W3C

– DRAFT –
(MEETING TITLE)

30 October 2025

Attendees

Present
carmen, ChrisCuellar, dean, howard-e, IsaDC, james, Joe_Humbert, jugglinmike, Matt_King, mfairchild
Regrets
-
Chair
-
Scribe
jugglinmike

Meeting minutes

ARIA and Assistive Technologies Community Group Weekly Teleconference

Matt_King: Review agenda and next meeting dates

https://github.com/w3c/aria-at/wiki/October-30%2C-2025-Agenda

Matt_King: Requests for changes to agenda?

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

Matt_King: Next AT Driver Subgroup meeting: Monday November 17?

Matt_King: Next CG meeting: Wednesday November 5

Current status

Matt_King: several test plans have newer versions, but they didn't require any new testing. I corrected some errors in which HTML features they are testing

Matt_King: Upcoming will be a quantity spin button test plan. IsaDC is still working on that

Matt_King: There will be a couple more which will advance to "recommended" soon, and we're getting three new people from Apple engaged in the testing process

App release 1.22

Matt_King: If you visit the "reports" page in the app right now, you will see two big differences

Matt_King: The first is highlighting the amazing work that you'll all doing. If you've ever wondered how much work you've actually done, at the top of the reports page and the home page, there is a new banner that shows stats about how many assertion verdicts we've generated, how many commands we've tested, how many tests there are, how many people have contributed, and how much of all that has been done in the past 90 days

Matt_King: We have 63138 assertion verdicts. That's a lot!

Matt_King: 9,493 in the last 90 days

jugglinmike: It's over 9,000!!!

Matt_King: This is motivation for us to see progress. It's also showing value to the community and helping them understand that this is fresh

Matt_King: The other change: every verdict is associated with an aria-html feature. There there are two new tabls: ARIA features and HTML features. There, you can see the breakdown of this data according to the assertions

Matt_King: If you click on THOSE, you can get the raw data for that. And all the data is available in CSV format

Matt_King: Which is pretty cool for analysis

Matt_King: There are still a few mistakes in that data. I'm working on getting that corrected before we highlight this at W3C TPAC during the week of November 10

ChrisCuellar: I'm excited to have this out!

Matt_King: I'm over the top about this one. This is going to be very impactful at TPAC; I'm positive of that

Matt_King: here are other changes in the app; you can find a link to the full change set in the release notes referenced from the agenda

Matt_King: Thank you to the whole Bocoup team for all the hard work on this release

Issue 1590: Failure to detect conflicts after editing previously generated reports

github: w3c/aria-at-app#1590

Matt_King: I haven't seen any comments on this issue, yet. I wanted to check in because I added some test plans that are in the test queue right now. There are these hidden conflicts where I edited one of the runs but not the other, and so conflicts should be showing up in the test queue, but they're not there

Matt_King: I'm pretty sure I know where the conflicts are (because I know what changes I made), but I don't want to erase those until we know the root cause

howard-e: If you know where they are and you want to move forward, you can. They are indeed invisible conflicts. There's no worry about data loss.

howard-e: The update on what's happening here: it's a little trickier than we initially thought, but we're still hoping to have a pull request out for review by either today or early Monday

howard-e: When that change does land, there won't need to be any re-testing. It will be an in-place correction

Matt_King: Got it, thank you

Running test plan for Tabs with Automatic Activation

Matt_King: I asked IsaDC to de-prioritize conflict resolution, so that's up to Hadi or I

IsaDC: I heard back from Hadi--he hasn't been able to contribute recently, and it's not clear if he will be back soon. I've replied to learn if/when he thinks he can return

Running test plan for Tabs with Manual Activation

Matt_King: NVDA is done

dean: Joe_Humbert and I are still working out one last conflict on VoiceOver

Joe_Humbert: In that conflict, the AT moved to the wrong tab. Dean marked the assertion that it said "tab" as "pass", and I marked it as "fail"

Joe_Humbert: This is for "shift+j"

IsaDC: Should it be "untestable"?

Matt_King: Did "shift+j" move backwards or forwards?

Matt_King: We mark it as "untestable" when it goes to completely the wrong place

Matt_King: I've never heard VoiceOver say "first form element" before. Is that really what it's saying?

Joe_Humbert: I can double-check...

Matt_King: So it went to tab 2 of 4. I'm sure we didn't start on tab 3

Joe_Humbert: And yes, it says "first form element"

Matt_King: When you do the "run test setup", which tab does it focus on?

Joe_Humbert: "Carl Anderson"

Matt_King: Ah, so it doesn't work. In that case, I think you were right to mark this as "untestable"

Joe_Humbert: This is test number 10

Joe_Humbert: I'm assuming that this is probably the same thing for test 12

Matt_King: That's right

Matt_King: That is the weirdest VoiceOver output...

