W3C

Accessibility Conformance Testing Teleconference

21 Apr 2022

Attendees

Present
thbrunet, Will_C, Wilco, kathy, trevor, Helen, Daniel
Regrets
Chair
Wilco
Scribe
thbrunet

Contents


<Wilco> scribe: thbrunet

ACT rules sheet and Survey Results

Wilco: Aria-hidden where are we?

thbrunet: Took a look through the two PRs, some more changes. Please review

kathy: PR for taskforce comments. Carlos has another PR and asked for assistance. Not sure if necessary for this rule, but can talk about later

Wilco: Is it something we need for feedback?

kathy: Related to live streaming

Wilco: I don't want to wait on that. Carlos can work separately on that.

kathy: Carlos can do as a separate PR?

Wilco: agree

thbrunet: I opened PR on the Pass to N/A

Wilco: There's one by Tom, one by Helen, I have one more to open

Open ACT pull requests

Wilco: I'll review 1835. Helen and Daniel also to review.
... 1834 already approved. Can go to call for review

Helen: I wasn't sure what to do. Same process?

Wilco: Yes
... 1833 - Unify expectations - Why combining two expectations?

Helen: Unifying it because there are two rules. Jean-Ives issue. This is about making sure the same format is used for rules on the same subject basically.

Wilco: I would go the other way. If there are rules that split it, split it.

Helen: He wanted that grouped.

Wilco: I'll assign myself as a reviewer. Anyone else? Kathy?

Kathy: Ok

Wilco: That was 1832
... 1831 - Will this needs some attention from you, some requests for changes

Will_C: Will reach out if I need help

Wilco: 1829 - I should fix typo. I would like some reviewers.
... Tom, Will, Daniel
... 1828 - Needs a review from Will.

kathy: I fixed something, so it needs a re-review

Wilco: 1827, 1826 are approved, so I need to send those out
... 1822 - I think this has been a week
... Tom changes requests on 1819, 1820. Please dismiss the reviewers next time.

thbrunet: Will do that next time

Wilco: Also true for the other I assume.
... Helen said 1809.

Helen: I'm not reviewing. Waiting on Wilco

Wilco: Will do tomorrow
... Busy working on WCAG charter.
... Anything else?

kathy: Do we want to look at 1661?
... Carlos said he's blocked. Need live stream example.
... He's run out of ideas

<Will_C> I just approved Kathy's change in #1828. It should be good to go

Wilco: I know I've looked at that too. I wrote those originally. Brian and I looked and found nothing. We talking about this on the task force before. Does anyone know of free to use live streams that don't have lots of terms of use on it? That's what we need. An always on stream that you can embed anywhere that is royalty free. Short of that, not sure what we can do other than not have examples.

trevor: All I can think of are some music channels that are just on constantly

Wilco: NASA has some feeds. Webcams. BBC has some things. But, none of it is free to use.
... always some terms that prevent you from embedding

Helen: Are they allowed to have adverts?

Wilco: Not sure how they could. Needs to be a stream. Don't want a bunch of 3rd party code. We just need the audio stream.

Helen: When it's free, there are always adverts that break it up

Wilco: Usually not just a stream, but a player with it that complicates the examples.

kathy: I was looking at this one because of the note under the play button. This seems to try to fix it in many places for different roles. I was thinking I'd ask Carlos to pull it out of this PR. I'll ask him how to best do that.

Wilco: This play button definition?

kathy: Yes.

Wilco: I think we take this definition out and put it in its own PR. Adds another rule. That's great, but don't need to wait on this to move forward.
... I don't know what answer to give about the audio stream. There's live radio, but again terms of service.

kathy: I'll let him know we talked about it.

Helen: 1832?

Wilco: I did miss that. Sorry. I'll assign myself. Any other takers?

kathy: Can add me

Wilco: Thank you. I think we're good
... any other PRs?

ACT rules gap analysis - what does it mean for ACT rules format 1.1

Wilco: handing it to Trevor for this.

trevor: Won't be super long. Will assign some homework
... Last time we started talking about gap analysis.

