W3C

– DRAFT –
ARIA and Assistive Technologies Community Group Weekly Teleconference

12 February 2025

Attendees

Present
howard-e, IsaDC, James_Scholes, jugglinmike, Matt_King, Michael_Fairchild, mmoss, stalgia
Regrets
-
Chair
-
Scribe
jugglinmike

Meeting minutes

Review agenda and next meeting dates

https://github.com/w3c/aria-at/wiki/February-12%2C-2025-Agenda

Matt_King: Requests for changes to agenda?

IsaDC: I'd like to discuss issue #1195 for the "disclosure" test plan

Matt_King: Sure

Matt_King: Next AT Driver Subgroup meeting: Monday, March 10

Matt_King: Next Community Group Meeting: Thursday February 20

Current status

Matt_King: Same 11 plans. We have some issues that we're working on resolving to move some of those forward

Matt_King: The draft review and refactoring--that's current as of today. Four plans in "Draft review" (five in the test queue, but the "disclosure" plan is just a re-run)

Matt_King: And two "slider" plans which are very close

Matt_King: In terms of testing schedule for the next few months, IsaDC and I are still working on that. We hope to be able to talk about it next week

Testing status

Matt_King: We can review the test queue in order

Matt_King: Starting with "disclosure navigation menu" example re-run

w3c/aria-at#1195

mmoss: This one stumped me a bit. I wonder if it's due to differences in macOS...

Matt_King: Is the issue primarily with the output?

mmoss: There was specifically the test where you navigate from an expanded disclosure button to the page link using "tab"

mmoss: Similarly, when you use "shift+tab" and move from the collapsed disclosure button to a link in the dropdown

mmoss: Tab and "shift+tab" did not seem to actually function for me. Which is bizarre

Matt_King: Yeah, that's what your AT responses look like to me

Matt_King: That almost seems slightly unbelievable

mmoss: Right. I've re-run the test several times because I thought I was missing something

mmoss: It's not that it does nothing. Instead of the screen reader's cursor moving into the menu, it moves from the button you're currently on to the next button

Matt_King: Ah, so, skipping over all the links (because the links appear in between)

mmoss: Right

mmoss: This is test 9, specifically after "tab"

Matt_King: I'm encountering a different behavior. This doesn't feel like a VoiceOver problem, though

IsaDC: I ran it, and it goes to the link after pressing "tab"

Matt_King: Hm, this time, it worked! It went to "overview"

Matt_King: Now, I am not running with default VoiceOver settings at this moment.

Matt_King: I just re-ran it, and this time, it went to "administration"!

mmoss: That's what I've been observing

Matt_King: Oh, I'm in Chrome! I'm not supposed to be in Chrome!

mmoss: Yeah, I've been testing Safari

Matt_King: Well, that doesn't bode well for Chrome, but that's a separate concern.

IsaDC: I've run it twice, now, and it goes to the correct place

IsaDC: I'm using Safari version 18.2

mmoss: I'm using Safari 18.3

Matt_King: First time in Safari, it worked for me--it went to the "current page" link

Matt_King: And two times in a row, it demonstrated the expected behavior

IsaDC: I'm installing an update, now

IsaDC: To see if I can reproduce the behavior that mmos reported

mmoss: I checked Joe's results. Joe was running with Safari version 14

IsaDC: Really? That shouldn't be accepted by the app because I specified the version...

Matt_King: If we require a minimum version, and then you put in a version that's older than that, you shouldn't even be allowed to run the test

IsaDC: I don't know what's going wrong

Matt_King: There might be an issue with the app, then, howard-e

IsaDC: This is the "disclosure navigation" test plan

IsaDC: I set the minimum version to 15

IsaDC: And Joe was able to run with version 14

howard-e: I'll make a note and check out what happened there

mmoss: There was one more conflict

mmoss: Test 14, "activate the link"

mmoss: When I do it, it takes a beat, and then I hear the activation click, and then it says "you are currently on a region"

mmoss: Joe reported the sound

IsaDC: I don't think that's correct for us to record because it is a sound; it's not output

Michael_Fairchild: we don't record "Earcons"?

mmoss: No; I think Joe just typed that in

Michael_Fairchild: I'm curious why we don't consider earcons

Matt_King: We've never encountered a situation where that would be considered...

