W3C

- DRAFT -

AGWG Teleconference

16 Jan 2024

Attendees

Present
Jennie_Delisi, Laura_Carlson, bruce_bailey, Rachael, dan_bjorge, wendyreid, Ben_Tillyer, jeanne_ec, ShawnT, GreggVan, Graham, Azlan, Wilco, kevin, giacomo-petri, mbgower, kirkwood, julierawe, Carolina_Crespo, JustineP, jtoles, mike_beganyi, jeanne, Bri, AWK, Makoto, Detlev, Frankie, tburtin, ashleyfirth, ljoakley, Poornima, DanielHE, jon_avila, jaunita_george, Francis_Storr, alastairc, GN, scolto, \, maryjom, GN015
Regrets
JakeA, ToddL, DuffJ, DJ
Chair
alastair
Scribe
Jennie, mbgower

Contents


<Jennie_Delisi> scribe: Jennie

<Jennie_Delisi> AlastairC: We will get started in a minute

Announcements (inc CSUN)

<Jennie_Delisi> AlastairC: We had a few introductions last week. Anyone new who has not introduced themselves?

<Jennie_Delisi> ...(no one came forward)

<julierawe> present_

<Jennie_Delisi> AlastairC: CSUN in March - no official group activity.

<Jennie_Delisi> ...There will be people from the group going to CSUN.

<Jennie_Delisi> ...If you are interested in dinner, contact Kevin

<Jennie_Delisi> Kevin: email is best.

<Jennie_Delisi> AlastairC: Gregg has an announcement

<Jennie_Delisi> Gregg: I was speaking with my previous co-chair, and she has retired from Google

<Jennie_Delisi> ...If anyone wants to put a thank you note together, it would be appreciated.

<Jennie_Delisi> ...We had over a 1000 comments come in, and I never found anyone who worked so hard for WCAG

<Jennie_Delisi> ...I think she is amazing and an undersung hero

<bruce_bailey> +1 to GreggV comment on Loretta's hard work!

<Zakim> jeanne, you wanted to ask for a personal email address for Loretta

<Jennie_Delisi> Jeanne: just to ask for her email

<Jennie_Delisi> Gregg: I will share it

<Rachael> +1 to sharing. In Zoom could also be used to keep it out of IRC

<Jennie_Delisi> AlastairC: Thank you Gregg. I am sure there will be quite a few people.

Card sort of outcomes

<AWK> +AWK

<Jennie_Delisi> AlastairC: This is a card sort of the outcomes

<Jennie_Delisi> ...Our aim: work out which outcome people associate with each other.

<Jennie_Delisi> ...This will help map out the information space.

<Jennie_Delisi> ...This is trying to bring together the work of smaller groups.

<Jennie_Delisi> ...These have been developed somewhat independently, so we can learn how they relate.

<alastairc> Card sort: https://uxd-library.optimalworkshop.com/optimalsort/062osikf

<Jennie_Delisi> ...When you go through it, you will get some instructions.

<Jennie_Delisi> ...Drag or select each card, put it in the right hand side, group things you think should go together.

<alastairc> Mapping document: https://docs.google.com/spreadsheets/d/1fw0-q4_iZPZ44_PmvUbJBH8CtQtxwXEXlJ3jac1T5JY/edit#gid=1073399112

<Jennie_Delisi> ...Each card is showing the outcome short name and description.

<Jennie_Delisi> ...This mapping document has the ID from each card.

<Jennie_Delisi> ...I have tried it. I need to look up a few.

<Jennie_Delisi> ...You can look up the user need to help you know where it should go.

<Jennie_Delisi> AlastairC: When you sort things, if you get a large group of things, there is an accordion to open and close it

<Jennie_Delisi> ...Resizing may cause issues

<Jennie_Delisi> AlastairC: 1st page gives a welcome and instructions.

<Jennie_Delisi> ...Continuing there are more instructions.

<Jennie_Delisi> ...Aim: group what you think go together. Give them a name if you can for the groups you make.

<Jennie_Delisi> ...If you leave it overnight, it may time out.

<Jennie_Delisi> ...You may lose some

<Jennie_Delisi> ...a partially completed sort is better than none.

<Jennie_Delisi> ...Hit finish even if you have not done all the cards.

<Jennie_Delisi> ...You can leave a comment, possibly just 1.

<Jennie_Delisi> ...You can pre-create a group for things you think will be there, or you can create groups as you go

<Jennie_Delisi> Rachael: I think 4 hours of inactivity triggers the time out

<Graham> yes it says it when you open the first screen that it is 4 hgours

<Zakim> kevin, you wanted to comment on Finish

<Jennie_Delisi> Kevin: I found you can't finish if you have unsorted items. Is there an option to allow it?

