Meeting minutes
<maryjom> rain: interesting discussions with different thoughts on how visual disabilities were affected by glass collection - perceived as not blocking. However, went with a lower score.
<maryjom> glenda: When there was disagreement, did you change your personal score but allow the group score to be different?
<maryjom> rain: Some we did change scores and aligned, and some we kept our individual score but agreed on a different combined score.
<maryjom> juanita: When you're a person working alone you might have a different perspective but as a group we benefited from hearing other's score and supporting thoughs. Makes me a better evaluator.
<maryjom> Rachael: Home page, search, and sign-in
<maryjom> Jan Jap: Reads out scores from spreadsheet
<maryjom> Rachael: What problems did you find
<maryjom> Jan Jap: Lots of keyboard navigation issues, carousel had no play, pause hide and the video
<maryjom> patrick: Sometimes quite difficult for specific success criteria to determine with a group of experts whether it is pass/fail
<maryjom> ...depending on how someone judges good/bad granular scores caused greater variation.
<maryjom> ... it was ambitious to give us 6 pages to assess. Qualitative gives more differences in how you're judging the score.
<maryjom> juanita Whose opinion do you believe. One person's opinion is different from another - this creates a moving target.
<maryjom> glenda: we have an issue description library which helps with judging. If we had this type of list, it would help people to judge a web page better.
<maryjom> ... was still using a WCAG perspective on testing. Each issue should have a tie-back to disability
<Zakim> alastairc, you wanted to say this isn't the proposed WCAG 3 methodology!
<maryjom> ben: It was difficult to not think about WCAG2 when assessing. Some things weren't technically covered by WCAG, such as spaces between letters, mixing of fonts.
<maryjom> alistair: This is a gut check to test whether we can recognize problems that may not already be covered in WCAG.
<maryjom> rain: rating scales themselves are difficult for some cognitive disabilities, such as dyslexia.
<tburtin> +1 Rain
<maryjom> rain: To echo Ben, we were instructed to not use WCAG as the rubric, but it is sometimes hard to detach from that.
<maryjom> rachael: This exercise, though hard to do quickly, really did help us to understand what it feels like using a scale to rate websites and we'll take that feeling to tomorrow's exercise.
<maryjom> Glenda: we didn't get to the other pages
<maryjom> Julie: Read out scores for the blog. Some control should be given for video.
<maryjom> ... intersectional - the page had confusing navigation, especially with zooming pictures
<maryjom> ... physical - unexpected extra tabbing
<maryjom> ... cognitive - confusing to understand where you were. Quite a few cognitive things
<maryjom> ... sensory - uncontrollable zooming. Discussion on how bad was the uncontrollable zooming.
<maryjom> ... one error message said "congratulations" was confusing
<maryjom> juanita: For speech disability - didn't know if it was voice control/voice input or for a speech disability
<maryjom> Others indicated they also had that confusion.
<maryjom> rachael: Sports page
<maryjom> lori: This was a small page, not too bad, not the worst.
<maryjom> ... focus disappeared sometimes. Focus was there (most of the time)
<maryjom> ... no auditory content - N/A
<maryjom> ... keyboard functions worked ok, not too bad
<maryjom> ...cognitive - you could get lost sometimes when the focus disappeared
<maryjom> ... screen reader - repetitive speaking was annoying, but information was there so not too bad.
<maryjom> ... speech - N/A didn't find any speech input on the page
<maryjom> ...overall this page was scored a 3
<maryjom> ... There are a couple of sub-pages where you can favorite an image. Color contrast was bad. When you clicked on the heart, there was a pop-up and it wasn't well understood what "undo" or "dismiss" would do.
<maryjom> ... Couldn't get the dialog box to reappear.
<maryjom> Julie The dialog would also disappear quickly.
<maryjom> alistair - Volunteer page was not good. We weren't able to figure out how to judge intersectional - combined score from other disabilities to fill that one.
<maryjom> ...sensory some said best and some said worst - not sure why. didn't get
<maryjom> ... Technology we didn't have enough time on
<maryjom> ... Reads through the scores on that page from the spreadsheet
<maryjom> rachael We'll pick up on this more tomorrow.
<maryjom> ben: Things I didn't consider: 1) If the page was bad for everyone I gave it a score of 1 in my head - because it was bad for everyone. 2) Didn't consider the reason why people would be going to the page - and gave as much time in links in the header and footer as the links to go to the page.
<Zakim> alastairc, you wanted to mention that under a WCAG 2 conformance model, all the overall scores would be 6.
<maryjom> ... Next time would think about the user journey and what would be main path vs. more obscure and judge based on that
<maryjom> alastair: under a WCAG 2 conformance model, all the overall scores would be 6.
<maryjom> tiffany: had difficulty with the categories, and think category definitions would be helpful.
<JenStrickland> Thank you, tburtin! Great insights!
<maryjom> Rachael: Thanks everyone for their work on this and working through the difficulties in making these quick assessments.
<maryjom> ... We'll use what we learned today in our conformance model discussion tomorrow.
<Chuck> order didn't work
<Chuck> We will begin in a few minutes
<Jennie_Delisi> * Just got booted from IRC and re logged in, so probably not the best scribe today
Chuck: Welcome to Tuesday. Let's talk about yesterday's events. Shortly after yesterday's event one of the chairs did test positive for COVID, we discussed approach and decided that in an abundance of caution, the chairs will present remotely
… Room layout is still mostly the same. Everyone else involved has tested negative. Queueing will need to be more predominantly via IRC
<alastairc> Just a reminder that any whispering in the room comes across very clearly to the remote participants, we hear you!
Rachael: Clarifying, the chairs have already spent a couple of days together, which is why they're considering themselves higher-risk and exercising caution
Chuck: Chairs will still come down for snacks/lunch; Rachael and Chuck will not be attending dinner
<Ben_Tillyer> +1
<maryjom> +1. what time is it planned for?
<Glenda> +1
<dj> 0
<kevin> +1
Chuck: call for +1s for who still wishes to attend dinner
<JJ> +1
<MelanieP> +1 for dinner
<dj> i need to figure out transport to the airport afterwards, but assuming i can +1
Chuck: will still go to reserve some space but reduce from original expectation
<JenStrickland> +1 for dinner tonight -- finally got in.