<trevor> https://github.com/act-rules/act-rules.github.io/issues/1836

trevor: Started talking about stateful rules. I threw a lot out at the group without a lot of background. Spent some time putting together what had been done before.
... I compiled what I see as the main resources that we have to build off of as we look at more state in the ACT rules.
... To go through what's in here, my preference would be to let you all look through it for a week and come back after everyone has had a chance to look through and has questions.
... the couple of things in the issue is the ACT state draft. Myself, Jean-Ives, Carlos, Will, Lionel sat down for several session and discussed what it would take to put state into ACT rules
... Inside of there, we discussed about how we need stateful elements for applicability and expectation because you might need to describe state you start and end, and expectation is that the transitions between state exist and you meet some expectation
... Just for historical reason, I included this issue by jean-ives that is the furthest back I could find on this discussion. He has a lot of good thoughts.
... Can get you started and understand the basis.
... And finally, just to show examples, for some of the changes proposed. Both Carlos and Jean-Ives proposed some that follow this format.
... Somewhere in here it has the state definition. In Jean-Ives. Subheading of applicability, what states it's applicable for.
... For this one, we did not have to change the expectation to work with the stateful rules.
... Please read through these resources on what state means and what we've looked at in the past. I don't think me talking about state for a half hour will get that across. Some questions I thought, I've posted at the bottom of the issue. 1) What rules are going to need state? I think this question is trickier and where I got us confused last time. Some rules require state explicitly. Some can use it implicitly.
... Blocks of content, I can foresee state changing, but it's implicitly handled. I'd like to find that line where state is happening vs implicit and feel okay about it. Tangential to that, if there are rules that are not using state where it would be better to make explicit. With block of repeated, would it be easier if we explicitly include states. Would be a good discussion to have.

<Will_C> got to drop, sorry

trevor: And with the subjective applicability, in our draft, we only used CSS pseudo classes for state definition. We said we could not handle JavaScript manipulation because too subjective. With subject applicability considered, opens that up. Challenge - still need to figure out how to structure the language to describe those JavaScript manipulations or it becomes a free-for-all
... I'm of the opinion everyone should do some background reading and we circle back to this when everyone's on the same page.

Wilco: We can send it out to the mailing list to read for next week's meeting

trevor: I may not be at the next two week's meetings

Wilco: Let's not schedule until you're available again.
... When we're reading for that conversation, we'll send out a notice. A week we don't have surveys would be good also.

trevor: I'll let you know. I might be around, but uncertain

Wilco: I want to ask though. This topic seems a little nebulous. When we have these conversations, it gets quiet on the line. Is this problem making sense?
... Or too vague and need concrete examples?

thbrunet: I think I need to read through that issue and get a sense of what's already covered.

Wilco: Example I like related to state is color contrast that doesn't account for state. Just says 4.5:1. Challenge there is if you're actually strict about how you're doing this and you have an animation, during that fade, contrast will drop below 4.5. So there's this actual weird exception that everyone understands exists, but it isn't written anywhere that you don't test during animation - you test after it's done.
... I think one of the topics that the subgroup touched on is that there's a transition period and need to understand what a transition is and say that rules don't apply during a transition - maybe some don't and some do. Not sure which do.
... Does that frame one angle?

thbrunet: Is something like hover covered here?

trevor: Carlos's issue is sort of along those lines. If you have a link in a paragraph and it's hovered, what do you do?

Wilco: The difficulty with that. I'm digging a little deeper. WCAG doesn't say to apply it to all states. Should we really have rules that tell you? The requirement isn't to test it in those states.

trevor: I'd think that a naive in a simple way interpretation. State transition and there are flashes, still can't break the seizure rule.

Wilco: Great example. That suggests there are rules during a transition.

trevor: Be all end all, regardless of the standard, there is real impact on people. If you tab and link changes, that become unusable.

Wilco: I agree. But more importantly, AG has agreed. There are techniques that say applies to all states.
... it's explicitly written in the understanding. Why I'm asking is, is it a question of responsibility?
... Is it the responsibility of the SC to say what states to test? Or the responsibility of whoever is performing the test to find all of the states?

Helen: One example about screen reader. User will know to take out verbosity. But, how do you know? Can't assume everyone has the same knowledge. I'm careful about assumptions regarding the users.

dmontalvo: Going to add +1 to that. We're talking about interactions. When we talk about complex widgets and we talk about selections, that may have impact on screen sizes and reflow. We would try to get to those on what people should be testing, etc.

Wilco: Kathy, what's your view?

kathy: Because Trusted Tester breaks out into different categories, we're able to put some changes of state under forms. But, we don't cover it everywhere it needs to be covered. We look more for change of state and instructing the tester - if you expanded something that was collapse, then do this.
... The change of state is more focused than looking at a state and knowing what the new state would be and providing instructions.

