Meeting minutes
<WilliamTennisNFCU> presetn+
<WilliamTennisNFCU> I am 850 and cannot access the zoom video meeting
<WilliamTennisNFCU> I get a 403
[New Issue Triage](https://bit.ly/3uLYC88)
https://
<spectranaut> jcraig: to explain: various image file formats have the ability to include different text in the file format. svg has "title" element, could be used as "alt-text". likewise there is some accessible-description fields available for jpg and gif, and the new format high efficiency image format has "alt-t". the goal is ot have something in the acc name spec, and it hasn't been hidden, but it has alt text embedded in the image file, we should
<spectranaut> use it.
<spectranaut> jcraig: I don't think it is specific to html
<spectranaut> matt: does it go into core-aam?
<spectranaut> jamesn: not core-aam, not an aria thing
<spectranaut> jcraig: the spec of each file format should define this
<spectranaut> jcraig: maybe there should be a note that points to all the file format
<spectranaut> bryan: acc name spec does call out svg specifically?
<spectranaut> jamesn: I'd like scott to take a look
<spectranaut> scott: sure
<spectranaut> jamesn: no milestones right now.... because accname doesn't have milestones
[New Issue Triage](https://bit.ly/3uLYC88)
[New PR Triage](https://bit.ly/3lgCDTP)
jamesn: anyone aware of PRs?
Scott has one
jamesn: I accidentally broke the query
https://
Scott: would also like to discuss fig caption. have heard many saying fig caption for name is "awful".
matt_king: it's being handled like a container not a name
Scott: will add fig caption to agenda for next week
[Deep Dive planning](https://bit.ly/aria-meaty-topic-candidates)
matt_king: would like to lead a grid/listview discussion in a deep dive... timing important with TPAC coming up.
should have Sarah H, Stefan, Corey(sp?)
matt_king: could wait until after TPAC... will arrange a time.
cyns: Accname #145 would be a good issue for after TPAC
cyns: use 9:00 slot for TPAC prep
<jamesn> https://
jamesn: breakout sessions one week, then joint WG meetings in the next week
matt_king: planning breakout session for screen reader automation
Scott: There is a WhatWG issue for `focusgroup` attribute... ARIA group should familiarize themselves with it before TPAC
https://
Bryan: If AccName talks needed, I will NOT be available Oct 20–22.
jamesn: issue with ReSpec came up. The license wasn't what we were expecting...
<jamesn> This Working Group will use the W3C Document license for Recommendation track deliverables and the W3C Software and Document license for Note track deliverables and licensed non-TR publications.
Current charter ^
Most other charter use Software & Document License.... difference is in the ability to fork the work.
MichaelC: You can't make a derivative license under our current license.
jamesn: should we keep current or move to the more permissive license?
WilliamTennisNFCU: Would worry about opening ARIA license.
MichaelC: Derivative license not as concerning as Derivative+Conflicting.
If we were to change, we'd need to re-charter.
jamesn: ARIA, APA, and AG are the only WGs I've found that use the W3C Document license.
does anyone strongly prefer one or the other?
jcraig: would need to review to represent Apple's opinion.
cyns: also true for Google and others
<MichaelC> https://
MichaelC: Charter went to the AC already
Also to horizontal review
jamesn: charter the same... just moving the AAM specs to living standards
WilliamTennisNFCU: Why does Marcos want the license change?
MichaelC: likely so he won't have to support both... Phillipe doesn't have a strong opinion so it's up to us as the WG.
last plan of record is that these licenses are intended to co-exist
jamesn: unlikely that someone would use the license change to create a different, conflicting standard
MichaelC: and W3C licensing is available in other ways
https://
* [Expectations for aria-checked on HTML checkbox](https://github.com/w3c/aria/issues/1622)
Scott: in testing, I've found aria-checked=mixed is not support on native HTML checkbox. makes sense, IMO.
aria-checked doesn't match the boolean nature of the HTML native checkbox.
boolean `checked` is in direct conflict with aria-checked=mixed, so the HTML attribute "wins" as allowed by the spec
potential solutions, 1. add note to aria to use the native indeterminate IDL attribute to expose the mixed state (my preference)
2. special case for aria-checked in the conflict w/ host lang section
<jamesn> add a note to ARIA indicating that if authors need to expose a native checkbox in the 'mixed' state, that they need to use the host language feature to do so (this is my preference).
