W3C

– DRAFT –
ARIA and Assistive Technologies Community Group Weekly Teleconference

18 September 2025

Attendees

Present
ChrisCuellar, dean, hadi, IsaDC, james, Joe_Humbert, jugglinmike, Matt_King, mfairchild
Regrets
-
Chair
-
Scribe
jugglinmike

Meeting minutes

Review agenda and next meeting dates

Review agenda and next meeting dates

Matt_King: Requests for changes to agenda?

Matt_King: Hearing none, we'll stick with the agenda as planned

Matt_King: No CG meeting: Wednesday September 24

Matt_King: Next CG meeting: Thursday October 2

Matt_King: Next AT Driver Subgroup meeting: Monday October 13

Current status

Matt_King: We went from 17 plans in candidate review to 18 plans

Matt_King: Thanks to all the people who helped the latest plan advance, "Disclosure of Answers to Frequently Asked Questions"

Matt_King: We had previously placed "Disclosure Navigation Menu Example" on hold, but I believe that's ready to move forward, now

Matt_King: We could use someone to review w3c/aria-at#1299 . Perhaps howard-e because he previously worked on the builder that concerns the processing of that JSON

Matt_King: But we can re-assign to someone else at Bocoup if necessary

Running test plan for Disclosure Navigation Menu Example

Matt_King: For this one, NVDA and VoiceOver are complete, but we're looking for two testers for JAWS

Matt_King: I ran the bot last night, so the test plan should be ready to be assigned

Matt_King: Do we have any volunteers?

IsaDC: The JAWS bot is done, so I can assign the test plan to a human

Joe_Humbert: I can take that and begin work after I've completed my current testing

Matt_King: That would be perfect, thanks

IsaDC: I will initiate a second bot run now that I've assigned the existing one to Joe_Humbert

Matt_King: We'll need another tester for JAWS

Hadi: I can take that

IsaDC: I will assign the new test plan once the bot has finished collecting AT responses

Matt_King: Thank you Hadi; it's so good to hear your voice!

Running Switch test plan

Matt_King: Here, we have JAWS and NVDA complete

Matt_King: We're waiting on results from Elizabeth

IsaDC: She isn't present today, but I can e-mail her

Matt_King: That would be helpful; we haven't had her for a few meetings

Matt_King: If she is no longer available, we may need to ask someone else to pick it up

IsaDC: Though we have three testers assigned: Joe_Humbert, Elizabeth, and dean

dean: I just signed up for that this morning

IsaDC: But I need to assign the bot's results to you so that you don't have to enter the AT responses manually

Matt_King: Let's wait until we hear from Elizabeth, though, since we don't want to ask you to do work that isn't necessary

IsaDC: I'll reach out to Elizabeth and let you know if we'll need your help with this one, dean

Running test plan for Switch Example Using HTML Button

IsaDC: Why do we have two JAWS versions?

Matt_King: I thought I fixed that last night...

Matt_King: I only see one row for JAWS

Matt_King: The bot says that 8 tests are complete but that zero responses have been recorded

ChrisCuellar: That looks like a bug

ChrisCuellar: There is an open bug describing how you can currently enter a test plan for the same AT version multiple times. We have a fix that prevents that

ChrisCuellar: But this behavior that we're reviewing now seems to be a new bug

ChrisCuellar: We can file a bug for this. It seems like a high-priority bug to fix

Matt_King: Maybe we can wait on testing this.

Matt_King: Does anyone on this call have JAWS (besides Hadi and Joe_Humbert)?

IsaDC: It's a short test plan, so I can take it on

Matt_King: Both testers reported "no output"

dean: Joe_Humbert reported "negative side effects" and then specified that the negative side effect was that there was no output

dean: Is that repetitive?

Matt_King: I don't think we need to do that

Joe_Humbert: Okay, I can change mine

dean: It's the same conflict in Test 6

Joe_Humbert: I will update my results on Test 6 as well

dean: As a follow-up from when I was reviewing NVDA tests: NVDA will focus programmatically on the element (and change its status from on or off), but it does not visually change focus

Matt_King: That sounds like a JavaScript or CSS bug

Matt_King: If you're in browse mode, then the behavior you're describing is normal

james: Well, it's normal when you navigate. But when you click, then you should focus it

Matt_King: NVDA does, I thought, try to drag the focus along as its navigating

james: No, it does not, and the setting to enable that behavior has recently been removed

Matt_King: That's a huge deal!

Joe_Humbert: Regarding the conflict: I marked the command as untestable; that's why I had to specify a negative side effect

Joe_Humbert: So I should not mark this as "untestable", right?

