– DRAFT –
Accessibility Conformance Testing Teleconference

ACT Standup

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

<dmontalvo> Wilco: Working through the WCAG feedback and writing responses, not a massive ammout

<dmontalvo> Helen: I opened a PR that has been approved by Jean-Yves, 2074 update table header rule

<dmontalvo> ... Waiting on the discussions that will happen once we resolve a related discussion that Trevor is leading

<dmontalvo> ... Updating the transcript rules

<dmontalvo> Wilco: That is part on me, too many things on my plate

<dmontalvo> Suji: Not much new after the annual review

<dmontalvo> Wilco: You could have a look at the issues I mentioned

<dmontalvo> Tom: Working on ARIA rules, parent child relationships with generics are changing for 1.3

<dmontalvo> ... Do we fix test cases for 1.2 or we wait for 1.3

<dmontalvo> Wilco: I'll bet the ARIA WG will want us to focus on 1.3

<dmontalvo> ... I don't think our rules should be inconsistent with 1.3

<dmontalvo> Wilco: Are you sur that is supposed to pass?

<dmontalvo> Tom: They say it's supposed to ignore generic

<dmontalvo> Wilco: We should raise this with the ARIA WG

<dmontalvo> ... Especially if browsers are doing what we say they are doing

<dmontalvo> Tom: I'll do that

<dmontalvo> Wilco: Feel free to loop me in

<dmontalvo> Daniel: Repo maintenance and Bikeshed error fixing

<dmontalvo> Trevor: Subjective applicability stuff, hoping to get to my open PRs tomorrow

<dmontalvo> Catherine: I finished the annual surveys

TPAC planning & registration

<Wilco> https://www.w3.org/2023/09/TPAC/registration.html

<dmontalvo> Wilco: Sign up for registration, that is open now

<dmontalvo> ... Meeting Thursday and Friday

<dmontalvo> ... We will be meeting a little later in the day to make it easier for US folks to participate

<dmontalvo> ... There is a Waiver program

<dmontalvo> Daniel: Also discounts for hotel that you can get to from the venue page

Open pull requests and issues

<Wilco> https://github.com/act-rules/act-rules.github.io/pulls

<dmontalvo> Wilco: Trevor suggested that to help prioritize

<dmontalvo> ... Can I get a third review?

<dmontalvo> Catherine: Yes, I can take a look

<dmontalvo> Daniel: I think we should close that, not the work the group feels need to be done

<dmontalvo> Wilco: This is a small PR

<dmontalvo> Daniel: I can take that

<dmontalvo> Wilco: Transcript rules is on hold until I can get to that

<dmontalvo> Wilco: 2050 has changes requested

<dmontalvo> Trevor: I think it's pending some of the subjective applicability talks

<dmontalvo> Helen: We are waiting for the process to be agreed on

<dmontalvo> Wilco: We could write rules while we discuss that

<dmontalvo> Helen: The feedback we've got is related to the outcome of that discussion

<dmontalvo> Wilco: 1959?

<dmontalvo> Trevor: I'll have a look

<dmontalvo> Wilco: Essential text changes needs some work

<dmontalvo> Trevor: I still need to make some changes

Subjective exceptions in the applicability

<dmontalvo> Trevor: Two topics to talk about today

<dmontalvo> ... First is part of JEan-Yves comments. Is subjectivity in the expectations any better than in the applicability?

<dmontalvo> ... Second. Can we start clasifying some of the types of subjectivity we may want to allow?

<dmontalvo> ... Jean-Hves took the different types of rules that we can write and established what happens with the different places were we can allow for subjectivity

<dmontalvo> ... Nice framework for authors to clasify their rules

<dmontalvo> ... He suggests if we should treat subjectivity the same everywhere? If so, that would apply in the expectations and in the applicability

<dmontalvo> ... Almost all of our approved rules do not have a subjective expectation

<dmontalvo> Wilco: Part of that was based on funds for us to write rules that were easily automatable

<dmontalvo> ... You are right that they are generally easier to write as well

<dmontalvo> ... Should we explicitly avoid writing rules in one category that could be in another? Maybe even requiring separate rules

