W3C

ARIA and Assistive Technologies Community Group

10 Jun 2021

Agenda

Attendees

Present
rob-fentress, Matt_King, joehumbert, Sina, James, Alysa
Regrets
Chair
James Schoels
Scribe
Matt_King

Contents


<scribe> scribe: Matt_King

Merge conflict resolution

Seth: The PR in aria-at is 441

Howard finished changes discussed in last call. in review now.

Would be good to have review from test writers.

One separate f/o task would be to have a local serve command for local preview.

Thinking that should be separate PR.

Another f/o issue is to have a smarter diff-based system to build only what has changed.

That should also be treated separately.

James: Yes, those 2 are definitely stretch that can be separate.

One question: will this address merge conflicts with existing PRs.

Howard: After this is merged and then existing PRs are rebased on master, you should be able to procede as normal.

Rebasing will run the github action

James: merge master into each PR branch and then push?

Howard: yes

Seth: does this need to be coordinated?
... after merging PR that changes process, wwe'll try merging oldest approved new test PR to see what the process needs to be.

Issue 382: need better way to discern if NVDA is in browse mode

<rob-fentress> https://github.com/w3c/aria-at/issues/382

James: current instructions tell user to press escape, but sometimes escape is passed through to the page

Instead, instruct user to press nvda+space until user hears browse mode.

But, in default settings, it does not say browse mode.

Sina: The default should be words, not sounds.

James: I propose we have testing settings that are set up for a testing session.
... I think escape generally works so we can leave this issue unresolved

People are less likely to make a mistake based on sond.

James will rename the issue and direct it toward the automation concerns.

We still need this, but it is not urgent for manual testing.

Hint text

Issue 365

mk: We discussed that essential info should not be conveyed in a hint

James: should the hint contribute to passing or failing of an assertion

Rob: Should we consider disabling hints

James: like on Mac, VO doesn't tell you a radio is notchecked, but it has hint that says how you can check it\

mk: if we decide to diverge from defaults, would that cause uproar, would it be a problem.

James: we are sort of making decisions like this already.

examples include JAWS smart nav and VoiceOver quick nav

the only difference is hint text is opt out instead of opt in

Should also consider what this means for developers.

Rob: This is part of what we communicate about "what works for users"

Developers need to know what the role of hint text is in the experience.

They need to know if it matters.

James: sometimes developers rely on hint text when they should not because it may not be heard by many users.

Rob: what about text communicated by aria-describedby

rsagent, make minutes

We will discuss the pros and cons of not using default settings in the issue.

Summary of Action Items

Summary of Resolutions

    [End of minutes]

    Minutes manually created (not a transcript), formatted by David Booth's scribe.perl version (CVS log)
    $Date$