Meeting minutes
<Zakim> Chuck, you wanted to introduce myself
ACT stand-up
Carlos: No progress, other work
Gaicomo: Lots of PRs that are ready for review
… Working on new rules. One on required accessible names
Dan: Left some reviewed, have an issue on the meeting's agenda
Helen: Waiting on Kathy to review my PR. Don't want more than 1 PR at a time
… The PR on transcript, 2064?
Tom: On vacation, getting caught up. Wilco can look at the last open issue?
Wilco: Will review Giacomo's PRs tomorrow
Task Force update
Helen: TF & CG meetings have been merged
Daniel: We've published ACT Rules Format 1.1. first public working draft
https://
… Rest of W3C groups now get to review. This is the start of the process.
… There are still a few steps before the draft can become a recommendation. Now is the time for everyone to review
Tom: For merging the two groups, it looks like the regular TF meetings are still on the calendar
Daniel: The planing team should discuss if the merge will continue through to the end of the summer
… I'll mark these as tentative
… For June we've merged. We'll continue for the summer, we'll get back to it in the planning meeting.
Carlos: Regarding the rules format, what's the timeframe? When will the comment period end?
Daniel: There's a 2 month period for horizontal review. Then if everything's resolved we can publish another working draft
… To get to rec that's a longer period.
… If we want further changes, the sooner the better. That way we can then publish another working draft
Dedicated call -- Intro to ACT
Carlos: In the last few meetings we've had new faces. The goal was to ask if the new members would like a dedicated introduction call
… We don't have people here who require it
Wilco: I'd suggest we just e-mail these new members to ask
Carlos: Agreed, even if it's one or two people
"Zoomed text node is not clipped with CSS overflow": "link to a full version": source unclear act-rules/act-rules.github.io#2163
Dan: There's a WCAG issue update. This has been clarified, I was misreading the understanding doc.
… A link to a full version is okay in all cases, not just in limited cases such as user generated.
Carlos: So if we have a link to a compliant version it would pass
Dan: That's not exactly the wording. They were focusing on without loss of content.
Wilco: CAV includes no loss of content. It has a few more requirements, and is for every criteria.
Dan: So this isn't in this criteria, it applies overall.
Carlos: If this is just about CAV, should we have specifically for this rule, those kind of passes
Dan: What's been discussed so far is not for alternative versions
… I think the link to the full text is covered by the rule's assumption
… I could submit a PR that elaborates on the assumption.
<Chuck> Wilco: Do we have real world examples where we see this fail?
<Chuck> Wilco: Generally I think we want assumptions to only be known edge cases, but not happen in practice. If it happens in real sites, common scenarios, we want a rule to cover it.
<Chuck> Wilco: Seems like the quickest fix.
Dan: This might be important too. I don't understand why pass 2 is a pass
… pass 2 has ellipses and a link to the full version
giacomo: I have examples of product sites that force a full product name onto one line
… If you the move to the product page you get the full product name
Dan: That sounds like the thing I've seen too. We should address this
<Dan_Tripp> https://
<Chuck> Wilco: From reading of it, it passes because expectation 1 is true. We are explicitely excluding where the clip is intentional. We know it's a potential problem, if designers accounted for it, they know to sort it out elsewhere.
Dan: So this passes because of expectation 1, and the link to the full reason is irrelevant to the rule.
… So removing that link would still pass the rule but fail the SC
Carlos: so the link to the full version might help people understand the example better.
Dan: Maybe I can clear this up before boundary examples go live
Do branch publish for easier reviewing of PRs act-rules/act-rules.github.io#2183
Carlos: This was brought up by JY, a consequence of the rules being written. Reviewing rules can be quite complex. Some examples are very hard to review.
… JY came up with doing branch publish to let us more easily review what's proposed
… Mark came up with an easier solution to let us generate the rules page from any branch
Mark: We can publish whatever markdown's checked out. You can also use it to preview your own PRs.