<dmontalvo> Language rules: there is the third one where both applicability and expectations are subjective

<dmontalvo> Trevor: I thought all had objective expectations

<dmontalvo> Trevor: Depends on how we define common input aspects (language)

<dmontalvo> Wilco: These seems like boundaries that we are building

<dmontalvo> Trevor: We would rather you be objective in both, if not, be objective on the applicability, then on the expectations, and so on

<dmontalvo> ... The enforcement of that is what I think is more difficult

<dmontalvo> ... We wanted to be very clear on what our test targets were, that's what we wanted objective applicability

<dmontalvo> ... But not sure if not konwing what the expected results are is even worse

<dmontalvo> Helen: It is subjective in the fact you can interpret it in different ways

<dmontalvo> ... Are we trying to reduce the different interpretations?

<dmontalvo> ... For example, decorative images tend to be subjective

<dmontalvo> ... If the description is in the following text then it is easier to make it less subjective

<dmontalvo> Trevor: In most of the rules where we have decorative we've pushed it into the expectations but it should really be on the applicability

<dmontalvo> ... Can we just put in the applicability that it does not apply to decorative content?

<dmontalvo> Helen: For example an empty alt may be the applicability if in the context of an SVG

<dmontalvo> ... Expectation should match what you are expecting to achieve, how to get there is what can be subjective

<dmontalvo> Trevor: Let's discuss the subjectivity types

<dmontalvo> ... We do not want to allow rules like "applies to any type of non-text content"

<dmontalvo> [Screen sharing]

<dmontalvo> Trevor: These are my initial thoughts for categorizations

<dmontalvo> ... First -- something is styled as a [role]

<dmontalvo> ... Secon: more interactive, in addition, it has to operate like a [role]

<dmontalvo> ... Third: something that expresses something subjective (a text node is decorative / expresses something in human language)

<trevor> https://github.com/act-rules/act-rules.github.io/discussions/2061

<dmontalvo> Wilco: I like the idea of trying to refine these

<dmontalvo> ... "Operates like a check box" does not seem subjective to me. There are rules for that

<dmontalvo> Helen: Then you get to mobile

<dmontalvo> Wilco: There are still some expected behaviors

<dmontalvo> Helen: You can't define a check box and a button on mobile

<dmontalvo> Trevor: I could use a more complicated widget instead of check box

<dmontalvo> ... But the point still stands that there could potentially some degree of subjectivity

<dmontalvo> Tom: Is it "operates mostly like a check box" but it may not b a check box?

<dmontalvo> Trevor: You may have a div and some JAvaScript to make it operate like a check box but it may be difficult to programmatically determine that

<dmontalvo> Wilco: I'd call this "functioning"

<dmontalvo> Trevor: I changed that because "operate" was more WCAG

<dmontalvo> Trevor: I was looking at more interactive things

<dmontalvo> Wilco: Something can look like a heading and very clearly not be a heading. If you give a paragraph a heading style, that does not necessarily look like

<dmontalvo> Trevor: Either "styled as a heading or looks like a heading" is actually a heading

<dmontalvo> a heading

<dmontalvo> Trevor: Manual rules, consistent navigation. We could say that "any [...] that functions as a navigational mechanism"

<dmontalvo> ... Similarly to focus related when opening modal windows, although the trigger to open a modal window is still not very clear

<dmontalvo> ... Also we can qualify the expectations through exceptions,

<dmontalvo> Wilco: One example we've talked about in the past is transitions. Kwnowing when something is transitioning and when it is not

<dmontalvo> Trevor: If we were to start allowing subjectivity into rules and we said that you need to use one of these predefined types, we'd have templates for the language to use. Does that help in trying to pin down what people can use in the applicability?

<dmontalvo> ... We could create some predefined templates depending on the subjectivity types

<dmontalvo> Wilco: That is an interesting exercise

<dmontalvo> Helen: And exhaustive

<dmontalvo> Trevor: We miss types and we'd need to do some further tweaking, but we could formalize this

<dmontalvo> Wilco: What if we create input aspects?

<dmontalvo> Wilco: Let's think about it some more