Meeting minutes
New Issue Triage
jamesn: ARIA 2236
Sara: Headings within feeds. Headings could be months or years and then beneath that you could have the posts
aardrian: Do you mean the feed role?
Sarah: It would be hard for on-demand loading
jamesn: It seems overly restrictive to only allow articles in a feed
Scott: Because content model is not defined, things like this are assumed to be invalid
… We may need to take on the work of stating the specific roles that are not allowed, thus the others would be allowed.
Sarah: Maybe we need to create groups, make it more fine-grained
… Also feeds can have arrow keyboard navigation within articles
… I've tried to talk to screen reader vendors to trigger focused articles, especially fro browse mode
jamesn: That's a big ask, I'll agenda it though
New PR Triage
Daniel: ARIA 2225
giacomo-petri: I put together a proposal for what we discussed last week
… I think this is a better place to put author responsibilities
aardrian: Is it id reference processing what you mean?
giacomo-petri: Yes.
aardrian: Is this looking for reviewers?
giacomo-petri: Yes
… Since ids are unique, that seems much like an author's responsibility
aardrian: I can review
filippo-zorzi: I can
WPT Open PRs
jamesn: Nothing new. Is this one from you Adam still needing review?
Adam_Page: It needs me to invert the test expectations
Deep Dive planning
Rahim: I am still dealing with the whole business of "undefined"
… Maybe the week after next
jamesn: 27th?
Rahim: Yes
jamesn: What specific issue I can put in there?
jamesn: No deep dive next week, we can potentially do another one on the 20th
Discussion tracking for ARIA Notification proposal
<Rahim> deep dive issue: w3c/
Doug: Deep dive recently. We did not have a lot of feedback. We have had some chatter in Microsoft around some issues
… We may need to get the basic ARIA-notify out and then work on the problematic functionality based on how people starts to use it
jamesn: aardrian posted a lot of questions yesterday
Doug: Didn't have time to respond yet
jamesn: I need to have a look at this myself
Dogu: Some issues are around interpreting what we currently have, and they come when implementing it
aardrian: I am not sure if I put them in the right place. Is that the meta issue to discuss?
Dogu: We can come back and comment there
Old Issue Triage...
jamesn: ARIA 651
pkra: This sounds like a CSS-AAM discussion
Skott: Transfer that to HTML-AAM
jamesn: Since we don't have MathML yet, can we close #660?
Skott: Do we need a MathML ARIA?
jamesn: We will never have a complete Math representation in ARIA
Sythia: An OS platforms also may not have that
Pkra: We started a CG, but that went nowhere. I'd close this
jamesn: @@@ issue number
… Is this something ARIA-notify can deal with?
… Any proposals on how to deal with this?
Pkra: It feels like an unsolvable problem. Sometimes you may want the same text replaced and repeated and sometimes not
jamesn: Should we have something along the lines of repeating it?
Adam_Page: I can always delete the text and readd it as an author
Skott: That's what I've told people to do
jamesn: Some browsers you have to do it, some you don't
aardrian: Not so sure it's a big concern here
jamesn: Should we put it in the authoring practice in the upcoming live regions section?
Skott: We should not assume this is how it still works
Skott: One of us could test this and verify if it's still an issue
… And then we could send this to APG for author guidance
jamesn: The spec should probably say what happens
Skott: We'd be saying what AT does
jamesn: No, just what browsers should do with sending the events
aardrian: I can test it
Jamesn: 683
… I totally agree with Steve
Skot: Wasn't there a new issue?
jamesn: I think we do have a problem here, but we may want to close some of them
Skott: I'll take a look through the issues and choose one that'd be the primary issue for this, link the related ones and close them
jamesn: Thank you so much Skott
<aaronlev> aria-checked also allows mixed
Jamesn: Issue about Tristate toggle
… Do we have use cases?
Siri: Yes. I raised similar issues before, mainly for expanded and collapsed which won't be completely press or unpress
Adrian: Turning plane mode on and then activating wifi is another use case
Rahim: Pressing a button changes the visible label and thus the accessible name, not sure if they are using aria-label
Adrian: aria-pressed is check or uncheck, and also it canbe put into anoter things because of other UI elements
jamesn: When you press it you can only put it into true or false, but it can become mixed by something else
Matt: I don't see a reason for such a limitation
… If someone is using a toggle button instead of a check box for a tristate then I would
… It seems weird that there is inconsistency amongst them
jamesn: There are UI elements where you have to toggle through several elements
Matt: They don't generate a state event
Adrian: I'd pause here. ome of us didn't know about the third state
Rahim: Is there desire to have parity with aria-checked and aria-pressed?
jamesn: Consider a token role for next version
… We've already put an Open UI lbel for this
Matt: I think this will help and reduce the need for things like the list view
Skott: I'd suggest we close it. I don't think ARIA should create a new role for ssomething that does not yet exist in HTML
… We should push this through to Open UI or HTML and see if that gets traction there
Matt: +1
<StefanS> -1
Adrian: +2
jamesn: Closing. This should be first in HTML
<aaronlev> I believe you can set input.indeterminate=true on <input type=checked>
Stephan: I'd be fine this interesting if we could make this resemble a list. These are objects typically, they can be deleted
<aaronlev> https://
Stephan: There should be patterns to show how to treat it
<aaronlev> there's a css pseudo class for it
Skott: Someone can step up and provide it
Matt: We have high priority things in APG that we need to do first. Please join APG and contribute code
2233 Presentational roles conflict resolution
giacomo-petri: Old issue was related to implicit presentational role. This one is about inherited presentational role
… Table with role presentation that contains data with aria-description, it is not clear from the definition id the td should be exposed as a table data or as a generic element
… Looking at how browsers expose it, they are exposing it as a generic element. Maybe it is not clear enough from the definition
Matt; If the global property causes the table role to be revealed as table, it removes role presentation because it's a global property
giacomo-petri: Are we expecting the td gets exposed as table data or as generic
<Zakim> scott, you wanted to react to aardrian
Skott: Because of the table being role presentation, all its children are generic. But aria-descciption may make it exposed as table data
… It may need some clarification here
… If aria-description was placed together with a table role then we would have presentational role conflict resolution
giacomo-petri: Agree.
jamesn: Me too. Are browsers doing the same thing?
giacomo-petri: Chrome Firefox and Safari exposes it as generic
jamesn: If that is the case we could create a PR with that
Matt: I thought description is not allower on generic
Skott: It's not prohibited right now
… The title attribute can be put in generic
giacomo-petri: I can create the PR
Skott; Put me on as a reviewer