Meeting minutes
Review agenda and next meeting dates
https://
Matt_King: Requests for changes to agenda?
Matt_King: Hearing none, we'll stick with the agenda as planned
Matt_King: Next CG meeting: Wednesday June 3
Current interop reporting status
Matt_King: There are no new plans in Candidate Review. That's the same as last week--we're at 21
Matt_King: There are fewer plans in Draft Review. The only thing in the test queue right now is not actually ready for testing
Matt_King: As for the other two that are in draft review: one is stuck because of an app bug (the "tabs" plan), and the other one is also not ready for review
Matt_King: The one I'm getting ready for review is the "minimal datagrid". That's coming up later in today's agenda; hopefully, it will be ready for people to work on by the end of the day.
Update on infra engineering training
Matt_King: We recorded it and will make the recording available to anyone interested
Matt_King: We're not publicizing the video link, though, in case they include any sensitive information
Matt_King: But if you'd like it, let me know, and I will share it with you
Matt_King: I'll share it at least with the meeting's attendees
Matt_King: The next meeting is scheduled for next Tuesday, May 26
Matt_King: We'll focus on updating one of the bots to use the latest version of its associated screen reader
Matt_King: Hopefully we'll have NVDA Bot updated soon
Matt_King: And the VoiceOver Bot will follow soon after, and then the JAWS Bot
Matt_King: And as a reminder: for anyone who will attend that, if you can get the bots running locally, that will help. A bonus bit of homework would be to get the ARIA-AT App running locally
Draft review of minimal data grid test plan
Matt_King: I want to find out who will be available for testing
Matt_King: This isn't a long test plan. It has 11 tests. But it's probably the most complicated test plan I've worked on in terms of authoring the test
Matt_King: Navigation is very nuanced. Because of that, getting the assertions right for each command (column number, row number, content, etc) is ridiculously complicated and kind of annoying (from the perspective of writing the test plan)
Matt_King: Running it should be straightforward enough, though. I strongly encourage testers to be extra critical of these tests!
Matt_King: I'm currently working around a preview-related bug, but I hope to merge it later today.
Joe_Humbert: Is there a completion timeline?
Matt_King: No, we don't have a tight timeline
Joe_Humbert: I'm asking because this weekend is a holiday weekend [in the US]
Matt_King: Oh, right
Matt_King: I expect it to be in the two-to-four week range. So within the next two meetings
Matt_King: I don't know if I'll have any more new test plans ready by the time of the June 3rd meeting
Matt_King: So I would say that we have until June 27th
mmoss: I can help. My preference is VoiceOver, but I have access to JAWS and NVDA
Joe_Humbert: I can help on this one, too. I don't have a JAWS license, but I can work within the trial period
dean: I can help
elizabeth: I can help, too
Matt_King: For mmoss and elizabeth, I would prefer to use your time for engineering work.
mmoss: I don't have to split my time; I can do both
Matt_King: Thank you!
Progress on proofing auto-updated reports
Matt_King: I'm assigned one, and Joe_Humbert is assigned seven
Joe_Humbert: I sent an e-mail about this; I completed all seven
Matt_King: Did you run into any details--any spots that I should pay special attention to? Or was it all straightforward?
Joe_Humbert: On the non-radio-button tests, there were a handful where it seemed to read what it captured from the top of the page
Joe_Humbert: It looked like JAWS was starting at the top of the page and that's what it was capturing for the test output
Joe_Humbert: I didn't mark them specifically because there was so much testing to do
Joe_Humbert: There were quite a few where the bot incorrectly reported "No output" following "U" / "Shift + U" (plus potentially others)
Joe_Humbert: There were a handful where it looked like it was navigating through a handful of radio buttons. It should only output a single radio button, but it read all the radio buttons
Joe_Humbert: Those were mostly for the commands that included multiple arrow keys.
Matt_King: That's a very strong indicator of the same problem we've had with NVDA as well, where the correct cursor or mode wasn't invoked. I bet it was actually operating in "forms" mode when it should have been in "virtual cursor" mode.
Matt_King: jugglinmike, I know we've had some past discussion about what the bot does to ensure it is in the right mode, but I don't recall any details of those discussions
Matt_King: Is this something we should be thinking about incorporating into the training?
jugglinmike: [A long-winded and highly speculative of what might be going wrong]
Matt_King: We should be sure to document workarounds we're currently forced to do in the harness
Joe_Humbert: There also were a handful of bot responses that I had to modify slightly. Some omitted tutorial text. And in the radio buttons (I think), it was capturing "1 of 3", but the new version was leaving out that information. I updated the response to include that text because it was actually spoken
Joe_Humbert: Those were cases where the bot response were for an older version of JAWS than what I was using
Joe_Humbert: When you view the automated updates for some of the test plans, it says "required reports complete", but for others, it says "required reports missing". Why is that?
Matt_King: That's because they are in the "Recommendation" phase in the working mode, and for that phase, we are supposed to have Firefox results, as well
Joe_Humbert: Got it. I just wanted to make sure it wasn't failing to receive some reports
Matt_King: Nope, that is all expected behavior
Matt_King: We're going to have the JAWS Bot updated to 2026 soon. If everything is great, then maybe Vispero was able to use the bot to capture those changes, and maybe they fixed them already!
Matt_King: Okay, I think we covered it all!