Meeting minutes
<chlane> scribe:
New Issue Triage
ARIA 1844,
not a required owned element per JAws test
scotto: not a required element, sarah's pr make alleviate
wilcos other issue fall under same bucket
scotto: will respond
aam 445
special description computation
scotto: get on agenda
related to updated score accname desc
special conditions for html elements, name becomes description
why do these elements get special treatment?
New PR Triage
pkra: how shall we treat issues with updates?
look at issues with updates, the token thing, could be deep dive, editors meeting
spectranaut: plus the issue
Pr aam 15
updates on mappings
reviewers?
html aam 444
further alignment w/accname desc updates
scotto: did second PR to align with core aam description PR
we should be able to merge them, this one and updated core aam pr
aam 443
reference to placeholder in naming comp
it should be html aam 443
jcraig: draft pr
spectranaut: no way to end up back in our triage agenda
jcraig: how do you take it out of the new pr triage
spectranaut: it just looks for new prs
jcraig: its fine to leave it on list
pkra: updated prs
as a secondary step
Deep Dive planning
pkra: flow to discussion
spectranaut: next week?
conflicts ....
sheduling
alt time 4pm pacific?
jcraig: can do next week
pkra: midnight for peter
jcraig: 12/8 normal time?
scotto: it has been this way for a while
scotto: flow to 12/8
ARIAMixin and ElementInternals
spectranaut: can anyone introduce this?
jcraig: additions with IDL in ARIA spec
spectranaut: not a lot to discuss
pkra: theres going to be work on the week of 12/7
in the other issue matrix chatlog
cyn aom meeting this afternoon
short notice
putting on pending agenda items
scotto: per domonics comment in html issue, there doesn't seem to be a spec that identifies how aria checked different states values be mapped
done in aria, html aam , core aam,
jcraig: have to look into possible a resuable pattern referenced from html
stream reflection, token list
enumerables
prefer no specific processing markup algo
rather in a broader spec like html
Define how custom elements are exposed to AT
spectranaut: similar, also from anna
scotto: i don't think it belongs in accname
either aria issue or html aam or core aam
comes down to how the other issue is resolved
cyn scheduling sync with Anna
spectranaut: PR against accname for a month
PR 170 need reviewers
melsumner: working on it
some discussion about alignment, need more feedback
meeting later today
spectranaut: accname 170
jamesn: its a different thing
melsumner: bryan an i talked earlier made a list
jamesn: meeting is morphing
melsumner: need opinions for context
jamesn: if there is an agenda add it to the meeting
melsumner: weekly meeting to get a handle on what should be there and what we need opinions for
spectranaut: this could still be reviewed by other requested reviewers
deep dive on accname today
4pm
Support aria-description
updates from scott
scotto: context, there was original PR, went through html aam's description computation, previous PR simplified it, more alignment, took out the rows in the middle that called out specific html features replaced by a single row
svg aam, math aam ... will be included in the future
spectranaut: mel bryan time to take a closer look
melsumner: sounds good
1.3 blocking issues agendabot]
jamesn: definitely review ask what was intended
jamesn: looking at 6 blocking issues
1150, sarah put out an awesome gist
jamesn: is on me to resolve unless anyone else wants it
spectranaut: now is the time to take one from james
jamesn: will get to it sometime
1159, inconsistent use of 'hidden' valerie won't be able to get to it until testing is over
1161 inconsitent use of owned by 1161
jamesn: we discussed it
2 PRs
sarahs work makes this irrelevant, by PR for the required owned elements
to allow owned elements
scotto: most issues will become irrelevant
jamesn: they will be new issues
<jamesn> https://
matt, close issue with no action?
matt asked for specific list of three changes are what people are expecting in a Pr
if we do one
scotto: also not sure I understand why this needs to be changed
that makes aria controls do what aria-expanded does
jamesn: matt, follow up with wilco and close
jamesn: give wilco a few weeks
matt, downstream work on user agent side
jamesn: benefits and costs
jamesn: changes of what roles ... could break the web
have path forward
leave it as it was
add editors notes
about naming correctly
work with devtools to flag
that they are doing error correction on accessible naming
rather than following the spec
cynthia, how do we get that done?
matt, we do not want the computed acc name in devtools include error correction
jamesn: it should be the same as exposed to a11y apis
matt would he awesome
cyn, happy to take that issue to devtools
assign to cyn
jamesn: or a new issue
is there a better method
cyn best way to start
jamesn: hoping we are happy maintaining status quo with editor docs
scotto: need to provide evidence that things dont actually break
matt, james you are going to right text
jamesn: yes, will make 1.3 is the same as 1.2
1824, presentation role conflict
all prohibited states and properties are also global
scotto: related topic
scotto: touching on, imgge, aria-label
or alt with text
setting role none presentation
should be treated as aria-hidden=true
meaningful text with role=none, presentation pruned it from a11y tree
is treating role=img as aria-hidden=true the same correct?
Matt_King: focused elements, it is error correction, mixed accessibility signal, do the least harm for user
I see why it is related
<melsumner> I really don't think we should auto-correct author errors, can we explore a console error?
Adam_Page: why role-none on the image?\
jcraig: author error
scotto: aria spec says it
guidance says its the same thing
siri: if you have aria label used on div and screen reader doesn't read it
why on div element
jamesn: authors must not
scotto: because authors. [will do it.]s are the answere
Matt_King: feels like same issue conflicting a11y signals
jamesn: if prohibited does that make it be global
it should still reverse presentational aspect
matt, we could write it but we should have for clarification in the presentation role section
none is already sooo complicated
jamesn: section to explain prohibited states and properties, still global for conflict resolution
matt include example
jamesn: they would be bad examples
jamesn: funky, should not be copied
scotto: makes sense
jamesn: conflict resolution is a mess
matt, involve normative user agent statements
<melsumner> it would be great to have a good example of code in our spec, and then say "note here's a weird edge case" but make it look like "hey you should avoid this"
scotto: will file issue
scotto: negate author issues
jcraig: added, a tag 'needs automated test'
whatever we decide, we should file a separae issue, great one to automate
jcraig: anything that relys on label or role computation can be tested
jamesn: is there documentation on how to write those tests
spectranaut: its the issue for writing tests in the accname repo
jamesn: use webdriver