Meeting minutes
<ray-schwartz> +present
<aardrian> +present
everyone: a round of introductions
New Issue Triage
spectranaut_: aria 2268 roadmap
jcraig: came across it while reviewing new charter
jamesn: let's talk at chairs meeting.
spectranaut_: html-aam 552
… should this be put on an upcoming agenda?
aaron: I think I want to build an experiment first.
… there's nothing illegal about using title and I think we should do something with it. description makes most sense when name is prohibited.
jcraig: first impression makes sense. The interop link has more discussion.
<melsumner> I'm not super psyched about the idea of putting a title attribute on everything. That would be so noisy and disruptive.
aaronlev: might be some pushback when there is a lot of names despite it or if it introduces a lot of noise.
scotto: I added a link to an html-aam issue where I suggested the same.
spectranaut_: ok, let's keep it going.
… next: accname 241
… agenda'd today
… next: aria 2266
jcraig: wasn't this one of Alice's use case for crossing the shadow dom?
spectranaut_: let's ping Alice on it.
aaronlev: we've had long term goal to make aria-owns solid. it's been difficult and I wouldn't want to spend the same effort again for shadow roots.
… it seemed natural to forbid this, also to avoid risks.
spectranaut_: I'll ask Alice.
jamesn: you can't cross right now anyway, correct?
aaronlev: yeah, but I'd like to avoid it in the future if there are no use cases.
jcraig: to clarify: it's about affecting the tree, right?
aaronlev: yes.
jcraig: ok, because I think right now we're on course to allow it.
jamesn: will be a bit odd to allow idrefs but not this.
New PR Triage
<aaronlev> Agree that it might seem weird to others, but they didn't spend a year of fixing aria-owns bugs
spectranaut_: PR 2273, html-aam
<aaronlev> Agree that it might seem weird to others, but they didn't spend a year of fixing aria-owns bugs (regarding previous discussion)
scotto: porting over from old repository
spectranaut_: isn't there another one?
scotto: yes, but keeping it separate for implementations so we an merge them quicker.
spectranaut_: reviewers?
scotto: the previous PR has reviews. It just needs to be done.
spectranaut_: 2271 PR looks good to go.
… PR 2269
… has reviews
… PR 2264 html-aam popover
… reviewers please?
keith: will do
clay: will do
spectranaut_: PR 2262 popovertarget - old one
WPT Open PRs
spectranaut_: seems like we're up to date.
jcraig: if you're a reviewer, please take a look. ping me if you need assistance.
Deep Dive planning
spectranaut_: orphan roles deep dive today, what's our next topic?
scotto: what's the result for today's?
jcraig: we'll take the line from core-aam, make it a note, make it an author error but permit browsers to keep what they're doing.
… I listed all cases I could find on the issue. I think they're all passing in WPT right now and then we're not planning to change that.
… if it's consistent across browsers. But there's also a list of roles on the issue where we will not enforce it.
scotto: does that include HTML implicit roles?
jcraig: e.g. for list item without list, rendering with bullet: we will say it's an authoring error but we won't require that it should be exposed as list item. But if we think it should because, say, css requires it to have a bullet, then we can deal with this.
… but before the discussion, that would have been disallowed.
scotto: I'
… am sorry I wasn't there.
spectranaut_: any new deep dives?
TPAC Reminders: registration fees waved for Invited Experts, and early bird ticket price ends on Friday
spectranaut_: just a reminder.
… early bird ends.
<Francis_Storr> "Current Fee: Early Bird – until 16 July 2024 00:00 UTC: USD 360" https://
spectranaut_: invited experts have registration fee waived. There are other waivers, case-by-case.
aardrian: what's our schedule?
spectranaut_: there are a lot of cross-group meetings.
jamesn: main is Mon+Tue but we have cross-meetings on Thu+Fri.
… the w3cc calendar already has meetings in there.
<aardrian> https://
jcraig: are there already agendas for some of these?
jamesn: a bit too early.
daniel: I'll keep an eye on these within staff contacts.
Accname testable statements include prohibited name situations
<jamesn> Info about TPAC Diversity fund - https://
spectranaut_: mel suggested closing the wiki page?
melsumner: yes, archiving it, making it clear it's historic, and link to WPT page going forward.
spectranaut_: I can edit the page.
… aaronlev, does this conflict exist elsewhere?
aaronlev: there are some in WPT test. I'm currently doing an experiment behind a CLI flag.
… wanted to try to move name to description field instead.
jcraig: that seems reasonable to me.
aaronlev: if people like it, then we'll have to change WPT to allow this.
spectranaut_: ok, then we can close this after the wiki edit.
rahim: I was just wondering what the wiki page was
<melsumner> what do you mean by "like it" though? I do not like the idea of aria-label being valid on generics
melsumner: it's old stuff done by Joanie.
scotto: from 2018, before we had something like WPT
rahim: ok, makes sense. WPT now does the same job.
jamesn: yes. and beter.
jcraig: these are technically a step beyond what WPT can.
jamesn: yes, but not necessarily important for accname.
WPT test fails if I enforce name prohibited
spectranaut_: that's effectively the same discussion?
aaronlev: I found a way to do it without changing the tests right now.
… if we end up upgrading the flag, then I'll need to deal with this.
jcraig: but perhaps we can come to a decision anyway.
… if this is a test we want to remove, then let's do this as soon as possible.
aaronlev: this assumes the browser is loose.
jcraig: it seems we need more discussion and experiments. So perhaps pull the test temporarily and mark on it. Then once we have it resolved, we can go back and fix any agents that need it.
… I'll link from those 5 tests to this issue.
aaronlev: so the tests won't change?
jcraig: I wanted to leave them but comment them out.
aaronlev: ok
… that would unblock me.
Should aria-valuenow be required if aria-valuetext is used?
spectranaut_: patrick left a detailed comment.
scotto: yes, we were talking about different things. He tackled progress. I had done sliders etc.
… it's important. I'm not sure implementations with a particular role should prevent moving forward.
jcraig: there's no way for AT to trigger browser action - but only for ranges?
… just trying to think what VO would do, the value change can trigger an event listener, changing valuetext. So you're saying, a custom slider with value now and value text, because they would control by keyboard or synthesized events, then value now is not needed. Is that the gist?
scotto: yes. I was porting over from Wilco's issue on axe-core.
… from his testing, value now was not being exposed, just an unnecessary request for authors to update despite users not benefiting from it.
… I didn't test myself.
jcraig: API level or end user?
scotto: end user.
jcraig: I think we might announce the percentage. would have to check.
scotto: I've seen confusing examples.
… not making it clear for the user at all.
melsumner: this feels like an implementation bug. I'd rather not tell authors to not do it because users don't see it. It seems to me that it should be exposed.
scotto: but value text is supposed to override value now.
aaronlev: is this about value now being parsable?
… get e.g. sonification?
scotto: makes sense for progress. sliders though?
aaronlev: maybe percentage?
jcraig: we have a feature to get more information.
spectranaut_: please add those examples to the issue and let's continue another time.