W3C

Accessibility Conformance Testing Teleconference

22 Nov 2018

Attendees

Present
SteinErik, Anne, Trevor, Wilco, Shadi, Jey
Regrets

Chair
Wilco
Scribe
SteinErik, Trevor

Contents


"Optional items" section + renaming "Assumptions" #305

SES: introduction to proposed section: adding an optional information section to the structure

WF: Should accessibility support be an optional - it is using the word should
... optional things are useful things, but the should part absolutely should be there, but exception if certain information is missing

WF We shouldn't use a must on parts that could be unknown

SAS: Let me clarify. We MUST have the accessibility suport section in every rule, but it may be empty or incomplete. Or these are the accessibility support limitations that I know of as rule author, but that would be a must requirement. If it is completely optional I could have a rule without that section at all.
... Same applies for assumptions. If there are no assumptions, do I leave the section out or say there are no assumptions

WF: See what you mean

SES: It is fair to say a section is compulsory, but it is possible to let it open

WF: I think accessibility requirements and support should be a MUST, but possible to leave blank or with none

RESOLUTION: ATH and SES will change pull request

SAS: We could add a sentence saying that it is possible to leave the section blank
... I don't particularly like having this optional section. There will now be two things left after the agreed edits. Suggests removing the part about may include in the structure section
... Would be a good thing to put (optional) after headings that are in fact optional

ATH: Asks for clarification

SAS: The section called optional items with the sub level.

ATH: We could just leave it out and remove the extra layer.

<Wilco> https://github.com/w3c/wcag-act/issues/301

Use Cases https://github.com/w3c/wcag-act/issues/301

WF: Tried to break down this discussion to separate questions.

WF: First one is best practices and relation to WCAG SC. Things that are considered good practice.

WF: Can we use the background section for this to resolve it

ATH: I'd like to see something that allows structured data, not free text fields..

WF: What is the relations hip between a best practice and a SC

<shadi> scribe: trevor

shadi: My question is, when we talk about relationship. Relationship between outcome of the rule and thing we are testing, or thing we are testing and success criteria
... Lets say we are testing no heading levels have been skipped. We all agree this is an advisory technique.

anne: Couldn't we talk about both, the types of relationships will be similar. And we don't need it to be perfect

wilco: I see two categories of relationships. Either directly testing a thing, or thematically related to thing
... Failing rule means you failed the thing. Thematic is everything else such as same types of technologies, similar techniques etc.

anne: I disagree there are only those two types of relationships. For example, we have the html passing. If we have a passing requirement, there are specific ways to fail, but if you pass the advisory you may pass another.
... Just saying that because a rule fails means the success criterion failed is too simplistic

wilco: We defined a rule as being failure test. Failing a test means failing a sc
... Allowing rules as passing things is a significant scope change for the rules format

anne: I agree, but we have changed so many things in this direction I am surprised that is still our scope

<Wilco> https://w3c.github.io/wcag-act/act-fr-reqs.html

<Wilco> https://w3c.github.io/wcag-act/act-fr-reqs.html#rules-test-for-failures

wilco: This is the requirements document from way back when we started

shadi: what is the consequence of changing the scope? What would we break

wilco: I don't know the answer, which makes it hard to change
... An example would be having valid html. You could have valid or invalid html that passes the parsing requirements

shadi: Would that be a problem or a change?

wilco: That would be a problem, it confuses the actual requirement
... If you had a rule that said the page has no video, it could be a valid rule for the video success criteria
... If it has no video you pass, if it does you fail. If we drop the failure stuff, you could write rules like that and you could mix up applicability with expectations
... We would be releasing a restraint we have been using to achieve consistency in rules

anne: I don't see it as such a problem

stein: still two types. Explicit and more loose relations. Wilco has a fair point. In the case of the video you wouldn't list any sc, but could have another section. Having people create creative rules isn't a bad thing

wilco: I do see the use case for what Anne is pointing out. There is an interesting one for extended audio and contrast where a passing a rule means you pass a lower level rule as well.
... I am nervous to let go of that completely, since we have definitions of what is in applicability and expectation. How many constraints may we lose?

anne: I don't think there is any problem with that.

stein: I agree that we should change applicability and expectation.

wilco: If we let go of the constraint, it allows you to do the same thing in more than one way.
... I think the more variations in how you allow a rule to be written, the harder to standardize the rules

anne: I think we already have that today
... As I see the rule format is one thing, but there should be some vetting process for rules added to WCAG

stein: Lets see if we can get closer to resolution. Should we change words in expectations and applicability

wilco: No I think thats good
... I think sticking with background info and mapping failure to failure

stein: Could it be helped by adding more to background information

anne: The accessibility requirements section still needs a rewrite correct?

wilco: Yes, we have reached an agreement that changes will need to be made.

anne: My issue with all of this, is that the failure to failure relationship does not stand out as better than a pass to pass relationship

anne; I think the accessibility requirements only contains those, needs to explain why that mapping above all other mappings

wilco: That we can do.

anne: I don't know, the accessibility requirements list doesn't really say anything about it. Its not just the accessibility requirements. These are specifically the non-conformance mappings
... I think it could be made clearer, maybe even in the title of the section

shadi: I still don't understand why the non-conformance is more important. But I understand why wilco wants the rules to be clean.
... there are examples for when passes that are reflected onto other rules can be useful

wilco: I will write this up better later. But i see it for consistency.
... There are some rules that do not map directly to sc but are required for wcag.

anne: I completely agree, but my vision of a solution is that people write the rules as ACT rules, and then look at their rule suite applies to wcag test

stein: I agree as well. If you can make it clear that their is a clear relationship of passing and then have to take extra steps to make sure people are writing high quality rules.

shadi: I think there are failures that equally cause problems. With fail to fail, there are still some methodologies that have incorrect interpretations
... I think we need to address both aspects of documenting things the group agrees on that are failures and passes
... I think the pass will be much less than the fail, but would still be noticeable

wilco: Let me finish up on this, and let me think if I am okay dropping this, and we need to follow up on this next week.

stein: We should also think about some compromise solutions.

Summary of Action Items

Summary of Resolutions

  1. ATH and SES will change pull request
[End of minutes]

Minutes manually created (not a transcript), formatted by David Booth's scribe.perl version 1.154 (CVS log)
$Date: 2018/11/22 16:41:49 $