<Jennie_Delisi> ...I put them into a group called "couldn't sort"

<Jennie_Delisi> Detlev: Should we call it up in the browser and do it individually?

<Jennie_Delisi> AlastairC: We are not expecting everyone to do it, but if 10-20 do it, it would help.

<Jennie_Delisi> ...It would be better not to do it during the meeting.

<Jennie_Delisi> ...Ideally in the next week.

<Jennie_Delisi> Rachael: We talked about a week

<Jennie_Delisi> Detlev: If there a way to save the results of a sort or how does it work?

<Jennie_Delisi> Rachael: I don't have a way to do a partial save. So, it is do as much as you can.

<Jennie_Delisi> Detlev: how do we document what we have done?

<Jennie_Delisi> AlastairC: When you save it, we get access to the results.

<Jennie_Delisi> Rachael: And if you cannot finish everything, create a group indicating this, and put everything else in there.

<Jennie_Delisi> AlastairC: Each card has an id. If you go to the spreadsheet, scroll to the ID. It gives me access to the associated scratch pad.

<Jennie_Delisi> ...I can look up the term.

<Rachael> Thanks to Jeanne for all her work getting this together.

<Jennie_Delisi> Detlev: it is pointing to different places within the scratchpad document?

<Jennie_Delisi> AlastairC: That is the aim

<Jennie_Delisi> ...It gives the user need, and sometimes that was all I needed

<Jennie_Delisi> Bri: This google excel sheet has the categories.

<Jennie_Delisi> ...So we should not use it?

<Jennie_Delisi> AlastairC: That is the scratchpad where they came from, but it may not be where you want to categorize it.

<Jennie_Delisi> ...Use it only if you don't understand the requirement.

<Jennie_Delisi> Bri: The goal is to get consensus about how we think it should be categorized?

<Jennie_Delisi> AlastairC: Partially, it is also to get ideas.

<Jennie_Delisi> ...When others hear about a possible structure, when others hear it, they may agree.

<Jennie_Delisi> ...We are exploring the information space.

<Detlev> +q

<Jennie_Delisi> JulieRawe: I am concerned about the time limit.

<Jennie_Delisi> ...Do you recommend looking at the Google doc first to give a sense of what will be in the card sorting document?

<Jennie_Delisi> AlastairC: The 4 hours is an activity time out. So, you could do it over the course of a day. Hopefully it is not stricter than that.

<Jennie_Delisi> ...Don't leave it overnight.

<Jennie_Delisi> ...I would not recommend going through the spreadsheet first.

<Jennie_Delisi> ...I would have it on hand to look something up.

<Jennie_Delisi> Detlev: I realize that looking at the spreadsheet - it is not possible to widen the columns.

<Jennie_Delisi> ...Is that something that can be done?

<Jennie_Delisi> AlastairC: The shortname and description should be completely visible in the card.

<Jennie_Delisi> ...In the spreadsheet you are just looking for the ID number. The rest of the information is in the scratchpad - from the link.

<Jennie_Delisi> Dan: You can make a copy, then format however you like

<Jennie_Delisi> Graham: You can get your categories set up with a column after saving a copy.

<Jennie_Delisi> Jennie: Can people participate more than one time?

<Jennie_Delisi> AlastairC: The order of cards is randomized, so people do not receive the information in the same order.

<Jennie_Delisi> ...If people cannot do that, the spreadsheet is potentially a back up.

<Jennie_Delisi> Rachael: For anyone who cannot use the optimal sort - email the chairs.

<Jennie_Delisi> ...We will email you a document, and you can email a sorted version.

<Zakim> bruce_bailey, you wanted to ask about 2 hour guess on welcome screen -- is that estimate from app or chairs ?

<Jennie_Delisi> Bruce: This initial splash screen mentions 2 hours. Does that feel about right? And did that come from the chairs?

<Jennie_Delisi> AlastairC: That is how long it took us

<Jennie_Delisi> ...Several of us tried it.

<Jennie_Delisi> JulieRawe: can we have 2 weeks to complete this? Given some have a short week at work.

<Jennie_Delisi> AlastairC: If it will help you fill it in then yes.

<Jennie_Delisi> ...Any other questions?

<alastairc> Card sort: https://uxd-library.optimalworkshop.com/optimalsort/062osikf/sort

<Rachael> Aiming to close this on Jan 30th.

<Jennie_Delisi> Graham: is there a link to keyboard controls?

<alastairc> Mapping: https://docs.google.com/spreadsheets/d/1fw0-q4_iZPZ44_PmvUbJBH8CtQtxwXEXlJ3jac1T5JY/edit#gid=1073399112

<Jennie_Delisi> Rachael: I will find it

<Jennie_Delisi> AlastairC: If further questions, please email the chairs

<Jennie_Delisi> Mike G: you can tab across the top, then it goes to the items on the left side.

