Meeting minutes
New Issue Triage
spectranaut_: JSON roleInfo.js not updated
pkra: these only get updated on recommendations
… but we can do them more efficiently after PR
jamesn: is it possible to publish as part of TR?
… if so, that could be the versioned copy
spectranaut_: let‘s discuss in editor’s meeting
spectranaut_: aria#2233
jamesn: this seems like a dupe
… is it aria#1389 ?
spectranaut_: will you confirm that this is a dupe, giacomo-petri?
giacomo-petri: yes
spectranaut_: aria#239
aardrian: Chrome may be changing — will dump link into this issue
<aardrian> There is recent activity on this Chromium issue, though not sure it helps:
<aardrian> https://
sohara: the CSS WG needs more a11y participation
jamesn: it’s hard for them to know when a particular issue has an a11y impact
aardrian: I volunteer
<aardrian> I appreciate Adam simplified my response.
spectranaut_: core-aam#234
Rahim: do we have documentation on the relationship between content in the DOM and the a11y tree?
jamesn: it’s too difficult to document it in a standardized way for all browsers
spectranaut_: core-aam#234 is editorial
spectranaut_: then we have some follow-up tasks for the monorepo migration
spectranaut_: aria#2232
… previews will take a little longer to load, but we’ll be fixing that soon
… skipping over some more editorials
… pkra’s made an issue for editors to review and finalize
… another editorial that we‘re trying to remove references to custom css files
spectranaut_: aria#2225
giacomo-petri: ARIA spec is clear from a user agent perspective, but not for authors
… when IDREF references an invalid ID
… I recommend it is important to clarify that there is an author issue if the ID is missing or has not matching element in the page
Rahim: I raised a similar issue on `undefined` and see some overlap
… how an author provides or doesn’t provide a value
… empty string, undefined, null, etc.
… there’s ambiguity in the spec on those assignments
spectranaut_: I’ll link that
… do these proposals conflict with that?
Rahim: the empty string case is clearly specified, IMO
keithamus: you said there’s a discrepancy between undefined and null. I’m curious about that
… if it’s serialized HTML, it has no concept of undefined or null
New PR Triage
spectranaut_: a ton of PRs moving to the ARIA spec
… nothing to discuss
WPT Open PRs
Deep Dive planning
spectranaut_: any deep dives to schedule?
aardrian: I missed the ariaNotify and want to catch up. How can I?
jamesn: please send me an email and I’ll send you the recording
… and don’t distribute
daniel-montalvo: you may need Google Drive access
… just request it if true
Additional editor of HTML-AAM - Rahim
<TheoHale> jamesn can you send me the link to the meeting/deep dive.
spectranaut_: please everyone welcome Rahim as a new HTML-AAM editor!
CSS: handing the a11y of anchor positioning
<jamesn> TheoHale: please send me an email
sohara: we don’t have a place to to outline the CSS a11y semantics
… anchor positioning is a new feature meant to be used for popovers
… currently, position:fixed in top layer of browser and you have to use JS to position it correctly
… this CSS property will associate the elements together so you don’t have to micromanage the location
… we had talked about there being an `aria-details` relationship with this property
… so that if someone used it on a non-popover element
… we could add an aria relationship to communicate that
… we don’t have CSS-AAM, so maybe we can carve out a section in HTML-AAM
… after the monorepo migration
spectranaut_: any questions?
… is there anything controversial here?
aardrian: not for us, but need to read it
… may prove challenging for devs who are not familiar with surrounding concerns
sohara: part of this is making sure there’s a minimum layer of a11y
aardrian: predictable, if nothing else
sohara: please poke at it
… post in the issue
… I or aaronlev will respond
spectranaut_: we spitballed making a CSS-AAM but ended up decided to stick with HTML-AAM
sohara: for now
… might still be a good idea to spin up, but then definitely need a bridge to the CSS WG
Issue: CSS Highlights Not Clearly Documented in AAM
spectranaut_: may not be anything further to discuss here
TheoHale: I need to do more work here
… but this is a case where CSS did do some good a11y work, but didn’t document it
spectranaut_: should this issue actually be in HTML-AAM?
… I’ll move it there
Triage Issues...
spectranaut_: aria#482 from Nov 2016 (!!!)
jamesn: we don’t need a feature for the href attribute
pkra: this ties into attribute parity
… is this something that will come up again?
… should we have a marker for saying “maybe, in the future, but right now, no”?
… but I do support closing this
jamesn: why would anyone ever want to set a different href for the a11y api?
sohara: one use case I can see, and this is not a priority...
… someone makes div role=
<pkra> apologies for jumping the queue!
sohara: someone makes div[role="link"] and wants to set a href
… they wanted to make it as close to a hyperlink as possible, but don’t want to use href
… most people using role="link" are using tons of JS, and this wouldn’t help them
sarah: I agree with pkra
… “in theory, this might be needed, but it isn’t now”
… in sohara’s example, other functionality is sabotaged, like opening link in new window from context menu
spectranaut_: so we need to make it easy to find for someone who might log a similar issue in the future
… so have a label like “not a priority”
aardrian: I would still close it
spectranaut_: let’s pick a label
… let’s go with "wontfix"
spectranaut_: aria#483
pkra: is this for html-aam?
spectranaut_: “not mapped”
pkra: is there something to map it to?
spectranaut_: we’ll close as "wontfix"
spectranaut_: aria#484
jamesn: close it
spectranaut_: closed
spectranaut_: aria#517
pkra: blocked by jcraig’s comments
sohara: there are platform roles now
… I don’t think we need this
jamesn: it was only ever for test reasons
pkra: "wontfix"
jamesn: we could just use aria-roledescription, for example
spectranaut_: closed as "wontfix"
spectranaut_: aria#521
… roles for drag and drop
TheoHale: is it actually helpful information for AT users?
sohara: propose closing
spectranaut_: closing as "wontfix"
… aria#524
… alllow aria-setsize on container elements
sarah: I can speak about this
… maybe this can be folded into the other PR
… for toolbars
… ref either aria#2158 or aardrian’s PR
… aria#2170
… so we could close this in favor of the open PR
spectranaut_: closed
spectranaut_: aria#542
… new attribute aria-readonly for role=form
sohara: I’d argue we don’t need this, it should start in HTML first
… and HTML doesn’t allow a form to be readonly
… recommend closing
<aardrian> +1 to scott
sohara: if you want it to be readonly, just use text
… or readonly inputs
giacomo-petri: I have an example
… checkboxes don’t support readonly
… in learning use cases, they sometimes want to show a checkbox but in a readonly state
… and disabled doesn’t work well
… edge case scenario
spectranaut_: maybe this should be pushed into HTML
jamesn: I agree; there are already options to make it readonly
spectranaut_: closed
spectranaut_: aria#574
… role="switch" should have additional attr `aria-valuetext`
sohara: I have a dupe for this
<sarah> w3c/
spectranaut_: will close this as a dupe, since there’s more chatter on the new one
spectranaut: we closed every issue we looked at!