Rachael: Starting with some takeaways from yesterday's conformance exercise
<Rachael> slide deck: https://
<Lisa> (morning only)
Rachael: One thing that we got out of the notes is that accessibility across all aspects differs. Teams differed as to how they scored, e.g. severity of error to most affected group, location of error
… the more subjective the exercise, the more variable evaluation seemed to be
… questions: do percentages matter, how to handle "N/A", focus of core document in WCAG 3
Sheri: I thought the categories weren't granular enough
… questioning use of "Intersectional" - once you start multiplying, the numbers get too large
… need super-clear instructions on what constitutes "worst" / "best"; groups seemed to look at it differently, and it's important to apply the same criteria the same way
Glenda: Reinforcing what Sheri said RE description of best-worst; suggest clearer rating scales
<Lisa> sorry that I missed the afternoon meeting yesterday. but what was the conformance exerersize?
julierawe: Echoing comments: found intersectional the most challenging, felt like I was guessing and could make an argument that any time you noted a concern under any other column (or 2 coilumns) then does it factor into Intersectional?
Kevin: Ratings scale and groupings were really to get people to start thinking about these things, not like we're actually going to go down that route
<Ben_Tillyer> @Lisa - check out https://
JenStrickland: Julie hit on some of what I was going to jump in about - intersectional is great to include to get people to think about it, but including some instruction about it would be helpful, and make sure it notes not only disabilities but other identity characteristics that affect equity of access
… would be helpful to have a column for comments for each row for people to leave notes
tburtin: I really appreciated the Intersectionality because it made me feel included. e.g. I use dark mode, screen readers, font changes, zoom, all at the same time.
Lisa: I wasn't here yesterday afternoon; I feel like intersectionality is necessary to include for it to be meaningful to think about inclusion. e.g. diabetes can cause more than one disability, and often solutions no longer work when dealing with more than one disability at once
<Patrick_H_Lauke> not going to lie, i do worry about an author or auditor being able to account for the huge potential range of factors once we factor in intersectionality
<Jon_avila> Some groups of conflicting needs - something that benefits one group may be problematic for another.
<jeanne> Pat, I think it doesn't become fully practical until we get better authoring tool and browser support
Lisa: It's fine to do things in steps but we haven't necessarily tackled the full space until we tackle intersectionality. Having multiple disabilities e.g. related to aging or diabetes is common
Ben: Although the solutions might not be suitable to people with intersectional disabilities, do you think that there are new problems caused by intersectional disabilities that auditors haven't come across?
Lisa: Yes I do. e.g. in COGA there are a lot of things to do with pronunciation of screen readers, e.g. if numbers are read in a long list, it won't make sense to someone with problems with short-term memory problems, whereas someone who was blind from birth may not have any problem with the same thing
<Jon_avila> One example is being having low vision and being hard of hearing - there are unique needs for display of captions.
Lisa: there's a lot of nuance in where you put commas in between numbers, which can make a big difference. Someone may use text-to-speech from a browser add-on. There are many different things; ideally we should look at all disabilities and also across all combinations. Currently we look at each in subgroups
julierawe: I think it is important to think about these intersections - I just didn't know how, and felt like I was guessing. Felt like I needed more guidance. Hoping that's something we can clarify to help everyone think about it
<kirkwood> +1 to Julie
<kirkwood> +1 to Julie
<kevin> +1 to Sheri
Sheri: As others alluded to, there's definitely some key intersections where we know that things are different, e.g. deaf/blind being common. I think we need to focus on assistive tech use and not diagnoses, since even within a single diagnosis there can be significant differences
<Lisa> +1 to julie
<Jennie_Delisi> +1 to Julie
julierawe: If we just look at AT use, we may be missing folks that don't rely on AT. There can be issues even notwithstanding AT use, e.g. confusing layout
<Jon_avila> Most of the access features I use are not assistive technology - or at least I don't consider them assistive technology.
<shawn> WAI uses "assistive technology and adaptive strategies"
<jeanne3> +1 to Sheri and Julie - avoid the medical model but include people that don't use AT
<Zakim> shawn, you wanted to react to JenStrickland
JenStrickland: Understanding of AT needs to be updated. e.g. I use a mobile browser to access the web in order to limit the amount of stuff coming at me. That can be considered AT. Need to dispel the still-too-prevalent illusion that "if you're not using a screen reader, you're not disabled"
<tburtin> +1 to JenStrickland Choosing a mobile device over a computer to navigate. I do this too.
Shawn: WAI resources use "assistive technologies and adaptive strategies" to cover these other things
<Sheri_B-H> +1 to the WAI definition - that was what I was trying to say, but didn't know it had been defined that way
Lisa: We use diagnoses sometimes to start looking in a direction, but then focus on the end result, e.g. impaired short term memory which could stem from one of multiple diagnoses. Diagnoses are not the way to go; function is. It's not all about AT.
<kevin> +1 to Lisa's comment on considering functional needs
Lisa: Let's say someone has an impaired short-term memory or a focus disability. They may be employing a variety of browser plugins; what's important to website authors is what not to do, in considering functional needs.
<Zakim> JenStrickland, you wanted to bring up symptoms vs diagnosis
<kirkwood> +1 to lisa: we should be concerned with “functional needs”
JenStrickland: I had meant to mention, we talk about symptoms rather than diagnoses because people have a range of symptoms that shape their unique experience. Aligning to symptoms may be more helpful/inclusive.
<kirkwood> +1 to lisa: we should be concerned with “functional needs”
<kirkwood> +1 to lisa: we should be concerned with “functional needs”
<Frankie> +1 to Lisa & functional needs
JenStrickland: I'm very thankful to Tiffany for sharing her experiences. I don't mean to dismiss anyone's identities
Rachael: This has been a really rich conversation that we should discuss in more detail in the future; if it's okay I'd like to move on and parking-lot it
<Jon_avila> Yes, I agree functional needs rather than disability is the way to go as people have different needs in different situations.
<tburtin> I would be willing to be a part of that future conversation.
<kirkwood> can we have link to presentation?
Rachael: Reorienting RE the models from yesterday: Baseline plus % based levels; vs. Prerequisite, Baseline, and % of enhanced
<Ben_Tillyer> @kirkwood https://
<kirkwood> can we have link to the presentation?
<kirkwood> can we have link to the presentation?
<Chuck> Thanks for providing links, I am limited to two screens, and cannot chase down content :-)
<Chuck> One sharing and one IRC. I'm use to four screens to support my chairing.
