<scribe> scribe: Jean-Yves
Carlos: 2 one week calls from
Wilco.
... one being merged now.
... 2017, using webauthn for autocomplete, finishes next
week.
... 1 two weeks, changes to text contrast rules
Carlos: got some time to work on ACT. Lots of reviews. Need reviews on my PRs (1845, 1560, 1655).
Wilco: reviewing rules in the TF, making changes on them.
Helen: not much on my side, need someone to look at 1979
Wilco: now, only chairs can merge because we push to WAI website directly.
Giacomo: 3PRs assigned, need more review on 1907. Still working on others.
Wilco: 1907 has a failing test.
Carlos: if you update the branch, the test re-trigger.
Giacomo: 2007 is tricky to include "ambiguous to users in general", we're making progress.
Dan: nothing going on currently.
Jean-Yves: need review on 1923 and 1994.
<Wilco_> https://w3c.github.io/wcag-act/act-rules-format.html
Wilco: Editor's draft of rule
format 1.1
... please review and give feedback.
<Wilco_> https://w3c.github.io/wcag-act/act-rules-format.html#secondary-requirements
Wilco: going through the ARIA rules.
Jean-Yves: working on a report for a semi-automated implementation
Wilco: we are also working on states.
Helen: ...
... I am a bit delayed currently. Not much progress.
... I want to work on something for 2.4.3
Carlos: if someone want to help
for manual test rule, that would be great.
... this came up while JY was working on these rules; I noticed
we used "computed value is not important", but important is
only used on declarations, not values.
... we need that concept because if important is on an
inherited value, this is OK.
... we almost need to replicate the value computation to find
the correct declaration
... we could create a definition for "important value" and
explain what happens.
Jean-Yves: we can probably still rely a bit on the CSS cascade definitions.
Wilco: it is a problem. Adding
this def sounds like a solution.
... maybe we can chisel the Applicability to remove the
declaration with inherit/…
Carlos: ARIA 1.2 changed the
conflict resolution so that focusable element always ignore
presentation role.
... so the rule checking that becomes obsolete…
Jean-Yves: what about the similar rule with presentational children?
Carlos: Scott from ARIA thinks it's a bit premature to deprecate the rule.
Wilco: I agree with Scott.
Jean-Yves: should we remove the WCAG mapping and only keep the ARIA mapping?
Carlos: I agree to add the ARIA mapping.
Wilco: it still creates
accessibility issue. So we should keep WCAG mapping. Background
likely needs to be updated.
... does that mean you can affect an element's role from the
shadow DOM or an iframe?
... it is not about the parent being hidden and the descendant
focusable, the ARIA 1.2 is about the same element.
Dan: how do we now which version of ARIA a page should be interpreted?
Wilco: essentially the latest, with some bits of drafts, because browsers do that.
Giacomo: can I hide an iframe with role="presentation"? Now it will trigger the resolution and be exposed?
Wilco: I don't think it is possible to make iframe fully invisible to all ATs. role="presentation" would work on Chrome or Safari; but not in Firefox.
Giacomo: Some UAs requires to "enter" the iframe, but that doesn't make sense if it's just content streamlesly included from another side.
Carlos: the rule checks that the parent is presentational, and a descendant is focusable.
Wilco: that scenario does not
trigger the conflict.
... this is on a section on the role of presentation, so I
think it triggers when there is "role=presentation"
explicitely.
Carlos: the conflict mention *inherited* presentational role.
Wilco: inherited presentation is table/cell, or list/listitem; not presentational children, or any children of role="presentation".
Jean-Yves: we should add specific
question to ARIA group.
... I'll try to open an ARIA issue.
Carlos: 4.1.1 is removed from WCAG 2.2. What should we do with our rules mapping to it?
Jean-Yves: "attribute not duplicated" actually never was a 4.1.1 failure. We should remove the mapping, but keep the rule for linters.
Wilco: should that be a secondary
requirement?
... parsers used to behave differently on duplicate attributes
(first or last), so it is a 4.1.1 failure.
... HTML5 never fails 4.1.1
Jean-Yves: the rules still have value as best practice.
Wilco: thinking about failing
them with lesser severity.
... we could have an ARIA rule looking specifically to id that
are used in aria-labelledby, …
... I think we should deprecate both.
Carlos: deprecating the rule doesn't prevent anybody from using them.
Wilco: at least we stop
encouraging implementers to fail nonsense.
... deprecating the rule allow tools to stop implementing it
without dropping in consistency numbers.
... on my todo list to show the reason for deprecation.
<SEan> need to drop.
Wilco: I will take the discussion with the TF.
This is scribe.perl Revision VERSION of 2020-12-31 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: Irssi_ISO8601_Log_Text_Format (score 1.00) Default Present: CarlosD, giacomo-petri, dan_tripp, Jean-Yves, Helen_, Wilco_ Present: CarlosD, giacomo-petri, dan_tripp, Jean-Yves, Helen_, Wilco_ Found Scribe: Jean-Yves Inferring ScribeNick: Jean-Yves WARNING: No meeting chair found! You should specify the meeting chair like this: <dbooth> Chair: dbooth WARNING: No date found! Assuming today. (Hint: Specify the W3C IRC log URL, and the date will be determined from that.) Or specify the date like this: <dbooth> Date: 12 Sep 2002 People with action items: WARNING: IRC log location not specified! (You can ignore this warning if you do not want the generated minutes to contain a link to the original IRC log.)[End of scribe.perl diagnostic output]