Joe_Humbert: On Test 13, it travels to some hidden tab panel

Matt_King: I observed this, too

dean: It's one of the tabs which is present but that is hidden, so it's not supposed to receive focus

Joe_Humbert: I tested that two times just now, and I'm seeing the behavior consistently

dean: Got it; I'll run the test again

james: How is the content hidden?

Joe_Humbert: It's a tab panel that is hidden with "display: none". And it's empty

Matt_King: Just use what Joe_Humbert wrote, dean

james: What does the setup script do?

IsaDC: It activates and focuses the fourth tab

dean: I'll duplicate whatever Joe_Humbert's got on test 13

Joe_Humbert: I was experiencing 502 errors quite frequently last night

dean: And I can't get it to load at all right now

james: If you carry on going, does it go to the other three?

Joe_Humbert: Yes

Matt_King: So that is VoiceOver conflicts

Matt_King: The Bocoup team is working on the 502 errors

carmen: We're working on a solution: we're increasing the RAM on the server and considering other options

Matt_King: Joe_Humbert and IsaDC have one conflict in the JAWS run

Joe_Humbert: It says "tab panel" but doesn't say anything about "tab list". Isn't a group of tab reported as "tab group"?

Matt_King: Yes

Joe_Humbert: It's test 13 for the "tab" keystroke

IsaDC: I think I made a mistake

Matt_King: This assertion is a "MAY" rather than a "SHOULD" or a "MUST"

Matt_King: I think Joe_Humbert got it right

IsaDC: Yes, I agree. I'm going to update my results

Matt_King: This is awesome. Both of the tab test plans are just down to catching up on a few conflicts whose resolution we all agree about

Matt_King: The next highest priority in the test queue (and where Joe_Humbert has already done a ton of work) is the switch with HTML checkbox

Matt_King: We're waiting on results from mmoss and dean

dean: That's next on my agenda

Matt_King: Fantastic. You're signed up for both VoiceOver and NVDA, dean. Mmoss is signed up for JAWS

IsaDC: mmoss is not present today

Matt_King: We won't bother talking about two-state checkbox today

Feedback on plan for tabs with manual activation

github: w3c/aria-at#1318

Matt_King: I think this is feedback on the bot. I wanted to verify with IsaDC before moving it to ARIA-AT App

IsaDC: I think I tagged Carmen, but I'm not sure about that

Matt_King: You did tag Carmen, yes

Matt_King: I don't see the bad output in the issue

IsaDC: It should be recorded in the issue because I didn't modify it before

IsaDC: it should be in the result

Matt_King: the output isn't in the issue

Updating reports to latest screen reader versions

Matt_King: Joe_Humbert, if you have time to help with VoiceOver: if you go to the tab labeled "automated report updates" in the test queue...

Matt_King: Oh, my mistake. Not that tab

Matt_King: On the other tab, use the filter named "automated updates (21)"

Matt_King: I'm looking for VoiceOver ones that I want to prioritize

Joe_Humbert: You can send me a list asycnhronously

Joe_Humbert: I didn't want to update Safari to the new version 26, but I think it will be difficult to go back. I'm guessing the automation system is not using Safari 26, yet

IsaDC: we always recommend running the latest version

ChrisCuellar: I think updating is okay.

james: If the macOS version is 15, but the browser version is 26, that in itself could cause some differences

Joe_Humbert: Agreed!

james: But I also think that there could be a lot more differences when the Safari version matches the macOS version

Matt_King: I don't think there's any advantage to updating to Safari 26. If you stay where you are, you'll have the screen reader version that matches the bot right now

Joe_Humbert: Okay!

Joe_Humbert: If the output is wrong, do I correct the output, or do I file a bug?

Matt_King: Well, actually, it would be good to file a bug

Joe_Humbert: Okay, I'll go through and do all the VoiceOver ones and check the output compared to what my machine outputs, and if there are any discrepancies, I will file an issue

Request for change to alert test plan

Matt_King: We can't do this issue justice in the three minutes remaining in this meeting. I'll place it at the top of the agenda for next week's meeting (which will be the final meeting before TPAC)

Minutes manually created (not a transcript), formatted by scribe.perl version 248 (Mon Oct 27 20:04:16 2025 UTC).

Diagnostics

Succeeded: s/focuses/activates and focuses/

Succeeded: s/>/?/

Succeeded: s/2021/(21)/

All speakers: carmen, ChrisCuellar, dean, howard-e, IsaDC, james, Joe_Humbert, jugglinmike, Matt_King

Active on IRC: howard-e, jugglinmike