Meeting: October 3, 2023

[cookiecrook]

Previous meeting was: #54.

This meeting will be October 3rd, 2023 @ 12PM Pacific. The meeting is usually recurring the first week of the month.

Please let @zcorpan know if you'd like to participate in the meeting and he can add you to the calendar invitation. Use the Agenda+ label on issues in this repo to add agenda topics.

New Meeting:

• Meeting: November 7, 2023 #62

[cookiecrook]

Some Agenda:

• WPT Interop 2023 Accessibility Breakout Session (W3C TPAC 2023) #59
• Additional TPAC updates summarized in this comment: Computing the Accessible Backing Node, or a similar test-only interface WICG/aom#197 (comment)
• Splitting up remaining issues from Scoring criteria for Interop 2023 Accessibility Testing Investigation #3 (ideally to finish by year's end)

[cookiecrook]

web-platform-tests/wpt.fyi#3494

[adampage]

Having trouble connecting to the Zoom link in the meeting invitation.

Ditto.

[zcorpan]

Meeting minutes:

Topic: recap from TPAC sessions
James: testing accessibility trees and selectors in webdriver
… note from jgraham that most proposed features are possible in both classic and bidi, but bidi would be easier for some. Can start with the features that can work in classic, since bidi isn’t supported in webkit.
Chris Harrelson: status for bidi in webkit?
James: ask gsnedders
… plan to work on a few features for classic (link to comment)
Chris: are you prototyping in webkit?
James: planning to see how much it would change a spec (core or extension), and work with webdriver engineers at webkit to see that it’s achievable. Ideally multiple impl mid next year.
… people seem on board with continuing investigation area next year
Chris: which specs would need changing?
James: either core or extension (accessibility webdriver extension)
Chris: webdriver spec is the core spec?
James: yes
Simon: would the choice of where to spec it impact implementations?
James: I think no
… might make a difference for the syntax. Different pattern for extension spec
Chris: I got a message from Alex Rudenko who was present at TPAC. He mention that on Thirsday there was a discussion…
James: yes I talked to him
… discussing issue 203, and AT Driver. Was there a third one?
Chris: don’t know. He said it would make more sense as the core spec. Also mention that there’s a need for event-based interfaces. More convenient in bidi.
James: wouldn’t be challenging to simulate system events that generates events for the page. The reverse is more difficult.
Chris: for those that aren’t a good fit for classic we should do in bidi.
Simon: in bidi you have access to browsing contexts, can execute scripts for example. So different feature set between classic and bidi already.
James: one doc or multiple docs between classic and bidi?
Chris: one doc seems fine.
Chris: I can ask Alex to help with the spec.
Topic: scoring criteria
James: pretty close now. Some open PRs. 5 outstanding PRs, missing 3. Would get us to 95%. Last 2 are SVG and DPUB (low impact). Seems to be on track.
James Nurthen can review
James: others can review also who are familiar with wpt
James: 87% is the current score
Topic: boaz’s report web-platform-tests/interop#459
James: investigation area or focus area?
Simon: focus area (for the written tests but also investigation for the ongoing interface improvements)
Chris: all tests?
James: exclude graphics role
Simon: exclude SVG, DPUB?
Chris: propose all tests, then exclude some in the first week
Adam: some interop difference in some accname test I wrote. (Something with visibility)
James: I think Firefox is right in that case
James: similar as aria-hidden=”false”, but can’t make that work on content that isn’t displayed
James: I think aaron and I agree that it should be accessible
Adam: yeah, tho I couldn’t find spec language to support this
James: I think aaron filed an issue. Several spec ambiguities from this work.
James: Simon and Chris, what’s on the table for a focus area… what happens if someone adds new tests?
Chris: tests are listed explicitly. There’s a process for test changes. This year there were a bunch of changes.
Simon: subtests can affect the score, it’s only included on a per file basis.
Chris: should propose changes to interop group before merging
James: tests seem reasonable for a focus area. There was another proposal for focus area, with display…
web-platform-tests/interop#512
#60
Simon: part of investigation area for next year?
James. Sure
Chris: unless it can land in the next week. Is that feasible?
James: no
Chris: Deadline oct 12
Simon: maybe we can link to PRs and the testing plan, and promise to write those tests in time
James: exclude graphics aria role, SVG, and DPUB tests from the focus area
James: lukewarlow proposed prefers-reduced-transparency media query as a focus area
James: low impact for accessibility, privacy concern (fingerprinting)
James: but he then proposed an investigation area to work out the privacy issues with matchMedia()
Simon: should happen in the CSSWG though. Interop project should ideally not be used to drive spec changes.
web-platform-tests/interop#515