Wilco: Last minut comments from Mary Jo - do you want to run us through it?
Mary Jo: I've added editorial comments
scribe: line 91: editorial change
Kasper: it definitely reads better
Mary Jo: line 98, what is it validation of? The e.g. in the middle was throwing me off
Kasper: Yes, it is validation of the HTTP
headers or unparsed html/css
... I can write "such as", but it seems we're missing something. But if
the intent is clear from Mary Jo's suggestion, we can go with that
<Wilco> of interest to ACT Rules. For example to perform validation of HTTP headers or unparsed HTML and CSS.
Mary Jo: It applies to all of the examples. I tried to fix it, but Wilco's way is probably better
Wilco: it's the same for all of them
Mary Jo: That was basically it. I don't have anything substantial, just want to make sure that it is clear.
Wilco: I will update it. Any other comments?
Charu: When we have all of these different aspects, do we see cause of conflicts because of differences between how browsers handle things?
Wilco: That should be handled through "accessibility supported"
Kasper: Also touched upon by "The means by
which the DOM tree is constructed, be it by a web browser or not, is not
of importance as long as the construction follows any applicable
specifications that might exist, such as [[DOM]]"
... if browsers haven't implemented it correctly...
Wilco: There were also questions about the examples
Shadi: I think that when there are lots of
examples in specifications, people will just copy the examples rather
than read
... but I don't feel to strngly about this
... [[CSOM]] can I see an example of that?
Kasper: We could include an example of a rule that tests color contrast
Shadi: It's just to make it more illustrative. There's a lot of text
Wilco: I will add an example
... Moe had a comment about the word "Aspect". She suggested some
alternatives
... Kasper do you feel strongly about the word?
Kasper: Not really. If someone can come up with a better word, that's fine
Wilco: I will update the editorial stuff and
merge in
... There is a bunch of editorial comments
... Moe had a comment "I'm a little confused about Subjective
properties. Are we saying that they must be defined objectively by the
rule, for example, a corporate logo is decorative?"
... I'm not sure I'm understanding what she's asking
... yes, the applicability has to be objective, it can't be subjective,
so because "cooperate logo is decorative" is subjective, it can't be in
the applicability
... does anyone not agree with me on that? No. Moving on
... Some more editorial... Then there is a question about "cannot tell"
... So if I just change this into "further testing is needed" is that
all right, or is the objection to the entire "can't tell" concept
Kasper: for me, it's the whole concept.
Because you cannot reach a cannot tell result with the way that the
expectations are now defined
... Alistair brought up earlier that if you cannot tell, you should
probably rewrite the rule. If it doesn't lead to a fail or pass, you
cannot execute it
Wilco: I suspect that we will have a lot of rules where you can automate some of the results. We need some sort of intermediate results
Kasper: I think we need a seperate format for intermediate results instead of muddying the results
Wilco: I think you might be right. We can
take it out, if there are objections against it. Anyone for keeping it?
We'll take it out
... The rest of the comments are editorial
<Wilco> https://github.com/w3c/wcag-act/pull/155#issuecomment-360190550
Wilco: There is one from Charu...
Charu: we are missing the word of that each expectation is discrete
Wilco: Let's discuss that. Kasper you had an example rule where one expectation led to the next one
Kasper: yes, but I would argue that we need that to be able to model hierarchies of expectations
Wilco: I'm not sure we want that
Kasper: Wasn't that the whole reason why we went away from the selector and steps model?
Wilco: This will still give you a possibility to build in steps, which is what you did, build in an execution order
Kasper: I just wanted to get away from the linear execution order. That was what was bugging me with the selector and steps approach. I wanted to show that we could still model this, but also tree structures or whatever we would need
Charu: Alistair's point was also that we need to keep the rules atomic, break them up into more rules
Wilco: We are at an impasse. We are close to draft publication. Charu, will you be okay adding this as an issue for further discussion?
Charu: yes
Shadi: It's the same, but completely
rewritten. I tried to be as specific as possible, but tried to leave
some room open
... It leaves a lot of room open for the community groups and leaves a
lot of room open for common sense
... trying not to be too prescriptive, but be specific
... objective, clear, short - but complete
... it's the whole workflow from the creation of a rule, right from
creating the pull request
... this uses what we discussed at the face to face at TPAC, that if a
vendor implements a rule, they can create a JSON file that tells the
results of the test cases. Self-declaration
... I tried to think about maintenance
... If it's a well-documented rule that there is general agreement on,
it should be easy for the working group (AGWG) to rubber stamp it
... let's leave room open for the working group to decide what to take
up or not
Wilco: Do we leave this on the wiki for people to review it or do we move it out?
Shadi: Maybe now is the time to put it somewhere
Wilco: We could put it together with the example rules in Github
Shadi: Yes. I'm just wondering where the more permanent home for it is
Wilco: We are a little behind on the publication. I want to send out a survey later this week
I'll add an editors note for rule example 2 with interlinking expectations
Wilco: Was there anything particular you wanted to get feedback on in the review process?
Shadi: No, it's all open for discussion
Wilco: There are still action items at the
bottom. But they don't look essential to get in before we ask for review
... Do you want to create github issues for these so that we can keep
track of them?
Shadi: Yes
Mary Jo: Do you want to talk about some of the open issues?
scribe: "Aspects under test" - I don't know what to survey
Wilco: We just survey'ed that one
Mary Jo: Ok, then it can be closed
Mary Jo, 138 was reviewed and assigned to Stein Erik. There are several that are assigned to people and still open. I'll send out emails to remind assignees
Mary Jo: Please try to get a resolution posted to your assigned issues, so that I can survey them.
Charu: I think 150 is ready for review. It's assigned to me
Wilco: That'll be in the survey, with the
examples
... R2 and R3 are ready for survey as well