<Jennie_Delisi> ...you can use arrow keys to do an x/y mouse control.

<Jennie_Delisi> Graham: I have a different issue.

Evaluation Criteria for Conformance Models https://github.com/w3c/wcag3/discussions/34

<Jennie_Delisi> AlastairC: (sharing #34 in github onscreen)

<Jennie_Delisi> ...We are not evaluating the individual guidelines themselves

<Jennie_Delisi> ...There was a symposium on metrics

<Jennie_Delisi> ...From the discussion there were 2 additional proposals

<Jennie_Delisi> ...Equity and Comprehension

<Jennie_Delisi> ...When we get to look at conformance models, how is one better than the others

<Jennie_Delisi> ...Wilco raised a comment about is it possible to come up with a conformance model that meets all these criteria?

<Jennie_Delisi> ...Rachael replied: it is a matter of coming up with a way to evaluate.

<Jennie_Delisi> ...It is a matter of working out the trade-offs.

<Jennie_Delisi> ...What are the measures we are taking.

<Jennie_Delisi> ...We had a few comments around adding equity as a measure.

<Jennie_Delisi> ...One comment: is that part of validity?

<Jennie_Delisi> ...If we are trying to correlate accessibility, equity across disability groups is part of it.

<Jennie_Delisi> ...There is a related comment: would an emphasis on test time mean we will cover

<Jennie_Delisi> ...certain needs in a less detailed way?

Equity

<Jennie_Delisi> AlastairC: Should we add this as a measure? What should the balance be between this and complexity?

<Jennie_Delisi> Shadi: The report was about can we measure accessibility?

<Jennie_Delisi> ...Like bronze, silver, gold - things like that.

<Jennie_Delisi> ...That is different than conformance

<Jennie_Delisi> ...WCAG 2 defines the levels, which you could see as a metric

<Jennie_Delisi> ...These criteria were not to evaluate the conformance model itself, it was to talk about scoring for accessibility

<Wilco> +1, these criteria feel out of place

<Zakim> mbgower, you wanted to say that you need to better define equity

<Jennie_Delisi> MikeG: I am always worried about a definition of equity

<Jennie_Delisi> ...Is it the # of criteria?

<Jennie_Delisi> ...A certain level of effort?

<Jennie_Delisi> ...I think we need to be more specific.

<Jennie_Delisi> ...It is a harder one to figure out what it needs.

<kirkwood> +1 to MG re defining term w/ term

<Jennie_Delisi> Wendy: Agreed - we should be clear in our definition of equity.

<Jennie_Delisi> ...I also agree with Shadi.

<Jennie_Delisi> ...It is tricky to use equity. I do think it is an aspect we should consider.

<Jennie_Delisi> ...When we have levels, like bronze, silver, goal - each level should create an equitable level for all areas

<Jennie_Delisi> ...We should probably be clearer and define it more clearly

<Jennie_Delisi> Wilco: My comment is about scoring, not equity

<Jennie_Delisi> AlastairC: Can we come back to that?

<Jennie_Delisi> Wilco: Yes

<Jennie_Delisi> Gregg: I think equity is a really important design principle.

<Jennie_Delisi> ...Then people with epilepsy are not considered as much - we would need more items related to them

<Jennie_Delisi> ...Everything was put where it could fit.

<Jennie_Delisi> ...It is that more time was spent on it, but if we couldn't figure out how to make it fit with the criteria for the other levels.

<Jennie_Delisi> ...We have the assertions going - that is a huge step forward.

<Jennie_Delisi> ...Some might say that all the cognitive ones are in assertions

<Jennie_Delisi> ...But it is not. If it meets an outcome, then it goes there.

<mbgower> "Whereas equality means providing the same to all, equity means recognizing that we do not all start from the same place and must acknowledge and make adjustments to imbalances."

<Jennie_Delisi> ...If we base equity on counting - I want to speak towards having equity as something we put in front of us to be sure nothing gets left out

<Jennie_Delisi> ...But I worry about it as a criterion when we have no way to talk about exactly what we mean by it.

<Zakim> bruce_bailey, you wanted to ask if adequacy definition is definition for sensitivity ?

<Jennie_Delisi> ...We don't have a way to measure seizures triggered by sound - this is because we don't have a way to measure it.

<Jennie_Delisi> Bruce: Sensitivity - seems like part of validity to me.

<Jennie_Delisi> AlastairC: It is useful to go to the full document for those.

<Jennie_Delisi> Jeanne: I want to propose - equity is a good idea.

