W3C

– DRAFT –
ARIA and Assistive Technologies Community Group Weekly Teleconference

02 October 2025

Attendees

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

Meeting minutes

Review agenda and next meeting dates

https://github.com/w3c/aria-at/wiki/October-2%2C-2025-Agenda

Matt_King: Requests for changes to agenda?

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

Matt_King: Next CG meeting: Wednesday October 8

Matt_King: Next AT Driver Subgroup meeting: Monday October 13

Current status

Matt_King: Since we last met, we've moved two test plans into candidate review

Matt_King: that's pretty awesome! We're up to 19 plans in candidate review and 4 plans in draft review

IsaDC: I'm working on a test plan for check box. It should be ready for review by Monday at the lastest

Matt_King: By Wednesday, we'll have two more new test plans in the test queue

Matt_King: Hopefully, we'll make some progress today and get some test plans out of the test queue

App updates to make bot more available

ChrisCuellar: Just this morning, we released some big updates to how you manage bots

ChrisCuellar: Now, testers can start bot runs

ChrisCuellar: If you look at any test plan run, you should see a "Start bot run" button in the "actions" column

ChrisCuellar: Right now, each test plan run can only have one bot run associated with it. That's more of a UI limitation

ChrisCuellar: The idea is that testers will generally start off with a bot run

ChrisCuellar: We want to improve the workflow based on Matt_King's feedback. But for now, when you start a bot run, you will see the typical status information on the run. Also, the "start bot run" button will change to a "manage bot run" button. When the bot is done, you can use that new button to re-assign the test plan run to yourself

Matt_King: Right now, there is still an "assign yourself" button. If you use that, then your name will be on a test plan run, and you won't be able to assign a bot run to yourself

Matt_King: For now, in this current iteration, the "assign test plan to yourself" button isn't particularly useful unless you don't want to use the bot

dean: I tried to use it this morning, and I am now signed up twice for it, somehow. I don't understand what's going on with that

IsaDC: That shouldn't be possible. Is that a bug?

ChrisCuellar: Yeah, I'm seeing that, now--for VoiceOver. I think this is a bug

Matt_King: Is the "unassign yourself" button, there?

dean: Yes

dean: I'll pause my work there, for now

dean: I just un-assigned myself from one of them, and they're now both gone

ChrisCuellar: You should be able to use the workflow Matt_King described, now

Joe_Humbert: The state looks corrupted, now, though.

dean: Right now, I'm not assigned, but I do have a "manage VoiceOver bot run" button

Matt_King: When that run is done, you will find an option to assign yourself in the modal dialog which is opened by that button

dean: Okay, I will assign myself when it's done

Running Switch test plan

IsaDC: I e-mailed Elizabeth. She hasn't been well recently, but she is going to run it as soon as she can

Matt_King: Great!

Matt_King: The regular "switch" test plan (the one to which Elizabeth is assigned) has no open issues

Issue 1298: Feedback on test of Switch using HTML button

github: w3c/aria-at#1298

Matt_King: I think we can close this

Matt_King: dean raised this issue

Matt_King: he is reporting that there is no VoiceOver output in a specific circumstances

Matt_King: Right, I think we were aligned here. There was no output, and we reported that in the appropriate Test Plan

dean: That's right

Matt_King: Okay, then I will close this

Running test plan for Tabs with Automatic Activation

IsaDC: Hadi needs to update his results on this one

Matt_King: right. We talked about this last time, during the meeting of September 18

Matt_King: I'm pretty sure we were aligned on Hadi making his report consistent with Louis's

Matt_King: But Hadi isn't here today, so I guess we can move on

Running test plan for Tabs with Manual Activation

Matt_King: We're just waiting on IsaDC for JAWS

IsaDC: We need one more tester for NVDA

Matt_King: Since Dean signed himself up for VoiceOver, we're covered there

Joe_Humbert: I raised issue 1300 before I finished testing, so maybe its not a bug

w3c/aria-at#1300

Joe_Humbert: I can view my own results. It isn't allowing me to see my results. I can see results from others

Joe_Humbert: I was at the tab panel, then it took me to the tab group, but the assertion expects that you arrive at the final tab

Joe_Humbert: So everything would pass if you did the key press one more time

Dean: the bot finished, and I just re-assigned that test plan run to myself

Dean: I went to start testing, and it gave me some output. It looks like it works

Joe_Humbert: But you didn't receive any feedback that you successfully assigned it to yourself

ChrisCuellar: That missing feedback may be a bug. There may have also been a lag

As for issue 1300

Matt_King: For JAWS we have three up arrows, and for NVDA, there is only one. Why is there only one?

Matt_King: I'm wondering about Test 4

Matt_King: I think that should also be three...

