Meeting minutes
<Chuck> +1 to Wendy!
[general chatter about the wonders of scribing and why we don't just use AI transcription features]
<kirkwood> i use it as well
Announcements, introductions, new topics
alastairc: Don't think we have any announcements
… do we have anyone new or with new affiliation?
… any new topics for us to tackle?
Publication CFC
alastairc: let's kickoff with the agenda
Lori: There was discussion some time ago, is WCAG going to say anything about AI generated text or images?
… there's a push for AI generated text or images to be marked as such, is WCAG going to say anything?
Chuck: In the short term, no, but in the CFC that went out, there are outcomes that do consider it
… considering possible guidance in WCAG 3, in motion
alastairc: The CFC discussion, we did send one out yesterday
… thanks to those who've +1'd
… we did receive some comments
<Rachael> w3c/
alastairc: Wendy had a few comments on the introduction requesting we provide an overview of the structure
… how people should send guidance, and some working
<Rachael> Pull request: https://
<Rachael> https://
Rachael: I am working on the comments, didn't quite understand the third so I'm hoping to work through it
… we should be thoughtful about the changes
… [quote from doc]
<Rachael> https://
Rachael: I thought it was the first part of the sentence, but there should be a change to the bullet itself
… we're in the introduction
… change the bullet to
alastairc: People are sending in votes, and if we make any tweaks, we make sure everyone who has voted already is aware
<Rachael> We also expect changes and clarifications of Outcomes,such as: The text of the Outcomes will change, and Only some of the Outcomes will be required to meet the base level of conformance.
julierawe: Rachael, the changes you were showing, some outcomes could be removed, but some could also be added if new gaps are identified
Rachael: Great point
GreggVan: Just a wording, the outcomes will be required, actually it's "only some outcomes are required", it's backwards in the wording
<alastairc> "... the Outcomes will be required at the base level"
GreggVan: only some outcomes are required at the author level
Rachael: Drafting wording in IRC
<Rachael> The final version of WCAG 3 will change. Some outcomes may be added and some may b eremoved. We also expect changes to the text of the Outcomes will change. Only some of the Outcomes will be required at the base level of conformance.
alastairc: Might be best use of time if anyone has questions or comments
… run final changes by people at the end of the meeting
wendyreid: In the editors note for the guidelines, we clearly explain what we want in feedback in research. In research we tell people to file a github issue or send email with research, but we don't do that for feedback. We should do the same for feedback.
Rachael: So just repeat or adjust the sentence
Wendyreid: Copy the sentence or adjust existing sentence and list out both research and feedback.
Nayan: Question on the conformance levels, is this the correct time to discuss this?
alastairc: We'll be coming back to conformance fairly soon, we have other things coming, and it's not part of the working draft changes
… but we'll come back to it soon
Rachael: If other people have comments on the changes I made to the About WCAG3 section, please let me know
alastairc: Looks like a big change
Rachael: It's mostly a copy-paste from the abstract
alastairc: Lines 60, 61 are from the Abstract
Rachael: I put in a linebreak for readability
<alastairc> https://
alastairc: If anyone has +1'd but disagrees with the changes, let us know, but otherwise we'll proceed
… any other questions or comments about the working draft publication?
julierawe: Question about conformance, it's not changing?
alastairc: It's not changing in this draft
julierawe: If you had a question or issue, you could report it?
alastairc: Yes, feel free to open an issue in the repo
<alastairc> https://
julierawe: I'm trying to understand the CfC, and how the conformance part relates
Rachael: To try to move things forward, we should focus on the parts of the document that are changing from the previous version
… it's not to downplay the questions people might have about conformance, but we will continue that discussion
… did that clarify?
