Silver Conformance Options Subgroup

21 Jan 2021


Bryan, jeanne, Jemma, JF, JohnNorthup, PeterKorn, Rachael, sajkaj, Wilco_

Meeting minutes

<jeanne> https://www.deque.com/blog/first-public-working-draft-wcag-3/

<jeanne> https://sheribyrnehaber.medium.com/silver-wcag-3-0-first-public-working-draft-released-15f765b9525e?source=messageShare-260ccf64b894-1611244201

Agenda Review & Administrative Items

Janina: Continue talking about use cases today. Looking at some additional cases, invitation is wide open to add use cases, extract candidate issues, etc.
… Need to cover as many use cases as we can find.

Peter: Wonder if we might shift to reviewing use cases. Reintroduce what we did to folks who missed last week.

<PeterKorn> https://docs.google.com/document/d/1GyUYTnZp0HIMdsKqCiISCSCvL0su692dnW34P81kbbw/edit#

Peter: We started by going through proposed use case examples. We got through the first one on principle #3.
… Example was Upcoming events calendar. The page includes a time zone converter. The use case is that the priority of the calendar is higher than the time zone convenience feature.
… Feature is available in other places, and people are often familiar with time zone conversion.
… Last week we also got into, but did not complete use case #2, on principle 5. A travel website that collects not only content from the site owner, but 3rd party reviews. Those contributions might include various formats.
… The licensed content is under copyright of the licensed owner.
… Suggest we start on use case #2

Janina: Yes, lets do go into 3rd party content.

Peter: Collected 3 issues related to the use case. 1. Whether the owner has permission to do so. 2. Whether a programmatic test can ever detect sensory information.

<PeterKorn> Whether and how effective guidance to the general public can be on creating accessible content (e.g. reference to sensory information)

Peter: 3. How effective guidance to the public can be on creating accessible content.
… Travellers might use sensory information in their travel log.

<PeterKorn> Yes Bruce!

Bryan: How does the impact to the process play into this. With WCAG 3 we talk about alt, and use in process to denote if it is more or less important to include an alt tag.
… It may be very important to the overall accessibility if third party content is accessible.

Peter: Fantastic additional use case. The use case speaks to third-party content that is generally static. Whereas what you talk about is more like a third-party app. Will start to write it.

<jeanne> +1 to having use case examples that are apps!

John: I think the notation of third-party content is too broad. A google maps is third party content, but there is also user generated content, which you have no control over.

<jeanne> +1 to splitting 3rd party content and user generated content

John: I struggle with including that in a conformance measure, because the site owner has no control over that.
… We can't penalise a site for not doing something they can't do.

Jeanne: I think we need to split out the things a site can do. There are things they can do, for example with CMS formatting, how they filter comments. There are things that can be done.
… Don't want to say across the board there is nothing that can be done.

John: Agreed. An example is when you upload an image into wordpress, you have to explicitly provide an alternative or check that it is decorative.

Peter: We have not made this explicit in WCAG 2. Principle 5 is that the way WCAG 3 treats third party content should be more thoughtful and nuanced, more than just making a partial claim.
… We need to make third-party content a first-class citizen.

Janina: It is likely a user might want to assert copyright over user generated content, or they might explicitly not. There is an accessibility aspect at either point.
… Our facilitators have pointed out that we have not yet addressed third-party content.

Peter: Part of my intent for the use case was to suggest it may be difficult to distinguish between what is site-owned content, contributed content, and user-owned content.
… It makes it harder for a conformance claim to exclude third-party content.

<Zakim> JF, you wanted to make the distinction between 3rd party content and user-generated content

John: Think it's really important we single out the two use cases. Third-party content, chosen by the site owner. But for user generated content the site owner has no control over the content.

Jeanne: Suggest we make a list of different types of third-party content. Break it down and have a use case for each. Would be powerful, then we can work out what we can apply to it.

<Zakim> jeanne, you wanted to suggest that we start figure out all the different types of 3rd party content and write a use case for each one.

<Bryan> +1

Jeanne: The wordpress example cam out of ATAG. If we can bring in ATAG to apply to third-party content, we can give structure for people to claim conformance on third-party content.

<JF> Suggest one distinction: "packaged" 3rd party content, and "spontaneous" 3rd party content

Peter: Completely agree, there is a lot sites can do to chip away at the user-generated content challenge. There is a lot that can be done today, but more in the future.
… Need to not limit us to what we have today.

Peter: I think this use case is real life, and less construed like a site that only has one or the other type of content.
… Don't want us to miss real-world use cases that mix different solutions.

<JF> +1

Wilco: Reminds me of test cases. Keep them simple, have one of each type, as simple as you can make them.

Jeanne: We don't need future tools. If you have a company that supplies content, you can put in the contract to require it to meet WCAG 3. But if you have more unstructured, we could scope out what parts of WCAG 3 you would have to meet.

<PeterKorn> +1

Jeanne: Maybe you don't need to meet all of WCAG 3, but list what parts you'd have to pass for different types of third-party content.
… Each type is probably going to have a different solution. I think we could have a really good solution, that allows big dynamic sites to pass.

John: Agree with Wilco, think about the primitives, then build out the complex.
… Doing work now with development agencies. They need a conformance statement like a VPAT, but when the site goes live and starts getting user comments, which effects the overall accessibility.
… I see a need for a distinction between packaged code, and contributed content.

