Minutes: December 1, 2014 ARIA Caucus

Minutes: December 1, 2014 ARIA Caucus

Web:

http://www.w3.org/2014/12/01-aria-minutes.html

Text:

scribe: jcraig
Topic: new meeting time.
RS: Thursday possibility
LW: Not ideal for me.
JC: Friday?
+Susann_Keohane
+Stefan_Schnabel
CS: Friday's problematic for Lisa (and EU callers)
RS: Need 1.5 hours for action items
RS: If people get their action items done.
JS: Propose move to 1.0 hour call on temporary basis
+James_Nurthen
RS: Need to move to ARIA 2.0, too. Could work in parallel.
zakim, who is on the call?
On the phone I see janina, LJWatson, Fred_Esch, Rich_Schwerdtfeger,
Joanmarie_Diggs, Michael_Cooper, Joseph_Scheuhammer, Cynthia_Shelly,
Jon_Gunderson, jcraig, Susann_Keohane,
RS: Supposed to have heartbeat by F2F.
MC: Heartbeat scheduled for next week.
http://rawgit.com/w3c/aria/master/core-aam/core-aam.html#substantive-changes-since-the-last-public-working-draft
MC: AccName going to FPWD, too.
Spec changelog http://rawgit.com/w3c/aria/master/aria/aria.html#changelog
+Bryan_Garaventa
JC: change to 1 hour would help attendance.
RS and JS: I don't think that will make a difference
+Matt_King
sees MichaelC on the speaker queue
RS: Thursday at 12 Eastern (less meaty stuff first)
JC: 9AM Pacific
JN: So the real meeting starts 30 minutes in?
JS: Next meeting times (skipping Xmas and NYD)
JC: December 11th (next week) will be first Thursday time.
MC: use code #ARIA (Zakim code #2742)
JS: PF name changing too
JS: it's been a year
TOPIC: Publishing next week
MC: ARIA heartbeat, Core-AAM heartbeat, and AccName FPWD
JC: December 11th publish ETA
JC: starting pub checks this week
TOPIC: Action Items
https://www.w3.org/WAI/PF/Group/track/actions/1440
action-1440?
is looking up action-1440.
action-1440 -- Joanmarie Diggs to landmarks section uses "region of page"
in prose even though "region" is not a landmark -- due 2014-10-06 -- OPEN
https://www.w3.org/WAI/PF/Group/track/actions/1440
JS:  http://rawgit.com/w3c/aria/master/aria/aria.html#region
sees Bryan_Garaventa on the speaker queue
sees Bryan_Garaventa, mattking on the speaker queue
sees Bryan_Garaventa, mattking, Cynthia_Shelly on the speaker queue
RS: windows SRs treat region as landmark
JC: but they shouldn't
BG: some value in having region as landmarks
sees Bryan_Garaventa, mattking, Cynthia_Shelly on the speaker queue
sees mattking, Cynthia_Shelly on the speaker queue
sees mattking, Cynthia_Shelly on the speaker queue
JS: q+ that the HTML5 <section> element is "mapped" to role region.
sees mattking, Cynthia_Shelly, Joseph_Scheuhammer on the speaker queue
RS: section is being used as landmark, too
sees Cynthia_Shelly, Joseph_Scheuhammer on the speaker queue
Joseph_Scheuhammer: you wanted to note that the HTML5 <section> element is
"mapped" to role region.
sees Cynthia_Shelly on the speaker queue
MK: question of semantics (?) just because SRs provide navigation to
regions doesn't make it a landmark
sees Cynthia_Shelly on the speaker queue
MK: It's fine for SRs to do this.
JC: I agree
MK: JAWS provides special landmark behaviors for Main. I don't see need to
call regions "landmarks" in order to achieve this navigation behavior
RS: Devs are confused by "landmark" terminolgy in AT
MK: this is why JAWS calls it a "region list" not a "landmark"
JS:  http://rawgit.com/w3c/aria/master/html-aam/html-aam.html#el-section
CS: Does HTML section element count as a region from the AT perspective
MK: Jaws counts unlabeled role=region as region, but not unlabeled
<section>
CS: Purposeful or not?
JC: Seems like a logical heuristic
MK: If we required "region" to be lableled in order to be treated as
navigational item, I think that would be an improvement
JS: http://rawgit.com/w3c/aria/master/aria/aria.html#region
sees Cynthia_Shelly, jcraig on the speaker queue
sees jcraig on the speaker queue
CS: I think that's a good change for both region and <section>
JS:
http://www.w3.org/html/wg/drafts/html/master/sections.html#the-section-element
action-1440 -- Joanmarie Diggs to landmarks section uses "region of page"
in prose even though "region" is not a landmark -- due 2014-10-06 -- OPEN
https://www.w3.org/WAI/PF/Group/track/actions/1440
RS: Is using "region" as prose a problem here?
jc: you wanted to ask if Matt wants label required on region in spec, or is
this an Authoring Guide note?
JS: current text is:  "Authors SHOULD ensure that a region has a heading
referenced by aria-labelledby. This heading is provided by an instance of
the standard host language heading element or an instance of an element
with role heading that contains the heading text."
CS: Don't use role names as prose text
JC: "region of page" should be "portion of page"
CS: Should be a UA must, not author MUST
FE: Is <nav> mapped to region? (no, navigation)
JC: Authors won't know to do something different
MK: UA MUST map regions if it's labeled, and MUST NOT map "region" (not
including subroles) if it is not labeled
RS: also worried about live regions
MK: live region is not a "region"
RS: but this req would mean <div role=region aria-live=assertive> would not
be mapped
MK: Okay. So the UA req would be to ignore the region role if not labeled.
sees jongund on the speaker queue
JC: but not a req on subclass roles.
MK: Would require change to prose and mapping table
sees no one on the speaker queue
JC: q+ to ask what the mappings are for role=region with and without a
label
sees Joseph_Scheuhammer on the speaker queue
