[[Levon introduces himself]]
[[Everybody else's introductions]]
Wilco: One response, survey is still open until Thursday next week
... This is a minor update, aligned with what requested
Wilco: No responses yet. Visible label is one of the WCAG 2.1 SCs, don't think we have seen this before
... Is the general idea of this requirement clear for people?
... It looks that the accessible name provided with aria-label matches the visible text
... These types of survey are the thing we are at now. What we do is we set up a survey based on the rules the CG has produced
... Then we look at the survey results and try to address issues
... After that we send it for W3C publication, or we send it back to the CG to fix the issues
... Tis specific survey lasts another two weeks, until the 24th
Wilco: This one we have looked at it before, this is to address feedback from a couple of weeks ago. I am hoping we can get this one done by next week
MaryJo: Opened it for 2 weeks based on process
Wilco: Let's see how many responses we have next week
Wilco: This is based on a conversation we had, a lot of draft specs are used. I went through all the drafts used in the CG rules
... W3C policy is not to point to draft until they are stable
... but most of them did not even make it for recommendation
Trevor: If there is no published version, we need to point to something. Yearly this should be reviewed to communicate the state of these drafts
Wilco: I think there is little we can do here
Levon: Do they need to be finalized to support the rule properly?
Wilco: Ideally yes. There could be more ideal things to link to but I could not find them
Trevor: Maybe having a look at who is working on those drafts to see if the changes they are making are substantial
MaryJo: We should probably engage with the chairs of these groups , if they have plans to finish them.
Trevor: We have included most of these because browsers have already implemented features. Could we include a note?
Wilco: We usually do, we could include these references in the link title
Trevor: I feel we are doing OK
Daniel: I also support these arguments but there is a strong policy issue here, we need oterhs'input on this who are not in the call today
MaryJo: We probably should get to talk with these groups to see what their plans are
... Groups are supposed to either complete or withdraw the specs
Wilco: I went through some of the rules. When we were looking at the color contrast rule, it was said that it pointed to 2 SCs, contrast minimum and contrast enhanced. If you have a fail-to-fail relationship, you have to list all the SCs that apply
MaryJo: The concern was the fact that the passing test cases do not necessarily pass all the SCs listed
Wilco: I went through others and I find a lot more like this one
... You would need examples that pass the high-level so that it passes both levels
Trevor: I generally like this way. The only alternative could be a composite, 1.4.3 in one of the atomic, but you would need to give a "more test needed"for these examples. To me is relatively easy to understand
Wilco: Maybe we could put something additional into the mapping to make this clearer
MaryJo: If the pass means a pass for all the referenced SCs it would be wrong because maybe it would not pass the stricter SC
Wilco: Color contrast checks only HTML, there is still more than needs to be tested anyway
... One suggestion that was made is to add notes to the examples, that is a daily straight forward thing to do
MaryJo: It has to be clarified
Daniel: Maybe the understanding documents is a goodplale for such notes
Wilco: If I had a rule testing 2.5 contrast, the mapping would not change
... and that seems odd
... It seems to me that we are missing some properties to express it
Trevor: If we start having passed examples that say this one passes but the other does not, it might feel like a second-class rule
MaryJo: Back to mapping, I see that it is OK to list, because a pass could mean either satisfied or further tested needed
<maryjom> https://www.w3.org/TR/act-rules-format/#accessibility-requirements-mapping
MaryJo: The example in the requirements show a bulleted list with outcome mappings.
... would doing that be more helpful?
Wilco: We are doing that, but we have to expand the component to see it
... There is a Must requirement in 4.4, you must list all accessibility requirements having a fail-to-fail relationship, it would be better to have a Should there, for flexibility with these cases we are dealing with in this rule
MaryJo: Is it possible that we clarify in the outcome mapping that the stricter criteria may need further testing?
Wilco: Technically yes, but the way we are using these technologies, it might not be easy. We are including standard text, we would need the way we generate rules
... I think I would prefer to take that out
... My preference would be to try to avoid the problem, maybe this is something we can change in an errata
MaryJo: My preference is not to list something for which the test cases are not strictly equal
Trevor: If we add that except clause, maybe there are other cases that we may run into problems with
Wilco: This is why I am thinking of a "should"that gives as more flexibility
Trevor: That's good with me
<Wilco> https://act-rules.github.io/rules/afw4f7
<trevor> https://www.w3.org/TR/act-rules-format/#accessibility-requirements-mapping
Wilco: We've had some pretty small updates to rules, maybe there is a better way to streamline the process. Surveys are good for new rules, but maybe not so for updates. Probably for corrections like typos or small changes we would better open a pull request and ask for CFC
... Would people be OK with having editorial changes in pull requests?
Trevor: I would be up for that. With the pr would be a lot better
... Maybe it is hard to decide what is editorial
Wilco: There is a definition in the CG. Anything that may change the outcome of an implementation is not editorial, but even that would be too strict for our purpose
... We could tweak it to suit our needs
MaryJo: Maybe even changing the length of the survey that we have when we review a rule that we already reviewed before, maybe giving one week instead of two to make the process faster
... Maybe for removing test cases or something we do need surveys, but other things may go into PRs
Wilco: What if we start putting PRs in the agenda?
... And then if in the call we decide that PR is fine, it would go through our CFC
[[General agreement]]
Wilco: I would need to go through some pull requests myself