not here.
ISSUE-618/ACTION-1371: (Cynthia) Investigate how IE implements item
de-selection in single selection container.
JS: Ditto.
ISSUE-635/ACTION-1372: (Cynthia) Create test case for rowgroup role.
JS: Ditto.
ISSUE-644: (All) Update ATK/AT-SPI mapping for aria-hidden="true" to
reflect elements should not be exposed/mapped.
<clown> issue-644?
<trackbot> issue-644 -- Update ATK/AT-SPI mapping for
aria-hidden="true" to reflect elements should not be
exposed/mapped -- open
<trackbot> [22]https://www.w3.org/WAI/PF/Group/track/issues/644
[22] https://www.w3.org/WAI/PF/Group/track/issues/644
<davidb> one of the bugzilla bugs is
[23]https://bugzilla.mozilla.org/show_bug.cgi?id=972845
[23] https://bugzilla.mozilla.org/show_bug.cgi?id=972845
<clown> scribenick: clown
JS: the point is the reality is it's in the the spec and
authors are doing it (using aria-hidden to remove accessibles).
JD: It can be done in the browser, or it can be done by orca.
... It's more efficient for the browser to do it.
... I want FF to do it (the more efficient of the two ways).
... And, I want the table in the UAIG to be updated as well to
say, "do not expose".
DB: I want to talk to Alex again on this issue.
... He's easier than he was.
s/JS: the point is/JD: the point is/
DB: there are some implementation concerns that Alex brought
up.
... I think I can answer them.
JD: if you look at the AG, it makes it clear, that you have to
be careful if you use aria-hiddne.
... What's in the UAIG is much more neutral.
... I would like to see the cavets given in the Mac mapping.
... I think that language is closer match to whats in the APG.
... I don't think the AT should have to look at a child and
then check the parent chain to find if a parent is hidden.
JS: aria-hidden is supposed to inherit from its parent, so the
browser could propogate it to the children.
... So the AT does not need to look up the parent chain.
s/JS: the point is the reality/JD: the point is the reality/
s/JS: the point is the reality/JD: the point is the
reality/
<joanie> scribenick: joanie
1.1 ISSUES and ACTIONS: (All) Start with ISSUE-540,
[24]https://www.w3.org/WAI/PF/Group/track/products/23
[24] https://www.w3.org/WAI/PF/Group/track/products/23
JS: This is a review of all the remaining issues. :)
Group: Let's move on.
be done.
RRSAgent regrets Cynthia
Summary of Action Items
[End of minutes]
__________________________________________________________
Minutes formatted by David Booth's [25]scribe.perl version
1.138 ([26]CVS log)
$Date: 2014-02-18 20:54:20 $
__________________________________________________________
[25] http://dev.w3.org/cvsweb/~checkout~/2002/scribe/scribedoc.htm
[26] http://dev.w3.org/cvsweb/2002/scribe/
Received on Tuesday, 18 February 2014 21:00:06 UTC