W3C

Accessibility Conformance Testing Teleconference

27 Sep 2018

Attendees

Present
Wilco, Romain, Trevor, Anne, Moe, Shadi, MaryJo, Charu
Regrets

Chair
MaryJo, Wilco
Scribe
Charu

Contents


Pull Request 277: Changes to Test Cases and Rule Accuracy sections https://github.com/w3c/wcag-act/pull/277

<Wilco> https://github.com/w3c/wcag-act/pull/277/files

Wilco: first comment from Anne, yes we should link to rule structure
... some editorial from Shadi, all fairly valid
... comment from Mary Jo, sound good to me
... comment from Ann about being consistent about ACT Rules, open a new issue for that
... From Ann, should this be under test cases? it is now
... don't feel strongly about it
... any objection to changing that?
... ok i will make that change

<Wilco> https://github.com/w3c/wcag-act/pull/277/files?utf8=%E2%9C%93&diff=split#r220437369

Wilco: moving forward, editorials from Shadi

Anne: it is about accessibility issues

Wilco: i will make the update, it is about avoiding accessibility issues
... Shadi is suggesting on accuracy of test cases
... don't understand why we want to test accuracy against known test cases

Anne: 2 types of accuracy and leaving too much for interpretation
... make sense to have both false positive and false negative

Cpandhi: there is always going to be edge cases

Wilco: its not about test cases
... Even if you pass all test cases, there will still be false positives
... So we calculate the % of false positives by revalidation by the expert

Shadi: i don't agree on how to calculate the % of false positives

Wilco: it is done by accessibility re evaluation

Anne: even if we agree on what is accessibility expert is, how is an agreement reached

Wilco: ok do you have a suggestion

Anne: We should mention 2 types of accuracy, or make it clear in the section that it is not about accuracy
... as it is now under test cases

Wilco: that should be changed
... Why did u think this was about test cases?

<Wilco> https://github.com/w3c/wcag-act/pull/277/files?utf8=%E2%9C%93&diff=split#diff-9ac0a6633720a5535b0a53cba04ababeR329

Shadi: We should mark the section non normative and continue working on it in CR
... We should fix it or agree that we should come back to it later

Wilco: Are we ok marking the section as non normative?

Shadi: and also we agree on it to fix it later

Wilco: i will make the change that accuracy is non normative for the ACT rule

+1

<anne_thyme> +1

<maryjom> +1

<romain> +1

<Jey> +1

<MoeKraft> +1

<trevor> +1

Wilco: need to do some more work on the pull request

Pull Request 276: Clarify definition of Aspect under Test https://github.com/w3c/wcag-act/pull/276

Wilco: another no go from Shadi

<MoeKraft> https://github.com/w3c/wcag-act/pull/276

<shadi> https://github.com/w3c/wcag-act/pull/276/files#r220444487

Shadi: I feel consistency is conditional, we are doing 2 things, putting both output and the result and how it relates to the SC in one

Wilco: why is it not consistent

Shadi: It is only consistent only if it fails and maps fully to the requirement

Shadi it is not consistent when it passes and maps only partially to the requirement

Shadi: the pass of the rule does not mean it passes the SC

Anne: Are you questioning consistent when the result fails?

Shadi: yes, the consistent to my understanding only applies to the subset

Wilco: any suggestions?

Shadi: This is not thought out, the outcome is the result of the rule
... and an indicator how it maps to the requirements
... indicate how the outcome relates to the requirements

Anne: How it passes the rule

Wilco: There is nothing stops you from doing that

Shadi: So we should make that more clear
... that way we do not need to know the internals of the rules and make the outcome clear
... You run a rule and get the result is pass, but we do not know if the SC passes
... So we need to determine if the SC passes

Wilco: but that is already part of this

Shadi: how?

Wilco: let me find it, it is in the second para
... ACT rules must indicate if the accessibility requirement is satisfied

Shadi: how do we do that?
... so we need to indicate that

Wilco: That's where things get difficult if we put requirements on implementations

Shadi: yes we need to have that discussions but before that we can indicate that in the rule
... have those indicators, 5 or 6

Wilco: there are only 2

cpandhi: i agree with shadi, we should try to explain it with an example, if a rules passes but meets the partial criteria

Wilco: A composite rule checks for HTML pages but is out of scope for flash
... It makes the rules complicated

Anne: As Shadi says if we make it more explicit, about the results of atomic or composite rules

Shadi: I don't think it makes the rule complicated, the information is encoded in the rules, we just need to indicate what happens when the rule passes or fails to the requirements
... Make it an explicit flag or property how the outcome of the rule affects the accessibility requirement

Wilco: I am hesitant

<shadi> Fail-on-Fail, Pass-on-Pass, Unknown-on-Pass, ...

Wilco: you are proposing how to explicitly map the results to the accessibility requirements

Shadi: So we need to add these indicators, not stuck on this but have some way to report this

Wilco: That is presuming 1 to 1 relationship

Shadi: what other relations are there?

Wilco: seems like we are going way out of scope

Shadi: so that is clear fail on fail
... what happens when the rule passes? there could be unknown or undetermined on pass or unknown on inapplicable

Wilco: why does that need to be part of the rule?

Shadi: It will make it easier, you don;t have the describe consistency and rule implementation
... we need to think this through

Wilco: ok, we are out of time

<Jey> -1

+1

<anne_thyme> +1

Shadi: The proposal is to provide a relationship of the rule to the requirement

<trevor> +1

<MoeKraft> +1

<maryjom> +1

<MoeKraft> And remove the MUST be consistent paragraphs


Summary of Action Items

Summary of Resolutions

[End of minutes]

Minutes manually created (not a transcript), formatted by David Booth's scribe.perl version 1.154 (CVS log)
$Date: 2018/09/27 15:02:52 $