Meeting minutes
[New Issue Triage](https://bit.ly/3wlqKig)
jnurthen: issue 1500
… do we call this a won't fix ?
… or anyone disagree?
bryan: I'll take a look
jnurthen: accname issue 132
… will assign to Melanie
… accname issue 131
… is it a duplicate?
bryan: will do.
[New PR Triage](https://bit.ly/3pOjIAx)
jnurthen: aria PR 1498. please take a first look
[Meaty topic for next week](https://bit.ly/3ivpVzm)
jnurthen: deep dive next week.
… Aaron can't make it for accname discussion, so probably not a good idea.
bryan: would be good to have as many as possible.
jnurthen: any proposals? Otherwise we skip
… e.g., work session on author tests
… ok. currently no proposal. accname deep dive for week after.
[Problematic mappings for <a> without href](https://github.com/w3c/html-aam/issues/333)
scott: comes back to "generic inline" etc.
jnurthen: it's not really the role that makes it generic inline or block.
scott: right. comes down to platform mappings. One currently says it maps to text.
… that's why I'd like to get consensus from the group
jnurthen: what does span map to>?
joanie: ATK maps it to generic.
… if my platform is the only blocker, I'm ok with changing it.
scott: a no href for IA1 maps role system text / text frame
… but we have different mappings, same with div
joanie: maybe we can check with NVDA people. IA2 are parallel to ATK mappings. So ask them if it's a problem.
… if they agree, then we can make progress.
… as a follow up conversation, make inline vs block generic a separate issue
… taking CSS into account
… I think we can take CSS into consideration for platform roles
scott: sounds good with me
aaron: do we agree that a without href can just map as span with id?
all: sounds good
<siri> agree as <span> as it is used a place holder when we <a> without href
aaron: speaking with James Teh, they use a different role for div vs span
… he agreed it was kind of nice but probably didn't matter.
… and we can tackle block vs inline (and other display) separately
[Mappings for body and html don't seem to match reality](https://github.com/w3c/html-aam/issues/330)
jnurthen: do we have what we need for this?
joanie: currently, we don't match reality. We could match but what would the mapping of body be?
… scott had suggested generic
… but aria-label then prohibited, brought in "UA not to expose it" discussion
… but authoring error and resolution as follow up
… which seems ok
scott: I'm still for generic. naming is not a good idea.
jnurthen: what about matt's recent comment?
joanie: I disagree. it's generic, authors shouldn't put labels on body.
… does anyone think it's a good idea?
cynthia: at most: iframe scenarios
joanie: same as with div
… if we think authors should not / must not do it, then validators will flag it
… but UAs can handle it
scott: and validators already flag it.
jnurthen: so we map it to generic?
… no objection
[Exit Criteria Testing: Need test cases and results for Authors MUST/MUST NOT statements in ARIA 1.2](https://github.com/w3c/aria/issues/1492)
jnurthen: a deep dive / work session would be great.
joanie: I could do that. Tried locally, throwing at aXe
cynthia: I could, too.
+1 from pkra
jnurthen: ok, then let's schedule a work session for deep dive next week.
cynthia: could use a refresh on test framework
joanie: couldn't find a great way to automate things yet. But yes, let's talk.
jnurthen: Harris should be able to help. Hopefully at the meeting, too.
… I won't be there though.
[When is hidden content taken into calculation of name and description?](https://github.com/w3c/accname/issues/57)
+1 to that
jnurthen: not a lot of comments on it so far.
… maybe start with joanie's tldr
<jamesn> https://
cynthia: I think I remember some history.
… the original idea for allowing hidden content was to have accessibility API only content.
… initially, all visually hidden was disregarded always.
jnurthen: is there a real world use case for this? I can make some up but I don't know a real one.
… children of something that's hidden not getting exposed.
… e.g. tooltip via describedBy, not visible when focus is moved (delayed) there but you still want it as description.
… you still want that child span to be part of the accessible description.
cynthia: I wonder why it has to be so complicated though.
siri: is this similar to screen-reader only text?
jnurthen: not quite.
<jamesn> https://
jnurthen: joanie's next comment.
<jamesn> <div hidden>
<jamesn> <i id="desc">This is a description.
<jamesn> <b aria-hidden="true">This should stay hidden.</b>
<jamesn> <b hidden>And this should stay hidden too.</b></i>
<jamesn> </div>
jnurthen: div hidden and being referenced, div with two children additionally hidden
cynthia: first reaction: everything but aria-hidden
scott: a situation where I wouldn't want hidden ones exposed would be a input with aria-required and a hidden label, I might want that.
<siri> + scott
jnurthen: but that'd be aria-hidden again
scott: right.
cynthia: the only scenario I can come up with is a dynamic tooltip. Use hidden to get rid of visual stuff and use aria-hidden to hide it from accessibility tree
scott: yes, I can imagine that.
siri: sometimes you put aria-hidden on chevrons etc. because they add too much noise.
jnurthen: so nobody expected it to work the way the spec says today?
… and what about implementations?
joanie: IIRC James had mentioned not wanting to accidentally dive into a huge tree.
… we should check on that.
… my colleague tested it and it seems to prune it all.
jnurthen: do we need a way to not prune that stuff? E.g., aria-hidden.
cynthia: what's chromium doing right now?
<jamesn> https://
joanie: I think Case 2 was only baz in the tree.
jnurthen: comment has some results.
… so chromium matches what you suggested
… firefox ignores both types of hidden
… webkit we'll have to double check.
… can we get agreement from implementors?
… real world use cases would greatly help here.
cynthia: maybe complex tooltips?
cynthia: so for labels, can we agree that aria-hidden is to prune it from accessibility tree, display:none from visual tree?
jnurthen: everyone, please comment on the issue.
… more comments will help inform these decisions, especially from real world examples.