<Laura_Carlson> Scribe: Laura_Carlson

<Rachael> present_

Announcements and Introductions

AC: welcome

Kevin: member meeting in Europe tomorrow.

<kevin> https://www.w3.org/events/meetings/9d269c34-9d05-439e-9ed6-316765e18f9d/

Kevin: welcome to join us. Web sustainability Guidelines

Rayianna Daniels-Wynn: I'm new to the group.

WCAG3 conformance review https://docs.google.com/presentation/d/1DlDxp8MCYXj3RWnFCCz13zsmM2fV4Wf8NbECKogdul8/edit

RM: This is a proposed conformance model based on all of the conversations we have had over the last. For consideration.
... Where are we? We are working towards a full developing* draft for publication in December.
... Developing = This content has been roughly agreed on in terms of what is needed for this section, although not all high-level concerns have been settled. Details have been added, but are yet to be worked out. Feedback should be focused on ensuring the section is usable and reasonable, in a broad sense.
... Nothing is “final” even after that publication
... AG often focuses on conformance but there is more to it
... Conformance is how closely a “thing” meets the technical standard
... Measured against 100% of technical levels defined by the standard
... A conformance claim must clearly define the scope and time of testing
... Reporting tells the story of how closely a product conforms
... Reporting can show progress over time
... Compliance sets legal levels for conformance and consequences if not met
... it is out of our scope
... but we can advise
... Compliance (legal) can allow more or less flexibility than Conformance (technical)
... Types of provisions: Foundational Requirements, Supplemental Requirements, Assertions, Best practices
... Requirements will be written so they can “age out” by being automatically met on mature technologies.
... Over time user agents, platforms and AT within the “Accessibility Support Set” may support an increasing number of requirements
... “Foundational” and “Supplemental” terminology likely still needs work
... Provisions will be marked with a unique alphanumeric code that does not relate to sequencing
... Provisions will be tagged so people can customize how they are presented eg, POUR
... We have an Initial draft for Functional needs categories.
... It is a starting point
... Levels of conformance
... Bronze = All foundational requirements

10-20% of supplemental requirements and assertions in each functional needs category

scribe: Silver All foundational requirements

50-60% of supplemental requirements and assertions in each functional need category

scribe: Gold = All foundational requirements

+80-90% of supplemental requirements and assertions in each functional need category

scribe: Supplemental Requirements that are Not Applicable (N/A) based on a technology or situation being present are counted as a Pass
... So far, Assertions are always applicable so never count as Pass by default
... It can be very difficult to verify or state full conformance against an entire websites when their size exceeds the scope that can be realistically tested.
... WCAG 3 informative materials will provide more explicit guidance on sampling and stating scope.
... Question? Should we add or call out several assertions that make it more likely that the accessibility of the sample is representative of the accessibility of the larger site?
... Reports include:
... Reports include: Scope tested, Accessibility support set, Level of, conformance reached (Fail/Partially Supports*, Bronze, Silver, Gold)
... % of foundational requirements passed by functional need
... % of supplementary requirements / assertions met by functional need
... Reporting against functional needs allows for clearer insight into accessibility, even when full conformance isn’t reached.
... Because % are based on which requirements and assertions are met
... Testers enter 1 set of results and the % against functional needs can be automatically generated.
... Note with suggestions to policy makers (Compliance) Some possible content - Ways of evaluating the selected scope of conformance claims
... Recommended sets of supplemental requirements and assertions
... Explore recommending dating conformance tests in some way
... Additional areas to explore - Suggesting that reporting provide both a snapshot test and the process to fix
... Defining a prerequisite set within foundational
... Reporting partial results
... Ways to encourage honest reporting
... Next steps: Next week, do we agree to publish this as our conformance model?
... What adjustments are needed?
... Subgroups and individuals discuss and revise provisions list
... Continue editorial cleanup
... Reorganize provisions
... Write crosswalk between WCAG 2 and WCAG 3
... Publish entire draft as Developing
... Begin process of revising provisions
... Integrate ACT Tests
... Draft parts of understanding documents
... Deeply consider wording
... Questions?

Gregg: incredible amount of work.
... Categories, functional categories, everything seemed to be going fine until you got to functional categories,
... don't know what that means.
... the functional categories was just, basically, it's categories, and then you have one of them, cognitive broken into, like, 8 or 9 categories,
... that's going to be very hard, because you can say, well, does this provision go in this category, or this category, or this category?
... I think we ought to be looking at general categories rather than getting all the way down to specific functional needs of subcategories
... so I don't think we should be using categories at all.
... slide, just talking about tags. And you can tag things lots of different ways.

