Editorial:

• First sentence: ...written in "a" format, not "an" format and "comparable" instead of "comperable"
• First sentence: Suggest you say "comparable accessibility standard" since not all are called "guidelines" or are a "recommendation", like the EN 301 549 or other non-W3C standard.
• You should probably provide an explanation of why there MUST be a warning if the (do you really mean test case or do you mean rule?) contains accessibility issues listed in the Non-interference section of WCAG.

Didn't address the last bullet. This is not a big important point, so I don't want to spend too much text on it. The point is that if someone has a test case that can be a keyboard trap, they are allowed to do that, but they need to warn the user for it.

"may" -> "might"

The second sentence is long and complex. It also uses the keyword "should", I think incorrectly. Consider this edit:

"If this involves multiple "failure conditions", each of these is provided in a separate atomic rule. The logic for determining (non-)conformance to the accessibility requirement according to outcome of the individual the atomic rules is specified in the composite rule."

The sentence "These rules should be kept small and atomic" uses the "should" keyword, and I think incorrectly. Consider "These rules are to be kept small and atomic" or just "These rules are kept small and atomic" or such.

"Atomic rules test if web content is correctly implemented in a particular solution" -> "Atomic rules test if test subjects correctly implement particular solutions" or "Atomic rules test for particular solutions within test subjects"

"Composite rules test if a combination of outcomes from atomic rules is sufficient for an accessibility requirement to be conformant or not." -> "Composite rules test if a combination of outcomes from atomic rules satisfy/meet the conformance for an accessibility requirement."

Not changing the first part. The group has gone over that sentence multiple times already.

"Only if the outcome of multiple atomic rules has to be combined in order to determine (non-)conformance of the Accessibility Requirements can composite rules be used" -> "Composite rules are used when the outcome of multiple atomic rules needs to be combined in order to determine (non-)conformance [of a test subject] to the Accessibility Requirements"

"web page" -> "test subject"
"must conform to" -> "conforms to" (drop the "must")
"Often organizations" -> "Some organizations"
"which may come from" -> "that come from" (drop the "may")

Test subjects aren't required to conform to WCAG, web pages are. The way I see it, web pages are a type of test subject. JSX or Vue files are examples of test subjects that are not web pages. Instead, they are used to build a web page. WCAG does not apply to those files, nevertheless those can be tested by ACT Rules. I think this sentence is correct.

I don't understand why you want to limit "accessibility requirements" to web page and WCAG only. As I understand it, accessibility requirements is more generic. It can be used for any requirement, for any test subject. We happen to only currently use it for WCAG and web pages in Auto-WCAG. Anyway, this sentence does not set a normative requirement and could be edited later, if we decide to.

"not satisfying the success criteria" -> "not satisfying the accessibility requirement" (we are not WCAG-specific here)
"may" -> "is" (drop the "may")

"may be of use" -> "can be of use" (unless you really mean "MAY")
"may need to operate" -> "can operate"

"that must be true" -> "that is true" (drop the "must")

"composed" -> "composite" (twice)
"may be" -> "can be" (drop the "may)

Suggestion: "If the atomic rule is used in a composed rule, the outcome of the composed rule may be passed when the atomic rule failed" -> "When atomic rules are used in composite rules, the outcome of the composite rules can be passed when the outcome of individual atomic rule is failed"

"that a website" -> "that content" (to be less website-centric)

"may not" -> "might not" (drop the "may")
"include a warning if there are" -> "indicate" (more concise; also not really a "warning", it is information like all other sections)

The name is a bit odd. I suggest "Accessibility Requirements Mapping" (see more below on why this name may be more apt)

The first sentence is already covered by section 4 "Rule Structure"; the second sentence is already covered by section 11 "Expectations"; and the third sentence is already covered by the definition of "outcomes". I think we can lose this entire paragraph. It re-defines existing normative requirements and definitions, which is very confusing.

"impacts" -> "relates to"

Replace the second sentence with the following:
"Typically failing a rule means the associated accessibility requirement was not met. At the same time, passing rule does not usually mean that the associated accessibility requirement was met, unless the rule was designed to address all fail and pass conditions of that accessibility requirement."

This example should also be dropped. This is rewriting the expectation, and seems unnecessary to me.

This is already described in the definition for outcome and can be removed from here.

I don't understand what this means but I think it is already sufficiently addressed by my proposed rewrite, and by the example above.