<Jennie_Delisi> ...But I don't think we should do it in a traditional measurement way

<Jennie_Delisi> ...In Silver: we had people who had expertise in equity, and point out areas where we were weak.

<Jennie_Delisi> ...I think that is a better approach, rather than a measurement.

<Jennie_Delisi> ...We don't need something counted.

<Jennie_Delisi> ...We could have paragraphs written by experts pointing out potential issues.

<Rachael> +1

<GreggVan> +1

<Zakim> alastairc, you wanted to comment on evaluating conformance models

<Jennie_Delisi> AlastairC: Going back to Shadi and Wilco's comment on how these are going to be used

<Jennie_Delisi> ...My interpretation: we have several proposed models

<Jennie_Delisi> ...We have created them.

<Jennie_Delisi> ...We can use them within the conformance models

<Jennie_Delisi> ...Once we work out how that would work

<Jennie_Delisi> ...We would use these principals for how each performs

<Jennie_Delisi> ...If you have a test page, you review for each set of outcomes

<Jennie_Delisi> ...How complex is it? Others. These are all things we have to balance?

<Jennie_Delisi> ...It is: do we have the information we need to assess?

<Jennie_Delisi> Rachael: We have been collecting information from previous conformance model discussions.

<Jennie_Delisi> ...We don't want to try again to have the conversation going over different topics

<Jennie_Delisi> ...We want to look at all the ideas that have come, with data on each model

<Jennie_Delisi> ...And review the trade-offs, pros and cons, of each model

<Jennie_Delisi> ...This is information brought together by a subgroup

<Jennie_Delisi> ...We want to evaluate which data we should bring back to this group to have the conversation

<Jennie_Delisi> Wendy: On the equity point - it is not about counting

<Jennie_Delisi> ...It is not about making sure we have 10 for each, it is about equity of experience.

<Jennie_Delisi> ...Also, the difficulty of testing a specific criteria should not be the deciding factor in isolation

<Jennie_Delisi> ...If it creates equity for an important group, that is why those things need to be measured against each other.

<Jennie_Delisi> Wilco: What I understood - why we are looking at separate models? We have a requirements document within WCAG

<Jennie_Delisi> ...If there are use cases or user stories we need to look at more specifically, that may be the way to review rather than these principles.

<Jennie_Delisi> ...4 of these use the word scoring - a conformance doesn't need a scoring model.

<Jennie_Delisi> ...This doesn't make sense on the face of it.

<Jennie_Delisi> ...User needs model may be a better way to proceed.

<Jennie_Delisi> JulieRawe: I want to echo what Wendy and Wilco are saying

<Jennie_Delisi> ...What is concerning when reviewing this criteria, especially if the group is moving away from equity

<Jennie_Delisi> ...as a criteria, and if reliability means anything involving user testing goes to assertion

<Jennie_Delisi> ...How will WCAG 3 be different than WCAG 2?

<Jennie_Delisi> ...Is it adding assertions?

<Jennie_Delisi> ...What is the big picture difference?

<mbgower> I don't want to move away from equity. I just want to define it better

<Jennie_Delisi> Gregg: I think that we should change the title: to considerations

<Jennie_Delisi> ...We don't know what the conformance model looks like

<Jennie_Delisi> ...These are things we should keep in mind

<Jennie_Delisi> ...We are still trying to figure out what a conformance model should look like

<Jennie_Delisi> ...If we made it considerations, then we can put equity squarely back in there

<Jennie_Delisi> ...It is something we want to keep in mind

<Jennie_Delisi> ...I much rather see us working on the conformance model

<Jennie_Delisi> AlastairC: We are coming back to the conformance model

<Jennie_Delisi> ...We have had discussions about it before

<Jennie_Delisi> ...On what basis are we considering the conformance model

<Jennie_Delisi> ...None of them take supremacy

<Jennie_Delisi> Gregg: Maybe we split it into 2 parts

<Jennie_Delisi> ...Requirements - we know they are true

<Jennie_Delisi> ...Design requirements - less concisely measureable, but we need to keep in mind

<Jennie_Delisi> Rachael: These are not meant to be requirements

<jspellman> +1 to Considerstion

<Jennie_Delisi> ...I think considerations is a fine word

<Jennie_Delisi> ...It is to frame the conformance models

<Jennie_Delisi> ...Happy to change the title

<Jennie_Delisi> ...They were never meant to be requirements

<Jennie_Delisi> ...To Julie's question

<Jennie_Delisi> ...We don't have a conformance model yet

<Jennie_Delisi> ...The whole point of this is to be a short conversation about what are the considerations to have around a conformance model

<Jennie_Delisi> ...To bring to the group what is needed to have a conformance model conversation

<mbgower> +1 to using "evaluation considerations"

<Jennie_Delisi> ...What makes them good? What trade-off spaces are we trying to have?

