Accessible Rich Internet Applications Working Group Teleconference

10 May 2018


AmeliaBR, melanierichards, Joanmarie_Diggs, jamesn, aaronlev, jongund, Bryan


<joanie> https://lists.w3.org/Archives/Public/public-aria/2018May/

<joanie> https://lists.w3.org/Archives/Public/public-aria/2018May/0004.html

<jamesn> agenda order 3,1,2,4

<aaronlev> Can someone msg me the meeting password?

<jamesn> scribe: melanierichards

<jamesn> https://rawgit.com/w3c/aria/charter/charter/index.html

<jamesn> WAI-ARIA Graphics Module 1.1

<jamesn> Graphics Accessibility API Mappings 1.1

jamesn: the latest version of the charter doesn't have Graphics notes. We've taken out WAI-Graphics 1.1 and Graphics-AAM, following Amelia's feedback we don't plan on working on that much because there's not really anyone to work on them

AmeliaBR: we do have the existing Graphics roles level 1, is a very simple set of generic roles which were chosen specifically to describe default semantics of SVG. Unless we decide to change things considerably as well wrap up SVG mapping, those are the roles we need for role parity
... going beyond that, from an end-user perspective, there's benefit to having more complex roles to describe more complex graphics. Data charts, maps, flow charts, where there is lots of complex info conveyed by the visual layout, which isn't currently very well described in role structure. The problem is the complex relationships in that they aren't described in any accessibility role structures.
... Creating the more complex structures requires a lot of complex work on multiple implementation levels, currently we don't have that multidisciplinary involvement

jamesn: Removing it means we do not have to deliver it, but does not prevent us from doing it if we find interested people

AmeliaBR: there is going to be a little bit of follow-up on level 1 Graphics
... We have WAI-ARIA Graphics published on CR

joanie: What we're waiting on right now is Microsoft, we want to publish Graphics ARIA and Graphics AAM at the same time, missing mappings in Edge
... If we don't get it, we can exit CR, remove all automation stuff, and re-enter CR
... if we can get MS Edge mappings this month, we can tie things up
... ARIA WG still wants to collaborate with SVG WG, even if it's not in our charter

melanierichards: MS Edge should be able to deliver Graphics mappings in May still

jamesn: if we don't have resources, if someone wants to get a community group going, that might be a good place to get the work going, and it can be consumed later

aaronlev: I'm working on getting the graphics roles in Chrome right, was going to communicate with Windows SRs to at least recognize them. Are we talking beyond these 3 roles?

joanie: level 2 with charts, etc, things that don't exist yet

aaronlev: maybe we could split these into smaller pieces to get an interesting demo working to get people excited

jamesn: not sure we even have the people to do that right now

aaronlev: Google would get a lot of benefit out of this, would be interested in making this a long term project, despite not being the highest priority

jamesn: maybe we should put something in our roadmap so it can be post-charter but we don't lose all mention of this

AmeliaBR: there's also "other deliverables". maybe we should create a proper scoping document to show how much work needs to be done (re: graphics level 2)

joanie: maybe Amelia would be the most realistic creator of this document, if she has time to create it. We should put it/mention it in the roadmap and link to it

AmeliaBR: can't take on new commitments without corresponding funding offers, don't want to make promises I can't keep. A lot of the people who were doing the work are no longer here

joanie: how about small mentions in the roadmap based on Amelia's email about what needs to be done?

jamesn: I appreciate Aaron's idea of splitting things into smaller chunks of work so it seems achievable

<jamesn> https://www.w3.org/WAI/ARIA/roadmap

jamesn: if you can, please read the roadmap to see if there's anything we're missing out on
... Michael/Joanie, when are we planning on submitting the charter?

Michael: we should have it all ready to submit by mid to late next week

jamesn: so we have to do our CFC soon, we should get it by EOW

MichaelC: yes

Charter updates

Summary of last week’s F2F Meeting

jamesn: we went through the first round of role parity, which we'll discuss later. Closed a bunch of issues from old trackers
... did a bit of AccName, talked about CSS and how we can get CSS taken into account (or not) in the accessibility tree. I have some follow-up to do. Talked to the CSS task force facilitator, I agreed to get a meeting together to get that work done

joanie: Brian Garaventa will be an editor of the AccName spec!