Rachael: editors group has shifted recently; is now chairs + Jeanne + Francis, thanks to everyone who has been / is / will be an editor
<Rachael> https://
Rachael: want to emphasize that what we have is a first pass, it may be throwaway
<Sheri_B-H> Had to request access to the blank sheet, the other one was OK
Rachael: encourage trying the exercise in this sheet out with a page
… Sheet maps guidelines to outcome statements (multiple each)
… we're not talking about how to divvy these up, that's a process that will take months, we're doing the first exploratory pass
… each outcome has an assigned level of prerequisite, baseline, or enhanced
Rachael: mark each outcome as pass / fail / not applicable
… once we did all of that we ran percentages and numbers against all of that
… grouped 183 outcomes under 52 guidelines; worked out to 30 prereq / 86 baseline / 61 enhanced
<Rachael> slide deck: https://
<JenStrickland> Requested access
<MelanieP> The "blank" spreadsheet is not shared
Kevin: I have a sneaking suspicion that the blank spreadsheet is not in the W3C space, hence the requests for access
Rachael: everyone should have access now
<Lisa> thank u
Rachael: lessons learned:
… Sometimes the prerequisite is the hardest outcome to meet, e.g. adjust color vs. meet color contrast
… Some outcomes have challenges to testing (e.g. AI Editable, how can you tell if it was AI-generated)
… we have task-flow-based items; is reading/browsing the content a "task" or should a task be defined more narrowly e.g. requiring interaction?
… decision was much easier to use than flat list from testing standpoint
<Lisa> task can be finding information like the email adress or address
<Glenda> What do you mean by, “We need to be clear which outcomes are cross-view?”
Chuck: Regarding cross-view, conversation came up yesterday regarding making determinations across multiple web pages
Alastair: The reason the decision tree is useful is you don't need to cognitively understand as many definitions vs. a flat list, helps focus the testing exercise
<Glenda> +1 to what Lisa is saying about “finding info” is a task. Which may fall into “browsing”.
Lisa: in COGA we've been considering e.g. finding pieces of information that someone might come to the site to look for as "tasks" (i.e. a user's objective in accessing the site)
Alastair: I understand what you mean, Lisa; when going through some of the outcomes that we've got that are related to tasks, there is sort of an underlying assumption that they are more complicated. Maybe "task" isn't the best word, but we need some kind of definition that separates e.g. browsing/navigation from following a particular process
<kirkwood> interaction ?
Lisa: Maybe "process" would be a good word for it? Going back to "task", navigating can be very relevant, e.g. can you get back to the home page, do people get lost? May be underestimating how much browsing/navigating _is_ part of the process
Rachael: "task flow" was the verbage we originally agreed on; can discuss terminology more in depth in future
… reviewing what breakout groups came up with for different surfaces on the Museum of Broken Things yesterday, vs. what the editors came up with against Outcomes
… worth noting that we had 19 "I don't know"s for the Technology page
… as well as 73 Not Applicable for Technology page
<Chuck> I will process queue once Rachel finishes reviewing the slide
throwing a direct link to the current slide in case anyone needs it (or in minutes) https://
Sheri: We didn't have a good way of representing a single failure vs. things being terrible across the board, e.g. one missing heading level vs. problems spread throughout
<bruce_bailey> +1 to sheri that missing H1 very different than no headings (on a page that needs it)
<MJ> Link to the test page? I missed yesterday's session.
<Patrick_H_Lauke> "you can't use it" depends on who the "you" is though, which is the tricky part...
Rachael: Technology page did not pass level 1 on either model. Want to walk through this one example in detail to discuss the percentages
<JenStrickland> List of Breaks in Museum Website: https://
Rachael: percentages are given both for excluding and including N/A - as in, "including" counts N/A as passing
<Sheri_B-H> with a motion sensitivity warning for the home page of the test code
<Jon_avila> Regarding not applicable - I think people should be rewarded for not doing certain things like not using flashing content or motion. If we remove those then other factors have a higher weight.
<Lisa> is there a link to the actual page
<Lisa> the websites being evaluated
<JenStrickland> Be forewarned that the links lead to dangerous pages: https://
<JenStrickland> The websites are listed on slide 21.
<Zakim> Chuck, you wanted to react to DJ to ask for a scribe change
<Lisa> thanks
alastairc: re DJ: (playing devil's advocate) maybe it should've passed baseline because we didn't give it a 6?
… how likely is it somebody will be able to get through?
… okay in some areas, not in others
… so is a flat fail good representation in aggregate?
… maybe baseline instead of prerequisite
<Patrick_H_Lauke> agree with alastc. a flat fail for the site seemed...a bit harsh. as i alluded to before, saying "you can't use it" is a sweeping statement that does not necessarily reflect reality for all users and which tasks they tried to perform
Glenda: i struggle with what to do with NA as well
… for internal products, i'm harsh
… +2 points for anything that says "supports" or "NA", "partial supports" is 0, "does not support" is -2
… overall score isn't useful because of so many different disability types
<bruce_bailey> +1 to Glenda for partial support being zero
Glenda: may be a fail for screen readers, but not a fail for Deaf/HH for example
… individual scores are very desirable
julierawe: it seems like the two models have in common is that level 1 is failing the most basic level
… and then the main difference is that one starts with baseline and the other starts with prereq?
<tburtin> +1 Julierawe
<Lisa> +1 to jenny
<julierawe> +1 to Jennie's question
<Zakim> DJ, you wanted to include N/A - UDL and to "6 was death"
Jennie_Delisi: 'give credit for not using multimedia' - wouldn't that decentivise multimedia (which would decrease accessibility)
<Ben_Tillyer> +1 to Jenny
<Jon_avila> I agree we need some level of scoring or understanding the impact other than pass or fail at the site level. Most sites are not 100% conformant on every page yet can vary in accessibility.
<kirkwood> +1 to Jenny
<Zakim> Rachael, you wanted to answer julie
dj: I think it is important to include not applicable as this might be needed for accessibility
Rachael: what multimedia is presenting is the tradeoff space
… if we exclude NA from score, ... we need to have that conversation
… because whether we give credit changes what the percentage looks like, and that's important
<Zakim> bruce_bailey, you wanted to argue for test site maybe meeting prerequisite
<Zakim> Rachael, you wanted to react to DJ
bruce_bailey: i keep thinking of prereq level as non-interference (from 2.2) plus a few other things
<Jon_avila> The differences is that when you don't include not applicable then the other failures have a higher impact on the score as there are less criteria used to calculate the score.
<ChrisLoiselle> +1 to Bruce. Are you expecting small businesses to use these type of models compared to larger business that may or may not have a team or may or may not use a 3rd party auditing firm that could somewhat automate or report against this? Must , should, could vs. baseline, enhanced , pre req , how would an end user outside of accessibility try to implement all this?
