– DRAFT –
AGWG-2024-06-25

<Chuck> I've joined the call if any of my co-chairs wish to join early and do some strategizing.

Chuck: Is there anyone who would like to introduce themselves or new role?

Chuck: A sub-group survey is going about

<Chuck> Subgroups (Questions / support) https://www.w3.org/2002/09/wbs/35422/subgroups-june-24/

<alastairc> Especially for Adjust Color and Haptic stimulation

Chuck: We're looking to start up some subgroups. It is open until July 8, and we encourage you to chime in and pick topics. We may not be able to fit you in in your primary interest; this will really help us do that as best we can.

Alastair: We are light on the haptic group

<Rachael> https://www.w3.org/2024/09/TPAC/registration.html

<jtoles> Can we adjust our choices? I don't think I picked either of those, but I'd be willing to join one if there aren't enough people.

<alastairc> WCAG 2.x issues email for this week: https://lists.w3.org/Archives/Public/public-wcag2-issues/2024Jun/0009.html

Rachael TPAC is in September. We meet on Monday and Tuesday

<alastairc> jtoles - yes, just adjust the survey

Introduce next steps in subgroups https://www.w3.org/2002/09/wbs/35422/subgroups-june-24/

<jtoles> Thanks

<Chuck> proposed RESOLUTION: Publish the WCAG2ICT Draft Note from the Editor's Draft

Chuck: Our first item is WCAG2ICT. Here is what we're hoping to achieve

MaryJo: It will be the editors draft based on existing PRs.
… All of the substantive feedback has been handled in PRs.

<alastairc> present

MaryJo: There are a few open issues but they are entirely editorial and can be handled after publication.
… This should be the penultimate version.
… We are getting very close to finishing the update to this note.

Chuck: Thank you to everyone who has taken the time to review.

<Chuck> proposed RESOLUTION: Publish the WCAG2ICT Draft Note for wide review from the Editor's Draft

<rscano> +1

<bruce_bailey> +1

<kirkwood> +1

+1

<maryjom> +1

<dan_bjorge> +1

<JakeAbma> +1

<Azlan> +1

<kevin> +1

<alastairc> +1

<ShawnT> +1

<laura> +1

<MJ> +1

<sarahhorton> +1

<Jennie_Delisi> +1

<Makoto> +1

<Ben_Tillyer> +1

<Gez> +1

<Francis_Storr> +1

<tburtin> +1

<Jen_G> +1

<Poornima> +1

RESOLUTION: Publish the WCAG2ICT Draft Note for wide review from the Editor's Draft

<Rachael> +1

Continuing the conformance conversation by exploring a baseline structure https://docs.google.com/presentation/d/17VJvnm5UQW4WUzIoo9QNPVGfePgaZa8ifZWs-wtmv7E/edit#slide=id.g2e4b93386d6_0_15

Chuck: The task force can do some celebration

<GreggVan> +1

Starting discussion on consolidating outcomes https://docs.google.com/document/d/1GL-jyZ7CuYb-fms_jTKEe1YTAUZAH0EXkU3yqbRg5f4/edit

<bruce_bailey> Just as point of comparison, "wide review draft" for U.S. ADA web rule was open for 60 days, see: https://www.federalregister.gov/d/2023-15823

<bruce_bailey> DOJ final rule is available at: https://www.federalregister.gov/d/2024-07758

Rachael: As we've worked through Text Alternatives, and Keyboard Focus Visible, something that has come out is the tree structure seems to be a direction we're going.
… I began grouping the outcomes together.

Starting discussion on consolidating outcomes https://docs.google.com/document/d/1GL-jyZ7CuYb-fms_jTKEe1YTAUZAH0EXkU3yqbRg5f4/edit#heading=h.8n8glns7lrs4

Some seemed to be inter-related
… Which related ones should our groups be working on together?
… This proposed consolidation is the result. I would ask that people look through this. This is an exercise I felt helpful.
… I'll open a github discussion on this. It would be helpful to have other people's insights. Alternative suggestions are welcome. The sort exercise took about 2 hours if anyone wants to try it themselves.

Chuck: Just to emphasize that this is not a solution. It's an exercise to drive a more thorough conversation.

Rachael: The goal of this is to know what to hand the subgroup as a starting point.

WCAG2ICT

Conformance https://docs.google.com/presentation/d/17VJvnm5UQW4WUzIoo9QNPVGfePgaZa8ifZWs-wtmv7E/edit#slide=id.p

Chuck: Sometimes these conversations can be challenging and difficult.
… We wish to clarify vocabulary.

Going forward we hope to address issues of pass or fail, and how to motivate people to go beyond basics.