Joe_Humbert: Test 4 wouldn't matter because that's the one where it goes into tab panels

Matt_King: There's no way to get VoiceOver to do the right thing. I'm just saying that we should have the correct number regardless of VoiceOver's behavior

Joe_Humbert: How do we know what the correct number is if VoiceOver isn't behaving correctly?

Matt_King: Well, if you load that example, then by default, the first tab is selected

Matt_King: If you just tab to the link in the first tab and then use "ctrl+opt+left arrow", then the bug isn't present

Matt_King: But as soon as you activate any tab, then the bug appears

Matt_King: And it took three presses in the APG example

Joe_Humbert: So that's two tests that I have to re-run and two tests that Dean may have to change

Matt_King: Yes, to record the appropriate (but incorrect) output

Joe_Humbert: IsaDC is going to change something, and that will remove some responses from your test plan run

Matt_King: You can skip test 2 and test 4 for now, and come back and finish them once IsaDC has made the changes

Matt_King: Thank you for catching that, Joe_Humbert! Given how VoiceOver is behaving, it's kind of amazing that you recognized the problem

Matt_King: This is exactly why we want the test plans to pass through "draft review" because IsaDC, James, and I can't catch everything on our own. This is the process working as intended

Updating reports to latest screen reader versions

Matt_King: So we now have this awesome capability: an "automated reports" tab in the test queue

Matt_King: We can instruct the bot to update any report that hasn't been run with the latest version of a given AT

Matt_King: There are 18 reports that we've completed with the September release of JAWS. I was noticing here that on the "automated reports update" tab, it seems like the only bot that is 100% up-to-date with the latest release version is JAWS

Matt_King: We might start with that. I don't want to run them all because that's going to create a lot of manual work all at once

Matt_King: I know there is a ".3" release of NVDA out. Do we have that on the list of bot updates to do?

Matt_King: There's also a VoiceOver update in macOS 15.6

ChrisCuellar: I think we just need to carve out some time for those. We can try to get that in as soon as possible

Matt_King: I would say VoiceOver is a higher priority than NVDA right now because that one is further behind and because I'm going to have more conversations with James Craig

ChrisCuellar: VoiceOver is now on 15.6.1, I think.

Matt_King: But it lists "version 15" on the UI

ChrisCuellar: This is a UI problem. We want to more clearly label what version is actually available. This is part of the "minimal versus exact" situation which has been confusing everyone for a long time

ChrisCuellar: It's a weird quirk of the system that causes it to be rendered in that way

ChrisCuellar: It's a problem that's specific to VoiceOver. We are running 15.6.1

Joe_Humbert: Every Apple OS is jumping to version 26

Matt_King: Okay. That's... interesting

Matt_King: So in that case, maybe we should run the VoiceOver bot

ChrisCuellar: That will use 15.6.1

Matt_King: Okay. That's kind of a nicer place to start because there are only seven reports which haven't been run. That seems like such a small number

IsaDC: Is it taking into account version 1 of the test plans?

ChrisCuellar: Should it be all of the tests?

Matt_King: Right now, the tab shows the reports that are out-of-date. It's only showing seven reports that are out-of-date and not on 15.6.1

ChrisCuellar: What should we be seeing?

Matt_King: Maybe we do have a lot of reports that have been run with 15.6.1 already

ChrisCuellar: Right now, we have something higher than 15 in the system. The list of reports that it wants to generate new reports for, they are all lower than 15

ChrisCuellar: Do you want to re-run the ones that were in the lower versions of 15

Matt_King: Yes

Matt_King: But if we run these seven, it will generate 7 new reports. That's not a huge volume, so it will be a good "soft introduction" for the team

Matt_King: And when the version thing is fixed, it will probably show another 10 or 11 that are below 15.6.1, but above or at 15.0

ChrisCuellar: We might have to do some kind of small fix to make that comparison possible. I'll need to check with howard-e

ChrisCuellar: I think it might require a little bit of extra back-end work to achieve that. Just let us know how to prioritize it.

Matt_King: If we won't ever have that problem with JAWS...

ChrisCuellar: Yeah, I think this problem is unique to VoiceOver, but I don't fully understand why. That's why I want to speak with howard-e

Matt_King: Next week, as part of the agenda, we'll be walking people through what we want them to do in this workflow. It's similar to "draft review", but with a little less work. Still, understanding the context is important here

CSUN and TPAC planning

Matt_King: There will be at least one CSUN proposal. titled "Who cares if your screen reader says the right thing?"

ChrisCuellar: Bocoup also submitted a proposal

Matt_King: Great. We'll discuss conferences more next time

Minutes manually created (not a transcript), formatted by scribe.perl version 246 (Wed Oct 1 15:02:24 2025 UTC).

Diagnostics

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

Active on IRC: Joe_Humbert, jugglinmike