<Jennie_Delisi> Shadi: I don't have a particular opinion.

<Jennie_Delisi> ...I feel I need to clarify: this group working on this symposium - we were not thinking about what conformance models need to accomplish

<Jennie_Delisi> ...or what are the criteria

<Jennie_Delisi> ...We were looking at a scoring metric

<Jennie_Delisi> ...If you say a score of "42" is what a web page receives, how difficult is it to calculate, do multiple people come up with the same score?

<Jennie_Delisi> ...It may or may not include a score

<Jennie_Delisi> ...You can have a score separate from a technical standard

<jspellman> +1 to Shadi — but the research behind the symposium was highly applicable.

<Jennie_Delisi> ...I think we are trying to use criterion developed for something else

<Jennie_Delisi> ...On the discussion on equity: my understanding is that it is part of validity

<Jennie_Delisi> ...If the score does not consider equity, then it is a less valid score than a score that better reflects the reality of different people with disabilities

<Jennie_Delisi> AlastairC: Jeanne's comment: but the research behind the symposium was highly applicable

<Jennie_Delisi> ...This is generally applicable to how you look at a conformance model

<Jennie_Delisi> ...They should correlate.

<Jennie_Delisi> Gregg: There are a couple of requirements we have

<Jennie_Delisi> ...That it be acceptable to the W3C - meets their guidelines and requirements

<Jennie_Delisi> ...And, be adoptable by regulatory agencies

<Jennie_Delisi> ...Usable, or in some fashion, otherwise it will languish on the side.

<Jennie_Delisi> ...We should keep those in mind

<Jennie_Delisi> AlastairC: These aren't requirements.

<Zakim> bruce_bailey, you wanted to suggest substitute "rating" for "scoring" -- but that seems inconsistent with how Shadi just described draft Note

<Jennie_Delisi> Bruce: For these 5 if I change scoring to rating, it looks like it works pretty well

<Jennie_Delisi> ...I concur with Alastair

<Rachael> +1 to adaptable

<Jennie_Delisi> Dan: I feel the discussion has been focused on how is a given evaluation of a given product going to fit into a score or rating

<Jennie_Delisi> ...But what is important to what is the definition of the conformance model

<Jennie_Delisi> ...How is the document we are writing going to be applied in regulation?

<Jennie_Delisi> ...What is the bar people are going to be required to hit?

<Jennie_Delisi> ...Controlling consistency of that bar.

<Jennie_Delisi> ...There is a meaningful difference

<Jennie_Delisi> ...Suggesting what the minimum score should be

<Jennie_Delisi> ...We should prefer conformance models that have consistency across different applications of the document - I don't see that represented

<Jennie_Delisi> AlastairC: The difference from the reliability one?

<jaunita_george> +1

<Jennie_Delisi> Dan: We have been discussing reliability in terms of - is a given product going to be assessed with the same level by different testers

<Jennie_Delisi> ...I'm talking about: are different countries going to set the same minimums

<Jennie_Delisi> ...We should be preferring models that will consider consistent adoption of the document

<Zakim> alastairc, you wanted to discuss a poll

<Jennie_Delisi> * scribe change?

<alastairc> Poll: (a) Add "Equity" as a criteria for conformance models (needs work), or (b) include an equity review of the conformance models separately.

<kevin> -1 to setting minimums for regulators

<shadi> b -- it's part of "validity" already

<AWK> b

<Azlan> b

<mike_beganyi> b

<dan_bjorge> both > a > b

<ShawnT> b

<Graham> a

<Rachael> a assuming it is the review results we'd bring back to the group

<bruce_bailey> b

<laura> both

<Frankie> b

<Wilco> c?

<giacomo-petri> b

<Makoto> b

<Detlev> b

<jspellman> a then b

<jtoles> b

<Ben_Tillyer> both

<kevin> b

<jon_avila> b

<kirkwood> b

<GreggVan> b

<Poornima> b

<jaunita_george> b

<Jennie_Delisi> AlastairC: looks like we are leaning towards b

<tburtin> both

<Jennie_Delisi> ...almost a middle option: as part of validity it shouldn't score highly if it is an inequitable result

<bruce_bailey> +1 to dan_bjorge , very interesting to think about conformance model resulting in a score but not minimums or rating tiers

<Jennie_Delisi> ...But in general, I think Jeanne's proposal of having a separate review by equity experts - I think that would be a useful thing to do

<Jennie_Delisi> Wilco: For me, this conversation - I don't see how this fits into the conformance model

<Jennie_Delisi> ...That's where you score the criteria

<Jennie_Delisi> ...Feels like a whole different conversation

<Jennie_Delisi> ...I think the focus on use cases - what are the problems we are trying to address

<Jennie_Delisi> ...makes more sense that this

<Jennie_Delisi> ...What do people need? What problems do we need to solve?

