Meeting minutes
scribe Adam_Page
… #1824 from wilco
… aria-label is defined as a global property, it’s just prohibited on some elements
… probably should make explicit what is prohibited in 1.2
New Issue Triage
spectranaut: can’t put it in 1.2
jamesn: will label as 1.3 blocking
… #1823
… should move 6.5 under 6.6 as 6.6.1 Global States and Properties
jamescraig: sounds editorial
<chlane> I can take it
jamesn: #1822
… this is too specific
jamesc: France and Italy are beginning to require accessibility statements but still only on a subset of sites
… maybe for conformance checkers, but still not good enough
… as an alternative, what about rel="accessibility-statement"?
… jcraig will take assignment
jamesn: #1821
… propose to agenda for next week
BryanGaraventa: ok
… would be good to get all three accname mappings in one place for Mozilla, Chromium, and Webkit
jamesn: #143
… related to sarahhigley’s work on allowed child elements?
sarahhigley: working on two things related to this
… will take assignment
jamesn: #437
… need to get auto-publication going
spectranaut: agreed
jamesn: #1818
spectranaut: Aaron should look at this
spectranaut: aria-hidden removes things from the accessibility tree
jamesn: and then aria-owns puts it back in
… we’ll agenda for next week
New PR Triage
jamesn: 5 PRs
… #1820
jcraig: some authors may _expect_ user-triggered changes in a live region to be spoken?
… is there opt-in or opt-out
… nevermind, we clarified in statement
jamesn: we can merge this, right?
ACTION: jamesn to merge
<trackbot> Created ACTION-2168 - Merge [on James Nurthen - due 2022-10-13].
jamesn: #144 and #171 are chores, can merge
… #170
… need reviewers for this
… who understands IDL? Aaron?
jcraig: I’ll review
jamesn: I’ll review as well
Deep Dive planning
jamesn: any deep dive proposals for next week, Oct 13?
… are we done with process for now?
… deep dive on what should go in a review checklist?
spectranaut: no lol
jamesn: if no one has an idea, there will be no deep dive next week
jamesn: no deep dive next weekl
AOM ID Reflection
spectranaut: we can push this to next week because it’s going to be discussed in an AOM meeting later today
sarahhigley: I was looking at this yesterday from an author’s perspective
<spectranaut> explainer: https://
jcraig: this is just for frozen array element reflections
jamesn: any library that is relying on aria-labelledby will need to have new logic to deal with this
jcraig: if something else is conflicting with library and using this other method, then yes, it would need to update. Unlikely to happen overnight
cyns: if you’re buildling up controls all in script, will you be able to have element references like this?
jcraig: if you use element references to assign them, you couldn’t access them through DOM property
cyns: seems less than ideal, what problem is this solving?
sarahhigley: the idea of reflecting at all would be the thing that could cause issues in modern JS libraries
… I’ve never checked an aria-labelledby attribute to see if it exists and what it points to
… what I see happening is the library sets the IDL reference and then somewhere else in JSX, aria-label is set to undefined
… library will first set IDL attribute reference, and then React will have a render cycle... race condition
… reflecting at all seems like it will cause problems
jcraig: even reflecting empty string?
sarahhigley: haven’t used IDL attribute much but whether it’s set to a stale ID or an empty ID either way doesn’t matter
jcraig: please comment on that
sarahhigley: we do extra logic based on label / labelled-by, so we have conditional logic
jcraig: this sounds like a JSX update will be needed
sarahhigley: that makes sense
cyns: need to look at this more, not convinced
jcraig: look back at least 2 issues cross-referenced
jamesn: Alice commented on this 2 days ago
<jamesn> https://
cyns: can bring up at AAM meeting
Consider new abstract role "phrase"
<jamesn> Looking at the roles we are discussing - this is very different from HTML Phrasing Content which includes all sorts of interactives. Using the same term would IMO cause confusion.
<jamesn> I've been trying to find a good term but haven't yet come up with one. Until we do I propose that we park this as changing abstract roles from one to another doesn't sound useful unless we can get the naming right.
jamesn: after working on this, concluded that this isn’t worth doing
jamesn: will leave issue open in case someone comes up with something that makes things clearer
jcraig: recommend closing as wont-fix
jamesn: no objections, so will close. Considered it, not doing it.
Matt_King: the benefit of abstract roles is limited to property and state inheritance?
jamesn: only useful for us as creators of the spec
Matt_King: the benefit to us is property and state inheritance
jamesn: if abstract role were not in spec at all, I’d almost be in favor of that
Matt_King: it does actually provide us quite a lot of value
… for example, all landmark roles
… division of roles between structures, widgets, windows, etc. is also useful and we leverage that in spec
ACTION: jamesn to close #908 as won't fix
<trackbot> Created ACTION-2169 - Close #908 as won't fix [on James Nurthen - due 2022-10-13].
chlane: doesn’t dialog also depend on abstract role?
Matt_King: yes, another example
[normative][author] clarify widget in trees can only be group and treeitem
jamesn: we have 1 approving review
… confusing and needs discussion with scotto and pkrautz. let’s move on
Support aria-description
jamesn: I’ve made this mergeable so we have a preview available
jcraig: it’s not ready yet, leave as a draft
jamesn: okay, will convert to draft
jcraig: okay to remove all other assignees besides me
1.3 triage
jamesn: I’m going to take initial pass, probably simple change
… Matt_King, any progress?
Matt_King: yes, closer to PR, very high on my list
spectranaut: won’t do #1159 for another month
jamesn: #1150, need to figure out each use of descendant
jamesn: we don’t currently define “descendant”
… flat DOM, shadow DOM, accessibility tree?
… problem in ARIA 1.2, so need to put it in 1.3
… please help