Meeting minutes
ARIA-AT App Updates
Seth: The app and repo now support up to 26 commands per test.
Previously limited to 5.
So, now breadcrumb needs retesting b/c it had a 6th command in some tests.
The candidate review queue is delayed a week.
We failed to include Bocoup all hands meeting in plan.
Expect the update for candidate review will be in sandbox by end of next week.
Matt: could some part of it be in sandbox by wednesday morning?
Seth: some portion could be available by wednesday morning.
Seth: Deployed multiple performance improvements on Friday.
Page loads were taking long time b/c data was being computed at time of page request.
Howard worked on system that pre-computes a lot of the data.
This improves both page loads and stability.
App feels much snappier and James Craig confirmed his experience is fixed.
Michael: Is there risk of perf degradation as we add more data?
Seth: Due to precomputation, experience is better regardless of data volume.
However, we still see more opportunities related to versions of tests that are no longer being used.
Other thing is that W3C I/T team re-sized the server.
That could be another option in the future.
Seth: Howard deployed partial fix for test plan updater.
Think it might be fixed, but there is still some doubt regarding the option that allows keeping old data.
James: There are some situations where we have dumped all data.
Sometimes we do need to keep the data.
Seth: we still may need to make some improvements there.
Implications of toggle button feedback
James: our tests require that the not press state be voiced. VO on macOS does not do that.
Apple concerned that aria-at might be working toward making speech congruity instead of interoperability.
Matt: we are not striving for homogenaity. Example: if VO said toggle off, that would pass, right?
James: yes
We don't say how the state needs to be conveyed.
Matt: I'm still OK with trying to move forward with the original position that the not pressed state should be voiced.
Temperature check with others: are others still on board with this.
Michael: I still support the position, more concerned about consistency between macOS and iOS.
Matt: they may fix that by changing iOS.
Alysa: I hope not.
Seth k: more concerned about distinguishing button from toggle.
James: they have different roles.
However, there are selectable buttons in the apple ecosystem where it is not possible to know that a button is a toggle and that activating the button just selects it.
Seth K: Can't think of other controls where two different elements have same feel but different behaviors.
James: maybe grid vs table.
The larger concern I have that it does not seem to consider the feedback of real screen reader users.
The position of not speaking not pressed.
Michael: From my POV, not speaking not pressed is not ideal, but as long as it is consistent across Apple platforms, it might not be a serious problem.
Matt: my biggest concernis not speaking state change
James: there are other examples of speaking state changes to states that are not announced.
Seth K: But, that is a user-made verbosity choice
James: I agree with the state change concern as well as inconsistency with checkbox and switch.
And, my tertiary concernis the cognitive load.
Seth K: there are other ways to do this; doesn't have to be verbal.
example: TalkBack indicates links with a earcon.
Alysa: as long is there a way to convery that on braille display.
With Talkback, is is in apps, not on web where it uses the earcons.
James: If apple has concerns about verbosity, that is where audio indicators, like earcons.