[Shares Terminology screen]

Chuck: This is how we're thinking about these words. "Prerequisite" is a smaller subset and includes safety and AT support.

"Baseline" is a larger subset which is Prereq. plus author provided outcomes not currently met by AT.

"Enhanced" extends the baseline.

[Shares 'How a baseline might work..' slide]

<bruce_bailey> thank you for trying to sort out prerequisites from baseline

Chuck: The key here is that prerequisites must be met

Graham: I don't think that level 1 should be on the prereqs. It needs to sit on the 50% enhanced outcomes

<Zakim> alastairc, you wanted to comment on levels vs conformance

Rachael: By putting in a prereq, it becomes level 1 whether we label it that way or not.

<Zakim> Rachael, you wanted to answer graham

Rachael: We'd like to stick with this vocab for today, just to help the discussion.

Alastair: We don't need to use this terminology in the doc, but if you look at the next slide, I think you'll get what we're starting to work with.

Bruce: Glad to have the discussion; I'm waiting for the next slide.

<alastairc> Suggest moving onto the next slide soon, before we get stuck...

jtoles: I think we should get away from level 1, 2, 3. It doesn't say what it is. Maybe "minimimu, "optimal", "enhanced"?

sarah: I'm a bit confused by the term "baseline". On the previous slide it seemed like it didn't include prereqs?
… Is this the same thing happening on the terminology slide?

[Title for slide 29 is updated]
… I think it would be useful to not label prereqs at level 1

<bruce_bailey> +1 that prerequisites only is level 0 at best

Chuck: Someone had said 'let's get away from levels'. For today we will continue using those terms, but will note that down.
… There are two columns on slide 29. The percentages are placeholders. They just give us an idea for discussion.

[Shows slide 30 "How levels might work...]
… In essence, this is how WCAG 2 works.

Graham: On the first column, is there a typo?

<Zakim> alastairc, you wanted to say why you'd choose between prerequisites and baseline plus scoring.

Graham: Based on the prior slide, the numbers compound, not what is shown.

Alastair: Baseline is all the prerequisites plus more [goes back to slide 28]
… The reason you look at this difference between the baseline and prereqs is for flexibility.
… If you are struggling to pass one requirement, you can focus on other outcomes to meet.

<kirkwood> so are we balancing other accessibility requirements of different groups against each other?

jtoles: I was confused. Is Level 1, 2, 3 instead of Bronze, SIlver, Gold.

Alastair: These are placeholder names.

<sarahhorton> When we talk about "people" in this context it would be helpful to talk about people affected by accessibility rather than people seeking conformance

<Chuck> ack +

jtoles: I like the baseline not being too difficult to incorporate. So you've met the minimum. You're going to get scored on how high above minimum you go.

<GreggVan> doubleback

Graham: It could just be me... We have prerequisites. We have a set of criteria that form a baseline. Then we add on more outcomes to get to the levels.

Baseline is analagous to A.

Rachael: These are 4 alternatives
… We recognize the group has struggled.
… What we are trying to do is set terminology. The prerequisites are things that must be done.
… We're trying to say the prereqs would be a small subset. OR we could go with a baseline, which is the prereqs plus other stuff
… So we're talking about 2 different things.

<MJ> "outcomes that aren’t currently met by AT": Are these defined anywhere?

Rachael: Then on slides 29 we are looking at some different options.
… The two columns are two different approaches. On left, the prereqs are the starting level. On the right, you need to do the prereqs and more at the starting level.
… Slide 30 is 2 more alternatives

<kirkwood> is each prerequisite tied to a specific requirement? (requisite)

Rachael: The column on the right is basically WCAG 2

<Zakim> Azlan, you wanted to say using baseline is less confusing for me as my expectation of a prerequisite is something you need before you begin

Azlan: Using baseline is clearer for me.
… Prerequisites are things you need before you begin.

<Zakim> GreggVan, you wanted to react to Azlan

<JenStrickland> What I'm about to write is an absolute detour to the conversation, so I want to note that and set it aside for another time. The idea of being the safety issues, AT needs, and things that can't be fixed… currently fails to account for some cognitive / trauma considerations that we haven't included in our WCAG criteria to date. It may be something

<JenStrickland> for a subgroup to document in the future, possibly with a couple of COGA members. This isn't entirely a COGA topic, as AGWG currently defines COGA.

Azlan: Because of that, the baseline model is easier to get my head around.

Gregg: 3 comments. When we do this we need to keep in mind things that are able to do with or without an assertion.

<kirkwood> baseline > requirements > prerequisites ?

Gregg: Secondly, I always worry about percentages or scores because people can game it. It's great for enhancements.