<Rachael> Our definitions of the outcome levels are on slide 15
<Ben_Tillyer> Definition to WCAG2x non-interference https://
shadi: my concern is that the numbers we were showing in these models do not actually reflect the real severity
<alastairc> Shadi - that's a good point on how a conformance statement would be scoped. These figures are aggregate for a particular 'site' (section really in this test case).
shadi: for example, if just one heading doesn't pass, then the whole outcome is nill, so sites with just one erroneous heading are marked just as bad as completely erroneous sites
… i think there could be an aggregated number that reflects the user expirence
Sheri_B-H: i have been ordered to remove videos because they made the vpat look worse
… getting credit for NA is just wrong -- people currently do game the system because of that
<JJ> +1 to Sheri's points
Lisa: +1 to Sheri
… real problem when this type of scoring removes accessibility from the site to increase conformance
<bruce_bailey> +1 to sheri for not giving points to N/A
Lisa: there is a way to do multimedia that enhances the existing site
… that shouldn't reduce your accessibility
… to shadi's point - the answer is user testing
… another thing is heatmaps based on access frequency
<bruce_bailey> i am in a different environment, but i have NO concern about a .gov site taking down videos because they don't want to find resources for captioning.
<Zakim> alastairc, you wanted to comment on scoring and avoiding things to get a better score. and to also say that including NAs means simple sites can't pass.
alastairc: [slide 32]
… i don't think the percentages are going to matter for baseline/prerequisite levels
… if we don't look at percentages for baseline as well as enhanced then this is more applicable
<Patrick_H_Lauke> +1 to alastc this was part of the comment i wanted to make
alastairc: also, including NA in score could make a simple site not pass because they don't have forms, for example
<Zakim> DJ, you wanted to react to alastairc
<kevin> dj: When we say include is that in the scoring or as passing?
alastairc: i think it's resonable to say you need to include multimedia, for example
<Rachael> Slide 37 may reduce confusion
alastairc: include in percentages
<Zakim> DJ, you wanted to half %?
<kevin> dj: A few people have mentioned it not accurately reflecting the actual accessibility. User testing is one approach but half points might also be a solution
<kevin> ... I definitely think we should be encouraging inclusion of multi media as it helps accessibility
Chuck: point of order:
<kevin> ... but also forms, encouraging interaction is not a bad thing
<alastairc> But why would a site that doesn't ask any information of users need to add forms in order to conform with WCAG?
Chuck: long queue on this topic; we're supposed to take a brake right now
… let's continue after
Rachael: [slide 37]
<Glenda> Consider adding a category for “A11Y Penalty” - to allow for situations where someone is gaming the system…and we can add example like: removed video instead of captioning it.
<Patrick_H_Lauke> happy to snack
<Jennie_Delisi> * concerns about NA as passes - seems to still reward not including something, unless I am misunderstanding the edits
<Ben_Tillyer> @Glenda I can't see companies admitting that, if the audit is done on an internal build
Chuck: break till quarter of
<bruce_bailey> +1 to alastairc comment (quarter past hour) that "... i don't think the percentages are going to matter for baseline/prerequisite levels"
<julierawe> Rachael Slide 36 says "Fails to meet baseline" several times. Does this mean "Fails to meet WCAG 2 baseline"?
<julierawe> Rachael Are all the phrases that appear on the same line as "Level X:" comparisons to WCAG 2? I'm still very confused
<kenneth> julierawe, I expect that means the baseline alluded to in "Baseline plus % based levels"
<kenneth> which...I think would be equivalent to things marked as "prerequisite" in the spreadsheet? assuming that baseline in model 1 = prerequisite in model 2
<JenStrickland> That could set the stage to open a set of conformance standards if X exists. The ultimate score output would be the same, but if X exists then an additional bank of possible points is considered into the total. In Baseline a total is 100 for 100%. In Enhanced the total might be 125 for 100%.
<alastairc> For those not in the hotel, we've had a power cut, not sure when it will be back
<JenStrickland> I'm in the hotel, but using a personal hotspot.
<julierawe> Hi, folks, while we're waiting for the official meeting to resume, I'm sharing a comment I put in the Github thread about Pull Request to publish first drafts of two WCAG 3 outcomes: w3c/
<julierawe> I huddled with some COGA members and we had some concerns about the template for the WCAG 3 outcomes, including red bullets and other formatting. Please take a look at our comment when you have a chance, thanks: w3c/
<jeanne2> w3c/
<Rachael> We are settling beside the kids slide. There are chairs and lights.
<Rachael> We also have hotspots so if everybody can find it their way here on the fifth floor we could call-in to the zoom meeting
<julierawe> WCAG 3 preview link for those two drafts: https://
<Lisa> lot of people on the zoom call. can you call in?
<alastairc> We are setting up outside, will call in asap
<Rachael> We have gotten together and will be onlineish in a moment
<Zakim> Rachael, you wanted to react to videos
<Zakim> Chuck, you wanted to say do we include "not applicable" in the denominator
<alastairc> do we have a scribe?
<jj> We (Shadi, Jaunita, Ben, JJ) are back in the room on 5G
chuck: supposed you have 5 multimedia prereqs and 5 non multimedia. You have no intent to have multimedia
Have you met 5 of 10 or 5 or 5?
<alastairc> jj - can you get the zoom back on?
Should there be the action that you are only scoring 50% or do we remove them so you are getting 100% of prerequisits?
… that is how Chuck sees it.
We are passing Jen and Ben as they have not joined us.
Julie: When we say fails o meet the baseline for WCAG 2 or WCAG 3?
<kevin> Rachael: at the top of the slide, it is baseline plus prerequisite.
<kevin> julie: So Level 1 is failing
<kevin> .. You could be doing nothing and failing everything and meeting Level 1?
<alastairc> If you fail anything, you fail level 1
<kevin> rachael: No you would be Level 0
<ChrisLoiselle> there should be a level 0 , achieving a level 1 reads as you've done something positive to achieve a baseline or plus 1. I think that is what is being raised in the discussion.
<kenneth> slide 16 might help RE the question, that's what defines the levels https://
<Patrick_H_Lauke> is power back in meeting room?
<Zakim> DJ, you wanted to react to julierawe to point of order
Rachael: We wrote down percentages for all levels on slide 37 for Technology but maybe that's confusing - if anything fails baseline in the first model, you fail level 1
<Jennie_Delisi> With or without prerequisites
Rachael: right now the names are...faulty, so we have Baseline + % based levels, and Prerequisite, Baseline + % of enhanced.
Proposed names: baseline plus, prerequisite plus
glendaL reminded of something in AIR. Ability to adjust score. judges penalty.
Ben: conformance scores. Not the time the audit is done (may be once a year or once a month). Maybe we should also take into account remediation plans.
… a captionless video that a site never fixes vs. a captionless video that the captions will be prsent next week.
… in both cases the person doesn't have captions but the impact is less.
Kevin: That may be us stepping too far into regulator space
… not of the standard.
