Meeting minutes
Rachael: Announcements - there are 2 CFCs out, please check them.
Introductions and Announcements
alastairc: Call for anyone new to introduce themselves.
Chris Gimley: Joined today; works for IBM. Supports Carbon System Team. Joining because he wants to ensure that new standards are baked into their new guidelines.
alastairc: Two CFCs open, one is the WCAG 3 draft, closes tomorrow. Another, updates to the WCAG 3 explainer.
julierawe: Looks like the CFC for WCAG 3, it appears there are two email threads.
alastairc: One should say just WCAG3 draft, and the other should be WCAG3 Explainer, please be cognizant of which you are providing comments on.
alastairc: Time zones are going to be out of sync this spring. US timezone changes March 8. All W3C's meetings are based out of Boston's time. For Europe and UK, meetings will be an hour earlier. If you are subscribed to the W3 calendar, then it will adjust automatically, but it tends to catch people off.
shadi: Are we meeting on March 8?
<kirkwood> I’m sorry an hour earlier than ‘normal’ ;)
alastairc: Will advise if meeting on March 10 (that's the week of CSCUN). We didn't find that many people were going to the US to go to CSUN.
Conformance sub-group update
Rachael: Hearing a lot of concerns about traveling to the US. Looking for a conference that is in a non-US location, and will keep the group updated as we do. This spring we will meet asynchronously.
alastairc: Update from the conformance subgroup is next.
alastairc: Will cover what we've been doing, and what is coming up. There are a lot of tabs.
… meeting with a lawyer is still coming up. Have been working through things that are in and out of scope; what we're doing as a group. There's a lot of discussion on Conformance Discussion. This is a multi-author document, and it's not all 'agreed' yet. Whenever we go through what people want to achieve, we tend to come back to a few core topics.
The biggest is how to deal with the 'binary nature' of WCAG 2 conformance.
… Ongoing topics tab: everything is documented there. What goes where tab reviewed. Conformance challenges work from a few years ago is also included. Working through topics documented on this page. Trying to scope in and out.
… Lisa ideas tab, AC - Meta-assertion tab, Jennie proposal, Gregg - Conformance Proposal, Logius proposal tabs. These tabs include a walk through of each proposal as written as to how to approach.
Jennie_Delisi: When looking at the comments that we've received in the past about challenges acknowledging as a group to make accessibility improvements. Proposing using a weighted type of scoring. References ISO standards and EN standards that are concrete ways of scoring. This could help a group move ahead and improve their accessibility.
alastairc: Proposal from Greg when was adapting how the WCAG 2 conformance model works. Uses the WCAG terminology, but also has a conformance progress metric.
hdv: Proposal includes what we think is the most important to improve from WCAG2 conformance. Added a few tests for folks to check the boxes to understand if you have a good conformance model. Tried to write it down as a test-drive approach. Tried to come up with something that solves the problem that accessibility isn't binary, but we want binary
answers for whether you Emet WCAG or not. Introducing three different levels: not conformant, partial, or conformant. Partial is to help people figure out where they are, and what what they need to improve.
alastairc: Conformance Checklist tab - Basic questions that we need to work through when we're evaluating.
… Coming up on week 7 of meetings. Will likely be closer to 10 weeks in to work through the proposals that can be put into the next draft, if approved by both the subgroup and the wider working group. Any questions?
Sub group working on rules
<alastairc> https://
<Helen> https://
<alastairc> ACT https://
<AWK> Did you say that there is a template for ACT rules?
alastairc: No questions. Next topic: Subgroups are working on rules. Past few weeks have discussed rules and will breakout into rooms by subgroup. Work on the provision you are assigned to. Link above has the tracking the list. There is a template. Wilco and Helen have been filing in the ACT FAQs. Key thing is to start off with examples. break it
down into scenarios - using Flashing as the example, one with size, one was frequency, one was user agents. What would be a passing and failing example for each of those? Rules tend to be granular, so it won't test a whole requirement, it tests a scenario for a requirement.
<AWK> got it - pretty minimal template. :)
GreggVan: Struggling to figure this out. For flashing, you said to break it down. You could test the area, and you can't tell if it passes or fails, because the whole screen could flash, but if the flash isn't bright enough, it still passes. I know how to test the success, but I have no idea how you pass pieces of it.
Helen: To have a rule passing, it does not mean that it passes the success criterion passes, it means that scenario passes. You can only say that a success criterion passes when all the rules passes (or assess that it's applicable).
GreggVan: Does a fail of a scenario means it failed?
