Meeting minutes
Review agenda and next meeting date
<Matt_King> See agenda at https://
<Matt_King> Next meeting on April 20.
Matt_King: We'll reorder the agenda based on jongunderson's availability
Current testing check-in
jscholes: VoiceOver and Chrome -- "Radio Group" PAC is still in progress, Joe has completed (and provided some feedback)
jscholes: "Alert" PAC is still in progress, Joe has completed
jscholes: "Link" PAC is still in progress, Joe has completed
jscholes: NVDA and Firefox: "Radio Group", PAC has completed testing, and Alyssa plans to complete it by the end of the week
jscholes: "Alert" -- all testing is complete with no conflicts
jscholes: "Link Example 1" all complete with no conflicts
jscholes: JAWS and Firefox" .. "Radio Group" all testing complete with 6 conflicts, "Alert" , all testing complete with no conflicts, "Link Example 1" all testing complete with no conflicts
jscholes: w3c/
jscholes: VoiceOver with Chrome, "command button". Navigate backwards to a button with the command "Shift+Control+Option+Command+J"
jscholes: Multiple testers have reported that this command on the Mac simply doesn't work
jscholes: The intent is that when you press the command, the focus moves to the previous element, which is a button, and reads that button
jscholes: In fact, it appears to be moving forward
jscholes: Except that jongunderson reports that it works as the test expects for him
Matt_King: Yeah, the response that jongunderson has reported looks perfect to me. Could his Mac be somehow different than Isabel's...?
JoeHumbert: Is this the same issue that I reported?
jscholes: Yes
JoeHumbert: In all of the test that I did for RadioButton, I started to wonder if it was because I was starting on a link instead of on a control. It should work regardless, of course
jscholes: I would also expect it to work, even if it was getting confused internally, because it should take the previous focus into account. In any case, I feel the command is broken
jscholes: I think we have to report the results that we've collected which demonstrates that the command is broken
Matt_King: I can reproduce the bad behavior outside of the test environment, on real websites
Matt_King: It doesn't jump forward to the next control; it jumps forward to the next button. If I press it again, then it starts going by all controls
Matt_King: Actually, it may be jumping to the *last* button
JoeHumbert: I think TalkBack uses the concept of "controls". There, it means all buttons and all types of inputs
Matt_King: That's what JAWS does for the "f" key
jscholes: Are we saying that the ask for jongunderson here is, "Can you re-run the tests that include this command and if you end up seeing the broken behavior, please update your results"?
jongunderson: Yup, I can update my results
Isabel: We've marked the checkboxes as "incorrect output" is that right?
Matt_King: Yup, that's better than the alternate ("no output")
JoeHumbert: It looks like WCAG uses the term "control" but does not define it
jscholes: I'm replying to the issue with the resolution for jongunderson
Matt_King: This is the same as w3c/
Feedback: "Navigate backwards to a checked radio button" (Radio Group Example Using aria-activedescendant, Test 16)
github: w3c/
JoeHumbert: Are we treating "selected" as the same as "checked" for radio buttons?
Matt_King: Yes
JoeHumbert: Thanks. I'll correct the one test that I left
Matt_King: The assertion says that it "conveys the 'checked' state", right? We worded it that way to accommodate this
jscholes: reviewing testers' results is cumbersome using that app's UI
Matt_King: I'm in the process of reporting an issue about that
JoeHumbert: In the future, it would be nice if there was a "reviewer mode" so someone like jscholes could see all of the results for all of the testers
Matt_King: We do need to be able to see the report for testing that's in progress. I'm very attune to the whole situation and working to find the least expensive way to get to a better state
Matt_King: I'm writing something up
jscholes: Right now JoeHumbert has marked "no output" for all assertions. It should be "incorrect output" for the role and name of the radio button, and "no output" for everything else
Feedback: "Operate a pressed toggle button" (Toggle Button, Test 24)
github: w3c/
jscholes: this is about state change when you press the "enter" key
jscholes: this is not the same as the issue we previously raised with James Craig
Matt_King: We need jongunderson to move on this, so we'll save it for later
Matt_King: It looks like we're caught up on all the conflicts that we can resolve today
AT Support Table Launch Update
Matt_King: this has been quite a week!
Matt_King: Last week, I had just finished the draft of the announcement and shared it with a lot of people, but not everybody
Matt_King: We got more and more stakeholders over the course of the week, and eventually the W3C communications team got involved
Matt_King: Eventually, we had to negotiate some things. This included things about the aria-at-app itself, and Alex_Flennikan has pushed up a new note during this very meeting
Matt_King: "Note: The ARIA-AT Project is managed by the ARIA-AT Community Group in coordination with the Authoring Practices Task Force of the ARIA Working Group. The W3C staff contact is Daniel Montalvo."
Matt_King: The announcement is just about done, I just have three things left to resolve
Matt_King: Dequeue has a post ready to go. So do Bocoup and PAC.
howard_e: the work on the app is done
Matt_King: So we're ready to announce on Monday with a W3C blog post, a Deque blog post, a PAC blog post, and a Bocoup blog post
JoeHumbert: Can you send out an e-mail with links to all four?
Matt_King: Yes; I'll send an e-mail to our mailing list with all of those links on Monday morning
Matt_King: It's been a full-court press. Thank you to everybody for helping bring all of this together!
JoeHumbert: Only 60 more examples to go!
michael_fairchild: This is a huge milestone
Matt_King: We haven't yet published our schedule for what's next. We're going to have to turn our attention pretty rapidly to the three we have waiting "in the wings", and then deciding what comes after those
michael_fairchild: On the APG examples, there's a heading for keyboard support, and our tables live under another heading for "assistive technology support." I wonder if it would be better to word "keyboard support" as "keyboard interactions" since we're not really testing keyboard support there
Matt_King: Actually, we are testing support to some extent
Matt_King: "Keyboard interactions" is how we label it in the pattern
Matt_King: I wonder if those headings should be the same... It's worth raising as an issue in the APG, michael_fairchild