Meeting minutes
New Issue Triage
jamesn: #1845
scotto: prevents possibility of img role="none" but with alt text from being exposed
jamesn: not exposed as a node in the tree at all?
scotto: from quick tests, yes, pruned from the tree. SR couldn’t get to alt text at all
jamesn: how to classify?
jcraig: would like interactive images to be _a thing_ in the future
jcraig: e.g., fully interactive SVG should be exposed as an image with content
jcraig: so want to make sure this proposal is compatible with that
jcraig: whatever the outcome is for this, it should not prevent a presentational image *with* exposeable content nodes
jamesn: need to clarify the “role="none" is equivalent to using `aria-hidden="true"`” language
scotto: this is triage
jamesn: we need clarification
jamesn: looking at 1.4 for this
scotto: I’ll take assignment and keep investigating
New PR Triage
Deep Dive planning
jamesn: deep dive on Dec 8
jamesn: do we want to plan any more?
jamesn: consensus: no
jamesn: can do more next year
holiday meetings
jamesn: we’ll meet next week unless people are out for Thanksgiving?
jamesn: we’ll keep on meeting up until Christmas
jamesn: will figure out which exact weeks we’ll take off
jamesn: if anyone has issues, let chairs know
ARIA F2F 2023
jamesn: set up survey for face-to-face
jamesn: have 8 responses so far, need more
jamesn: want to measure appetite, and rough idea for location
siri: lot of budget cuts
jamesn: will definitely have remote attendance options regardless
Other elements which may need special description computation conditions?
jamesn: scotto, this one’s from you?
scotto: this is related to work on accname and html-aam updating description table
scotto: in doing that, mentioned interesting callouts to Aaron
scotto: e.g., summary element has a weird use case where if it’s given an author accessible name, like aria-label, then subtree should be converted to description
… similar to button type with a value
… that value becomes a description
… kind of weird that we’re doing that
… other elements don’t have this behavior
… needs fixing
… either elements with special conditions should have those conditions removed
… or other elements should consistently get those same conditions
jamesn: or potentially we leave it how it is, and clarify the ones that have those conditions?
scotto: true, but some of these may actually fail Label in Name
jcraig: could add to characteristics table for those roles?
jcraig: would match current behavior now, and then we could modify table if needed
jcraig: would also confront us with the decision for any new roles that come in
jamesn: how would modifying this remediate the Label in Name issue?
scotto: you’re right, they would fail anyway
jamesn: sometimes the visible label contains extra content that gives a bad SR experience
jamesn: we could consider deprecating the weird behavior
jamesn: but people may be relying on it
jamesn: but I still grant that it’s difficult to justify; seems weird
jcraig: if we go with deprecation path, then we should definitely not update the characteristics table
jcraig: I support jamesn’s deprecation proposal
scotto: the only other elements that I investigated with potential could be hyperlinks, but now think that would be a bad idea based on this discussion
scotto: fieldset with legend is a squishy use case
scotto: option element with overwritten label text
scotto: this could expand into those cases
scotto: but if we do the deprecation path, it’s a signal that we’ll no longer do this
scotto: I just don’t get the use case for the elements where it’s been done
jamesn: although I proposed deprecation, I don’t know how we’d actually do that
scotto: need more people to weigh in
scotto: describe the 3 different options we’ve discussed and solicit feedback in the issue
jamesn: sounds similar to what James Teh has brought up
jamesn: this would be a description “guess” that the AT could choose how to expose
scotto: looping this back to triage, it’s a similar concept to img alt role="none"
scotto: maybe your point about having a new prop is the way to go
jamesn: `aria-shrug` ;-)
scotto: I actually like that one the most
scotto: could extend to other elements
<chlane> aria-any
<jamesn> aria-🤷♂️
scotto: becoming convinced that jcraig’s proposal is the way to go
jamesn: this issue does seem to be a repeating pattern
jamesn: we need to talk to AT makers and make sure they can work with it
jamesn: need to set up a discussion with Freedom Scientific, NVDA, etc.
<chlane> I like the additional property idea rather than making the concept of 'description' more complex for authors
jamesn: where is this on our priority list
scotto: to close this up, 2 paths forward
scotto: 1) potential deprecation
scotto: 2) potential new property to funnel this info into
scotto: need to see what appetite is
jamesn: and 3) do nothing, but explain why behavior is there for historical reason
scotto: I’ll write this up
scotto: I have enough context to shop this around
Support aria-description
scotto: needs reviews
jamesn: this is blocking other things in accname, so need to get this in
jamesn: this changes a lot; can’t merge in any other PRs
spectranaut: who needs to review?
jamesn: Bryan and Melanie
Bryan: I’ll review later today
scotto: you would never reference one’s self to get a description
jamesn: can add an editorial note to clarify
jcraig: is there any opinion on my editorial feedback?
… HTML caption is specifically referenced
… we should take that out IMO
… do others agree?
… line 239 in the source
scotto: can’t find in latest version of PR
jcraig: nevermind
scotto: maybe there’s a merge conflict?
jamesn: please move that one along
scotto: just need people to verify
jamesn: reviewers, please review
1.3 blocking issues agendabot]
jamesn: #1150 is on me
… #1159
… any updates?
… #1161
… depends on sarahhigley’s PR
… will look into that
… #1824
… what did we decide for this?
scotto: Wilco opened about aria-label and prohibited
jamesn: need an editorial note to confirm it’s global
jamesn: decided in last meeting, need to look at minutes
jamesn: let’s look at sarahhigley’s PR
<jamesn> https://
jamesn: pinged me for editorial review
… what else are we waiting for on this?
scotto: we should all take one more peek at it
jamesn: it’s not building successfully right now
jamesn: would like cyns, Aaron, and jcraig to give review
jamesn: need checkmarks of approval
scotto: I’ll do another quick review by EOD tomorrow
jamesn: this will unblock a lot of other stuff. Should be our next merge
jamesn: everyone, put this at the top of your list
Open PRs on ARIA
jamesn: let’s adjourn
jamesn: please fill out face-to-face survey, everyone