julierawe: Yes
dan_bjorge: I see that the individual outcomes within guidelines are alphabetical, but the guidelines aren't? Is that planned
alastairc: We haven't figured out a meaningful way to do it yet
Rachael: We discussed it, since we put it through the group in a certain order, we kept it, but we can change it if we want
julierawe: To vote on the CfC, is it simply replying to Rachael's email with a +1 or your concerns? Is that the mechanism
alastairc: If you're happy with what is being proposed, +1, or 0 if you're neutral or can live with it, and -1 if you have strong opposition, but explain why
GreggVan: I think grouping them makes sense, and alphabetizing the groups, but scattering them doesn't make sense
alastairc: They are grouped, but only the outcomes within the groups are alphabetized, not the groups themselves
GreggVan: I misheard, I thought we were talking about moving them out of the groups
alastairc: Keeping in groups
Chuck: We've had concerns from folks about concerns "can live with" or "can't live with", preference is for "can tolerate" or "can't tolerate"
<NatTarnoff> I think "accept" is the way to go on that.
alastairc: We had conversation about alphabetizing the groups, no strong opinion in either way, so we may leave it as is for the time being
Accessibility supported https://github.com/w3c/wcag3/discussions/53#discussioncomment-9128840
alastairc: move on the to the next item
alastairc: Accessibility supported, it's a topic we've discussed a couple of times
… trying to work towards a decision on how to define this
… working from the definition in WCAG 2
… how does the standard need to define what level of testing we need to do to determine support
… plenty of work on that previously
… setting up a database of test files
… do we need a similar concept in WCAG3?
… I think we came to the conclusion that yes we do
<NatTarnoff> I'm only one person and a survivor, but "can't live with/out" has some trauma effect. Even "tolerance" bears some stigma given the hot topic of current world events.
alastairc: the proposal was taken from the many points and collect them
… [sharing screen]
… I thought it would be worth going through this proposal and seeing if it addresses the concerns
… does it sit well with people as the approach for this
… the proposal was to tie it to the levels
… the level of author requirement increases with the levels
… as an author you write to the specs, if you do that properly, you don't have to thoroughly test all of the technologies
… at the bronze level, assume AT can support
… at silver, can involve a level of compatibility testing, but being mindful of our ability to take on AT testing
… write documentation for how authors can determine their levels
… gold would be similar to silver, but more usability testing
… silver and gold, the author takes on more responsibility
… we could put out documentation on how to determine what AT to support
… in the UK, the government provides a list of what they require testing on
… we could advise on how to create the list, not creating the list ourselves
… put a list together
… methods at bronze need to be generic
… at silver, some UA support
… imagine for a moment there's a new HTML element called <tabpanel>
… at the different levels, there's more integration
… and in addressing this proposal to try and answer the questiosn
… you may not like the answers, but we have answers for each of the questions
graham: My gut reaction to this is there are 2 issues
… we're introducing a barrier for countries that don't have the funding to test with all the combinations
… we create work on our side as we need to keep up to date with platforms and operating systems
… in 6 months my testing could be out of date
… vs previous WCAG versions where we test against a list of platforms
<Zakim> jeanne, you wanted to remind about use case identified where there are unfixed bugs in AT
<NatTarnoff> How much would we rely on specific versions? Browsers are updated every 6 weeks.
jeanne: First of all I like this approach
… what I was doing was looking at some of the use cases we started with around accessibility supported
… I think this addresses some of them
… the one it doesn't is where there is bugs in the AT, and the org can't get the AT manufacturer to fix the bug
… we need an exception for that situation
… the org has done what it can, but they should still be able to get silver gold
… I like the direction this is going
<Zakim> GreggVan, you wanted to say IN silver level suggest editorial change to " testing with user agents (e.g. browsers, assistive technologies, etc.) "
GreggVan: Two things, first editorial, what it's going to do is cause many more people to not look at all the cognitive language and learning things at silver
… if I'm doing this and I'm a teacher, small business owner, I'm going to look at this, see I don't have the technologies
… stick to bronze
… silver is going to catch a lot of things we want people to do but is more complex than bronze
… we want people to do silver, but this might push them down
… the other thing, conforming to specific technologies at the bronze leve
… we've tried to be technology agnostic
… we've put this in techniques, which have no level
… we can use them across technologies, as soon as we put methods near the normative information, we fall out of being technology agnostic
… it will be challenging to adopt across technologies
<kirkwood> +1 strongly agree to Gregg’s concerns
GreggVan: and will happen at a time when web technoligies will become more blurred
<Zakim> mbgower, you wanted to say can we measure against the a11y tree instead? how does the author report success/failure?