But, um, but I do worry about trying to,

Jon: the functional needs of someone who's colorblind are very different than someone who's totally blind, or someone who is low

<Glenda> Glenda wants to say: I recommend the use of the functional categories from EN 301 549. Categories are so valuable to anyone trying to understand what an ACR means to the human beings (customers and employees) using that digital thing.

Jon: vision. So, I think, um... even other standards, like EN301-549's separate those into 3 different categories.
... ...did we talk about units of conformance, and has there been any progress on that?

Shadi: I think what we should be looking at is really a conformance model.
... To explore different kinds of models that would allow more.

differentiation, more separation, more

measurement, and making the conformance itself

Julie: ...it might change people's thinking if to know whether what we currently have as foundational requirements are going to remain Foundational requirements.
... is there any way we can incentivize people?

<Zakim> Rain, you wanted to discuss Gregg's questions around mobility and fine motor

Julie: I find it interesting that the. the percentages don't cover a lot of the range of supplemental requirements.

Rain: maybe mobility needs a more specific name, but there are incredible differences in what you might design.

<kirkwood> +1

<shadi> [What I said: did we sufficiently explore, or do we want to explore, conformance models other than a single aggregated "score" of bronze/silver/gold? Maybe the labeling for electrical appliances could be an example -- for a washing machine it might show power consumption but also water consumption, noise level, and washing load etc, which allows for more

<shadi> granular understanding of the performance beyond a single aggregated score.]

Rain: ...Instead of leveraging percentages against a number of supplemental requirements. Maybe We might leverage percentages against a number of functional needs that are fully met.

<Zakim> Rachael, you wanted to respond to Gregg

RM: it's not a one-to-one relationship, like, A requirement doesn't support only one functional need.
... it's not that we would shove one requirement into a single functional need.
... the whole reason for percentages is to avoid the counting problem.
... question, mobility is quite literally the ability to move in space from point A to point B.

<kirkwood> addressing the functional needs of specific disabilities guides policy and compliance level adopted by the content publisher should be a lense used for this guidance

RM: Whereas large gestures are the multidimensional body movement.

Glenda: the functional needs categories addition.

is crucial to the success of accessibility at large.

<Adam_Page> +1 to Glenda

<kirkwood> +1 to Glenda

<Zakim> alastairc, you wanted to comment on reporting dimensions and to comment on the number of foundational reqs

scribe: I suggest we start with. and 301549, Annex B.

And this is going to help consumers. As well as companies making purchases, understanding the gap.

<Rain> +1 re: foundational requirements being requirements that just need to be there!

Ac: Whether the conformance model could sort of give an idea of how much something conforms, and there were examples you've got a washing machine, how much electricity it confuses, how noisy is it?

And you sort of get graded on those differently

scribe: I'm really struggling to see how that would... really work in an accessibility context.
... particularly on the foundational side.
... I do wonder if it would be better to have a smaller set of foundational requirements. And a larger percentage. of supplementals and assertions.

<LoriO> +1 to Alistair

Gregg: we should be looking at safety standards, not energy or other types of standards, where you have trade-offs.
... I think that you should go from a small number of large categories into your actual needs that you collected. Which are enormously instructive and informative.

Shadi: on the one side, we're saying.

Foundation should stay foundational, and there's no leeway around

scribe: and then at the same time, we're saying, well, there's too much foundational, we should make a smaller set.

<sarahhorton> Are the prerequisites within foundational requirements the "show-stoppers"?

<Zakim> Rachael, you wanted to disagree with tags

scribe: we previously talked about the ruler and the rule. The question is, how many notches do we want to have on the router?

RM: question to Shadi if you divide it into categories and you weren't dividing it by functional need what are some examples of categories you divided into?

Gregg: the category means that you have to sort everything into one category. If it can sort into more than one category, it's not a category anymore, it's a tag.

tagging for sorting. If you want any provision to relate to more than one disability, it can't be sorted into one disability or another. It has to be tagged to apply to both disabilities, and almost all of our provisions.

<kirkwood> +1 to percentages being a problem (policy maker hat)

Gregg: reporting: when people talk about doing this, people always say, well, nobody meets 100% of it, so that doesn't mean that you fail.
... It means that you have to report where you do fail
... I think that we need to not try and back off of the foundational.

