Meeting minutes
CSUN recap
<sashanichols0> +
Wilco: Jean-YVes and I presented at CSUN two years ago
… A couple people said they'd be interested in joining
… Will wait and see
Suggestions for rules to be sent for AGWG approval
Shunguo: Do we want to clean up examples? For instance, based on HTML5 summary element is only.l used as a first child of the details element
… Similarly to fieldset and legend
Wilco: I don't think we want to change the example here
… These are things we want testers to pass even if it's invalid HTML
… I do think it's worth noting that this is not valid HTML, maybe in the rules background
<Wilco> https://
Wilco: Thee rule I'm thinking of is the one in titles
… We have a similar thing going on where we include multiple titles
… We know it's not correct but it doesn't cause accessibility problems
Shunguo: The problem is if you have an empty summary element
Shunguo: Another is if the summary uses aria-label and summary happens to contain text
… If you give aria-label the text isn't available
Wilco: We have other rules for this
… We try to have our rules focused
Shunguo: Then that's fine
Shunguo: aria required owned element there is an inapplicable case that to me is a pass one
… They all have an implicit role
… Inapplicable example 2 is a native UI. The list item is implicit role
Wilco: This only applies to elements with explicit role
Shunguo: There are mixed cases
… Some element have implicit,others have semantic role
Wilco: The applicability here is about explicit elements. The expectation allows for explicit and implicit. We have separate rules for implicit scenarios. For example table rules for native elements
… Are you suggesting that these should be the same rules?
Shunguo: Fine. But to me the custom widget example would be mixed. How would you handle this?
Shunguo: You have explicit for some, for some you don't have because it's using implicit role
Wilco: Is that a problem with the rule?
Shunguo: This is tricky for custom widgets where it might not be easy to tell whether they are explicit or implicit roles
Wilco: Does anyone want to weigh in as to whether a distinction between implicit and explicit role is worth pursuing?
Kathy: Not clear what the ARIA spec includes
Shunguo: This is the ARIA in HTML spec -- div and span have implicit roles of generic
https://
Shunguo: It first marks an implicit role and then it says which other roles you could use
… For this rule we should consider implicit and explicit equally
… There is no difference in whether the role is used from the default mapping or specified by the author
Wilco: I was fully under the impression that we had separate rules for native elements and for ARIA
… But I am realizing I can't find these rules
… So I think you are right, I think we should have to supply the implicit semantics
… ul should have li elements
Kathy: ARIA required owned elements rule is mapped to 1.3.1
… Are we looking at whether these elements are present or just at the info and relationship?
Shunguo: I think this is required by WCAG. Other case for this is custom widgets that are created using ARIA
Kathy: The way the rule is written I am not sure if it'd fail. But I understand you are trying to go beyond to cover custom widgets
… I understand the implicit role concept, but just the way the rule applicability is written, the example does not have the explicit role, which is inapplicable
Wilco: I think the examples are correct for the rule as it's written. But this rule could cover also things that have implicit role, and Shunguo has a good point on that
… I think we should open an issue for this, it wouldn't be that complicated to change
Kathy: Even if another rule did exist that wouldn't meet this example?
Wilco: If we change the applicability to "any semantic roles" it would pass.
Kathy: Are you suggesting taking out "explicit" from this rule?
Kathy: OK, I'm fine with that.
Wilco: Do we need that update before we send the rule to AG?
Daniel: I wouldn't send something to the larger group knowing that we are about to change it
Wilco: We've done it in the past
Rachael: How long would it take to make the change?
Wilco: It depends on who is up to work on it
Shunguo: I can volunteer to take this on and ask for your review
Wilco: That's appreciated. You could open an issue or open a PR straight away
Wilco: I'll take on the summary group
[[Overall agreement on discussed approach]]
rules format 1.1 backward compat
Wilco: Trevor opened this a while ago. This is about backwards compatibility of the rules with 1.0
… The conclusion we came to is that 1.1 is not quite backward compatible with 1.0. The reason being that there are some requirements in 1.1 that are not in 1.0
… Specifically we are allowing the applicability to be subjective
… Daniel proposed adding something very explicitly in the rules format acknowledging that it's not backwards compatible
… Is this a reasonable way forward to everybody?
Daniel: I'll take silence as agreement -- I'll put a PR for this
satisfying tests
<Wilco> w3c/
Satisfying test
Wilco: The quesztion here was that the rules have passed and inapplicable examples. Passed means 99% of the times that it requires further testing. Rules almost never can tell it the SC is passed, they can tell for sure when something fails
… There is this mismatch -- we talked about this extensively when we wrote the first version of the rules format
… We came up with satisfying texts for when passing the rule means actually passing the SC
… The open question is if this address Tom's concern?
Kathy: I don't think we have any satisfying text
Wilco: I don't like the name anymore. It's a rule, not a test
Wilco: Maybe sufficient rule
Kathy: I think we've also added information when an atomic rule is part of a composite. That might be adding to what Tom is looking for
Wilco: I've been working today on whether we could be more explicit for some of the composite rules
Kathy: He also says "Better way to describe the impact of passed versus cant tell"
… His use of cant tell may be different from the cant tell in implementation results
Shunguo: There is no one on one mapping of ACT rules and SCs?
Wilco: No.
Wilco: Even if the rule's applicability is fully tested and passed, that doesn't mean there are other things on the page that also need testing
Kathy: Would our caption rules be satisfying tests?
Wilco: No because there are other ways of embedding video in web pages
Wilco: I think we've gone as far as we could go without Tom. I'll reach out to him and see if he's OK with the outcome of this discussion
Simplified navigation
Wilco: There might be a potential for ACT rules in that work, especially if it's programmatically determinable
… Wonder if we could bring this up with them
Daniel: I need to engage with Matt for a couple of comments, I'll do that.