Meeting minutes
Review agenda and next meeting dates
Matt_King: Next AT Driver Subgroup meeting: Monday October 7
Next Community Group Meeting: Wednesday October 9
Review agenda and next meeting dates
Matt_King: Next AT Driver Subgroup meeting: Monday October 7
Matt_King: Next Community Group Meeting: Wednesday October 9
Matt_King: Requests for changes to agenda?
IsaDC: Can we add GitHub issue #1226?
Matt_King: Sure
Current status
Matt_King: We're a little bit behind our goal--we're looking for 20 plans in candidate review or beyond
Matt_King: I still think it's potentially realistic, though we may have to speed up our work a bit
Matt_King: Right now, we have 7 plans in candidate review, and 5 of them are approved by Vispero
Matt_King: We had a great meeting with Apple--we'll talk about that more later in this meeting--but we recognized some bugs that are present and arranged for additional meetings
Matt_King: I think 10 recommended plans may be a bit of a stretch, but we'll get close at least
Matt_King: We have 3 in draft review
Matt_King: And a couple more plans waiting in the wings. I need to review those
Matt_King: I was really happy for TPAC; I think it's going to give us a boost toward these goals
Matt_King: Any questions or comments?
.
Joe_Humbert: Ho might we increase our speed?
Matt_King: Radio group plan should go faster than normal because there's a low likelihood that we'll need to make changes (due to its similarity to another plan which we've already advanced)
Matt_King: If the bots are working reliably, that should help speed things up as well. They will allow people to focus on the content of the plans rather than the task of collecting the output
Matt_King: Do you have any other ideas?
Joe_Humbert: Seeking more participation, but I know that's been a challenge for us so far
Matt_King: Agreed, that would be a great thing
Bot Run Assignment Failed
github: w3c/
IsaDC: The original issue was about the bots not getting assigned to the tests--any of the bots, neither VoiceOver Bot nor NVDA Bot
IsaDC: It worked in the sandbox server, and the VoiceOver Bot improved a lot--that's great!
IsaDC: Today, I tried again, and the bots received the assignments, but the test run is cancelled
IsaDC: They do not report a recording
howard-e: I submitted a patch for this yesterday, but it hasn't been merged, yet
howard-e: There was a regression which was causing the bot to attempt to run the test with the earliest AT version possible
howard-e: That was an AT version that the automation driver doesn't have
howard-e: The patch should address this
howard-e: A quick fix (which I do not recommend) would have been to create a test plan specifically for automation
IsaDC: I tried to do that, but it didn't work
howard-e: I think it's NVDA 2023.3 and VoiceOver version 14
IsaDC: I believe that's what I tried
howard-e: I will have to verify, then
howard-e: The patch is available on staging. I'll confirm again there, then send a message, and finally push to production
Matt_King: Is that regression something that you think we need a test for?
howard-e: There is a test on its way
howard-e: There was another critical issue due to this same regression. It was reported by Dean on the mailing list yesterday. There's also a patch coming for that
howard-e: Dean reported that they were unable to submit the tests that they've completed for the "navigation menu button". I verified that. The issue is with how assertions are excluded by the navigation menu button.
howard-e: The regression occurred when we made a change back in August, but we haven't seen it in any test plan we've worked on since that time
howard-e: I'm submitting a patch (a patch to fix this) for review right now
Testing action menu button with activeDescendant
Matt_King: It looks like we're waiting on Murray
Matt_King: Have you heard from Murray, IsaDC?
IsaDC: I can reach out to them
Matt_King: It might be a good idea to re-assign, especially if the test plan has changed since Murray started
IsaDC: We can re-assign to a "bot" tester, but can re re-assign between human Testers?
Matt_King: I don't know! Checking ARIA-AT App right now, it doesn't appear to be possible
Matt_King: Didn't we discuss this use case when we built the re-assignment feature for the bot, howard-e?
howard-e: I don't recall that discussion
Matt_King: How difficult would it be to allow re-assigning from one human Tester to another human Tester?
howard-e: I can check for an issue for that
Matt_King: We can delete this run and create a new one. There's only one test complete so far, so that wouldn't be a huge loss of work
Matt_King: If you could reach out to Murray, IsaDC, then we can decide whether we should find another person to run this
IsaDC: Will do
Matt_King: We haven't heard from Alyssa in quite a while; I wonder if she's still available for testing
IsaDC: I believe she tested with NVDA
Testing navigation menu button
Matt_King: Looks like we're 0% done on all of these tests
IsaDC: I believe we are blocked on the bugs with the bots
howard-e: Manual runs are persisted, but it's a bit confusing. So feel free to wait for the patch to be live
Matt_King: We have hadi assigned here. Hadi, didn't you have a question about this?
Hadi: When we tab to an element and then we hear something like, "list with four items, blah blah link"
Hadi: ...Can "list with four items" be considered a boundary?
Matt_King: Yes
Hadi: Even the link is not the first item in that link?