<Ben_Tillyer> Do you have a source for the "middle one" line from industry Greg?

Gregg: Industry has pointed out that when you offer 3 choices, policy makers tend to choose the middle one

<Zakim> Rachael, you wanted to say chair hat off benefits of combining some set of requirements with some selecting set

<kirkwood> +1 to Gregg, case in point: AA

Gregg: So we just need to be careful that the one that's settled on is achievable.

<alastairc> Ben_Tillyer - see behavioral economics and the buying a breadmaker experiment

Rachael: I'm interested in the second one on slide 29, chair hat off.
… The group has wanted to have industry-specific requirements
… Education-centered organizations could create some material that enhance students, as opposed to banking.

<Ben_Tillyer> +1 to Rachael

<Zakim> bruce_bailey, you wanted to ask about having BOTH prerequisites AND baseline

Rachael: Having the flexibility built in seems like a real pro.

Bruce: I'm with Graham because I don't see a conflict between these.
… We can have a level 0.
… ON slide 29, level 1 becomes 0, and up from there.

<Graham> +1 to bruce, this is how I envisaged it.

<dan_bjorge> Very strong -1 to the flexibility being a pro at any level we want the possibility of legislation/policy requiring.

<laura> +1 to bruce

Bruce: You don't have to choose between prerequisits and baselines. You can have both.

<kirkwood> +1 to Bruce

<laura> Prerequisites = 0

Lauri: I would like to see this turn out so that if you don't get these basic things, you do not pass .

<Graham> Another +1 to Lori, this was what I thought the "prerequisites" was for!

<Zakim> alastairc, you wanted to comment on examples of baseline/prereq/enhanced

Alastair: A quick response to Bruce. We could have prereq of a subset. I don't see the point in going down that route. if everything in baseline is required, what does it matter if there's a broken out subset?

[Goes to Consolidation proposal]

Alastair: We went through an exercise to see what outcomes would be in each category for 3.3 Error and identification

[walks through how the SCs are categorized ]

alastairc: Give examples of what might fall into different areas. What is required. What is required but flexible on what to choose.

<Zakim> Chuck, you wanted to react to alastairc to change scribe

Chuck: Please volunteer to scribe if you can

dan_bjorge: I maintain the same concerns that came up the last time we had this discussion. Equitability concerns anytime there is a level that allows the user to pick and choose AND we want that level to be a candidate for adoption as a legal requirement. We should not allow people to pic and choose what outcomes they choose. At least one level must be outcome based.
… if we think its realistic that they would want to pick a middle level, then I also prefer the middle level only contain requirements

<DuffJohnson> +1 to Lori. This conversation feels confusing because the complex relationship between priority and severity isn't (so far as I can tell) fully acknowledged. As a result this convo often seems to be trying to paper over a "mashup" between these concepts.

<kirkwood> +1 to Dan

dan_bjorge: the naming is not my concern. It's critical that we have a non% outcome based option.

<GreggVan> bronze silver gold platinum

<Poornima> +1 to Dan on pick and choose among outcomes to satisfy different levels

Rachael: My one concern going back to Graham... having a prereq level with baseline level, if we don't build in flexibility....

We are essentially building what we already have.

Dan: I agree iwth you and it is good and correct

Dan: I agree, that is comparible, that is good and correct.

<Zakim> Chuck, you wanted to ask Dan about ...

Chuck: Question for Dan to answer in queue. You are saying there should be a level that doesn't include any %. The first option on slide 30 has that. Does that particular version address your concerns.

Graham: I get where Dan is coming from. Can someone show me a WCAG AA compliant website? If we have pass/fail only, it doesn't work. If you use % then you try to meet more in order to guarantee you meet the minimum %

<Zakim> Jennie_Delisi, you wanted to ask about terminology

Graham: if it is % based, I can focus on making the pieces that are most important.

<Zakim> alastairc, you wanted to ask Dan whether you'd go for a wider baseline, or multiple levels (WCAG 2 style), or something else? Misses the requirements

Jennie_Delisi: The items that end up in enhanced would be considered extra. We need to make sure that we put in enough in baseline and prerequisite levels to make sure needs aren't missed.

alastairc: Dan would want to go with a WCAG 2 way. A level or two above and an all or nothing approach. We have to address the pass/fail issue. To be clear, we aren't talking about allowing for a % of items that pass. Its outcome level.
… the other requirement is how do we encourage people to go beyond the basics.
… We also recognize that there are things that are quite core and fundamental.

