Meeting minutes
https://
https://
1. We change the definition of "prohibited" under Accessible Name Calculation:
Change - "the element has no name" to "the element does not support name from author"
joanie: How many folks are as immersed in this as aaron, scott and me?
<Jemma> +1 for change
joanie: in ARIA 1.2 we made a few changes that we need implemented
joanie: 1 change is that there is a list of roles where accessible name is prohibited... texty roles
"prohibited: the element has no name. Authors MUST NOT use the aria-label or aria-labelledby attributes to name the element."
joanie: changed the code, nuked the name, broke the universe
joanie: seemed to boil down to interactive generics
joanie: div with tabindex and div with contenteditable
joanie: oh great - maybe we don't have to test - as an authoring thing and the validators can do that
joanie: jamesn pointed out that other stuff in 1.2
jamesn: I do not see "User Agents MUST NOT expose an ARIA attribute to the platform accessibilty API when the element's computed role prohibits use of the attribute."
aaron: I ran into a name generic one - where it wasn't the authors stated role
aaron: for role presentation and role region, if there is no name then it not a region.
scott: was talking about aria-roledescription
aaron: we never really covered the computed role
jamesn: the 1.3 text does cover that
<joanie> https://
Action: file a 1.3 issue for what to do with global properties that are prohibited on role none/presentation
jamesn: why did we introduce name prohibited
joanie: lets get resolution on the 1.2 issues
joanie: I don't have to test that because there is no prohibition
joanie: we should tripple check... 1.2 CR we have name prohibited and a list of roles....
joanie: Author MUST not
joanie: Also says "the element has no name"
what does that mean?
there are a bunch of tests that are failing
in terms of end users.... I don't think that a <p> with aria-label should have an accessible name
paragraphs that are non interactive - are they even nameable
scott: once focusable comes into the mix then nothing was getting exposed
melanie: I tell folks they are wrong if they do that
bryan: yes
joanie: if you are a good author... user agents try to make things suck a little bit less for end users
joanie: user agents are going to caluculate name from content
joanie: if a screen reader user tabs to it then have a nameless thing
joanie: try to correct
melanie: would it be possible to throw an error
scott: have issues with how permissive html is
jamesn: have browsers ever considered standards compliant mode?
aaron: -1 2 different modes increase complexity
scott: The element has no name... we were primarily concerned about stopping authors from doing this
<joanie> https://
jamesn: the accname change is the one that changes the naming calculation.
carmacleod: Joanie proposed change is clearer
scott: can we make that change without having to reenter CR
jamesn: Michael?
MichaelC: I don't see any reason not to
<joanie> https://
MichaelC: No need to re-enter CR
<Jemma> +1 this is editorial
scott: does this unblock you?
<carmacleod> +1 for changing "the element has no name" to "the element does not support name from author"
joanie: Authors MUST NOT statement is not in 1.2 CR - was added later, so I don't have to test that
<Jemma> + 1000
joanie: going to remove "the element has no name"
jamesn: nothing in the spec that prevents error correction
scott: didn't want to make it explicit at first
scott: can disuade authors from this antipattern
joanie: a reasonable implementation for non interactive, non contenteditable we don't aboide by aria-label but if interactive then user agent corrects