<jamesn> or, i would think that a special case for aria-checked needs to be made since it has this unique feature of combining two native states into one attribute.
<jamesn> OR, create an aria-mixed (or similar) attribute, deprecating the mixed value from aria-checked to mitigate this special case.
OR, 3. create an aria-mixed (or similar) attribute, deprecating the mixed value from aria-checked to mitigate this special case.
<jamesn> (from the issue - the 3 possibilities)
cyns: I think there is a fourth option... would HTML be willing to do a mixed state checkbox again
another issue has been raised for a native mixed attribute... also getting pushback.
Stefan: is it good to "mix up" thing here: apply aria attrs in addition to the HTML attr
Scott: generally, people should use the host language features... this one in particular confuses authors because there is not obvious way for authors to get a mixed state on a native checkbox
jamesn: strong native semantics section of ARIA for attrs
cyns: implementations have ways around.... e.g. <input type=checkbox role=checkbox aria-checked=mixed> works in Chrome
Scott: Firefox follows the spec more strictly and ignores aria-checked=mixed on <input type=checkbox>
happy choosing any of the solutions.
cyns: I think adding the note (option 1) is reasonable... Can we also purse native mixed (option 4)
<cyns> https://
latest discussion for native mixed ^
<jamesn> "Authors SHOULD NOT use the aria-checked attribute on any element where the checkedness of the element can be in opposition to the current value of the aria-checked attribute."
Scott: I think it should change to MUST NOT
wanted to bring to the group for opinions first
various: general agreement in favor of change to MUST NOT
no need to change ARIA spec because validators can fail on the HTML-AAM MUST NOT
<jcraig> s/AIRA/ARIA/g
Scott: should I add the note (option 1)?
jamesn: Add a note that authors should not rely on the Chromium behavior... need to file bug... Will also file WebKit bug for indeterminate attribute.
[Updating ARIA 1.2 due to IDL implementations (exit and re-enter CR?)](https://github.com/w3c/aria/issues/1598)
<spectranaut> jcraig: main objection is to the term reflect
<spectranaut> jcraig: lets put a roadmap that says we want to get to aria 1.2, but we will address it. for now, let the spec mirror the implementations
<spectranaut> jcraig: it sounds like it should be ok, if they object, they can make a comment here
<spectranaut> jcraig: I talked internally, and looked at our implementation, cynthia and maybe aaron had concerns about implementation complexity. I think it is a quick and easy change in all three implementations. the complexity was not about nullable strings, but about other IDL types, like non-string types
<spectranaut> cyns: I thought there was complexity around sprouting attributes
<spectranaut> jcraig: for 1.2, match what is currently there, don't use term reflect
<spectranaut> jcraig: for 1.3, we could consider making it non-nullable reflective dom strings
<spectranaut> jamesn: we don't want to do that tho right
<spectranaut> jamesn: aria-description has a usecase to override a title further down the chain, it should be an empty string which is different from it not being set
<spectranaut> cyns: and alt should be able to return an empty string
<spectranaut> jamesn: aria-description="" should override a title?
<spectranaut> jamesn: override bad tooltips
<spectranaut> jcraig: we should mention alt in this issue
<spectranaut> jcraig: we should say alt is broken in the html spec
<spectranaut> jamesn: alt is doing what the spec says, but it's not useful.
<spectranaut> bryan: I found a weird case where aria-description="" comes in handy. custom slider where aria-valuetext represents the time of day
<spectranaut> cyns: should we deep dive next week?
<spectranaut> jcraig: I can't make next week
<spectranaut> jcraig: we need a deep dive with ana or dominic or ??
<spectranaut> cyns: I can put one in for the other IDL thing, we can have an IDL session
<spectranaut> jamesn: I'm ok with aria-label being a nullable string, but for the ones we do believe, like aria-decscription. anything enumerated should have nullable dom strings
<spectranaut> jamesn: html should all it
<spectranaut> jcraig: calling out alt is a good example since Anne objected to crossorigin attr
<spectranaut> cyns: I'll make a tpac session! I'll need help with agenda
<spectranaut> jamesn: can we get rid of the term "reflect" jcraig ?
<spectranaut> there are general agreement to not take out the IDL section.