<kirkwood> nothing enhanced will be accepted by regulators. xcept in cases of specific litigation involving a specific entities

alastairc: that's how we ended up looking at these models.

bruce_bailey: Enhanced doesn't hit the right tone.
… We could have a soft option between levels that includes assertions. So prerequisites, assertions, baseline which is closer to WCAG 2 AA.

<kirkwood> +1 to Bruce

bruce_bailey: So you have more objective things before objective items.

<sarahhorton> I can scribe

<Zakim> Chuck, you wanted to react to bruce_bailey to ask for scribe change

<Zakim> dan_bjorge, you wanted to answer Chuck

dan_bjorge options on 30 better than 29, second one problematic, agrees that WCAG 2, good to have middle option
… one to 2 levels not % outcome based

dan_bjorge some percentage of elements on page have alt text that meeting criteria, within outcome

<Zakim> GreggVan, you wanted to say "ruler and rule" one image missing alt text does not make a site fail -- it makes that page out of conformance until fixed. bugs are not in our remit. bugs will always exist and should not (and will not) cause anyone to be sued or punished. If they refuse to make it accessible -- then and only then is it a faillure - but again that is up to them that make the rules

<bruce_bailey> +1 to dan's idea about Gold could be about higher quality alt (for example)

GreggVan Bug when sites have little things wrong, not sued/punished for that
… confusion between % provisions, not % pages
… start with idea that require testable things, some things testable but not required
… only way to have % at 2nd level, have to have things that aren't required
… assertions mean you assert it happened, only test is that you did it
… testing effect is different from assertion
… can we mandate assertions, not established

<alastairc> We had a lot of (negative) feedback about % of instances (e.g. alt text) because it requires counting everything.

<dan_bjorge> And folks that need alt text have negative feedback about failing to consider alt text for everything. Making folks count everything is a feature not a bug.

Rachael All proposals are similar, hearing support for different approaches
… need to pick 2–3 to explore
… chair hat off, if set of requirements providing equitable coverage
… what's concern with flexibility next layer up in level?

Ben_Tillyer Companies would like flexibility, progress over perfection
… trickier for conformance claims, indicator of what to do next to make more accessible
… some lean toward different disabilities, skew is inevitable
… Points about regulator adoption, standard is mindful for them, should focus on users' best interest

<alastairc> "Making folks count everything" - was seen as costly, increasing the expense of accessibility for little gain when it is fails that are the problem.

<Graham> +1

GreggVan Nothing prevents percentage being required by regulation as long as testable
… if assertions can be required legally, standard can include assertions

Chuck Looking at baseline and percentages, baseline is equitable
… appealing that sectors, e.g., education, based on clientele and users, add outcomes that benefit users
… ability to say with understanding of audience, add those more important to users

<alastairc> +1, and could gamify somewhat with badges etc.

Chuck [next slide — Possibility to consider: Aggregate assertion level]

alastairc View level conformance, separate question, big company, big website, how to get whole site to conform
… separate site conference concept with assertions, at product level
… could include maturity model
… assertions provide confidence in passing core requirements
… could have sample-based testing
… plain language, design system
… things we know help with product accessibility
… maturity and assertion based
… achieve requirements using another approach to mass testing

Chuck chair hat off, massive sites, style 1: static content, style 2: third party
… should 3rd party be consideration?

alastairc Easier to apply to static sites, big site with 3rd party content, could have ways to make more confident

Graham Likes it, also apply at components, e.g., tested table component
… would also cover site-wide, interesting concept, would pass/fail?

<kirkwood> seems we are attempting to address LOE or burden?

<kirkwood> could that be gamed?

alastairc Advanced accessibility statement, saying have done certain things, using components, tested sample of table and passed

<dan_bjorge> I'd be very cautious about "table-based component" type things; I'd say it's much more common for accessibility issues to happen when you integrate the component into a site, rather than issues in the component itself (though both happen of course)

<dan_bjorge> yeah, +1 to what Ben is saying

Ben_Tillyer Component libraries, many sites tell people using component library but component themes, variations
… could have same component, different instances, different accessibiltiy

Graham Currently in that pickle, 3 awful, 1 not bad, would have to be on-page tested
… how to turn into conformance model isn't clear, how often component used, hard to form model

GreggVan Talking about comment made earlier or approach, aggregate assertion level?
… all the outcomes have to apply throughout, concur with idea that assertions could fill, can't do usability testing on all websites in org
… works, most assertions would be aggregate assertions

<Zakim> alastairc, you wanted to agree we would need to be careful, but the point is that the assertions are about people applying a process.