<Zakim> alastairc, you wanted to comment on differentiating by percentage of req or instance? and to comment on multi-tagging

Gregg: that the fineness that you're trying to get should be in reporting.
... we should figure out how to do it without using percentages basically anywhere except for plus stars, or something like that, not for any kind of basic conformance.

AC: was a question from Shadi on granularity. My question back would be, would you be , What percentage of provisions, or what proportion of provisions are you passing?
... there are graduations in our provisions, if you've got things like the existence like captions, and quality of captions
... we have been multi-tagging everything at the moment,

<Zakim> shadi, you wanted to respond to Rachael's questions on categories

Shadi: thank you, AC. This was a very helpful clarification to me.
... would imagine these safety requirements, these foundational ones, would be a much smaller set of what's needed to actually have.
... Shadi: I am a bit weary of percentages because I don't know what it really means, 10-20%. what is it?
... So, we do have somehow a notion of what the purpose or topic is. So maybe we can use that as a lever, as a way to measure how relevant something is or isn't.
... Maybe labeling that I was thinking about could be tied to categories.

<Frankie> +1 to Shadi's point - 10-20% of functional needs met for something like captioning for bronze level is concerning. Context is everything. Let's say a video has captions, but they're AI-generated and not very accurate. That might be okay on a social media post, but it wouldn't be okay for a video within an educational setting, particularly around

<Frankie> material where one might be tested. It would also be concerning in a healthcare context where a doctor provides a video to patients about specific follow-up care. If there are stakes for accuracy, it can't be supplemental.

Shadi: I still think that we're still boiling it down to an aggregated score. Bronze, silver, and gold, which does not say much.

LO: In terms of the various levels, I am a very much a believer in that our foundational should be,related to safety, and it's either 100% pass on that level, or a 0% pass.
... For the foundational, and make that list very, very small.
... Percentages are a scary.
... I can pick out all the easy things to do and still have a web page.

that would pass at, say. 80% the silver level. And for all of my Mobility users, they can't use the web page.

<kirkwood> levels will never be adopted by policy makers if it leaves out particular constituent groups

<kirkwood> +1 to Gregg

Gregg: we should wipe off importance of things off the bat.
... the idea of blocker is always a problem, because it's gonna always come down to the fact that for people things that look like friction for one person are the things that stop for another.

<mbgower> WCAG 2.x does have wording allowing a video not to be captioned: "except when the media is a media alternative for text and is clearly labeled as such."

<alastairc> qv?

Gregg: I want to emphasize that I really think that we need to look at the actual functional needs, not these labels, like executive function and all these things.

Scribe change?

<Zakim> alastairc, you wanted to comment on percentages and to also say the bar should stay at the same place if the foundational are reduced

<mbgower> scribe: mbgower

Alastair: In reply to Shadi, and in general... It's impossible for us to score things in regard to importance of task or user
... What I would be suggesting is that we have a smaller set of foundational.
... We reduce the number of foundational requirements but keep the bar in the same place; so a greater number of supplemental

<Zakim> shadi, you wanted to respond to Gregg on accessible yes/no

<alastairc> scribe+ Laura_Carlson

Shadi: Thank you, Gregg, for the example of a ramp. Even the ramps that follow the highest standards are a blocker for me.
... Even the lowest inclination will be a blocker for someone.
... Not everything can be accessible to everyone. We're going to have different grades or levels.
... We can't keep pretending that everything can be accessible to everyone. That's one point.
... Moving from that, this is not about getting away with things. We put sign language at AAA. So in a way we already said 'captions are more important than sign language'.
... By the same token, text alternatives for decorative images are different than those for functional images. Maybe.
... I don't think we've explored nuances enough. Two websites could be at the same level, but we know from an accessibility perspectives they vary in performance.
... Maybe a different idea... WE have 10-20% from Bronze, then 50% from silver. If we are able to measure percentages, why couldn't these things be fluid?
... I think it is clear that some things are more accessible than something else. How can we capture those in a model?

<giacomo-petri> +1 to Shadi points

Venkat: No matter how you categorize things, engineers will try to game.

<Zakim> GreggVan, you wanted to say 1) if we have assertions at Bronze we have to first have determined that comopanies are allowed to make assertions. 2) Any use of percentages

Venkat: Bronze level should be designed to reduce that risk.

Gregg: We shouldn't put assertions in the bottom level.
... Assertions are about process. That's not a measure of accessibility.
... That's one comment. Another is any use of percentages trades off one functional need against another. I don't think we should use any percentages. I don't think we need to.
... Instead, if you want to report progress. You report which you pass and which you fail.

<kirkwood> percentages are very confusing for policy makers

Gregg: Then you can see which things are supported.

