W3C

Accessibility Conformance Testing Teleconference

30 Jun 2022

Attendees

Present
ChrisLoiselle, Helen, Todd, thbrunet, Wilco, Will_C, kathy, trevor, Daniel
Regrets
Chair
Wilco
Scribe
thbrunet

Contents


<Wilco> scribe: thbrunet

ACT rules sheet

Will_C: Need help from Wilco on multiple changes

trevor: Haven't seen further comments

ChrisLoiselle: Not sure if I have others to review right now

Open ACT pull requests

Wilco: First is grammar issue with changes requested
... reviewers Todd, Kathy, Will
... We have a couple of ARIA 1.2 pull requests
... reviewers Tom, dmontalvo, Wilco or Will
... Image button, needs one more approval. Tom to have a look at - may be interesting
... 1859 - has some reviews pending
... 1855 waiting on some change requests.

Helen: I saw those this morning, looking...

Wilco: 1845, 1831 has changes requests

ChrisLoiselle: 1873 - I need to sign the CLA from Oracle to post for first issue. Jean Ives made some comments, but before I accept, I think I have to sign the CLA. But, from what is mentioned, seems straight forward.

Wilco: Seems good to me

ChrisLoiselle: I agree

Wilco: Only part I'm eh on is "significant part of the visible pixels"

Announcing ACT implementations

Wilco: We talked about this last week where we're pretty close to a point where I think we can take the 'under development' banner off of the implementation table pages.

<Wilco> https://www.w3.org/WAI/standards-guidelines/act/rules/

Wilco: New as of this morning, banner updated. And we have the implementation page on the navigation bar now. Now formally have this content on the w3c site. Now the question is at what point do we take the 'under development' banner off of the pages.
... Daniel, Kathy, Trevor, Wilco talked about it earlier this week. Need to have all of the implementations +1
... Also, the wai email address is under feedback and need to replace with something else.

<dmontalvo> We could have group-act-rules as a group list

Wilco: Open question on how to deal with composing rules. Atomic rules that are part of a composite rule. As stands, the accessibility requirement is a little funky.

kathy: How frequently are reports updated?

Wilco: Couple parts. Reports pulled every 6 hours. I may bump that down to once a day.
... Website itself we have to preview. That's every few hours. Main site is a manual process. Maybe once a week.
... Plan is to get it done every week.

kathy: I updated the trusted tester results yesterday, but waiting to see when it will show.

Wilco: Should be in the preview, but checking.

kathy: Should be 12 if it's updated.

Wilco: I'll look, something went wrong there

kathy: Can we get a last updated note somewhere on the page?

Wilco: Good point. We have that information now. I'll make a note.

dmontalvo: Also, group-act-rules list as a private list. Could say if you want to provide public feedback, post here.

Wilco: Wanted to ask Tom about the numbers

thbrunet: We're reviewing. Should be better in a couple of weeks.

Wilco: We have a coming soon option
... That shouldn't be around after we've taken off the banner.

Helen: Could be an update in progress banner

Wilco: Maybe some more information on the implementation page.

dmontalvo: We don't want to give an impression that it's not ready yet.

Helen: Some indicator that results might change in the near future.

Wilco: I'm inclined to say let implementers get to a point where they're comfortable. If that takes awhile, could take them out temporarily.

thbrunet: Maybe can put a note in their yaml file description

Wilco: Okay, can have a vote

Rules Format and state

trevor: Regarding discussion on states, wanted to dive into transitions. I want to better define them
... And I want to look at some examples. Two cases we might want to use transitions and want to hear opinions.
... When we talk about a transition, we have a pair of states and we're moving from one state to another. Clicking, entering text, even waiting.
... There can be some nuance to it.
... The other thing interesting is these transitions can take place immediately, or have a search function where there's some delay where you see a loading icon and then get results.
... State in beginning, transition, and then end, or some other way to tackle this?
... Mainly want to focus on the middle two examples. First example is input field, enter text, trigger search. Loading icon appears. Do we want to model that as being a transition or it's own state. Is that something we need or want to test?

kathy: Question about transition. Does it matter how long that transition takes?

trevor: Good question. We've danced around the transition. But, think we have to set some time limit.
... In cases of a loading icon, would hope a second or two, but could be longer.
... But in case of flashes, might be less than a second.

thbrunet: Loading is tricky since it can be an arbitrary amount of time

Will_C: If it hangs, can be an issue also

thbrunet: If that's hung for everyone, that's a functional problem

Will_C: Sometimes you can see that the content has or hasn't change and have more information

trevor: Usually need some other indicator also to give feedback.