<Zakim> kevin, you wanted to react to DJ
Kevin: could be an assertion.
<GreggVan> +1
Ben: OUtside the conformance model conversation but, if somoene is doing an evaluation against WCAG 3, how small can the scope be? Can it be just one video?
Will those scores be valid over time?
Kevin: Scoping levels are there but you could conceivable scope at item.
<Patrick_H_Lauke> i'm downstairs at the moment, heading back up. my point for the queue was roughly echoing what alastc said earlier: the discussion on not giving bonus points for NOT doing things like having video, and how that is then worse for certain users. isn't the solution to that then to have requirements that clearly require use of, say, video. rather than blanket penalising sites for not using video/having it n/a, because that woul
<Patrick_H_Lauke> d then ALSO just be gamed ("need to add a video so i get the full score...")
Sheri: 3 things. First DJ made a comment about making videos required. An alternate thought would be that W3c is not prescriptive. We don't require pause/stop/hide button rather we require a pause/stop/hide mechanisms. Instead of requiring videos, require multiple presentations.
<Patrick_H_Lauke> +1
<Patrick_H_Lauke> coming up now
Sheri: 2. I feel very strongly that you need that we need a way to encourage people doing assessments to do more. Combining prerequisite and baseline reduces htat. You can meet prerequisite and celebrate that. Maybe extend the number of levels to 5.
<Zakim> JenStrickland, you wanted to say if video provides alternative content for content on screen use case, vs video only and to ask, do we ask if X (forms / multimedia) exists first? then, if it exists, is it accessible? and to ask: An Enhanced scenario is if video provides alternative content for content on screen. Where as if there is video only and there isn't text on the screen that is baseline (also the reverse of if text only and no video). Do we in the conformance flow, first ask if X (forms / multimedia) exists first? Then, if it exists, is it accessible?
3. Not applicable needs to be removed becuase otherwise people can never meet a level.
<Zakim> DJ, you wanted to you should put multimedia on the site (re chuck)
DJ: I think we are converging on shared idea. If we want ot require the use of multimedia, whihc I think we should. We shouldn't require use of multimeda but rather include that in teh guidelines. I like what Sheri said about having the multiple levels. Have prerequisite, then baseline and so on, maybe % based levels to encourage people to get to the next levels.
Lisa: People have intentionally broken things in order to fix them and show improvement.
Lisa: I think we have to be careful about taking steps that do you have a roadmap and how much have you improved. I've had teams and they purposefully broken features and then they get audit and show improvement to get their release. These kind of points for efforts can turn against you.
… there was something else about breaking accessibility for cognitive. About not causing people harm that focuses in on mental health. We've seen studies where people's anxiety and self confidence has deteriorated because they feel then can't complete the form becuase its more difficult than it needs to be. When you say content shouldn't harm, seizures are obvious but issues in the cognitive space can cause very real harm.
<tburtin> +1 to all that Lisa is saying. Constantly an issue.
Lisa: dark patterns, I have to ask someone else to help me. Even getting repeatedly locked out - saying they don't do harm is not true. Harms feelings of self worth.
… what do you mean when you say the lowest.
Shadi: For the baseline, if you miss one thing you miss the entire baseline. What then is different to the WCAG 2 model? Wasn't that one of the core things we were trying to address? What is the differentiation. Bruce was asking if the baseline was only non-interference or more than that? that brings us into the WCAG 2 model. Not what we were attempting to do.
<Zakim> Jennie_Delisi, you wanted to discuss use case
Jennie_Delisi: This may seem humorous because I brought it up, but I don't htink we should require multimedia. We need to put in checks nad balances while reviewing to accoun tfor different use cases.
<JenStrickland> Rather than "dark patterns," "deceptive patterns" for inclusivity.
Chuck: Snacks are here. Help yourself if her.
… we are now doing poolside AGWG.
GreggVan: 3 things. We need to remind ourselves that we are the ruler, not the rule. WE are trying to determine if something is accessible or not.
… How much time someone should have and related ideas is up in policy. If a site has videos and all but 3 are accessible, then all are accessible except for 3. If someone asks if the videos are accessible the answer is no.
… Similarly, if we ask if hte site is accessible and 1 video is inaccessible then the site is inaccessible. We don't need to get into that. We just need to figure out what causes something to be accessible or not. We shouldn't get wrapped around whether something is 98% accessible. The parts that are, are. The parts that aren't, aren't.
… If you have 0 vidoes, no videos fail. Think about it as "Do any videos on the site fail?" No, then it passes and you move on.
Sheri: It helped me crystalize my thoughts. I think that we've tried to assign things into prerequisites and baseline without adequately defining harm. For example, Count down clocks. They are difficult for certain disabilities. Do they count as harm? We need to rethink that.
Kevin: THInking about something gregg said about %. Our current models are looking at % of passes. That causes problems. Let's look at % of failures. It's the failures that matter.
It's not a good idea for me to shout, but regarding kevin's point: That's just a reversal of the percentage, it doesn't change the inclusion (or not) of NAs.
glenda: How old is that issue? I think by adding hte length of time the issue existed, and andrew suggested dynamic vpats off the JIRA issue queue.
… I went to a VPAT model but I like the idea.
acl ljc
Lack ljc
<kirkwood> +1 to Lori’s example
ljoakley: I have a site where the workflow that all the work goes through and it is inaccessible. Doesnt matter if the rest of the site is accessible. It misses the mark. The people who are doing the work can't get their work done.
<Lisa> that is why task based user test is best!!!!
<Jennie_Delisi> * have to drop off for the day. Wishing everyone on site light, health, and appropriate amounts of power!
<Lisa> if a wide range of users with a range of AT and functional needs can complete the main tasks the site is good
<Lisa> +1 to Jennies good wishes!
Ben: Two points. First, the % scoring. If you have two scopes, one for tasks and one for web pages might give you a way around that. Can all people get through teh main purpose of the page vs. how accessible is the site?
<Zakim> DJ, you wanted to % after baseline/prereq
DJ: The % would be after baseline or prerequisite so you won't get to the % until after you pass that.
Lisa - I agree that usability testing is good, but we can't use that for everything. It doesn't help people in the early design / dev stage do a quick evaluation against some guidelines.
Ben_Tillyer: In response to Glenda's point, having worked in huge orgs I have seen where scrums have closed tickets and reopened them to get around items. If we used tickets, we'd have to have requirements around how to use the ticketing system.
<Zakim> Ben_Tillyer, you wanted to talk to the thought I forgot about
Rachael: Question: we're looking at a model where we have bottom levels where everything is required, then the top levels have percentages, and the details are in how many levels are required vs. percentages. Does that move enough away from the WCAG 2 model to address requirements for WCAG 3?