<Zakim> alastairc, you wanted to comment on example for decorative images and to comment on alternative conformance approaches

<kirkwood> +1 to Gregg

<julierawe> I disagree with Gregg and would like to respond

Alastair: I was hoping for a bit of time for working through the exercise, but I'll cover off some points first.
... With chair hat on, Shadi, I think it would be worthwhile doing the exercise. You mentioned decorative images. We have those separately considered.
... We have more granularity in the provisions being proposed.

<Poornima> +1 to Gregg - following his comments, will there be requirements to identify which assertions make to which level?

Alastair: On the alternative conformance approaches, you mentioned WCAG hasn't been good at capturing variations. We had 2 sites, which came very close to each other on paper, but one was a complete car crash and the other was pretty good.
... Whatever individual provisions you look at involves the importance of the failing item to the user at the time. I don't think guidelines can capture that. It would need to be task-based.

<Jon_avila> and the frequency of occurrences can also make a big difference in whether something is usable.

Alastair: There are certain things we cannot solve with guidelines and a conformance model, and I think that's one of them.

Giacomo: Shadi's perspective aligns closely with mine. Conformance is a big challenge. I made an exercise, and one of the best sites I've ever seen; from one perspective it is very good, but from the representative sample, I detected a few violations.

Julie: I wanted to respond to something Gregg said. I think we should consider assertions at the bronze level.
... You can tell whether a list is properly coded. But you cannot test whether someone considered if it was appropriate to break up long sentences into a bulleted list.

<jspellman> +1 Julie

Julie: So that's an example that could be captured as an assertions at the base level.

<Adam_Page> +1 Julie

Poornima: One example of an assertion is to have a design system.
... Which are needed to meet any given level?

Alastair: I think the response is that with percentages, you get to choose which you choose to meet a level.

Shadi: When we previously explored options, I dont' think we had that level of granularity.
... I think we stopped those explorations too soon.
... We're saying things like 'this is impossible.' I'm not convinced we've thought more creatively. I also don't think we've thought about cut offs.
... If we have a missing text alternative, is it really progress to add an alt? Sometimes it isn't. It might not be in the way of anyone. Sometimes spending the effort somewhere else is better.

<Zakim> alastairc, you wanted to comment on percentages and trading things off

<shadi> yay!!

Alastair: We have been swinging between content (writing provisions) and conformance. That will continue, to some degree.
... I wanted to make a few comments. "Percentages shouldn't be used..." Percentages might be complicated, but where there are numbers some will report them as percentages.
... If you get to bronze, you report. Then if you then reach some more items, you can report that improvement.
... As Giacomo said, it's hard to meet the foundational. There are some foundational things that are hard to do.
... So reducing that hard pass fail on the foundational level will be more motivating.

Gregg: I'm going to do a number of short things.
... 1. SOmeone said captions are more important than sign language. That's not correct. It was technically not feasible.
... 2. At 4am, there is a stop light. So do we say you can disregard the stop lights because no one is around? No, there will be cases where you still need to. Same with alt text.

<Laura_Carlson> +1 to Gregg

<kirkwood> this approach my eliminate acceptance by policy makers

3. We keep talking about progress. You measure progress by getting more and more of them done.

scribe: It shouldn't be in foundational unless it's really important.
... And if it's really important, it should be in foundational.
... We checked on whether something is technically feasible.
... We have to focus on what's accessible, not whether people want to do it.

Shadi: I wanted to respond on the traffic lights.
... You look which lights are less frequented and you make them blink at night.
... You look at pain points and not make nonsensical rules, but respond with things in the right place. THat's what I'm advocating for.

Kirkwood: I'm concerned about this approach.
... As a former policy maker, if we no longer can say 'we guarantee this will be accessible' and say
... it will meet it for a certain percentage, this will be a problem

Alastair: Just to clarify, the proposal is to have WCAG 2AA are foundational. That's essentially the same model.
... Is that problematic?

Kirkwood: It's not moving the ball forward.
... Coming up with a proper categorization to meet specific accessibility needs, to widen the understanding of types of needs.

Alastair: I'm trying to work out how that isn't what we proposed.

<julierawe> I have to drop — thanks for good conversation, everyone!

Alastair: It has the WCAG AA layer. Then it has things on top of that, which you can score.

Kirkwood: I think that the score is a probelm.

<Zakim> alastairc, you wanted to comment on where to draw a line. And that WCAG 2 has been in force for years and isn't met.

Kirkwood: It's either accessible or not. It's binary.

