W3C

Accessibility Conformance Testing Teleconference

11 Jun 2020

Attendees

Present
Wilco, Trevor, Kathy, MaryJo, Charu, Shadi
Regrets

Chair
MaryJo, Wilco
Scribe
Trevor

Contents


HTML page language is valid - title change for the rule

<Wilco> https://github.com/w3c/wcag-act-rules/pull/24/files

wilco: This rule was approved pending update to the title.
... We haved settled on the title "HTML page lang attribute has valid language tag", there are a couple other small changes as well.

<maryjom> +1

wilco: +1 if you are okay with it merging

<Wilco> +1

+1

<shadi> +1

<cpandhi> +1

kathy: I have a question on line 80, passed example 1. We were only looking at the first subtag of the tag.

wilco: We are looking for tags that can be used by the browser, so a primary language subtag that can be followed by nothing or a dash.

<kathyeng> +1

wilco: I am going to merge this in then.

'audio' or 'video' avoids automatically playing audio: https://www.w3.org/2002/09/wbs/93339/ACTAutoPlay/results

wilco: First comment from me, a few issues with the applicability. Ambiguous if paused and muted attributes refer to IDL attribute or content attribute.
... Community group looking at a new definition of mechanism, may require some reworking.
... Next comment, kathy, should conformance requirement #5 be included.

kathy: I ask because I think the conformance requirements are part of WCAG.

wilco: 1.4.2 is under conformance req 5.

shadi: What are we trying to achieve?

kathy: I think when that if 1.4.2 is a failure then conformance requirement 5 has an issue. Mainly used to indicate this is a severe failure

wilco: I think you are right and needs to be included in the requirements mapping. It is is a requirement that needs to be satisfied
... If 1.4.2 is not satisfied then non-interference requirement isn't satisfied as well.

trevor: Wondering about how far up we follow it, or if we should be on the most specific level.

wilco: I guess we could list conformance requirement 5 as related to 1.4.2
... I think listing it as an accessibility requirement here would be good.

trevor: So each time we list one of the sub-requirements we will also need the conformance requirement.

wilco: I will take this back to the CG so we can build something to support that.

maryjom: Our tool reports on S.C., not on conformance requirements.

wilco: Neither does ours, but I don't think it matters, it is there to make the rule more consistent.
... My comment on vague WCAG about total length of audio. Is it 3 seconds consecutively, or can it be split up?

maryjom: We should ask for clarification

wilco: I will open an issue.
... Kathy has a comment on the definition of mechanism being unclear.
... For new rules, CG is using the word instrument and ensuring it must appear in the same location.

kathy: It is expected that it is one the web page, can that relate back to this rule. State it clearly that the mechanism is on the web page.

wilco: That might help, we will have to look into that.

kathy: For G170 comment, it says "near the beginning of the page", what does that mean when you are testing, how would someone know it passes the term near?

wilco: If there is no mechanism it is clear that it isn't in the beginning.
... We could hand this back to CG to better define "near", it will make it a more complicated rule.

kathy: If you just read 1.4.2, it doesn't specify where the mechanism is
... For accessibility you want the mechanism to be readily available, but the req doesn't specify where it needs to be.

wilco: Sneaky stuff like that happens where they put that into techniques

charu: Inside of G170 it should be keyboard operable and early in the tab and reading order. Still not clear as to how early

wilco: It is a recommendation but not part of the s.c.

kathy: for trusted tester we chose the first 3 tabs arbitrarily.

wilco: Do you think the rule needs to be changed with that? or okay with as is?
... We can just leave this for CG to look more into.

<Wilco> https://act-rules.github.io/implementation/axe-core#id-80f0bf

wilco: question of if these results count
... It reports cantTell on all but two of them

<Wilco> https://act-rules.github.io/pages/implementations/mapping/

trevor: We don't have any formal definition of implementation?

wilco: Only the CG has a document. The implementation conforms, but isn't very meaningful.
... We have seen this a few times where most of the answers are just cantTells
... It isn't inconsistent, but doesn't show a lot

maryjom: Mostly cantTells doesn't seem like an implementation kinda.
... Is it a sign that the rule is not well defined, or does it point out that the tools are insufficient.

wilco: It largely indicates that the rule is manual.

maryjom: There should be some manual testing methodology that can implement the rule correctly.

wilco: I think that deserves its own discussion so I am going to create an issue

shadi: example, situation a I have a button, tool doesn't recognize it, it should not return cantTell for all of them. it should only return cantTell if it recognizes the applicability that is true
... I think there is a difference if the tool can't say anything, or if it know something is there that applies but doesn't know what to do about it.
... Potentially we should have another response to distinguish the sentence.

wilco: It seems like we need an implementer that gets most of the answers correct
... I think a separate conversation is needed for implementations of atomic rules
... Kathy's comment on whether testers can determine the full length.

kathy: For passed example 2, the video gets stopped at 2 seconds, but there is no indication that it is that long.

wilco: Pretty clear cut that this rule needs more work before it can be published.
... My impression at the CG is that this may go on the shelf for awhile, there is a lot that needs to be fixed.

kathy: This rule included 2 atomic rules that we hadn't reviewed, should we have reviewed them

wilco: I did, but how do we want to do this.

trevor: I thought we would do the atomic rules independently, then do the composite rule at the end.

charu: agree, or we could have a batch that all gets reviewed at the same time.

Summary of Action Items

Summary of Resolutions

[End of minutes]

Minutes manually created (not a transcript), formatted by David Booth's scribe.perl version (CVS log)
$Date: 2020/06/11 14:15:43 $