Chuck: I think it does get us away from the WCAG 2 model enough. It does have a set that has latitude. I think there the current proposal does that.
Shadi: I think it depends on what is in the baseline. We don't have a good ruler yet.
… to measure how much friction there is vs. how much is compounded. The keyboard trap is a complete blocker. If everything is equally important to everyone, we are back to WCAG 2.
JJ: It seems like an improvement to have the baseline and percentages. Without that, one issue can prevent you from continuing.
… I am worried about how the baseline is accessible. If the base level is the same, then there is progress to be made there.
<Zakim> DJ, you wanted to prevent harm
Jaunita: I think there is a benefit. Things that are more subjective are part of the % partof the conformance model. I am concerned that everything subjetive ends up there. Because they are more difficult to assess, risk they end up in the %
DJ: We had a prevent harm subgroup and we had a large list of items that fell into prevent harm. Anything that can harm someone should be in base level. Also anything that prevents access.
Charles: I am skeptical that the model will help us. We will spend all of our time wrapped around the axle of whether something is in or out of baseline. We spend a number of years getting there. Rather, map the things you can't do or cause harm. Build a large collection of those pathways first. these people cannot solve this problem.
… these people will actively be harmed.
<tink> Update from the Team - the power cut was caused by an accident outside the hotel, and power should be restored in the next 15 mins.
Charles: looking at the way you report that. Its a more detailed view of what's in there. You don't get a pass result. It's more complex but we can have more effective discussions.
Rachael: think a lot of the discussion will be spent on what belongs in the baseline vs. not
Leonie: Agree with chals and Shadi. My question is if I understand there is a baseline required...
To Charles - We do have a rough definition of pre-requisite / baseline / enhanced, which should cut down the arguments there
Leonie: and then we get percentges. Why not enable people to track completion in the baseline too. We want to encourage people to do everything, not get stuck in working on baseline and failing all the stuff on top.
<Lisa> I am going to sleep. Goodnight all
Noting there is a big difference between percentages of outcomes, and percentages of *instances*. So far we've been talking about the former.
could everyone thinking of going to cheescake factory tonight please join #ag-dinner
Rachael: We had a good conversation online and offline despite challenges
… the chairs touched base and heard two themes
… one concept of the ruler vs applying it
… the other of separating the guidance and conformance sections
… release them separately and in different ways
… better define levels
… explore perceptions of harm
… what prerequisite really should mean
… the beauty of TPAC is getting diverse perspectives
… is there anything else people want to discusss?
… is everyone comfortable exploring those two topics
<Chuck> +1
<DJ> +1
<tiffanyburtin> +1
<wendyreid> +1
<GreggVan> +1
<Sheri_B-H> +1
<Rain> +1
<kirkwood> +1
<giacomo-petri> +1
<JenStrickland> +1
<Makoto> +1
Rachael: Let's talk about the possibility of separating conformance and guidance
… focus on publishing guidance, aim to take the structure and polish the guidance
… then create one or more informative documents on how to conform to it
… different models or potential ways to conform
… there are pros and cons
… just to start the conversation, the pros that come to mind include releasing guidance sooner
… we feel confident with our guidance, we can work on guidance
… other pro is custom conformance recommendations
… more targeted to different types of websites
… cons include reducing normalization
… just starting the conversation
<Zakim> shari, you wanted to point of order
Rachael: Before we do answers, are there any questions?
GreggVan: On queue to speak about both what we're talking about and what we're talking about
… I think we're confusing conformance with enforcement
… we can't have a standard without a conformance model
… all the discussed informative topics were enforcement
… whether someone has to do something
… on the first item, we need to separate conformance, what does the standard say, when it's accessible or not
… enforcement, when something is accessible "enough"
… it's a different topic
… the description felt like it was conflating them
… I do agee all the discussions about when it should apply should be seaprate
… it's out of our remit, but having a document that describes logical ways to enforce it will be helpful
<Zakim> DJ, you wanted to not if we decide what baseline means and to
DJ: Want to support the idea of separating
… as for the risk to normativity, that goes out the window if someone else beats us to this
<kirkwood> +1 to faster
jj: The baseline and percentages, I was thinking we could use percentages for the baseline
… limit is 50%, but can't get past 50% if you haven't completed baseline requirements
… incentivize companies to meet baseline, they can't cross the line until they meet everything
shadi: I think it may reduce normalization is an extreme understatement. We've worked hard to have WCAG 2.x to be an agreed common ground
… opening it up with here's some requirements, we'll tell you the bar later
… it's a dangerous approach to take
… the second perspective is that it's wrong to do for the conformance model to be on its own
… the conformance properties are interrelated
… thinking about conformance as we construct the requirements helps the requirements
… there's a category of things that do not belong in conformance or WCAG, for policy makers
… thinking about that in parallel helps us to understand what belongs in conformance
… timelines or grace periods, but we should understand
… if you don't do these things, they are blockers
… defined as part of the standard
<Zakim> Chuck, you wanted to ask about normalization
Chuck: I wanted to ask about normalization in this context
… someone mentioned harmonization and that helped my understanding
Sheri_B-H: One of the pros of separation is to let us do multiple iterations of the first before the second
… the con is what happens to people who want to adopt before
… or legislation adopting one before the other
… custom conformance, if we did that, there's differences between on-prem and cloud
… do not benefit users with disabilities
<Zakim> jeanne, you wanted to say W3C dropped the conformance requirement years ago
jeanne: I would like to follow up on Gregg's comment about standards requiring conformance
… W3C dropped the requirement for a conformance section, we're still allowed, but no longer required
… we don't need to by W3C requirements, that being said
… there is a lot to be gained by focusing on guidance vs conformance
… our ideas about how to do conformance will be strengthened as we work on guidance
… as we've seen this year as we've focused more on guidance have only strengthened our persepctives on conformance
… I'm not sure we should completely separate the two, but I do feel strongly that we focus on the guidance and it helps us make informed decisions
julierawe: I wanted to ask about need for speed
… delivering the guidance sooner
… someone else might beat us to it
… do we know other people are working on a competing standard
Patrick_H_Lauke: Just wanted to respond to JJ, about percentages, one of the problems is the categorization of prerequisite and such
… it would be nice to treat each (prerequsites, baseline, other) as a separate block of 100% instead of collectively
… all things need to be fulfilled, it gets away from the binary pass/fail
… the site owner can get a feel for the issues and scope of issues
<Zakim> alastairc, you wanted to say that if we provide a relatively simple conformance model (like WCAG 2.0), that's what people use.