<Jennie_Delisi> ...I don't know that this can be captured in this way

<Jennie_Delisi> Gregg: 1: having it reviewed separately, and reviewed by a separate group, are 2 different topics

<Jennie_Delisi> ...I think sending the equity review off to a separate group is a bad idea.

<Jennie_Delisi> ...Unless you take people from all aspects into that separate group, it will just loop again

<Jennie_Delisi> ...Echoing Wilco: the conformance model should be agnostic to disability in the conformance model

<Jennie_Delisi> ...It is how you conform to the standard

<GN015> Somehow I dis not understand whether equity is discussed to form a point in exploring and comparing conformance models, or whether it is discussed to become a part of the conformance model, that is, a point to evaluate the page (or set of pages).

<Jennie_Delisi> ...Then we try to figure out how to create the requirements and assertions - it is in that process that we want to provide equity

<Jennie_Delisi> ...Nothing will ever be exact

<Jennie_Delisi> ...It is not the number, but we know it is a general thing we want to do

<Jennie_Delisi> ...Fairness, or something

<Jennie_Delisi> ...What is equity?

<Jennie_Delisi> AlastairC: We will take this one away

<Jennie_Delisi> ...We will bring this back if necessary

<Jennie_Delisi> * scribe change?

Publication approach to WCAG3 https://github.com/w3c/wcag3/discussions/33#discussioncomment-8110163

<Graham> are we not going to talk about comprehension?

<laura> Scribe List: https://www.w3.org/WAI/GL/wiki/Scribe_List

<mbgower> scribe: mbgower

<Jennie_Delisi> * Thank you Mike!

Alastair: We've tried to define what each of the options would be
... We had some outstanding questions
... I thought it would be good to give 10 minutes to each options, and see how we would resolve each of these outstanding questions.
... These are things we've been trying to incorporate into WCAG 3. If we were looking at this option, how would we solve?

GreggVan: Is that saying we're going to have a new recommendation every year, or a new draft?
... If we're talking about a new draft for public review, I think we can do this. But it takes a year to do a Rec.

Alastair: Proponents can correct me, but I think the idea is you time box it so that you change a certain amount of material and say 'this is the 2024 release"
... From a regulator's point of view, you would tag a certain one as a potential pickup for US508.

<Zakim> alastairc, you wanted to comment on living docs approach

Alastair: I think the starting point is tactically important.

GreggVan: There is a monumental difference between a technical standard and a regulatory standard. The technical is voluntary. The regulatory is a long process, and it takes more than a year. Public comments...
... Maybe Bruce can speak to this in the US, but I can speak to it in Europe and other places.
... Once that's adopted, all your releases for the next 5 years are going to be ignored.

alastairc: Just for the sake of argument, if we release drafts every year but did a publciation every five years, would that solve that for you?

GreggVan: Yes, and you could recommend that anyone doing it read the new drafts. So you could push them towards the new material.

<shadi> +1 to separating yearly iteration cycles from W3C publication periods

alastairc: So my question to the proponents of this option is, would that work?

wendyreid: There are so many ways to approach this. They still require horizontal reviews.
... I think drafts are the best way. You could even push it to CR, but regardless of what it looks like on the outside, spending 3 months of the year on publications is a commitment.
... Even from CR to REC seems like unnecessary overhead doing that every year.

<shadi> +1 to wendyreid

<dan_bjorge> +1 on focusing on incremental vs from-scratch and not the specific update cadence

Wilco: I don't think we need to get super hung up on frequency. The main point is to start with what we have is do updates, and time-boxed updates.
... When there are frequent releases, there's less work to do.
... As opposed to a full REC process. But that's besides the point. The idea is to take a time boxed approach.
... We might not have a set for regulators to 4-6 years.

<Zakim> Rachael, you wanted to ask if this is an alternate to our maturity process then?

<bruce_bailey> I can affirm that historically U.S. Access Board has only cited finalized version -- but industry consensus standards vary as to publishing cadence and what they name their finalized versions

Wilco: And technology is changing pretty rapidly.

Rachael: Chair hat off, acknowledging what Wilco said. Starting point and time box.
... A year is longer than a quartlery, that we're working to now.

Wilco: I think that if something isn't mature, regulators aren't going to pick it up.
... I don't think we need to have regulators to pick everything up.
... Most in the US are still on 2.0
... EN has been updating every 2 or 3 years.
... If we try to rewrite everything and we say 'we arent' going to get to something regulators could use for many years'...

