W3C

– DRAFT –
ARIA WG

28 September 2023

Attendees

Present
Adam_Page, Cory, Francis_Storr, giacomo-petri, jamesn, jcraig, MarioBatusic, melsumner, pkra, Rahim, sarah_h, scotto, Siri, StefanS
Regrets
-
Chair
JamesNurthen
Scribe
Rahim

Meeting minutes

New Issue Triage

<melsumner> maybe a different label, future-consideration

jamesn: For https://github.com/w3c/aria/issues/2046, no real need to denote content is AI generated; no strong impetus to do something with this at the moment

jcraig: One main challenge is autogenerated descriptions; this is interesting because alt could be blank, or AI generated alt as separate attribute. Or, specify set of things AI generated about a particular (a11y) node. The important part is which parts of this page, or node, are AI-generated.

scotto: May have practical value in announcing ai generated content at the page level; if content is AI generated, it should be conveyed as such visually and programmatically. I'm OK with a metadata conveying something is AI generated however, must be exposed that way for everyone. There's a separate conversation about AI vs author generated text alternatives

<melsumner> +1 for this being an HTML attribute

jamesn: Need to present same information (that something is AI generated) to everybody

melsumner: Should be handled by HTML markup

New PR Triage

<jcraig> looks like the WPT PRs have fallen off the agenda

Decide what "draft PR" means

scotto: Times that I've submitted PRs and need help finishing them and then they turn into "draft PRs"; if you don't need a review, don't submit the PR. Could have a separate label

jamesn: Will assign to myself/Valerie for future discussion

aria-controls spec update

jamesn: Waiting on my review

jamesn: Could have a "not ready for merge" label

Clarification on aria-braillelabel SHOULD vs. MUST guidelines

pkra: Ultimately, I agree with this issue (that they should be MUST)

<melsumner> @jcraig yes will do, thanks for the reminder

<Adam_Page> jcraig, ack, yes will do

jcraig: Might be some scenarios where you want to mix braille and plain text; worth considering separating conditions for aria-braillelabel usage requirements

jamesn: Looking at condition 2 (the value of aria-braillelabel is not empty...); is there a scenario where a label is redundant because the value is known (e.g., currency)

Cory: Everything that's been a braille error was also a visual copy error

pkra: Worth testing braille pattern characters across platforms (VoiceOver, Orca)

pkra: Consider an image with an empty alt, how is name (and braillelabel) considered

jamesn: Move to a future (ARIA) version?

Rahim: No concerns on my end with leaving spec as is and closing it

Support aria-readonly on rows in grids

MarioBatusic: It's a good idea to have ability to put aria-readonly on an entire row; might have instances where you can see data but aren't able to modify it

scotto: Should aria-readonly propagate to child controls (e.g., form controls in a grid cell)

<melsumner> side question: are there good examples of reasons to use aria-readonly instead of the html attribute?

Adam_Page: RE: Scott's question on inheritance of read only state, does this apply to the disabled state as well?

scotto: There's many "worms" (edge cases) to consider, e.g., putting required on a <fieldset> and how state propagates

scotto: It's a real problem, putting required/disabled on elements and what the actual author intent is

Siri: We use row/grid on tables with interactive elements; if there are no interactive elements, does this automatically mean they are "read only"?

melsumner: Is it because a grid cell can be editable; in such case, you may want to use aria-readonly on a cell

<Rahim> s/No concerns on my end with leaving spec as is and closing it /Rahim: No concerns on my end with leaving spec as is and closing it

Minutes manually created (not a transcript), formatted by scribe.perl version 221 (Fri Jul 21 14:01:30 2023 UTC).

Diagnostics

Succeeded: s/(a11y) node/(a11y) node. The important part is which parts of this page, or node, are AI-generated./

Failed: s/No concerns on my end with leaving spec as is and closing it /Rahim: No concerns on my end with leaving spec as is and closing it

Succeeded: s/No concerns on my end with leaving spec as is and closing it/Rahim: No concerns on my end with leaving spec as is and closing it/

All speakers: Adam_Page, Cory, jamesn, jcraig, MarioBatusic, melsumner, pkra, Rahim, scotto, Siri

Active on IRC: Adam_Page, Francis_Storr, giacomo-petri, jamesn, jcraig, melsumner, pkra, Rahim, sarah_h, scotto, StefanS