I think we should remove this paragraph. This is not evaluating a rule, but actually an implementation of the rule. It is also not the evaluation instructions that are applied but the applicability and expectations. These provide the outcomes.

This may be confusing as this separation of responsibility is not always the case, as stated in the following paragraph: sometimes an atomic rule itself would test for conformance to an entire accessibility requirement, sometimes it requires a composed rule. My suggestion is simply to drop this paragraph, or make it a "note" block with a slight reword like

The separation (...) creates can be seen as a division of responsibility. Atomic rules test typically test if web content is correctly implemented in a particular solution, while composite rules test would test if a combination of outcomes from atomic rules satisfy the conformance for an accessibility requirement . as a whole.”

agree with the reworded suggestion.

Discussed on the call. We'll leave it in place, not as a note but taking over the editorial changes

I preferred the previous version, saying “The applicability is defined as” rather than “The applicability MUST be”: if we say it's defined as such, then there is no room for interpretation, and implementations don't have to guess what to do if the author didn't follow the spec and defined the applicability otherwise. If we use a MUST, then authors may (will) break this statement and implementations are left with an unspecified situation.

Decision from the call is to leave the "MUST" in here, it is a requirement for the rule author. We can work on rewording this to address the issue separately.

Suggestion: "The applicability of a composite rule is defined as the union of all the applicability sections of its [=atomic rules=]. This can be inferred from the list of atomic rules. However, rule authors MAY provide a description of the applicability for composite rules. This can be useful if it is difficult to express the combined applicability in plain language. In that case, the applicability of the composite rule provided by the author MUST also be the union of all the applicability sections of its."

• "decide if an accessibility requirement is (non-)conformant" does not make sense. Maybe "decide if a test subject conforms to an accessibility requirement"?
• Same with "logic for determining (non-)conformance to the accessibility requirement" -- needs "test subject" -- "logic for determining conformance of the test subject to the accessibility requirement"
• BUT if you define composite rules that way, then each composite rule MUST be related to an accessibility requirement. However, the requirement later on in section 7 is only a SHOULD requirement. You need to either change that requirement or this definition of composite rules.

• According to the definition of composite rules in section 3, they actually MUST identify the accessibility requirements they relate to (see corresponding comment above)
• There is no "Evaluation Instructions" section (anymore). I don't think that sentence is needed anyway.

Need to remove "evaluation instructions".

I suggest making that a sub-section 7.1 -- maybe call it "Rule Mapping" or such

We'll break it up, add 2 headings, one for acc requirements list, second for rule mapping.

Looks good!

Remove "the" between individual and atomic

• I really like the idea of saying "satisfying conditions" because it is not limited to "failure" or "success" but could be any. I think we need to reflect that phrasing in the atomic rules definition too. That is, "satisfying condition" instead of "failure condition".
• The last sentence seems to repeat the first two. Actually it contradicts them somehow -- is it "multiple" or "any one or more"? I'm not sure what the last sentence adds that is not already in the first two, and think it should be removed.
• Following the previous point, I think the word "multiple" is wrong in the first two sentences, regardless if you keep the third sentence or drop it. It should read "any one or more" instead of "multiple".

Suggest the last sentence reads, "The logic in the composite rule describes one or more of these satisfying conditions is used to determine a single outcome."

I'm not the grammar expert but "would" jumps out at me. I think it should be "can".

I would simply delete the word 'would'.

applicbility -> applicability

Suggested edit: 'ACT Rules developed by the Auto-WCAG community groups use..." What happens if a rule can be used for testing multiple WCAG criteria? Does that happen? If so, what should you list as the criteria number?

Also, to clarify understanding of the 2nd sentence...Do you mean, "They include the number corresponding to a WCAG 2.1 success criterion, combined with a handle that includes an element name or name being tested:"
Does that change characterize it more clearly without changing your meaning?

I think "affects" instead of "effects" is a better word.

Nitpicking here, but an editorial change. "The outcome of this rule maps..."

Last bullet should say, "implementations".

Use "meet" instead of "meets"

I'm thinking it should say "using the [subject property]" rather than "on..."

Used "with" instead.

Same as previous comment, "using" instead of "on".

Same comment, "using the [outcome property]" instead of "on".

Do we have 'incomplete' as a valid result defined in our list of outcomes?

No, incomplete is not an outcome recognised in the ACT Rules Format.