Graham: Bear in mind I'm still new. We seem very focused on conformance.
... I've always treated them as guidelines. Not laws.
... The 2.2 group is focusing on maintenance. The 3 group allows complete freedom to reimagine things so we can get it better.
... I just want to understand where we're coming from from the 2.2 perspective.
... Shouldn't we focus on what 2.2 doesn't cover? Focus on the guidance side over the conformance side?

alastairc: I don't think you've missed much, other than you are describing option 2.

Rachael: to give history, we spent the fall focusing on the guidance. Recently we've pivoted to the conformance.
... This is asynchronous work on conformance before we keep working on the guidance.

GreggVan: I think the idea was to take the same scope covered by 2.2 and do it in 3. If we're only going to cover things not covered in 2.2, then we're going to have 2 things to meet.
... So I think you want to cover the same content, but not in the same way.

alastairc: I think anyone doing the card sort will discover we've expanded scope.
... At least in theory, what we've covered in the card sort is pretty wide. I'm not sure it would all get in.

dan_bjorge: Are we saying folks will have to meet to different requirements for some time? I think the answer is yes.
... The outstanding question: how do we replace key underlying assumptions is key.
... I think people adopting are going to have to think about that. If we're doing incremental changes, I think that's how people are in practice going to use these guidelines.

<Zakim> alastairc, you wanted to move to option 2

dan_bjorge: I think we should set expecations about how to adopt these incremental changes.

alastairc: We've moving on to the second option.
... How do we publish things more rapidly than 4-5 year timeline, which is in our project plan?
... One way is a modular approach.
... We'd define modules: outcomes, assertions, conformance. We publish as they are ready.
... The first outcome, that achieves parity with 2.2, could take a few years.

[ reads out the bulleted questions]

alastairc: Can people come up with solutions to those questions, regardless of which they prefer?

GreggVan: The first option says 'starts with 2.2'. I think...

alastairc: Option 1 was strating with 2.2.

GreggVan: Okay, WCAG 3 requires a separate structure. I don't think you can do that incrementally
... If we look at option 2, it will overlap with 2.2 becuase we want to show how it covers the same concepts.
... I think a power will be to show how it can cover existing and draw in other things that we couldn't get into 2.x
... I think we have a better chance of working alongside it. When it completely covers 2.2 plus the new, 2.2 could be retired.
... I think that's a better approach. In regard to your question, how do you break up with modules. You will need to revise later.
... We'll discover things we thought would work, won't, and will ahve to be revised.

Graham: I love these discussions.
... I'm answering the first bullet.
... Each one can start out as guidance.
... Each one can be its own module and be out in the world early....
... I can't think of one, but I think I saw one.
... If we defer and find out one doesn't work, it means a rewrite. If we go atomic, folks are so excited about the latest stuff that we'll get real world feedback which will hopefully save us a lot of work.
... Starting on 2.2 would be detrimental to what we're trying with 3. Take 1.1.1. It's so hard to define where it ends in the real world. It would be hard to score.
... Being able to ignore 2.2 but steal its good parts would be useful.

<Zakim> alastairc, you wanted to comment on the modules

alastairc: Just as a reminder we're trying to sovle each one separately, not advocate for one or the other.
... As Gregg said, there will need to be revisions. If we are focused on guidance first, we could take a card. Focus on one or a few modules in parallel, and bottom out and start doing them in parallel.
... Just structured guidance, that is both draft, easy to update and easy to revise, and we start working across those modules, in parallel.
... If you start guidance first, it is a lot easier to get things up in draft.

Rachael: you could easily think of Making Content Usable as guidance.

bruce_bailey: I think your first and second bullet can be the same thing.
... You might have a regulator that says 2.2 plus Content Usable.
... If WAI or AG were to come up with guidelines for AI or machine learning, that would probably be pretty popular with the feds. So it could be 2.2 plus the AI module.

Wilco: maybe to ask more questions...
... What concerns me is that WCAG 2.2 is going to go sooner rather than later.
... AI is changing and is going to change really fast in the next couple years.
... I'm not sure we have the luxury to wait until we have a full rewrite for regulators.
... If we don't provide a solution to urgent problems in the next 2 years, regulators are going to spin off.
... I'm concerned enough with no adoption and a split off, than continuing with unified.
... How would we handle this?

alastairc: I'm going to try and combine Bruce and Wilco's point.
... I would be hesitant to do an AI module. I think it would overlap.
... You wouldn't want to do a tech-specific module separately.

<bruce_bailey> i agree that VR module is more feasible than AI module

alastairc: What we could do is pick a module like multimedia guidelines. Sans conformance module, but you're doing 2.2 plus these ones from these modules.
... If you're worried about other people, they're either going to be in the same position -- a big pie to take a bite out of, or they're going to write a few add-ons to the previous things.
... But they're going to have the same challenge of it not working very well in the 2.2 structure.
... we're not making a decision today. just settle on options.