<dan_bjorge> +1 to all of Gregg's concerns
mbgower: I'm wondering if we can measure against the accessibility tree vs AT
… for some of the reasons that have been highlighted
… I put in a link in the topic, ARIA-AT that is kind of in this space
<kirkwood> Does the proposal incorporate AI? seems oddly omitted to me. Is AI part of AT?
mbgower: they're writing methods for testing
… differences between ARIA and ATs
… similar to how we're looking at tests for checkers, I wonder if there is some way we can leverage the accessibility tree and further the validation without AT
… another point, at the moment we require accessibility supported, but there is nowhere an author can report it
… we have a very long outstanding issue where an application is usable by keyboard, but you turn on AT and keyboard stops working
… not clear how to fail that
… maybe you can fail against 2.1.1 or 4.1.2 but it's not clear
… we should be mindful of that in 3
<Zakim> alastairc, you wanted to comment on work for organisations at silver, is it different from now? and to also say that COGA ones haven't (so far) included AT, so less impacted.
alastairc: Comment earlier about the level of work required, my question is it more or less than what we have now
… or is it more obvious what it is
… we have this in WCAG2, I'm not sure it's different
… its just making it clearer
… to Gregg, I'm not sure I understand the cognitive part, it's either going to push out the cognitive requirements, but I'm not sure, due to the antitcipated level of work?
GreggVan: I'm looking at silver and bronze, silver says I have to test with AT
… I don't know how to do it
… so I have to hire someone, too much work, so I'll just stick to bronze
<kirkwood> very good point by Gregg
<NatTarnoff> +1
GreggVan: stop at bronze ecause they don't think they can get to silver
… silver should be approachable
… each silver is its own thing, so people can say "I've done x number of silvers"
<ben_tillyer> -1. In my ‘cup half full’ mind, I reckon some people might go for ‘bronze.5’ without having another medal to show for it
<kevin> +1 to Gregg's concerns
<avk> +1 Gregg
dan_bjorge: Mostly I think Gregg addressed my concerns, I agree but making silver only something achievable with AT and spending money on testing and such
… in WCAG2 we have an idea of accessibility supported at every level, we need the same in 3
… my inclination is we should do something like WCAG2 where we have something for ever level
… we require assertions and statements at levels like gold
MJ3: Agree with Gregg, gold seems the same as silver but with less guidance on the testing methods
… not sure if there is a pereivable difference, should be clarified
alastairc: Just to clarify, it should be [missed]
GreggVan: +1 to dan_bjorge, you should be able to conform without hiring someone, at least at bronze
… that's out of reach for many places
… we're misunderstanding what accessibility supported means, esp in WCAG2
… if you use our techniques, you don't have to worry
… if you want to do this a different way, then and only then do you need to worry about accessibility supported
… if I have a better/new/different way
… you do need to make sure it's supported
… that's where accessibility supported comes in
<kirkwood> In its current form Silver is likely to be determined to be an “undue burden” and therefore I feel there is a likely outcome is that it will not by adopted by regulators under the ADA per precedent.
GreggVan: when you want to go outside the accepted techniques
Nayan: when you say AT, authoring tools?