Matt_King: Right
Hadi: My confusion is that, when I heard the boundary, I just consider the items that are immediately before or after the end. For example, if I have a list of five items, when I shift+tab (or whatever) to item three (assume item 4 and 5 are not focusable), I don't know what to expect
IsaDC: We are changing those tests, so I recommend not running the rest of the test plan in order to avoid doing double work
Hadi: I thought you had removed it and were going to add it back when it was ready for me
Matt_King: Oh, no, sorry. We don't have a way to mark something in the test queue as "on hold"
Matt_King: I'm not sure how we would do that
Matt_King: Last time you were in the meeting and you volunteered to do the "navigation menu button", I thought you had agreed to start the work on that while we fixed the problems with the disclosure
Hadi: Yes; I just misinterpreted the presence of the disclosure as a signal that the issue had been fixed
Matt_King: You can run the tests on "navigation menu button" and it will store your data, but you won't be able to submit them today
howard-e: Just use the "next test" button--not the "submit" button
Matt_King: Later, after howard-e deploys his bug fix, you will be able to go back and press "submit" on every test
howard-e: There is also an issue to make the "submit" button work for all tests in the test plan, but that isn't a feature today. For now, you will have to submit each test individually
Hadi: Thank you!
Matt_King: Thank you!
CSUN Update
Matt_King: We had talked about submitting something for this conference
Matt_King: I was excited to do that, particularly due to interest from Vispero
Matt_King: I was working on something, but the timing wasn't great for me, and I wasn't able to complete it in time for the deadline (which was last week)
Matt_King: The deadline was in the middle of TPAC, and I also had some big meetings at work during that week
TPAC Debrief
Matt_King: This is a conference of a few hundred people. Many of the W3C working groups have participants meet there in person
Matt_King: On Monday and Tuesday, the ARIA working group met all day
Matt_King: On Tuesday morning, we gave an overall update of the ARIA-AT to the ARIA Working Group
Matt_King: The minutes of that update are available online: https://
Matt_King: That was about 20 minutes of the meeting. It was followed by a much longer discussion about the new "ARIA actions" proposal
Matt_King: The proposal has been in the works for about two years, now
Matt_King: The basis of discussion is available online, here: https://
Matt_King: We had representation (in some form or another) from NVDA, Apple, and Vispero
Matt_King: I think the screen reader's responsibilities will be pretty straightfoward. they'll announce when actions are present with some sensitivity to cases where there are a lot of actions available in a list
Matt_King: They'll also present all of the actions which are available for an element
Matt_King: And that was what a lot of the discussion focused on
Matt_King: Essentially, we talked about the authoring requirements and the browser requirements
Matt_King: This is something that we already have an example for in the APG, and we're probably going to want to draft tests for that example soon
Matt_King: I'd like to do it before the end of the year, but I also want to be careful that it doesn't get in the way of our other goals
Matt_King: So we'll figure out the right time to prioritize it
Matt_King: But in the mean time, does anyone have any questions about that effort?
Matt_King: Sounds like not! In any case, this is a sea change in the way we have been contributing. It will be the first time we will write tests before anything lands in the spec. It will hopefully help prevent regressions--prevent anything from getting worse
Matt_King: Moving on, Wednesday at TPAC was an "unconference" day full of breakout sessions with participation across working groups
Matt_King: Chris and Boaz from Bocoup hosted a session on ARIA-AT, and we had really great participation from across the working groups
Matt_King: It was a high-level presentation, and the slides are available online: https://
Matt_King: There was broad interest, and I felt really good about the participation
jugglinmike: I think the most novel perspective for those folks was how we are using tests to drive consensus
Matt_King: And they seemed agreeable to it; no one was pushing back, for instance
Matt_King: So the impact of our work is starting to be recognized more broadly. That's important for our mission!
Michael_Fairchild: I observed, and I thought a lot of good questions came up
Michael_Fairchild: I really appreciated the thoughtful engagement. I thought there were some interesting questions about internationalization of not only the languages in the project but also the assistive technologies themselves
Michael_Fairchild: It would be good to consider internationalization along these lines and how it fits into our roadmap
Matt_King: Agreed. It might not be our job to test JAWS in another language, but what about testing a screen reader that is "Japanese first"
Matt_King: One of the participants referenced a screen reader like that
Matt_King: I think a lot of the questions were a reminder that starting where we're starting is just the tiniest drop in the bucket of ATs
jugglinmike: Whether its internationalization or non-screen-reader ATs, I'm seeing a few kind of "dimensions" that we don't currently service. I'm really interested in getting those on the roadmap so we can make the necessary structural change, even if the medium-term result is only to extend the UI so that our aspirations for wider inclusivity are apparent
Matt_King: I think the next big leap is mobile (iOS and Android), and then after that, kind of dipping our toe into those other dimensions, like you say. Taking on some of those fundamental problems
Matt_King: Okay, that's it. Thanks for your time, folks!