jamesn/joanie: we ideally need a secondary editor for every spec

joanie: if anyone loves the AccName spec, [can get involved]

jamesn: we are planning on meeting at TPAC

joanie: please try to be present

<joanie> https://www.w3.org/2018/10/TPAC/Overview.html

jamesn; we try to do 2 F2Fs a year, one at TPAC and one at somewhere else

AccName – are we ready for CR

<joanie> https://joanmarie.github.io/test-results/accname/

joanie: I've been fixing bugs in webkit, I want to enter CR ready to exit CR. If our exit criteria is 2 passing results for every single thing being tested, I don't want to enter CR without that
... the only diff between the official results and what I pasted in was a couple items with listbox. my fix fixes 6 things that aren't committed into webkit yet

<joanie> https://joanmarie.github.io/test-results/accname/less-than-2.html

joanie: 4 things that are "less than 2", 2 of which are complete failures, 2 things that seem valid but no one's doing it

<joanie> https://lists.w3.org/Archives/Public/public-aria/2018May/0006.html

joanie: we have a test case that depends on what's in HTML-AAM. our spec does not dictate the naming convention of HTML elements. They actually haven't said anything about it, we have no passing results. The expectations we have may not be correct, anyway. Let's eliminate this test and depend on the smaller test
... I will probably generate results and skip that test so it doesn't show up in the implementation reports

<joanie> https://lists.w3.org/Archives/Public/public-aria/2018May/0009.html

[no objections raised on first one]

joanie: [2nd item] I've come to the conclusion that our expectations are bogus. We would never enter all the stuff from which the results are supposed to be coming from (from reading the algorithm)

jamesn: I think you're right. Because the child is hidden, you wouldn't go through it. I'm not sure that's what we meant to say. If you have something that's hidden, and you're referring to it, you would only be referring to that one thing and not the child contents. Not sure that's what we meant to say

jamesn; We are pointing directly to something that is hidden. Children would be hidden because parent is hidden

aaronlev: if you're pointing to a hidden thing, you want the hidden thing

joanie: and that's not what the spec says, if we want it to, we need to make that change

<joanie> https://lists.w3.org/Archives/Public/public-aria/2018May/0009.html

(technical discussion)

briangaraventa: I might submit a PR tomorrow with modified language

joanie: some language that basically says all the children participate in the calculation as if they were not hidden
... are we ready to transition to CR? [proposal]:if we cannot reach consensus on how to make this test what we expect it to be then, then we skip this test. No one is doing what the test says, and no one is doing the same thing
... we can't simply fix it to something else and now everyone's passing

aaronlev: I would take my time as far as trying to harden it into a CR. It sounds like there's some issues. I may be able to mention some issues no one's discussed yet

jamesn: get rid of this test in 1.1, log an issue, and address in 1.2?
... if people are working from editor's draft anyway, doesn't matter what version we're in

aaronlev: as an implementer, I probably want to use the latest thing in Github anyway

joanie: we are almost 2 years behind on getting AccName 1.1 out the door. It has to ship..
... we have to ship it, I'm raising things that will prevent us from exiting CR

aaronlev: the important thing about that doc are the use cases it creates and the consistency between those use cases. People should make sure the most common use cases work across browsers [not as concerned about edge cases]

bgaraventa: I don't see why there's any reason not go to CR

(logistics about sending out a CFC)

joanie: I'll send out a CFC

<jamesn> https://github.com/w3c/aria/wiki/Plans-regarding-role-parity

Role Parity Updates

jamesn: [gives summary of contents in URL]
... if there's anything we don't have consensus for where you feel we really need that, please speak up on that

[aaronlev to file a bug for a pattern that may not be represented by current ARIA roles]

Summary of Action Items

Summary of Resolutions

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.152 (CVS log)
$Date: 2018/05/10 18:12:14 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.152  of Date: 2017/02/06 11:04:15  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: Irssi_ISO8601_Log_Text_Format (score 1.00)

Present: AmeliaBR melanierichards Joanmarie_Diggs jamesn aaronlev jongund Bryan
Found Scribe: melanierichards
Inferring ScribeNick: melanierichards
Found Date: 10 May 2018
People with action items: 

WARNING: Input appears to use implicit continuation lines.
You may need the "-implicitContinuations" option.

WARNING: IRC log location not specified!  (You can ignore this 
warning if you do not want the generated minutes to contain 
a link to the original IRC log.)

[End of scribe.perl diagnostic output]