GreggVan: Assistive tech
<dan_bjorge> We definitely have published techniques that are not accessibility supported in regions where common ATs have poor ARIA support (eg, Japan)
Nayan: Regulators have been using 2.1 AA, silver is going to be mandated/required as a safe harbour for litigation
… I have an issue with the wording, sounds like insurance, I agree with Gregg and the others, most small businesses using tools like CMSs
… they should be able to conform to the baseline that is achievable
… that will increase conformance, if we require manual testing for a standard that regulators will mandate, it will impose a financial burden
… people will avoid dealing with it
… we do not pass any website without manual testing
… it is very difficult for any small business to achieve that silver level
… if testing is required, you're creating cost
… we've already had cases where orgs have turned off content because they can't afford to have it tested or remediated
<Zakim> alastairc, you wanted to ask whether 'meeting spec' is enough at Silver? and to also say accessibility supported doesn't work internationally at the moemnt
Nayan: whatever the middle is, the middle should be achievable by a large quantity of people creating websites/web content
alastairc: The alternative is meeting spec approach is at silver
… the current WCAG2 approach, doesn't work internationally
… one of the problems raised that we're trying to solve is that our techniques don't work in every country
… like in Japan, their main screen reader doesn't support much ARIA, we put a lot of emphasis on ARIA
… I was trying to incorporate into the proposal
… it's all on the author in WCAG2
… if we incorporate known information, and we know it works, we could incorporate into the method so that authors don't need to test that
… we got AT information that matches their market
… that wouldn't require work from authors
… I'm confused by the pushback at the silver lelve
<kirkwood> methods can be handled by AI these days… its all evolving.
jtoles: I agree with what's been said so far about requiring testing
… I don't agree that just meeting specifications is sufficient at silver
… what could someone do as a small business owner
… at bronze, you're using a code validator
… at silver, you're using x automated testing tools
… at gold, I've done AT testing, usability testing
… trying to insert some automation at silver
… automation doesn't catch everythign
wendyreid: I understand the concerns about silver and this level, and I agree with many including use case of small org or business owner that doesn't have resources for AT testing.
<kevin> -1 to specifically including automated testing tools as a level
wendyreid: We are forgetting that we are still holding on to the idea for recommendations for authoring tools and user agents, that should be part of the interplay. I agree that small business owners should not be responsible for large scale testing. But we should put requirements on tool vendors and platforms on behalf of small business owners.
wendyreid: And you are assuring as an author that you are using tools and platforms that are conforming.
<kirkwood> +1 to authoring tools point
<alastairc> +1 to investigating procurement assertions.
<jeanne> +1 to procurement assertions
GreggVan: Somebody said silver should be required for safe harbor. I don't think we should think of bronze, silver gold as A, AA, and AAA. If we make silver as something required then we would have to take anything out that cannot be tested. We shouldn't be thinking of silver as mandated.
GreggVan: Manual testing ... there was a question about no testing at bronze. Of course you need to look at your content. It shouldn't require "external" testers. AT testing does usually require external testers. That requires you find someone who knows how to use AT. Manual testing is required. User testing should be at Gold. It should
always be done, but not everyone has access.
<Chuck> GreggVan: There are many degrees of disabilities, and AT testing is spot testing. Finally, techniques don't work. Alastair, you had a great idea of "sufficient", "this is sufficient to meet group A AT and browsers".
<kirkwood> it is in NYC
<Chuck> GreggVan: Something that says these techniques work with AT. Or we have known ones that doesn't. We shouldn't have any rules that can't be met with any techniques in countries that have their own screen readers. We do have to think... if you have a country that speaks a unique language, with a screen reader put together by a grad student...
<Chuck> I don't think WCAG would be required in that country.
gregg: Trying to think international can be problematic.
<Zakim> mbgower, you wanted to say that not all the techniques are accessibility supported; just think of general techniques and to say that not all the techniques are accessibility supported; just think of general techniques
mg: thought bronze was = to WCAG AA
<kirkwood> NYC Urdu had to be legally accessible.
mg: did we expect mom and pop businesses were going to try for silver?
<Zakim> Rachael, you wanted to ask if we have an underlying assumption based on A, AA, and AAA
<ben_tillyer> can i ask what the thinking was behind Gregg calling out ‘that country’ with many Urdu speakers as not needing to comply with accessibility standards? Having seen the efforts India has gone through recently; the comment has irked me…
rm: difficult conversation.
… bronze would be A an AA.
… need to think about how well something can be met.