Alastair: We draw a line. That's what WCAG 2 has done. The line is drawn at AA or it fails.
... The proposal is to draw a similar line and let people build on that.

Kirkwood: Is that the case? I hear conflicting opinions.

WCAG3 provision evaluation task (continuing) https://lists.w3.org/Archives/Public/w3c-wai-gl/2025OctDec/0013.html

Rachael: This particuarly proposal does not go below WCAG 2.

Alastair: It really helps when people are more familiar with the requirements we are proposing.
... My impression from going through this with a website...
... Oh, decorative is foundational...

[sharing screen]

[Reviewing WCAG 3 guidelines testing sheet]

<Rachael> Link to template sheet: https://docs.google.com/spreadsheets/d/1vFgaAY7F1lIuQX8z2O4BrXdKC1Y-cfnda8EESct6iho/edit?gid=0#gid=0

scribe: If you have to meet 100% of foundational and supplementl is on top, what does this mean? My impression is that it's going to be harder to meet foundational, than to score a percentage.
... If they're going beyond foundational, what do they focus on?

<alastairc> https://drive.google.com/drive/u/0/folders/1b8ccocczCONPFokE6y6bAMckMcQMHJqT

scribe: We have this folder
... The idea is people take the testing sheet, and make a copy with your name. You can download it or save it to the same folder.
... We are going to draw together everyone's results.
... That will help us in terms of refinement.
... For those thinking about alternative conformance approaches, now we've go this stuff, how do we make it easier?
... For instance, if you have an assertion that you have a style gude for image text alternatives, and it's followed, maybe you can pass some of the rows that immediately precede that?

Giacomo: It was more than 4 hours!

Kirkwood: I was wondering about feedback from the group if they understand the exercise.

Glenda: I wanted to make sure I understood the exercise. I make a copy and use column G to decide whether it is foundation, supplemental or an assertion.

Alastair: No, with a website in mind, don't necessarily audit it, but for one you're famiilar with, would it pass or fail?

<kirkwood> some were confused by the directions. Per Glenda’s question

Alastair: The Notes is to say "i don't understand" or "not feasible" or "not applicable in this situation"

Giacomo: I did this exercise based on feelings. But doing it with real examples made a difference. For example one had a policy on alt images.
... Without performing an audit, it is too abstract. If it is "superficial" the results will not be real.

Alastair: I have to cut you off, based on out of time. I think what you're raising is important.

trackbot, end meeting

This is scribe.perl Revision VERSION of 2020-12-31
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: Irssi_ISO8601_Log_Text_Format (score 1.00)

Succeeded: s/did we explore/did we sufficiently explore/
Succeeded: s/whould/should/
Succeeded: s/wide/widen/
Default Present: alastairc, Laura_Carlson, kevin, RichardM, Azlan, jspellman, Rachael, Lenb, julierawe, filippo-zorzi, Adam_Page, mbgower, kirkwood, Nicholas, maryjom, Poornima, LTSzivos, Jennie_Delisi, giacomo-petri, Eric_hind, Charu, todd, Jon_avila, AlinaV, Rayianna, Rain, sarahhorton, Frankie, Glenda, BrianE, graham, shadi, LoriO, Jen_G, GreggVan, venkat
Present: alastairc, Laura_Carlson, kevin, RichardM, Azlan, jspellman, Rachael, Lenb, julierawe, filippo-zorzi, Adam_Page, mbgower, kirkwood, Nicholas, maryjom, Poornima, LTSzivos, Jennie_Delisi, giacomo-petri, Eric_hind, Charu, todd, Jon_avila, AlinaV, Rayianna, Rain, sarahhorton, Frankie, Glenda, BrianE, graham, shadi, LoriO, Jen_G, GreggVan, venkat
Regrets: Makoto, Hidde
Found Scribe: Laura_Carlson
Inferring ScribeNick: Laura_Carlson
Found Scribe: mbgower
Inferring ScribeNick: mbgower
Scribes: Laura_Carlson, mbgower
ScribeNicks: Laura_Carlson, mbgower

WARNING: No meeting chair found!
You should specify the meeting chair like this:
<dbooth> Chair: dbooth


WARNING: No date found!  Assuming today.  (Hint: Specify
the W3C IRC log URL, and the date will be determined from that.)
Or specify the date like this:
<dbooth> Date: 12 Sep 2002

People with action items: 

WARNING: IRC log location not specified!  (You can ignore this 
warning if you do not want the generated minutes to contain 
a link to the original IRC log.)