<jeanne> +1 John and what I think that we can tackle these distinctions with the basic structure we have, but I think it will take a lot of work to get to solutions.

Peter: I don't know we need to restrict the use cases to single-issue use cases. Wouldn't shy away from compound use cases, but no objections to having a complete set of primitive use cases.

John: Agreed.

Janina: Use cases in W3C accessibility sometimes give people a hive. We're talking more about user scenarios.
… We might get the same primitives out of multiple stories.

Peter: Do we want to write use cases that describe a scenario that might exist, or use cases that more describe the nature of the problem.
… Personal preference is for real-world use cases, as they are more understandable.

Wilco: Suggest to keep them simple, but not so simple that you make them difficult to understand

<jeanne> Embedded content from 3rd party under contract

<jeanne> Included content that is copyright and doesn’t belong to the host

<jeanne> Content that is structured but otherwise is user generated

<jeanne> Content that is unstructured user generated

<Zakim> jeanne, you wanted to say I started a list

Bryan: I like real-world examples. Like the direction here. Would like direction on when third-party content is included in scoring and when it is not.

Jeanne: That's what I'm thinking

Rachael: Seems like there are different types of third-party content. Like advertisement vs licensed content.
… Feel like we should not start with the combined examples.

<jeanne> I added advertising to the list. That's an important one

John: This use case is focused on the output, what the consumer gets at the end of the process, but the use case brings into focus the authoring environment as well.

Janina: Reminds me of a question from TAG on the personalisation TF. It might be that advertisement can become a barrier. Would like to find a way to keep that kind of opportunity in our issues.

Peter: Wonder if we should jump to another use case?

Janina: Yeah, and when do we want to start breaking down this doc.

Peter: Suggest we go to principle 4.

<PeterKorn> Proposed Use Case example [for discussion 21Jan21]: A large eCommerce site has a secondary process in a section for the rating of customer experience, containing a five point Likert scale

<PeterKorn> from “very bad” to “very good” implemented as a row of five radio inputs. Labels are not available programmatically due to the use of ‘hidden’ on the label and lack of programmatic

<PeterKorn> linking of the visible labels marking the endpoints and the midpoint of the scale. Selecting one of the inputs (for non-sighted users without knowing what they mean) causes the scale to

<PeterKorn> disappear with no way to change the selection. The (possibly unintended) rating has been submitted.

<PeterKorn> A. Automated testing would easily identify the existence of five radio inputs without accName. Would the automatically detected failure in a secondary process prevent the main process from claiming conformance?

<PeterKorn> B. Feedback indicates the customer experience has been rated, the context changes, the process is irreversible / incorrectible so it can be said to fail SC 3.2.2 On Input (and of course 1.3.1). How would these Failures impact on the rating of the main product purchase process?

<PeterKorn> C. Would the same issues in the secondary process affect the conformance rating differently if the process appeared in a popover dialog that users would have to call up explicitly to submit feedback on their experience?

Peter: My thought on the use case it is highly technical, and hard to read as a use case. It is not clear to me how much of the use case is inherent in what the site is trying to do, and how much comes from the particular implementation.

Janina: It would be relatively easy to scope out the radio buttons. It is not why people go on that site, but would hate to argue use comments are out of scope, even if it is not the primary function of the site.

Peter: What I don't like, is it simply because of the choice of how this was implemented? The use case might be better written as having some secondary content, the interaction of which isn't accessible.
… It doesn't matter what they did to make the content inaccessible.

Janina: Generalise the mechanism.

Wilco: Would suggest making it less technical, but not to make it too generic

Jeanne: Use case came from Detlev

<PeterKorn> Proposed POTENTIAL Use Case example [for discussion 14Jan21]: Based on website development industry research [citation needed], more than 50% of dynamic websites at any

Peter: Another use case

<PeterKorn> given moment contain known “blocker bugs” [definition needed] that have been present on the site for more than 24 hours, even though by definition a website update should never be

<PeterKorn> published with blocker bugs. That same research indicates that nearly all dynamic websites contain a substantial number of known “critical bugs” [definition needed] that have been

<PeterKorn> present for more than 1 month, even though by definition critical bugs impact customers ability to use a feature of the website.

Peter: (reads out comments)

Peter: What I'm trying to get at is a use case where a site has bugs, some accessibility bugs. What should WCAG do with that, particularly around the principle that whatever we do in WCAG 3, we should not over-require an absence of accessibility bugs in a world where bugs exist.

<jeanne> +1 to agree with JN Note "Accessibility-related bugs should not be overrepresented in the bugs overall, tolerated more than other bugs, or given low priority for fixing."

Janina: Looking for statistical data on bugs. Filling that out seems a useful step on how to quantify if accessibility defects represents a higher portion than they should.

John: In Peter's commentary he makes a distinction between a blocker and a critical bug. Given this call is about conformance, our current model treats them on a score 0 to 4. Now we have to make a distinction on what is critical, what is major, etc.

Peter: It's what exists in the real world. We'll have to deal with it.

Minutes manually created (not a transcript), formatted by scribe.perl version 127 (Wed Dec 30 17:39:58 2020 UTC).


Maybe present: Janina, John, Peter, Wilco