<kirkwood> we did

alastairc Separate and additional, baseline + percentage, would be doing that for conformance
… this idea is to package up assertions and make sure they are in a pack, all or set based on context
… have tested design system, have tested sample (EM methodology for choosing sample)
… would have these assertions, not view level (has to be there)
… idea to look toward as addition to help big sites, confident that most of our stuff is accessible

Strawpoll: 1) Prerequisite plus % based levels 2) Levels with prerequisite and % 3) Baseline plus % based levels 4) Set levels only 5) Continue discussing possible models

Chuck Added labels for models, want to know what group is comfortable trying

Graham Couldn't we apply a rule, over X pages do testing on sample of X pages

alastairc That could be an assertion
… select sample, all pages conform, then report on sample and results

Strawpoll: 1) Prerequisite plus % based levels 2) Levels with prerequisite and % 3) Baseline plus % based levels 4) Set levels only 5) Continue discussing possible models

<JenStrickland> 1

<Zakim> Chuck, you wanted to poll

<kirkwood> 1

Rachael Key point is to pick a few options — 2, to do first pass exploration
… try it out, review results

<GreggVan> 4 , 3

<dan_bjorge> Option 3 (but with 2 levels of non-%-outcomes) > option 4 > option 3 (2 I would object to and 1 is the same as 3 with more confusing naming)

<Ben_Tillyer> any apart from 4

<Graham> 2 - but with guidance "do not do other tests without doing prerequisites" (i.e. move that to "process")

<Azlan> 4, 3

<alastairc> 3, 2, don't think we should use 1, happy to use 4 as comparitor.

Chuck Are there favorites to try out
… put in 1st and 2nd

<rscano> 4,3

Strawpoll: 1) Prerequisite plus % based levels 2) Levels with prerequisite and % 3) Baseline plus % based levels 4) Set levels only 5) Continue discussing possible models

<Graham> 2,1

<filippo-zorzi> 2

<bruce_bailey> 5 as i am missing difference between (2) and (3)

1

2 and 3 (4 would be a good competitor)

<rscano> 2

<jtoles> 3, 1

<scotto> 2, 5

3,4,2

<bruce_bailey> 3, 1

<kirkwood> 3,1

<ShawnT> 3, 1

<laura> 1

<GreggVan> oops 4,3 if all midlevel are testable

<dan_bjorge> I don't love the format of this strawpoll because I feel 1 and 3 are the same option with marginally different temporary terminology (ie, overrepresented)

<Zakim> alastairc, you wanted to suggest 2+3 by including labels for prereqs and baseline

alastairc clarify difference between baseline and prerequisite
… could differentiate during the process

+1 to combining as it would potentially save work later.

GreggVan Prerequisites base of baseline
… not going to have level that's only prerequisites

acl bri

acl bri

bruce_bailey Also don't want only prerequisites, can get consensus because not looking for equity
… baseline should be equitable, longer discussion

<Zakim> Graham, you wanted to say +1 Alastair, that may also help us nail down what this looks like at scale as we will see how many "baselines" we have and that could drive this decision (i.e. we have 20 prerequisites and 10 baseline vs 20 prerequisites and 40 baseline tell different stories

<kirkwood> +1 to Bruce

<Francis_Storr> 3, 1, 2

Graham Don't need to decide today, go through and label as we go, add them up, will help us see and guide path forward

Chuck Confused by proposal, subgroups would identify what are baseline and prerequisites, refactoring what subgroups do

GreggVan Ones that have to do with AT compatibility, work for all disabilities
… labeling great idea, rather than sort out model, look at the pieces
… let's label, what do we have, baseline, prerequisites, assertions, higher level, testable

Rachael Poll helps with what to do with models
… present templates to subgroups next week, good to mark in templates, will add in
… explore these and other possible models

<Zakim> alastairc, you wanted to suggest 3 (and identify prereqs/baseline) and 5 - two levels + %

alastairc Alternative option worth doing comparisons, choose some models (see 5 above)

<Zakim> bruce_bailey, you wanted to note baseline items might also be scored in quality

<alastairc> Dan, does 5 get at what you meant?

<kirkwood> to me a prerequisite is a prerequistie to a requirement. for example on may be that a site has over 3000 pages one must…

<kirkwood> +1 to quality

bruce_bailey Sorting, might be other dimensions, if outcomes are good for evaluating quality (gold alt text)

<dan_bjorge> Yes, 5 is what I meant (and I think what Gregg also wanted in his strawpoll vote?)

<dan_bjorge> With that 5 in place, my vote would be 5 >> 4 > 3

Chuck Enough to move forward?

alastairc 3 and 5 to test next

Rachael Moving back/forth between conformance and outcomes, both inform each other

Chuck End of agenda

Do I need to do anything with the minutes?