Patrick_H_Lauke: do we need to fix some things or start from scratch
alastairc: There's a couple of points
… in WCAG 2 we have a relatively simple conformance model
… following up on it might not change how people use it
… it might be good to separate
… Shadi said something around the tight integration
… I thought that too, but as we started working on the guidelines and outcomes, we have some assumptions in place
… we have levels and structure, they are not as closely tied
… the things we would need to define would be scoping, conformance statements, I have a strong sense that it makes more sense to look at a product or section and the result should be at a higher level
<Zakim> DJ, you wanted to having criteria will make them easier to organize
DJ: To answer Julie's question
DJ: the 2030 projection and other a11y organizations and trends in industry
<JenStrickland> And the European Union's accessibility publication.
DJ: the G stands for guidelines, we really need to focus on the guidelines part before focusing on conformance
… it'll present to us more intuitive ways to organize them
… it's not clear yet
<Patrick_H_Lauke> (evinced's attempt was more of a PR puff piece, don't think it had any real-world buy-in from any serious players)
GreggVan: "Beat us to it" don't worry about it
… it's so hard to put these things together, so many people try
… people have claimed this in the past, it hasn't worked
… even just aiming at HTML, it was similar
<Sheri_B-H> G3ICT did their maturity model, but I think both of these examples are much smaller than what we are talking about here
GreggVan: the other standards are based on WCAG
… the consensus model, everyone looks to this group
… another version would need to clear the same standards we hold to
… we're revising EN 301549 to be 2.2 right now
… don't focus on worrying about others
… for conformance, it is true that not all standards have conformance
… anything to do with regulatory needs to
<kevin> Sheri, Level Access and Business Disability Forum have had maturity models for a while
GreggVan: most of the W3C is voluntary
… people follow or they don't
… they fall into a different category
… the reason for "Guidelines", it's W3C, we call our docs "Recommendations", everything we do is voluntary
… IETF calls theirs RFC "Request for Comment"
… anytime someone talks about percentages I cringe
… it's possible to game the percentage models
GreggVan: I make everything shallow, it's possible to game them too easily, it's not clear to user the meaning of them either
Rain: As much as I love the idea of splitting
… I love it because it makes our job now easier
<kirkwood> +1 to Gregg
Rain: but I worry about the fact its so hard for the people trying to do this work right now
… everyone here, people working in tech to do this work
… what do I need to do, what can I prioritize, deprioritize
… it's already very complex
… splitting it may make it even harder
… people need to go to two different spaces
<shadi> +1 to Rain
Rain: conversely, we could invent a way to make it easier
<kirkwood> +1 to Rain
<Zakim> shadi, you wanted to respond to Alastair and to respond to DJ
Rain: instead of starting from "we need to do this faster", we go from "how do we make this easier to use"
<Patrick_H_Lauke> easy resources to address very complex problems...that's the tension we have
shadi: Similarly to Gregg's comments, alastairc you're saying that the outcomes is not that interrelated to the conformance percentages
… the exercise we have done in the last few days suggests otherwise to me
… we've identified issues
… I don't agree with this
… to DJ, if you're concerned about others getting ahead, instead of addressing the hard questions, avoid addressing the hard parts
… kicking it down the road
Makoto: Regarding separating, I would like to point out that it might be a good idea for internationalization
… it will allow different counties to establish different policies based on their circumstances
… ex. Japan doesn't have legal requirements
… they may need a different approach than countries that already have conformance
… global companies will be required to use the same standards in different ways
… pros and cons on this issue
<Zakim> DJ, you wanted to GreggVan
DJ: Wanted to respond to Gregg, my point with g is for guidelines wasn't that it wasn't our place to do conformance
… if we separate them, it's not helpful to release an initial set without conformance
… if we don't release anything, there are challenges there too
… it's worth treating them separately but releasing them at the same time
… there's a few things we have to do for this work
<Zakim> alastairc, you wanted to comment on percentages, and the difference between % of outcomes and % of instances.
alastairc: What I was saying is that we have assumptions with outcomes that it has a level and scope
… with those constructs in place, you have a bunch of potential conformance models
… we aren't thinking of taking out the levels, we have prereq, essential, enhanced
… in a guideline we have a hierarchy
… some things more important than others
… there's a big difference between % of instances
… and %s of outcomes passed/failed
… no counting in that sense, if it fails the outcome, it fails
… one potential approach is to include a very basic conformance model similar to WCAG2, page level
… you could do that basic model, but you have a separate document that specifies more recommendations based on size of website or testing regime
<Zakim> Rachael, you wanted to explore about numerous level option or breaking apart by functional need
Rachael: Chair hat on, I did want to acknowledge that we're in the exploratory phases here
… there are lots of questions
… we're rotating between conformance and guidance
… the question is whether we split them as deliverables
… chair hat off, one way to think of it differently is more levels
… people may think of it like WCAG, do we increase/decrease number of levels
<GreggVan> +1 interesting
Rachael: use that as a way to change perspective
… in the first public working draft, we had it, but we can change our approach
… functional levels, etc
… I do expect we will continue balancing and whether to separate deliverables
Ben_Tillyer: Wanted to go back to competing standards, controversial, my goal is to help shape the web into a better place for people with disabilities, if someone else does it, good
… if another standard did come out, I would like to think we could embrace that, or bring it into our work
… look at our mission statement
… we will have a problem if there is a competition, but I don't think we should be scared of it
<Zakim> kevin, you wanted to say that it need these could be developed in parallel
Ben_Tillyer: for thebenefit of users
kevin: Wanted to come back to Shadi's point
… creating a set of guidelines without the appropriate bar to meet, we can develop them in parallel
… the challenge with thinking about this is overloading the outcomes
… thinking about things like minimum number of issues, impact on users, we're overloading them on outcome
… if we look at it as just the outcome, then looking at impact, location, number of instances
… one of the biggest values of splitting
… stops us trying to overload or overtthink
<Zakim> Rain, you wanted to recommend we consider a mini-sprint model
Rain: What Kevin just shared is a nice cognitive model towards helping us make the progress we're trying to make
… we're getting into circles
… not into a focused set of questions, we can build on a scaffolding
… I know there was exploration of models in the past, I am hearing more here
… what this has me wondering if we take the conformance questions, regardless of format, plan a mini-sprint, an environment like this,
… force people into a smallgroup setting and have those people test out a variety of conformance concepts and see what happens
… come back to the group with "here's what we learned"
… looking at this problem in a really focused, time-bound way
Many facets to consider to solve the previous problems: Scope of a conformance statement (views, sections etc); which outcomes fail on which view; whether it is in an important place on the page; whether you can even test all the views (if you have millions of pages); probably more...