Matt_King: Correct. Just "no output" (which leads to a simple failure)

Matt_King: I guess that's everything related to the Switch pattern

Running test plan for Tabs with Automatic Activation

Matt_King: Hadi finished this one this morning!

Matt_King: Last night, we had five conflicting JAWS results

Matt_King: Was this the one where you, IsaDC, were going to make some corrections to Louis' results?

IsaDC: Yes, and I made those corrections

Matt_King: This is the one where JAWS says "selected selected"

Hadi: This does not seem like "excessive" verbosity to me. It is just one extra word ("selected")

Matt_King: We're calling it "moderate"

Matt_King: We decided to only have two levels and to place "minor" issues in the same bucket

Matt_King: So it falls in this very minor-to-moderate bucket, Hadi

Matt_King: I think Vispero would want us to call it out

Hadi: Should I report that these had negative side effects?

Matt_King: Yes, if you could enable that and specify that there was redundant speech, and mark it as moderate, that would be good

Matt_King: Looks like the problem I had in mind got fixed, so I can mark the VoiceOver test plan as "final"

IsaDC: Yay! That always makes me excited

Running test plan for Tabs with Manual Activation

Matt_King: Let's see... We needed one more JAWS person, but I guess Hadi already signed up for disclosure, and Joe_Humbert's already done this one

Matt_King: So I guess we'll have to wait on that

Matt_King: As for NVDA, we don't have anyone, yet

dean: I'm available for work

Matt_King: I ran the bot run once, and then I assigned that to Joe_Humbert. Next, I ran it again (so I could assign the new bot run to someone else)

Matt_King: Is it going to end up making two different JAWS reports for the same JAWS version? Is it okay to leave it this way and assign the new JAWS bot run...?

ChrisCuellar: Let me ask howard-e to see

Matt_King: How have we done this in the past? Would it have worked differently if Joe_Humbert had already started the work?

ChrisCuellar: You can delete the report for the first run

ChrisCuellar: For the run that has Joe_Humbert as a tester, you can assign that to JAWS Bot

Matt_King: Alright, that's awesome

Matt_King: In the future, dean will be able to assign himself and then tell the bot to run it, and then when the bot is done, dean will be able to take it back

ChrisCuellar: That's right. I don't think dean will even have to assign it to himself, first

Matt_King: But he could assign it to himself and then change his mind later, right?

ChrisCuellar: yes

App Release

Matt_King: We currently have two tabs on the Test Queue page: the manual queue and the automated reports

Matt_King: There are currently 42 reports that can be automatically run

Matt_King: But because not all of those have historic data, we're going to see a lot of output which doesn't match the prior output, so we're going to need people to go through and see if the output is equivalent, and if so, set the assertions accordingly

Matt_King: This will be a lot of work

Matt_King: Everyone has plenty of work assigned to them right now, so we won't begin that effort, yet

Matt_King: But we will be addressing it in the near future

Matt_King: It's actually really exciting that we'll be able to collect data so efficiently. So big "hats off" to the Bocoup team!

CSUN and TPAC planning

Matt_King: I have a proposal for a CSUN Session. It's titled, "Who Cares if Screen Readers Say the Right Thiing?"

Matt_King: That question will be answered from five different points of view: users, web developers, standards authors, screen developers, and accessibility consultants

Matt_King: I now have confirmed who the presenters for each of those will be

Matt_King: However, speaking at the session will not be limited to those presenters

james: I don't believe PAC will send anyone to CSUN

Matt_King: I'm super excited about this presentation

Matt_King: Bocoup is also submitting a session

ChrisCuellar: We're still discussing our session proposal. It will likely include topics on the infrastructure we've built. It may include more speculative topics such as our AI research

Matt_King: As for TPAC (which is in November in Japan), there are three main things to discuss

Matt_King: One is a workshop for the ARIA working group to teach people how to write test cases and test plans (really atomic tests)

Matt_King: Another is about bringing ARIA-AT into the charter of a Working Group. That doesn't mean the Community Group is going to go away, but it will bring more formal W3C support to what we're doing. It could end up being a new working group that brings in ACT and possibly WPT work as well.

Matt_King: Daniel is taking point on this

Matt_King: It's going to be formed as a joint meeting between AGWG and ARIA

Matt_King: And then ChrisCuellar and jugglinmike will be present and leading a breakout session

Matt_King: We had awesome participation at the breakout we organized for last year's TPAC, so I'm super-excited about that

Minutes manually created (not a transcript), formatted by scribe.perl version 244 (Thu Feb 27 01:23:09 2025 UTC).

Diagnostics

Succeeded: s/disscuss/discuss/

All speakers: ChrisCuellar, dean, Hadi, IsaDC, james, Joe_Humbert, Matt_King

Active on IRC: jugglinmike