Wilco: 215 pull request
<Wilco> Sometimes there are multiple plausible ways that an accessibility requirement can be interpreted. For instance, it is not immediately obvious if UTF emoji characters should be considered "text" or "non-text content" under WCAG 2.0. Whatever the interpretation here is, this MUST be documented in the rule.
Wilco: This is a comment from button house
name rule? Adding a situation which was not there.
... Would like input on the example.
... Do people feel this is a good example.
Stein Erik: Good example.
Anne: What is UTF?
Wilco: Unicode.
Anne: Maybe put in a link to UTF?
Stein Erik: We don't have to be that detailed.
Anne: I don't really understand the text.
Wilco: It may not be obvious.
... Would it be OK to say Emoji characters.
Anne: It is also a AAA requirement to define abbreviations.
Stein Erik: Remove UTF reference.
Anne: Would work for me.
Moe: Last statement interpretation - maybe we could just drop the word "here".
Wilco: Committing it...
<Skotkjerra> +1
<Jey> +1
<maryjom> +1
<MoeKraft> +1
<trevor> +1
Wilco: vote to merge
<anne_thyme> +1
Wilco: Moe any updates on this?
Moe: Started on it this morning, referring
back to minutes from last week.
... We stated that we would link to the overview pages, but from minutes
I was not sure.
Wilco: It is the overview pages from the WAI specs.
Moe: Thanks.
Wilco: Walter will review. Could you complete for today?
Moe: Tomorrow.
Wilco: What is the status of this?
... Did this get updated since last week.
Stein Erik: Minor change last week.
Stein Erik: There are a few grammar issues.
Wilco: I can do that change.
... Please read whilst I update the document.
<Skotkjerra> +1
<Jey> +1
<maryjom> +1
<anne_thyme> +1
<trevor> +1
Wilco: Vote for merging.
Wilco: This could be a bigger topic.
<Wilco> https://github.com/w3c/wcag-act/issues/214#issuecomment-393162258
Stein Erik: In general a good suggestion.
Wilco: Last week we proposed what to do with
rule groups.
... What came out of this was the proposal which was posted yesterday -
the proposal was atomic rules; and composed rules - which aggregate
atomic rule results.
... Multiple ways - you need two or more passes from atomic rules for a
composed rules.
Stein Erik - Atomic rule tests web components?
Stein Erik - We need to be clear between an atomic rule and a composed rule.
Wilco: Jey - you were having difficulty with this proposal.
Jey: I shall read it again.
... Examples would be good.
... For a new comer it is a bit difficult.
Trevor: x number for a composed rule to
pass.
... or is it specific.
Wilco: you are asking - are there other
aggregation methods.
... It is usually one or two in a group which need to pass.
Anne: Site Improve concluded that if one rule was needed to pass that was an atomic test.
Trevor: If you have atomic tests in a
composed rule - you need to keep track of all the atomic rules.
... It could get messy.
Jey: One of many rules has to pass.
<Wilco> https://github.com/auto-wcag/auto-wcag/issues/149
Wilco: Wanted to share an example.
This rule has a composure procedure.
<MoeKraft> Alistair: I understand where this is coming from, We're saying WCAG criteria is passed if one of these is passed? Correct?
<MoeKraft> Wilco: No we are not saying that. If you fail a rule, you fail SC
<MoeKraft> Alistair: Video transcript, if not there are other ways to pass. We are saying if one is passed then the rule is passed. This is tricky. We are saying it has passed based on what we think. Many different ways to pass
<MoeKraft> Alistair: Need to test for failures. We can't say we definitely found that you passed.
<MoeKraft> Stein Erik: But we are not saying you passed but you didn't fail.
<MoeKraft> Alistair: Test target passed in one of the following rules. If I have video transcript, I passed the test.
<MoeKraft> Alistair: If I fail any or all I fail.
<MoeKraft> Wilco: It's a scope question
<MoeKraft> Alistair: But it
<MoeKraft> Alistair: It's constraining. Not how WCAG is designed
<MoeKraft> Wilco: Disagree. Not automatable
<MoeKraft> Alistair: You are testing 3 options. What if there are 5 options?
<MoeKraft> Stein Erik: WCAG helps us document conformance
<MoeKraft> Alistair: If I fail all 3, I may still pass
<MoeKraft> Wilco: Yes. WCAG says what you need to do. It should cover everything
<MoeKraft> Alistair: But also states non-exhaustive
<MoeKraft> Alistair: We need uniformity across SC
<MoeKraft> Wilco: Not every SC is testable or automatable
<MoeKraft> Alistair: We are constraining people
<MoeKraft> Stein Erik: Not following
<MoeKraft> Alistair: 2 ways to pass SC. 1. Sufficient techniques, we can say we passed but constrain to sufficient techniques. 2. There could be other sufficient techniques that we are not testing.
<MoeKraft> Alistair: We only know that we have failed the given tests
<MoeKraft> Anne: I agree with Alistair. Rules should only test failure techniques. Should each rule group have another atomic rule to ask if there is another sufficient technique that needs manual test
<MoeKraft> Anne: With atomic rules, we test for failures. But with composed rule we are testing if we passed.
<MoeKraft> Stein Erik: Depends on SC
<MoeKraft> Anne: What if we use a technique not on the list?
<MoeKraft> Alistair: This is a big problem really. We didn't want to go down the sufficient technique route
<MoeKraft> Wilco: When you do not use any of these 3, you fail SC
<MoeKraft> Stein Erik: Some SC easy because limited. Others not so easy
<MoeKraft> Alistair: Breaks down mechanism
<MoeKraft> Wilco: Just because it doesn't for all SC, doesn't mean it isn't good
<MoeKraft> Alistair: How? Mix and match Not uniform
<MoeKraft> Anne: How would you handle Multiple Ways? We have Sufficient techniques but also have other ways to meet
<MoeKraft> Stein Erik: This includes auto, semi-auto and manual
<MoeKraft> Alistair: Customer should be allowed to select techniques. Otherwise, testing becomes a constraint
<MoeKraft> Alistair: Don't want to be constraining factor
<MoeKraft> Wilco: There are good ways to write rules. If there is a good technique, should be accounted for
<MoeKraft> Alistair: How?
<MoeKraft> Wilco: If we do not have a finite set of solutions, then this will not work
<MoeKraft> Stein Erik: This is a WCAG design versus rule design problem
<MoeKraft> Anne: Yes. But not what rule groups are doing
<MoeKraft> Alistair: Not testing for failure but passing of techniques. Concerned that we are constraining
<MoeKraft> Wilco: Give me an example
<MoeKraft> Alistair: SC 1.1.1 Not only do you need all types of content, need all different alternative techniques, alt, aria-label, etc. Need to test zillion different techniques to be confident.
<MoeKraft> Alistair: No video transcript and no audio transcript, does it fail?
<MoeKraft> Wilco: Yes. Need 1 of 3
<MoeKraft> Some videos are descriptive and do not need audio descriptions
<MoeKraft> Alistair: Not allowing all options but constraining them
<MoeKraft> Stein Erik: Composed rules are meant when needing to consider many different factors are met.
<MoeKraft> Alistair: Do we pass if we have a video transcript?
<MoeKraft> Wilco: I agree we are focusing on passing and this is not correct
Alistair: I'll take back the minuting...
<MoeKraft> Stein Erik: Let's send back to committee and have another go of it.
<MoeKraft> Thanks
Alistair: Thanks Moe.
Wilco: Proposal for next week.
... Rules should indicate none conformance.
... We should have another pass on the rules format to make sure that
that is right.
... Nearing the end. Opening to other comments...
... We can schedule another meeting to update the format.