<scotto> coming in a bit late - but my initial reaction to seeing that people could be given different grading levels based on whether someone used a native HTML pattern vs an ARIA pattern seems troubling. especially if the end result is indistinguishable.
rm: bronze everything that can be met internationally. Gold would be experimental.
… need for an onboarding process.
<Zakim> GreggVan, you wanted to say -- I HOPE that mom and pop will try to meet as much of silver as they can.-- because of all the CLL things there... but here is another way of thinking of silver...
ac: had assumed AA would go into bronze.
Graham: high up latter we go less of this sort of thinking.
… feels restricts you backward.
np: would be conformable if Bronze was = to AA
… maybe have a questionerr if you are relying on 3rd party you can make an assertion.
<GreggVan> @Ben Sorry --- my understanding of the use of Urdu showed -- so pull that comment back. Forget that I ever mentioned Urdu as a language. It is actually spoken by more people than german, japanese, korean, italian and 30 of the top 40. It is in fact the 10th most spoken language on earth. So thanks for the correction Ben - and noting it for the record.
<ben_tillyer> Thanks GreggVan, acknowledged
<Zakim> alastairc, you wanted to comment on maintaining techniques/method compatibility
kevin: Any form of validating assertions is not something that w3c would get involved with.
<Zakim> kevin, you wanted to react to nayan to respond to course/validation idea
ac: assertions could perhaps work on procurement outcome.
<Zakim> GreggVan, you wanted to say -- I HOPE that mom and pop will try to meet as much of silver as they can.-- because of all the CLL things there... but here is another way of thinking
<jtoles> Provide a link to the accessibility conformance report for your CMS in your site's accessibility statement
<graham> haha. easily done!
gregg: Hope every mom and pop look at silver.
<avk> Rachael personally didn't have A/AA/AAA underlying assumption - didn't think bronze/silver/gold translated directly in that way. But it was unclear what the relationship was. Thanks for clarifying, maybe add to the proposed descriptions?
gregg: don't think of silver as something off on the side.
… have it up front. Think of it of something more valuable.
<kevin> +1 to embedding 'here's how you could do it better'
Todd: may be setting the bar too high. Mom and pop will use an overlay as it it cheaper.
<kirkwood> +1 to overlay point
<Francis_Storr> +1 to Todd's overlay comment
Dan: What is our goal for silver?
… encourage practical accessibility.
… hesitant to investigate 3rd party.
<Zakim> alastairc, you wanted to comment on what we can maintain, having methods that are known to work.
ac: thinking more of things like shopfy.
… needs more thought and investigation.
… wcag2 is relatively flat.
… we could add more nuance.
<Zakim> sarahhorton, you wanted to ask about whether we would still require some level of accessibility support when creating the requirements in the first place
sarah: If something was bronze we would have done due diligence.
… yes. sometimes internationalization is the issue.
np: agree with the last statement.
… code vs content.
<alastairc> https://
<kevin> -1 to separating code/content and proposing responsibilities
<Rachael> https://
ac: please discuss more on github. Or make your own proposal
<Rachael> https://
rm: please check https://
WCAG 2 issue resolutions https://lists.w3.org/Archives/Public/w3c-wai-gl/2024AprJun/0053.html
<Zakim> Chuck, you wanted to test queue
<alastairc> https://
<mbgower> https://
mg: ac is sharing project board
… some things were removed target size unintentionally. and is being restored
… second one is 3518 non normative.
… 1481 using CSS instead of square brackets.
… 3272 is a new technique for focus visible.
<bruce_bailey> list is here: https://
mg: 1651 is under discussion.
… is in flux and is likely will change.
ac: have a look at the email and comment or give thumbs up in GitHub.
<ben_tillyer> Thanks all
<alastairc> "Kevin just mentioned that he was trying to lower his hand and accidentally asked a question. He also clarified that he was using an AI companion that was able to summarize the meeting and ask questions, and found it valuable. He planned to continue using it."
<MJ3> Thanks, all!
<GreggVan> Yes -- good. Can he confirm that the summary was accurate? they always sound good - but good to know if they are actually accurate as well
<alastairc> https://