Meeting minutes
<jamesn> May 13: Data Viz Pt 2
<jamesn> May 20: Accessible Name Prohibited
[New Issue Triage](https://bit.ly/3vKCtGM)
https://
https://
jamesn: Scott's discussion on slack
Scott: not necessary now
jamesn: 1.4 or not at all
https://
already in 1.3
https://
also triaged
[New PR Triage](https://bit.ly/33juvIs)
<jamesn> https://
Scott: PR after last weeks discussion
jamesn: reviewers?
jcraig, siri, and mark m
[New Issue Triage](https://bit.ly/3vKCtGM)
[Meaty topic for next week](https://bit.ly/2RwBvic)
deep dive: dataviz next week
"accessible name prohibited" #324 the following week
Scott and Joanie, please help clarify the intro/prep for that discussion
[Feedback on ARIA in HTML](https://lists.w3.org/Archives/Public/public-aria/2021May/0000.html)
jamesn: thanks Cynthia for reviewing this doc
do you propose these as personal or WG comments?
cyns: Leonie asked for group comments
and create issues
<jamesn> * Why can an <a> with no href take any role but <area> with no href can take no roles? Shouldn't these be consistent? Seems like <area role=button> would be useful.
<jamesn> https://
cyns: and anchor with no href can take any role, but an area with no href cannot
jamesn: sounds like a reasonable issue to file
<jamesn> * Should <iframe> support role=dialog?
Scott: agreed
cyns: should iframe support dialog role? these are used as dialogs sometimes.
Scott: cross-origin iframe should have the close button outside the iframe rather than one the frame
<jcraig> s/Mozilla-specific problem/there are Mozilla-specific implementation problems/
jcraig: if people are using this for non-x-origin dialogs, the question stands
<jamesn> * should <img> support role=figure?
Scott: yes, file the issue
jcraig: and with SVG where child nodes can be accessible
Scott: is it okay to mask the default role of img?
group: yes file
group agrees to file any editorial issues
cyns: found the TR text confusing. would like feedback here.
(TR = table row in that context)
[1.3 triage](https://bit.ly/3en3k5W)
https://
aria-haspopup needs to spec default of false "if the attribute is not present"
GitHub process discussion
cyns: assign to me (thanks Cyns)
https://
https://
jamesn: both to 1.4
https://
<jamesn> https://
Scott: html is not case sensitive, so author *could* use camel case.... Presumably it should just work.
If it works in all user agents, we need to make sure the validators also accept it
cyns: allow but don't require
Scott: would be useful, with testing, to remove the authors MUST NOT
jamesn: can you link an ARIA in HTML? will push this one to 1.4
https://
jamesn: this is not resolved by aria-label/labelledby being prohibited. Any other global attr would do the same. We've recently moved this to "resolving author errors" section... Does that resolve Wilco's issue?
Matt_King: "prohibit" means "MUST NOT" Error versus the validator SHOULD NOT warning
jamesn: moved it to handling author error...
Action: jamesn will discuss #1288 with Wilco...
https://
jamesn: 1.4
https://
Matt_King: tied to NVDA bug... might be overcome by events
comment to Sina or James S
https://
jamesn: 1.4
https://
jamesn: assigning to myself
https://
1.4
https://
Matt_King: editorial wording change
StefanS: aria is missing a ~"grid-like layout role" maybe "layout grid"
https://
1.4
https://
1.4
https://
Matt_King: could not find justification to keep it.
jamesn: assignee?
Action: jamesn to comment #1369 that William will take assignment
Action: MichaelC to add William and Jaunita to the org and group
<jcraig> s/action: jamesn to add/action: MichaelC to add/
https://
1.4
https://
Scott assigned for 1.3
https://
assigned to cyns