Meeting minutes
New Issue Triage
<github-bot> I can't comment on that because it doesn't look like a github issue to me.
<spectranaut_> w3c/
<Daniel> w3c/
<melsumner> why is it a separate repo (html-ruby)?
<spectranaut_> w3c/
New PR Triage
<github-bot> I can't comment on that because it doesn't look like a github issue to me.
<pkra> w3c/
Deep Dive this week: CSS ::interest-button
<github-bot> I can't comment on that because it doesn't look like a github issue to me.
<jcraig> ::interest-button deep dive minutes here https://
AAM testing update
<github-bot> I can't comment on that github issue because it's not in a repository I'm allowed to comment on, which are: w3c/aria w3c/aria-common w3c/core-aam w3c/accname w3c/html-aam.
spectranaut_: Acacia AAM testing has moved to a new test type, with all new tests being written in python.
spectranaut_: this makes things easier for several reasons, one of which is that all of the platform APIs were in python from the start, so now we don't need to make bindings for the testharness.
dgrogan: What is the history here?
spectranaut_: see the issue linked in the PR
jcraig: I feel like we should prioritize the machine readable AAM mappings before folks start writing tests for acacia, otherwise there may be a migration issue.
Potential heuristic standardization: how does/should event delegation play into screen readers conveying "clickable"
jcraig: This topic came up from a gecko bug ... due to everything being reported as clickable, we have a heuristic to determine if something should actually clickable.
jcraig: I commented on the gecko bug with some information about what we do in webkit, we are open to working to make this heuristic more widely available or standardized.
jcraig: Should this be a priority?
keithamus: I'm a little worried about heuristics, especially regarding the table heuristics we have gathered. Table has a lot of history.
<spectranaut_> Table heuristics: https://
keithamus: My fear is that since frameworks can do some strange things around event handlers, I'm worried about how we can end up chasing our tail with this kind of work.
keithamus: I'm also worried about the lack of agency developers will have here.
jcraig: I was just using table as one example of a heuristic, I don't think we have really needed any changes since this was first introduced.
jcraig: we could have done this at the voiceover level but webkit has the information we needed.
keithamus: is there any special case for table or list?
jcraig: I don't beleive so...
<Zakim> pkra, you wanted to ask where "clickable heuristics" would fit
pkra: Where would these heuristics live?
jcraig: perhaps a wiki page that has a list with the general details each.
jcraig: if we have something solid we can then consider moving them into a spec.
Matt_King: In the case of clickable, since clickable isn't a semantic we have standardized to begin with...
Matt_King: its kind of a hint that something might work, but as a user I end up clicking everywhere, not paying attention to what is clickable.
Matt_King: the answer of if we invest in this should be determined after we have assessed the impact on user.
jcraig: I'm thinking something less format that we can all share and use, like a wiki.
Matt_King: It would be useful to know what browsers are doing
jcraig: With get accessible properties and acacia, we can now test all these things that are specced, if there is an issue then we can investigate what is being done.
<pkra> ack
keithamus: This wiki sounds a lot like a spec, thinking back to the table problem, up and coming browsers would appreciate a single source of truth rather than wikis and such.
keithamus: If we do this, I strongly suggest this goes into the spec.
jcraig: having this be something shy of a spec lowers the barrier to entry, right now updating the spec with this is going to be more than what anyone might be wiling to do.
spectranaut_: Can we have a non-normative version of the spec?
keithamus: For the table spec, my next step is to land the interoperable bits now, and then go from there .
keithamus: If it is compelling enough for two browsers, then it is compelling enough for the remainder.
<Zakim> melsumner, you wanted to ask why do we need clickable at all, though? Shouldn't the announcement of the element indicate interactivity??
jcraig: I'm just using clickable as an example. This is something that we decided to do about 20 years ago, this is a way to convey to the user that something is clickable.
jcraig: If a node has a click handler then webkit considers this to be "clickable: yes" and so we expose this to the screen reader as clickable. A user feature to cover an author anti-pattern.
spectranaut_: maybe a wiki just to document?
jcraig: Any opinions on where this should live?
RESOLUTION: jcraig will start a wiki on Aria when he has time
Presentational Role Inheritance, host language content models, and aria-owns
pkra: This is an edge case observation that comes down to "what happens when presentation role inheritance meets aria-owns reparenting?"
pkra: there are a few combinations for what happens, and we should add some tests to see where the inconsistencies are.
Matt_King: Some of the behaviors seem strange to me, and the behaviors that are happening might not be what we want.
jcraig: There are quite a few sub-tests related to this interspersed
jcraig: one of them is about orphaned roles, and about two years ago we updated what should happen for orphaned list items.
jcraig: Because computed role exists in WPT we can already test this. This would be a good place to continue this investigation.
jcraig: One of the ways we flesh this out is by writing a new tentative test file where we don't put in an expectation, and then get a sense of what each browser does.
jcraig: That information would help to further this discussion.
spectranaut_: Adam, didn't you work on aria hidden, aria owns, and the intersection?
Adam_Page: I can find some links to that work.
spectranaut_: Definitely an edge case, first step is to write tests to see what browsers do.
jcraig: Its possible some of these tests have already been written.
spectranaut_: assigning to adam