… We'll use Hugo to launch the web server. That can have changes appear live in your browser as well.
… It's about 45 lines, and all you need to do is install Hugo
Carlos: So this requires every review to install and run it
Mark: You could push the output somewhere.
Carlos: To make it useful for everyone, a solution that could be integrated with the build system might be more useful. Of course we could use this imiediately.
<Chuck> Wilco: Getting into the build system, the tricky part is that the files we write are not the files that go to W3C website. There's a conversation step.
<Chuck> Wilco: there's a template that gets incorporated, and we don't have that. I'm certain there's a way to do it.
<Chuck> Wilco: Maybe build locally. How close do we need the previews to match the final published content? Do we need to open code snippits for example? HTML snippits ok in one file?
Daniel: We're basically using two websites. A CG website, still running on Gatsy. We could try to run Gatsby in Netlify, it should do most of the things
… We can't use W3C's Netlify, it's a different W3C organization
<Chuck> Wilco: Not clear to me. If we want the markdown display, github does that for us already. Do we need more than that? What do we need?
<Chuck> Wilco: Does everyone know that github can do that?
<Chuck> Wilco: This kind of view <link example above>
Carlos: The issue is that reviewing examples is difficult.
Mark: There are a lot of files. Knowing how these interact is hard to do.
<Chuck> Wilco: Can't solve that problem, code snippits not combined in meaningful way.
<Chuck> Wilco: Code snippits are not rendered, and that's what we are discussing.
<Chuck> Mark: Looked hard to modify, but if one has never looked at code before, probably looks that way.
<Chuck> Wilco: We have a lot of what we need already set up. The thing we are running before, we would need to run that and publish from that. That generates the HTML files already. We can stick that in a preview. If we can stick ... into Netlify that would do the job.
<Chuck> Daniel: There is a free tier.
<Chuck> Wilco: We can start there.
<Chuck> Wilco: Question is who is going to build it.
<Chuck> Carlos: Much is already built.
<Chuck> yes
Carlos: I might have a look when I'm caught up
Boundary / Verbose examples act-rules/act-rules.github.io#2185
Carlos: This suggestion came from Dan. An additional set of examples that would not be pass/fail/inapplicable, but that would help illustrate edge cases.
… For example to show how a rule might pass when a criterion fails.
Dan: One question is what we want to call these. JY has called them boundary examples, but there are some uses that mean they're not boundary examples.
… I think there are other example that are not boundary cases, but they are more verbose. We'd benefit from having more verbose examples.
Carlos: I think JY places them before inapplicable?
<Chuck> Wilco: They are all at the bottom, there's a new boundary section.
Carlos: They're not normative, so they should be after the examples
Giacomo: I think that boundaries and verbose examples have two different meanings.
… One highlights edge cases, the other is something that's too complex, but it's not an edge case. I don't think we should put them together
Carlos: Would verbose examples need to be normative?
Dan: I think so
<Chuck> Wilco: 2 points. We really need to be careful with "normative". Rules are not normative. Need to find another term.
<Chuck> Wilco: I'm not a huge fan of either solution. Maybe something like advanced examples. Some generic term that puts them out of scope of the fault/pass fail inapplicable may be the way to go here.
<Chuck> Wilco: We should also be careful about how many examples we have. Every example added is an example we need to maintain. There's a reason we limited the examples.
<Chuck> Wilco: It all needs to be maintained, and is extra work if they go out of date.
Daniel: +1 about normative. I struggle with verbose and boundary examples
Carlos: I was thinking about additional examples.
… Regarding normative, I think if we just remove normative, it should be enough.
Wilco: I like Miscellaneous examples better
… that sets them aport from the other categories
Daniel: These go into different topics
Helen: Maybe we can have this as a topic for next time
Tom: How would you specify the expected result of these?
Carlos: I'm not sure I understand the question
Tom: They're not part of the normal pass/fail. The idea is not to report results to them. Even if they're reported we would ignore them