Meeting minutes
New Issue Triage
definition of navigation,
jamesn: don't think we should define it
someone else should define it like the HTML spec
melsumner: seconds that define at base level
navigate is defined in app history api and html standard, session history and navigation
https://
jamesn: leave open, not define in aria
1.4
New PR Triage
#148m
scott: label should behave the same as labelledby
jamesn: accname issue?
html aam doesn't devolve to accname without label or labelledby
scott accname computation is not lang specific
scott: add to html aam, a note for hidden sources
aaronlev: draft PR to look at
aaronlev: for hidden stuff
changeing license, no review needed
1690 PR mel and Jamesn
1.3 blocking issues
1159, mel has it
1177 jamesn pinging in issue
1365
jamesn: up for review
Deep Dive planning
no dd next week or the 24th jamesn on vaca
Announcement - Minutes Archive
jamesn: patchy
partially a warning that these minutes will be google searchable
people will finding old minutes more often
scott: promies a shirt
to Michael
scott: Roberto has joined welcome
siri works at apple lol by irfan
we won't say hey siri
TPAC 2022 - Who might attend in person? Vancouver from 12 to 16 September 2022. Non-binding straw poll?
hybrid in person remote meeting
how many would come to vancouver
?
folk are in favor of meeting and attending
some won't be able to travel, facilitate remote participation
jamesn: webaccessibility slack channel?
wcag slack channel is free plan not archived
web-a11y.slack.com
have to be invited, thousands on it
jamesn: can send one around, we have enough people to do it
lol
scott will bring tshirts, more than a feeling
orgs should pay expenses
micheal: diversity fund, targeted to invited experts with no sponsors
contact micheal cooper
Charter Update?
new charter, recently joined have to rejoin
we are kicked off, sign up again, once every 2 or 3 years, new patent policy
march 10th is the deadline
AC rep should do it
AC rep has to rejoin for you
they have to agree to new patent policy
joanne: expedite recommendation status?
MichaelC: updated CR
recommendation snapshot every 2 years
move aams to that status
do snapshot when we have everything that supports base version
jamesn: more people doing implementation and testing
we have joannie and valerie
Handling Author Errors: form & region roles
scotto: Sarahs point in review,
regions not announced as such without names
this happens in html
sara found role was announced without name
good idea to have parity to html
but aria is more specific
joanie: 1.1 - 1.2 changed spec to say authors must provide a name
user agents need a backup plan
best way is to not make it a landmark
like html
contention
form doesn't want to be generic if it is not a landmark
choices - landmark, generic, something in between
"this is a form"
"tell me all the things in this form"
IA2 role landmark
<carmacleod> https://
<carmacleod> Should form-with-name map to IA2_ROLE_LANDMARK?
issue for both core aam and html-aam
need several people to look at this mappings table
will provide Pr
<aaronlev> Chrome/IA2 exposes <form> without name as IA2_ROLE_SECTION, and with name as IA2_ROLE_FORM
and issue
Matt: screen readers would ignore it
if role=form has an effect
any at all
author is expressing intent
Matt: can't imagine why we would tell an author to do that
Joanne: give it a name when using role=form
<siri> like for section element give a name to announce as region
jamesn: larger issue
tackle in parrallel with the one Scott put on the agenda
joanie: James Craig, I am unsure of apple mappings
jamesn: directon of this?
a div with role=form should be exposed as if it was not there
<aaronlev> You can see how Chrome maps nameless and named forms to all platforms here:
<aaronlev> https://
<aaronlev> The first form is nameless, the second one has a name
siri: main element, nav element AT exposes
<form>should expose it
jamesn: just expose an unnamed form?
matt: but not as a landmark region
jamesn: why not
?
Matt: region not exposes as a region without a name
matt: creates unintentional landmarks
big problem
carmacleod: github forms are goofy
form elements
scotto: they are form elements with names
jamesn: coding problem
Matt: if we have naming requirement
no name then not a landmark
scott did it for html forms
form element is not exposed as a landmark unless it is named
Roberto: landmark element with role=region?
region is just one of the landmark roles
there are 7
native landmarks don't require namesw!
siri html doesn't have section element
siri: agrees no need to have name
names will be helpful
with many forms
melsumner: forms never seen div role=form
has anyone seen that?
need more info, method action, other stuff
how do you get values to the backend
should be allowd to fail
Harris: benefit to using <form> for enter key submission
Sarah has seen this
legacy code
Robert, in AEM
scotto: use most appropriate landmark
jamesn: there are other things than just form elements on the page
<melsumner> for clarity, I advocate for this scenario being allowed to fail because the author is failing on multiple levels here
jcraig:
do we need to map this doen'st look like it looking for guidance
matt: div role=form no name to role=generic
?
jamesn: role=form map to navigation with no name
carmacleod: form is the only role that matters
carmacleod: maps to native element that is form
map to whatever the base html element if role=form is not there?
jcraig: fallback roles
account for in PR
Jamie had a strong opinion
carmacleod: needs help on this
<siri> Why to use role=form why can't we role=none if it is generic
rrsagent make minutes