Meeting minutes
New Issue Triage
A bunch of dpub issues
jamesn: open them and see
jamesn: if editiorial don't worry
spectranaut_: #45 is about roles
jamesn: aaron is dealing with it
spectranaut_: great
spectranaut_: accname 180
testable statements, bryan will look into it
spectranaut_: test cases in wiki?
spectranaut_: will send it
author errors as opposed to user agent errors
New PR Triage
dpub-aam #23
need new editor from dpub
<MarkMcCarthy> chlane: maybe I could help, I have some related prior experience
jamesn: good to have a least 2 or 3
<MarkMcCarthy> spectranaut_: let's have a chat, chlane :)
spectranaut_: pr 1855
jamesn: not sure what is being attempted
Daniel trying to generate respec
jamesn: no merging
spectranaut_: coreaam 161
open pr to discuss msaa roles
if interested take a look
coreaam 160
just updated
html aam 454
scotto
related to popover
peripheral to content editiable
more reviews, introduces where attributes have been added
no longer generic
bumped to minimum role
the safe defauls
some user agents want to go further
div with tabindex should be a button
expose info important to a11y
not just say it's generic 'whatevs'
Adam_Page_: will review
dpub-aam 20
Deep Dive planning
jamesn: we were meant to have one today
sohara:
jamesn: schedule for next week?
sohara: retouching base on aria-haspopup
sohara: make it's use on buttons less problematic
argument to think about it again
no visible indicator of a popup its not needed
get everyone back up to speed on it
reaffirm resolutions
sohara: want us to agree and not confuse us
sohara: next week
spectranaut_: great next week
spectranaut_: any others
TPAC Announcement - Sep 11-15 Sevilla, Spain
bryan, accname traversal property, need to ping james teh
ACTION: jamesn to ping james teh
<trackbot> Created ACTION-2170 - Ping james teh [on James Nurthen - due 2023-01-19].
bryan is out next week
tpac announcement ...
sept 11-15 ins Sevilla, Spain
get your budgets in order
jamesn: pushed for warmer location
jamesn: outdoor time with covid
was great
spectranaut_: will follow up
jamesn: hotels are not expensive
in Segilla
sevilla
Step 2E isn't used for form field + name-from-label when it is the current node in an aria-labelledby traversal
spectranaut_: accname issue from Sara
sarah_higley: it's around whether 2C
which is around what happens when a form field is included in aria label traversal
input participates in a label
form field where the labe is like
'do this until time'
to participate in label of another control
…
true for all browsers
<jcraig> aq+
bryan, agree is doesnt say you should restrict the value
jamesn: the spec is not clear
jamesn: differenet numbering in editors and published version
the part that starts 'otherwise ...
jamesn: parentheses is a clarification
sarah_higley: any element direclty referenced is referrimg to the label
jamesn: needs clarifyinng
jcraig: discuss with AT team
and QA team to determine what they want
especially since UAs are in agreement, it goes to ATs
jcraig: we can change if we get general concensus
keep it as is if value is not getting in the way
Matt_King: more of a clarifying question
what it is that you think we want
my interpretation is that we want what the browsers are doing
after x hours
thats what is the value
not part of the label
we want to include the value in the label
Matt_King: good examples,
<jcraig> Notification Reminder
<jcraig> - repeat once
<jcraig> - repeat forever until dismissed
<jcraig> - repeat x times
issue related to form fields open forevery
my understanding is that we want the value
jcraig: notifications,
do you want to repeat it once or forever
repeat x times or ?
sarah_higley: talking about the inverse, only when control is directly referenced
Matt_King: if i want to compose with multiple referencers aria-labelledby
lets say x, that control is directly referenced
also directly reference a span that is a portion
composing a string out of a portion of a controls label
i want a way to get value
Matt_King: if we can word it in such a way, get value and/or label
sarah_higley: getting values needs refinement
sarah_higley: standalone components, more common with web components
was making a combobox, any field with secondary control
where labe is not required
i was doing tags
selected items within standalone widget
widget A share label with widget B
found workaround
but it is useful to say
this list of items has a ...
no knowedge of where accessible name is coming from
bryan had to do the same thing
birthdate as name of input
…
it would be helpful to know you are looking at the birthdate datepicker
sarah_higley: related items ..
Matt_King: workaround for select only combobox using value text
value text is not the value
ignore the value use the value text
bryan, a link has text content
Matt_King: name and value is the same thing?
ARIA doesn't have a value concept
bryan, set the value property in a11y tree ...
if referenced by aria-describedy you get the value
if something you reference has a specific name its supposed to be returned
<Zakim> jamesn, you wanted to say I don't want to change something where all browsers agree as it would seem like a breaking change
jamesn: as browsers are aligned
change would be breaking
interested in solving the problem
changing is dangerous
bryan, change spec?
spectranaut_: should we update the spec to match the implementations?
jamesn: that is one approach
jamesn: we should consider
sohara: I think it is interesting it behaves this way
opposite to how other things work
if <article> has accessible name
you get the name not the content
is there something that needs to be added, mechanism
can aria-labelledby be able to get a value
grab name or content
bryan there is always a workaround
Adam_Page_: sarah cant, need aria-hassamelabelas
Matt_King: if you can get the computed name
you can get what you want
Matt_King: that is in the AOM roadmap
sohara: is that the solution?
Matt_King: that's what I'm wondering
spectranaut_: that is the long term
Matt_King: spec is out of alignment with reality
then solve Sarah's problem
Sarah there are workarounds
<jamesn> the labelledby bit in C/E was added based on this issue https://
Sarah can make a PR
Clarify use case for aria-rowindex on cell/gridcell elements
spectranaut_: do we want an issue for when we want the value not the label
sarah_higley: fine with answer of getting it the value
aria 1853
sohara: ACT reviews
using aria-rowindex on a cell
odd
it is allowed in spec
asked by dev who was trying to not use rows
at all
headscratcher
looking back at spec
it was there
had it at the row level
it could be a way to add redundant or conflicting info
Matt_King: I wonder if anyone can remember, rowindex is new
what the reason for supporting on a cell was
Matt_King: can't imaging the justification
<spectranaut_> jamesn: can you see who added that example?
seems obvious, you could have a single row ...
error correction
sohara: dev wanted to use rowindex instead of rows
<jamesn> is this a duplicate btw.... https://
table reconstruction
Matt_King: if we could reflow tables anyway we want would be awesome
but for aria not lining up with html
complicating that would be suicidde
jcraig: anything that affects performance but chips get faster
Matt_King: deprecate rowindex because of chatGPT [laughs]
sohara: practicality doesn't seem to be there
sohara: i said footgun
spectranaut_: no one can think of a reason to keep it
jamesn: we could deprecate it
if someone set it should still work
jamesn: if you set it on the cell you shouldn't have to set it on the row
ARIA ACT Rules review - deadline Monday Jan 16th
spectranaut_: please review
yes