JG: Would this affect <section> elements or other region-like elements?
LW: HTML5.0 <section>
http://www.w3.org/html/wg/drafts/html/master/sections.html#the-section-element
JC: what if I pout labelledby on role=article?
RS: should have a label on articles
MK: suggest leaving article out of this discussion.
MK: Proposed text: "A region is not a navigational landmark. Content
authors SHOULD use standard document landmark roles for regions which are
meant to be
MK: treated as landmarks."
JC: Current text: "When defining regions of a web page, authors are advised
to consider using standard document landmark roles. If the definitions of
these
JC: regions are inadequate, authors can use the region role and provide the
appropriate accessible name."
MK: I think the current text is better
JS: http://rawgit.com/w3c/aria/master/aria/aria.html#region
JC: I think this behavior would be fine if "landmark" were made an
non-abstract role
JC: legacy could be marked up as role="landmark region"
JN: q to ask what is the point in region if we are not expectig AT to be
able to navigate to them?
MK: still think labeled regions should be treated as regions
JC: I agree
JD: To propose we delete the current text (last paragraph) under Region
sees Joseph_Scheuhammer, Joanmarie_Diggs on the speaker queue
JN: q+ to ask what is the point in region if we are not expectig AT to be
able to navigate to them?
sees Joseph_Scheuhammer, Joanmarie_Diggs, jamesn on the speaker queue
MK: In addition to allowing region to be labeled, but we should also allow
landmark as concrete role
JC:  <section> w/o label should be ignored
JS:
http://www.w3.org/html/wg/drafts/html/master/sections.html#the-section-element
JC: so unlabeled region should also have role ignored
RS: people are using labeled region to get it to show up in JAWS landmark
list
sees Joseph_Scheuhammer, Joanmarie_Diggs, jamesn, jongund on the speaker
queue
LW: What's the problem with using region as-is?
JC: Making region a landmark would also make grids and lists landmarks
RS:
http://www.w3.org/html/wg/drafts/html/master/sections.html#the-section-element
JC: pull this text out of region and add it to concrete landmark role:
"When defining landmarks of a web page, authors are advised to consider
using standard document landmark roles. If the definitions of these regions
are inadequate, authors can use the [landmark] role and provide the
appropriate accessible name.
JC: s/aking region a landmark would also make grids and lists
landmarks/aking region a landmark would also make grids and lists
landmarks, so it's confusing from an authoring perspective and a taxonomy
perspective./
RS:
http://www.w3.org/html/wg/drafts/html/master/sections.html#the-section-element
JS http://rawgit.com/w3c/html-api-map/master/index.html#el-section
RS: There is no requirement to NOT map seciton in hte HTML mapping guide
RS: region role (default - do not set), alert, alertdialog, application,
contentinfo, dialog, document, log, main, marquee, presentation, search or
status.
JC: So this is purely heuristic on the part of the Windows SRs
JC: s/the Windows SRs/JAWS and NVDA/
JS: http://rawgit.com/w3c/aria/master/html-aam/html-aam.html#el-section
RS: Proposal: Change the landmark role to not be abstract.
+1
+1
+1
JC: s/not be abstract./be concrete./
JC: s/be concrete./be concrete, not abstract./
MK: +1
RS: +1
MK: if it has a label
JG: What If I use an unlabeled landmark?
JN: we should make an accessible name a MUST for a landmark
action-1440: Also make landmark concrete (not abstract) and add this to
changelog
MK: could map <section> to region or none if not labeled, and map to
"landmark" role if labeled
1. we are changing the landmark role to be concrete
2. we are not requiring the user agent to do anything special if landmark
does not have a label
3. we are saying that if a region does not have a label it will be exposed
as a region
4. if a region does have a label it is exposed as a landmark
JD: thought #4 was solved by #1
JS: <div role="region" aria-label="it's matt's fault">  exposed as
landrmark?
JC: I disagree with that, but fine if AT want to add labeled regions to a
navigation list
MK: #3 will break some ATs (original problem JAWS was trying to solve)
JS: 4. if a region does have a label it is treated as a landmark but is not
exposed as a landmark
JS: 4. if a region does have a label it is treated as a landmark by AT, but
is not exposed as a landmark
MK: Does anyone have a problem with ARIA spec implying "if a region does
have a label it is [treated, not exposed] as a landmark"
4. if a region does have a label it is treated as a landmark by AT, but is
not exposed as a landmark by UA's
LJWatson has to drop off now.
RS:
1. landmark role to be concrete
2. Recommend in role region that for regions intended to be landmarks that
authors use a role of landmark with label
Zakim: leaving.
As of this point the attendees were janina, Fred_Esch, Rich_Schwerdtfeger,
LJWatson, Joanmarie_Diggs, Michael_Cooper, Joseph_Scheuhammer,
Cynthia_Shelly, Jon_Gunderson,
... jcraig, Susann_Keohane, Stefan_Schnabel, James_Nurthen,
Bryan_Garaventa, Matt_King
RRSAgent, make minutes
I have made the request to generate
http://www.w3.org/2014/12/01-aria-minutes.html richardschwerdtfeger
                                     [
                                  1:37pm
                                     ]
           fesch left the chat room. ("ChatZilla 0.9.91 [Firefox
                         31.2.0/20141011074935]")


Rich Schwerdtfeger