Helen: Yes
Helen: You have in applicability either that logic to allow for making sure that one of them has to pass for it to not be a failure, or you'd make it into a composite role because of the complexity. Steer away from complexity because they are harder to write. How you approach it might just be that way. Start small, group things together as it makes
sense.
GreggVan: I could take any one of our guidelines and name 20 different situations and the fact that you require this to always be true, but I guess it would be useful as failure examples. It doesn't help me test the rule, just test the rule to see that it fails. Are we creating a bunch of things that are failure tests, but they aren't enough to
sufficiently say you passed?
Helen: This is a can of worms, It's more complicated when you get to the more subjective rules. The idea here is to help people understand how to test the success criterion. Scenarios are meant to refine how to adequately test to know if it's a pass-fail for the scenario. Can't answer definitively. When it comes to making this accessible, there are
so many variations, we're trying to rule out people not understanding how to fully test.
GreggVan: Atomics are only useful when it is comprehensive. Are the only ones that are really helpful those that tell you if you pass or fail helpful?
Helen: If you have an atomic rule that fails, then the criterion fails. It's never going to say that success criterion passes, it's just saying that scenario passes in the example that was written.
Helen: A composite is one rule which merges, whereas atomics with dependencies are atomics. Then you have atomics that don't have dependencies, and they either pass or fail.
<Zakim> bbailey, you wanted to request screen sharing from group or chair who is making fair progress
alastairc: For the purposes of this exercise, work through the scenario, work through the passing, failing, and not applicable examples, because that helps us write the provisions.
bbailey: Asked if anyone would volunteered to show the work they've done as an example.
giacomo-petri: Thinks the ACT group was more focused on defining atomic rules to consistently identify what constitutes a violation. For your scenario, if you have a composite rule, you may need to test more than one rule, but if one if failing, then the requirement is failing even though it's a composite rule.
… Interesting example where this is working well: page view title. Link above.
… One rule is that title is present, so there should be at least one title in the root of the page. Another one where title is descriptive. Another rule is that the composite rule, and when combined, they ensure if you are passing both, it means you are passing the requirement. If you are failing one of the two, it means you are failing the
success criterion. This was simple and straight forward.
<giacomo-petri> composite: 1. https://
<Zakim> alastairc, you wanted to comment on an example
alastairc: Helen shared another example around content as alternative for visual reference, which is under sensory characteristics. It has applicability, which is any text note, visible or included in the accessibility tree. It has a number of exceptions. It goes through passing and failing examples. Might be a reasonable complicated example that
hasn't been tackled by ACT yet.
GreggVan: Thinks he figured it out: ACT is looking for tests for failures. Passing doesn't mean you pass, but if you do a composite and it really is comprehensive, it might, but it's mostly looking for failures. Take what is dependent on it, and put it in as preconditions for your test, and then all atomic tests are tests of failure. ACT helps you
identify things that would fail. Passing them all doesn't mean that you pass because it's not comprehensive.
alastairc: Confirmed Gregg's statement
shadi: Nearly correct, and think it's because its the way WCAG 2 is written; the requirements are so big that you can hardly have a fully thing that maps. There is a section that you list conformance. Secondary, you can test, you can put any kind of requirements that you want to map to. Maybe your internal guidelines something to do with
accessibility. What happens when the outcome fails? You fail the criterion. If it passes, you still need further testing. You need multiple atomic tests in order to be able to make a judgement on the actual requirement and how it maps.
<Zakim> alastairc, you wanted to comment on wider docus
giacomo-petri: As a group, we need to understand how to write examples. The goal of the ACT is to handle scenarios that are edge cases. Going back to page title example, having two <title> tags is not a failure, and these scenarios need to be included even if it wasn't a best practice. There are scenarios that are the best practice in how to do
things, not just sufficient to pass a criterion. ACT wants to be consistent and consistent behavior across technologies to determine what is passing, what is failing, and what is inapplicable.
<shadi> +1 to alastairc
alastairc: Remember/reminder, these are the most granular part of the informative documents under each provision. Under the provision, we will also have a high-level, generally how to test this. Then there will be methods (techniques) these are going to be the positive ways to pass something. A known way of passing it. This covers good ways for
recommending doing things. The ACT rules sit underneath that as the most granular. Our purpose is to break down the provisions into scenarios and think about pass/fail examples.
alastairc: 8 provision based subgroups; join the breakout rooms. Treat it as a sub-group meeting.
Rachael: Those that don't have a subgroup, please stick around and one will be assigned to you.
<Zakim> bbailey, you wanted to ask for examples of (probably composite) ACT Rules which result in "success criteria is satisfied" rather than "more testing is needed"