GreggVan: We should think about AI like we think about software.
... you talk about what you want to do.
... 2.2 is much more robust than you think it is.
... programmatically determinable. AI will be able to interpret images of text. You may not have to mark it up in the future.
... AI can recognize a heading as well as a human being can. At that point, you won't have to mark up the headings.
... I'm on the EN committee. There is no way... It is trying to integrate 2.2. It is not trying to tackle this on its own. I wouldn't worry about someone doing their own thing.
... No one is tackling this as intensively as we are.

<wendyreid> https://www.irccloud.com/pastebin/i5J4Z9z7/

wendyreid: I'd love to get other examples of technologies for spin offs.
... The publishing activity in w3c is also working on things that are parallel.

<Detlev> have to leave early...

wendyreid: There are considerations that are unique to epub and reading systems. So we just wrote notes about it. Not just from industry but from EU. They all love guidance.
... Industry or segment specific guidance is valuable.
... Myself and Matt Garish are trying to work back and forth. Industries are needing to write their own documentation.

ACT-Rules Format 1.1 FPWD https://github.com/w3c/wcag/issues/3616

alastairc: One of the hopes in 3 is that that would be possible.
... The ACT task force has requested our review of the FPWD.
... It looked like one of the big changes was a secondary requirement. Is that correct? Can you give an explaner?

Wilco: Yes.
... We did a copule of things that aren't fully completed.
... Briefly, this is a document that outlines how to write rules for testing, such as the WCAG and ARIA authoring.
... We are trying to get a stricter definition for what it means to be consistent with a tool.
... We found that there are rules that are primarily testing more than 2 requirements, because of overlap between requirements.
... WCAG has 2 criteria for text contrast.
... By defining the SC in some ways, it allows someone to report against 1.4.3 and 1.4.6

alastairc: Thank you for posting the diff.
... Does anyone have concerns with this update?
... If you haven't had a chance to look through it, that's okay.
... I will draft a resolution

<alastairc> draft RESOLUTION: Accept the updates to ACT Rules format 1.1

<alastairc> +1

<Rachael> +1

<laura> +1

<bruce_bailey> +1

alastairc: If you haven't had a chance to look through it, you don't have to put anything. There will be 5 days after to respond.

Wilco: I think it's a little bit more complicated than that. This will need to CFC at some point.

alastairc: That's fine. We start with the review at a meeting first.

Wilco: There will be some changes

Rachael: I suggest this resolution be for this version, and we get feedback on any changes.

RESOLUTION: Accept the updates to ACT Rules format 1.1

trackbot, make minutes

<alastairc> https://www.w3.org/WAI/GL/wiki/Upcoming_agendas#Upcoming

Summary of Action Items

Summary of Resolutions

  1. Accept the updates to ACT Rules format 1.1
[End of minutes]

Minutes manually created (not a transcript), formatted by David Booth's scribe.perl version 1.200 (CVS log)
$Date: 2024/01/16 17:56:41 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
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/but industry standards/but industry consensus standards/
Succeeded: s/taht/that/
Succeeded: s/darft/draft/
Default Present: Jennie_Delisi, Laura_Carlson, bruce_bailey, Rachael, dan_bjorge, wendyreid, Ben_Tillyer, jeanne_ec, ShawnT, GreggVan, Graham, Azlan, Wilco, kevin, giacomo-petri, mbgower, kirkwood, julierawe, Carolina_Crespo, JustineP, jtoles, mike_beganyi, jeanne, Bri, AWK, Makoto, Detlev, Frankie, tburtin, ashleyfirth, ljoakley, Poornima, DanielHE, jon_avila, jaunita_george, Francis_Storr, alastairc, GN, scolto, \, maryjom
Present: Jennie_Delisi, Laura_Carlson, bruce_bailey, Rachael, dan_bjorge, wendyreid, Ben_Tillyer, jeanne_ec, ShawnT, GreggVan, Graham, Azlan, Wilco, kevin, giacomo-petri, mbgower, kirkwood, julierawe, Carolina_Crespo, JustineP, jtoles, mike_beganyi, jeanne, Bri, AWK, Makoto, Detlev, Frankie, tburtin, ashleyfirth, ljoakley, Poornima, DanielHE, jon_avila, jaunita_george, Francis_Storr, alastairc, GN, scolto, \, maryjom, GN015
Regrets: JakeA, ToddL, DuffJ, DJ
Found Scribe: Jennie
Found Scribe: mbgower
Inferring ScribeNick: mbgower
Scribes: Jennie, mbgower

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.)


[End of scribe.perl diagnostic output]