Wilco: Something to be said for saying loading is its own state
... flashing can't happen. But, change of context might happen in the loading state.
... There's a SC on notifying when in the loading state

trevor: Sort of enumerating the transition states. Are there others we might include in well?

Wilco: One thought on transitions. Often animation - fading in and out. Can get complex. Especially in games and media.

trevor: I was trying to move the wording away from loading state to transition state

thbrunet: Perhaps a definition of transition state that can be generic and used in the rules?

trevor: I'd like to avoid having to worry about transitions all of the time.
... The other thing I wanted to look at is we may have cases where ACT has a pair of states and one is a result of the other.
... I can imagine having a rule like clickable button can be activated. Start in state 1 and show we can get to state 2.
... Feels like we'd be adding something to the applicability

Wilco: Please explain further

trevor: Way we were testing for keyboard was we'd use mouse, then do same action with the keyboard, and if we didn't get same state, we'd consider it inaccessible. That's how we used pairs of states to find issues.
... thinking we have these pair of states to show accessibility in some cases.

Will_C: If I'm thinking of this correctly. Maybe an accordion. Failure would be something like missing a tab index or the role is wrong.
... Feels like it's covered in WCAG. Cannot be activated by keyboard alone

thbrunet: Case of aria-expanded toggle in an accordion?

trevor: A little more than that. Debating if aria-expanded is just a proxy.

ChrisLoiselle: For a toggle button, on off. At Oracle, we have admin rights. Toggling a read-only on/off. The actual admin tool, you're triggering this at page level. That state for that setting of that read-only input. Is that what you're talking about? When you're triggering, you're providing information that toggled on off?

trevor: Kind of. I was trying to be more simplistic. If I can click something with a mouse that occurs. I can also do it with a keyboard. Maybe sometimes it's a specific attribute like a checkbox on/off. If all of those get treated the same. Still, posing it inside of a rule. Seems you need to specify the two states and how you get between them. Not sure how you specify the rule without that transition exists.
... When we talk about applicable states up here. It was go to a state and do a test on that state. This other is do something and expect a state to happen.

Wilco: What's the next step? Do we need to formulate a rule?

trevor: Couple places to go. Can play with formulating a rule. I think at some point, we need to have this applicability discussion and how we're going to talk about these things.
... Can we create a list to draw from so we're not just using whatever words. Those are the two topics areas to get into.
... Maybe the applicability, test section, and a test case or two

Difficulty with composing rules and accessibility requirements mapping

Wilco: Atomic rules part of a composite rule
... For atomic rule, maybe all the test cases fail a rule, but the atomic rule is expected not to map to a particular success criteria. However, the composed rule does.
... We now have an issue where if you're reporting you failed 2.1.2, but atomic rule expects no mapping, then you're reported as partially inconsistent.

Will_C: Why doesn't the rule doesn't map to an SC?

Wilco: Two ways to pass the composite rule. Failing each atomic rule doesn't mean you fail the SC because the other method may be used.
... Jean-Ives idea - can have a list of SC's that you 'may' map to.
... So in this case, you're allowed to fail 2.1.2 or nothing and still be consistent. You wouldn't be allowed to fail other SC's.

thbrunet: Another way of saying this is there are consequences of the rule (must map) and related SC's (may map)

Wilco: So if you map to a may map, it wouldn't be inconsistent.
... Solves other problems. Notable one is ARIA roles. Could say ARIA issues 'may' map to certain SC's.

Summary of Action Items

Summary of Resolutions

[End of minutes]

Minutes manually created (not a transcript), formatted by David Booth's scribe.perl version 1.200 (CVS log)
$Date: 2022/06/30 14:49:29 $