James_Scholes: I think we have (like the toggle button case), but we don't have anything in the app

James_Scholes: There's nothing in the app which allows us to describe an earcon

Matt_King: I could imagine us doing that, but when you're running a screen reader with its default state and default verbosity... At least so far, we haven't encountered any situation where an appropriate default response would be to only make a sound

Michael_Fairchild: I think there might be some scenarios where we do encounter that as the expected output, but maybe we just haven't reached those scenarios, yet

Matt_King: The way to get rid of these conflicts--it seems on the surface like they may all be related to version mismatches. Getting everyone on the same version is probably going to be our first step toward getting these resolved

Matt_King: So we have "radio group using active descendant." A brand new version

Matt_King: We have some work that's already been done--let's see how much of that has been preserved

IsaDC: Some of that has very old data

mmoss: I'm happy to start over on that one

Matt_King: Right now, for JAWS, we have Hadi, and it shows 8 of 15 tests complete. Are you concerned about the 8 tests completed by Hadi which might need to be re-run

IsaDC: Yes. My concern is that on the NVDA table, we recently re-ran them and have very difference results

Matt_King: It looks like your run of JAWS also has 8 of 15 complete

IsaDC: But I had zero before--I hadn't run it. Where are those eight coming from? From a very old version?

Matt_King: Hmm, okay. It's possible that you need to validate those results, then

Matt_King: We could delete Alyssa's run, here. That means we would need to get another person for NVDA

Matt_King: For JAWS "radio group active descendant", we have IsaDC and Hadi

Matt_King: Should we keep those assignment as-is?

IsaDC: I'm happy with that. We would just have to double-check that the results still match

Matt_King: We could ask Hadi to just run the seven and check for conflicts

IsaDC: I can e-mail Hadi

mmoss: I can run the NVDA test plan

IsaDC: We can use the bot for that, first, and then assign you

mmoss: Thanks!

IsaDC: I'll delete Alyssa, run the bot, and add mmoss and myself

Matt_King: Then, for VoiceOver, we have IsaDC, Joe, and mmoss

IsaDC: But that data is incorrect

Matt_King: mmoss can you do both VoiceOver and NVDA?

mmoss: Sure thing. I'll wipe my previous results and run through the test, again

IsaDC: Joe had results, but the results he got imported into the test plan are not the same

Matt_King: I think we want Joe to get his Mac updated before we ask him to run more tests

Matt_King: So, IsaDC, we might need you to run the VoiceOver one unless you want to write Joe and ask him to get his mac updated

Matt_King: Dean said he is available in the note where he said he couldn't attend today

Matt_King: I'm wondering if--we could have Dean and mmoss do VoiceOver for this

Matt_King: But I think we should focus on getting one of them done. Let's do the "active descendant", first, and get them all done. Then, if we have more bandwidth, we'll go to the next "radio" plan

Matt_King: I guess we're just a little short on testers at this moment

App issue 1162 - Generating updated reports for new AT releases

github: w3c/aria-at-app#1162

stalgia: We had some time to internally align on this

stalgia: With the addition of a new panel and the need for that panel to be public-facing, we agreed that the most appropriate place was the "data management" page

stalgia: I'm excited to do some rethinking about how to reduce complexity on some pages, but that doesn't need to happen at this moment

Matt_King: Are you thinking an "automated run history" or something like that as a separate table?

stalgia: Pretty much. I think you wanted an interactive UI for initiating runs to be above the table, but I think we can put it below the table

Matt_King: As long as we have something that allows us to see that administrative action is required

stalgia: Great, that makes sense

Matt_King: Will you update 1162 with a revision of that top comment? A sort of description of what the UI will be?

Matt_King: I think we can stub it out there and then we can review in the actual UI

stalgia: that makes sense. We can do that

Matt_King: This is going to be a really big deal! Especially when we have the latest versions of all the screen readers so we can re-run all the test plans. So, thank you!

Minutes manually created (not a transcript), formatted by scribe.perl version 242 (Fri Dec 20 18:32:17 2024 UTC).

Diagnostics

Succeeded: s/plan/page/

All speakers: howard-e, IsaDC, James_Scholes, Matt_King, Michael_Fairchild, mmoss, stalgia

Active on IRC: jugglinmike