<maryjom> https://github.com/w3c/wcag-act/issues/250
Mary Jo: We have 3 issues on the agenda today. We are looking into a tool to view pull requests, and expect to be ready for survey next week, when it is easier to review
scribe: This first issue is from Shadi's comments
Shadi: I think most of the text in the issue is just copied out from the spec
Anne: Shadi's comments are just the two in the bottom. It might be easier to open the attached document to see what the comments relates to
<maryjom> If it's easier to read from Shadi's document, here's the link: Accessibility Conformance Testing.docx
Shadi: It's section 7
<maryjom> https://github.com/w3c/wcag-act/files/2293796/Accessibility.Conformance.Testing.docx
Shadi: I had put in some proposed edits. What Moe has put in seems to be the changed text
... So she did integrate the changes
... The question here is SHOULD or MUST
... This is about identifying the requirements. So is it possible to write rules that does not list the requirements they are related to?
Anne: I think this is about atomic rules that are part of composed rules, but doesn't map directly to the accessibility requirement itself
Shadi: This might be what Moe has tried to do
... I don't know what Wilco's comment means, if he favors it
Kathy: Are we talking about that atomic rules doesn't always satisfy an accessibility requirement?
Shadi: If there are rules that are preconditions, they should be listed as exceptions
Anne: (line is breaking up)
Shadi: it sounds like there are valid exceptions. We may have to write a MUST for atomic rules and a MUST for composed rules. Or list exceptions
Kathy: Composed rules lets you combine atomic rules with an OR. I tend to require that the accessibility requirement is listed in atomic rules. But maybe remove the pass requirement
Shadi: The case here is that the rule is not consistent with, but still relates to a requirement
... My preference would be to list in this case... you MUST, in this case... you MUST - or list exceptions
Kathy: Are you saying that there is way to list that they are related?
Shadi: It can be provided in the background information. That is a side thing
... I am trying to be a bit more stringent on the requirements
Anne: My points have been covered
Shadi: I think the intent is the same, but I think we should rephrase to make it a bit more clear? Mary Jo, can we send it back to the editors with these comments?
Mary Jo: Yes, but I think Moe will need a bit of background information, since she is not here
Shadi: I don't know the bakground for the SHOULD, but I am happy to work with Moe on this
Mary Jo: I am trying to summarise on the issue
Shadi: There are two requirements here. The conditions for each of them has to be clear. This section only has two SHOULD NOT, so I can actually write ACT Rules without listing any accessibility requirements
Mary Jo: I have added the comment to https://github.com/w3c/wcag-act/issues/250
<maryjom> https://github.com/w3c/wcag-act/issues/243
Mary Jo: Okay, so "composed" or "composite" rules
<agarrison> +1 for Composite
Anne: The comment is from my colleague, who's writing Auto-WCAG rules with me
... and who is native in English
Mary Jo: I am neither for or against
<shadi> composed
<shadi> composite
<shadi> multi-part
Shadi: Let's brainstorm
<shadi> combined
Trevor: "composite" feels better for talking about the rules, but "composed" says more about how they are put together from atomic rules. I like "composite" better as the adjective
Mary Jo: I am fine with "composite". I like it more and more
Shadi: Anyone feeling strongly against "composite"?
Mary Jo: This will be an easy one, then
scribe: I will put in the meeting notes that "composite" it is
<maryjom> https://github.com/w3c/wcag-act/issues/237
She wants an improved example for Aggregated results (section 16)
I think we discussed the first part, renaming the aggregation definition. Then it's just the last part left
(note for above: first Mary Jo, then Shadi)
<shadi> https://github.com/w3c/wcag-act/pull/247/files
Mary Jo: So it
scribe: 's asking for a better example
Mary Jo: "Example: An expert evaluates a success criterion to fail on a specific page. When testing that page using ACT Rules, there are two rules that map to this criterion. The first rule returns no results. The second rule finds 2 test targets that pass, and a 3rd test target that fails..."
scribe: "In this example, the first rule is inapplicable (0 results), and the second rule has failed (1 fail, 2 pass). Combining this inapplicable and fail, means the success criterion has failed."
(from Example in https://www.w3.org/TR/act-rules-format/#output-aggregation)
Anne: So the pull request solves the fact that "aggregation" is used for both composed rules and aggregation of results by removing "aggregation" from composed rules
... but the pull request doesn't do anything about result aggregation in section 16, so the example there might still be a bad one
Shadi: What Annika says is that the example mixes both instances and rules, so checking different things
... I am not exactly sure what the issue is here
Mary Jo: Would it help to write that it is different atomic rules?
Shadi: Maybe the easiest solution would be to rename it to result aggregation
Anne: I read it as if Annika thought the example was too complex, trying to do too many things at once, whithout it being clear what it is supposed to illustrate
Shadi: I think there is a point in this, that the examples could be broken down even more
<shadi> Example 1: aggregation over different test rules
<shadi> Example 2: aggregation over occurrences of a rule
<shadi> Example 3: all the above (aggregating occurrences and rules)
Shadi: So basically the same example, just broken more down, allowing you to follow it bit for bit
... I am still not sure what the requirements here are, though. Are there any specific requirements here? E.g. that is should be in line with the accessibility requirement.
... It is kind of implicit here, maybe we should make it explicit
Anne: and maybe if it is a requirement that you do aggregate your ACT rules
Shadi: I'll open an issue about this
Mary Jo: I'll add a comment to https://github.com/w3c/wcag-act/issues/237 about breaking down the example.
scribe: I think it will make it an easier read if we break it down
... Next week we will go over pull requests with some of these editorials, and then maybe also hit one or two issues
Shadi: Feel free to tell Moe to contact me for discussion