Wilco: For the hover and focus links example, would Trusted Tester say to look at contrast?

kathy: We've avoided that because we find contrast difficult to meet. Hovered, visited, unvisited, etc. Difficult to meet them all together.
... so we avoided it.
... But we have discussed all of the different states for links particularly and haven't figured out yet if that's a requirement that can be met when compared with each other and not just against the background

Wilco: Yes, that's almost impossible if you have more than two states, with you absolutely do have in links

Helen: And what if customizable...

Wilco: Lots of different states on links.

Helen: A lot are set by the browser though? Browser defaults...

Wilco: Yes, but you can set them explicitly.
... Not sure why web has so many states.
... Is this topic starting to resonate? We're going to be starting to talk about this a lot.

kathy: Would this apply to checked / unchecked and different possibilities?

trevor: Depends on if it causes other external changes. If you have an error message that pops up, definitely new state.
... Just the check/ uncheck, maybe, if it changes the color contrast.
... Depending on how we require state reported in the rules, maybe only need larger changes, but up for discussion.

Wilco: One more thought, then want to move on. Have we talked about timing in this? Do we need to? One of the things I'm seeing is there are several rules where we've added expected this to happen within X time where that time is completely arbitrary, but we needed something to test consistently. Is that something to explore?

trevor: I think so. We have those meta refresh rules, except the time horizon is long.

Wilco: Yes, that one comes from WCAG, but others where we need something, so..

trevor: Might fall into might not require states, but better explained with states included. My initial thought at least.

Wilco: We'll pick up when trevor is certain of availability again

AG recharter & ACT

trevor: I know these state discussions get rough on the scribe.

<Wilco> https://raw.githack.com/w3c/wcag/charter-2022/charter.html

Wilco: This sort of came out of nowhere. But, AG chairs were working on WCAG 3 related update to their charter. AG is rechartering. Nov 1 need new charter approved by W3C to continue working. As part of that, we sent a list of things we're hoping to do.
... Rochelle emailed me Thurs after last week's call asked me to write something. Survey closed Tuesday.
... Wanted to bring this back so everyone's aware that we have some things that the next charter drafted related to ACT and want impressions.
... Two parts to this. In 2.2 Non WCAG 3 Scope, there is develop ACT Rules Format 1.1
... Item 2 is right below that, provide guidance and test materials to support 1.0

<Wilco> develops Accessibility Conformance Testing (ACT) Rules Format 1.1, to keep ACT rules compatible with WCAG 3, and enable writing rules that have proven difficult to write under version 1.0,

Wilco: That's for next Charter Nov 2022 to Nov 2024

<Wilco> provides evaluation guidance and test materials to support verification of conformance to normative guidance (such as Accessibility Conformance Testing (ACT) Rules Format 1.0),

Wilco: If anyone not cool with that, need to bring it to AG chairs soon.

thbrunet: Is there a thought on what needs to change for WCAG 3 yet?

Wilco: Not really yet. WCAG 3 still exploratory. Work happening that I'm keeping a close eye on how ACT relates. One area on how to map ACT to requirements. If WCAG 3 has a rating scale, ACT needs to be reporting to a rating scale rather than Pass / Fail. Don't know yet, but that one thing. Also might incorporate ACT rules more directly.
... Every method might be it's own rule. That would require some changes as well.
... Keep up and make sure we don't deviate.

trevor: If we figure out the stateful applicability, and if we want to change for WCAG 3, would this prevent us from starting that next working draft? Don't necessarily want to wait to be constrained on waiting on WCAG 3.

Wilco: I'd be surprised. Getting to a recommendation is a length process. We'd need almost a complete draft within a year. Can always start working on a 1.2, don't think we need a charter for that. Just need charter if we're expecting to publish within that charter period.
... We have Rules Format 1.1 as a deliverable within the charter period, but we can work on other things too
... Daniel, would it be good to put out a CFC to approve these portions of the charter?

dmontalvo: I agree

Wilco: That way we have a formal agreement.
... We are out of time, so, anyone disagree with putting it out as a CFC? I'll need to check with AG chairs for okay. I'll probably send out a CFC to approve the ACT part of the charter.
... we're at time. Thank you!

Summary of Action Items

Summary of Resolutions

[End of minutes]

Minutes manually created (not a transcript), formatted by David Booth's scribe.perl version 1.200 (CVS log)
$Date: 2022/04/21 14:07:57 $