GreggVan: Interesting thing Rachael said, percent of provisions vs percent of ??? on a site
… the idea of having more levels, is interesting
… I have a concern with percent of provisions, you'll get people saying "I won't do these"
… "I just need to have enough to meet the %"
… having levels gets around it, control over what people do less of
… I do worry that any of these where people are picking, and we're picking, we'll end up back into "why did this go here?"
We would have to define the levels very very carefully but we likely need to do that anyway
GreggVan: different groups have different needs, and discussion will focus on differences and trying to decide what goes where
… they are all severe to the person experiencing them
… harmonization doesn't mean things are identical, we should learn from others
… they need to not contradict, but not identical
… but I don't see a major competitor to WCAG
… in EN 301549, they cite WCAG, but not as one item
… they put all of the provisions in, they draw the provisions in with the permission of W3C
… they pull them in for web, for other types of documents
… for web, they are in verbatim
… point to our guideliens
… % of provisions concerns me, % of assertions does not
<Zakim> DJ, you wanted to my concern is that it won't
GreggVan: I can see us in endless discussions about what goes where
DJ: One last thing about harmonization, I'm not concerned about another org beating us
… if it happens, not a bad thing
… it's a concern when a respected org creates a bad set of guideliens
… we're discussion separating them becuase of the concern relating to weighting things or grouping
… that was my understanding
… but am I wrong?
GreggVan: To me my concern that conflating conformance and enforcement
… if we're just saying "this is the definition of accessibility", but "this is what is required for regulation"
… we shouldn't try to tell them how to write regulations
shadi: Last last point on harmonization
… lesser concern about one org publishing something
… vs 15 orgs
… different countries with different laws
… different states in the US
… I wanted to come back to %s, Rachael's idea about function, but I think this shows we're early in the process
… if we're working on them in parallel, where's the speed gain
… I agree with Rain, this could be tactical
… where do we put our focus
… at the end of the day, we want to provide something useful
… no matter the number of documents, we need to work on them at the same time
… I'm not seeing where the speed gain is
<Zakim> Rain, you wanted to expand more on functional need approach
Rain: I want to go back to Rachael's concept of functional need
… where that would fit into levels
… for this functional need, are you conforming
… it's not levels
… it's based on the functional needs
… it would have this compounding effect, leading towards shaming people for choosing not to conform to a particular need
<Zakim> Rachael, you wanted to wrap this conversation
<wendyreid> +1 to functional needs
<kevin> +1 to exploring functional needs as an approach
Rachael: We've reached a similar point with publication sub-group, the way we are working works, and we can come back to the question of what we publish later
<shadi> +1
<Chuck> +1 to functional needs
The teams I work with want to know: What's wrong, how do I fix it, and how do I prioritise those fixes. I don't think having 10+ categories of requirement helps with that.
Rachael: I appreciate us taking the time exploring this
<Glenda> + 1 to what Rain just said about listing a11y by functional need. I totally think this will humanize WCAG and help motivate better behavior.
<SteveF> +1
Rachael: I see us suggesting the functional needs approach
… Sheri suggested levels without %s
… the FPWD focused on needs, but it was tied to %s of errors, most of the feedback was the work required
… it sounds like we should return to functional needs,
… it's not something we can handle today, but for a future meeting
… do people have thoughts?
<Glenda> From a buyer/purchasing perspective….I need to know (when buying 3rd party software)…who is going to hit a11y blockers. What reasonable accommodations do I have to plan for (if there is not accessible version of this type of software available).
<Chuck> wendyreid: Scribing is intense, you are all very smart. Strong+1 to Functional Needs. One aspect to Rain's idea, is if we take our outcomes and tie them to functional needs, some outcomes will apply to multiple functional needs.
<Chuck> wendyreid: You can have a huge lift to the accessibility of your project. Gamifying for optimum achievement is not a bad thing. Usability of WCAG... if we can identify the things that set the bar high for achievement, it almost is self-motivating.
<Patrick_H_Lauke> +1 to the functional need classification/idea
alastairc: Not entirely sure how categorization by functional need to work
… most teams want to know what to fix and how
… prioritize accordingly
… Thus people considered A, AA, AAA as prioritisation, wrongly.
… having categories may not be helpful, avoid downloading complexity to the users
… of WCAG
<Patrick_H_Lauke> i think the functional need categorisation helps in making an understandable CLAIM about the accessibility of your product
alastairc: having levels, anything can be games
… people won't say "we don't care about [group]
… it's not about that
Glenda: Want to reinforce what Rain and Wendy said
… when Deque looks at software to procure
… I don't look at criteria
… I look at what disability types are impacted, how, are there risks, is this providing a core function to someone's job
… as a consumer, this would help and humanize the requirements
… if you fail 2.4.11, what does that mean
<jeanne3> +1 Glenda
<Zakim> Chuck, you wanted to ask for scribe change
Glenda: saying "you are blocking [group]", the impact is profoundly awesome
Kevin - also still have the problem of differentiating where on the page/task the barrier is.
Ben_Tillyer: thank you for putting on record how smart group was, Chuck
<kirkwood> +1 to Glenda
Ben_Tillyer: difference between what an organization is saying about the accessibility of the product than understanding what groups are impacted
… if you as an organization know that you have met the functional needs of a certain group, why not write an assertion that you think you have met the functional needs of that group.
… Then when you are going for higher levels, you can gain points by getting the subset of outcomes for a certain need. This comes back to reporting.
<Zakim> kevin, you wanted to say that functional needs is good but problem of impact and % pages still would exist
<Patrick_H_Lauke> i mean the functional criteria are effectively what's (to an extent) in a 508 and/or an EN 301 ... ACR
Ben_Tillyer: If you look at a conformance report, the mature organizations will say what they are aware of and the user groups they are failing. Something to be explored.
kevin: functional needs concept is really good, would motivate some companies. Some won't care whatsoever. Don't know if it solves the problem of the number of failures and impact of failures.
<Zakim> DJ, you wanted to don't we have that now though?
kevin: the failures are still there.
DJ: it is a hepful method to be able to have a report that says you are failing specific people.
<alastairc> +1, we already do that.
DJ: Why should that be in WCAG rather than in the conformance reports themselves?
<Zakim> jeanne, you wanted to say the original request for functional needs in WCAG3 was requests from companies for it for prioritizations
Glenda: if we delineate that, it harmonizes interpretation. We harmonize it rather than relying on inprepretation of the outside.
jeanne3: original Silver TF research surveys with usability specialists, they requested for purposes of prioritization that there be functional need information. When companies are being sued because they have someone with a specific disability that they need to fix the website for a specific person, they need to know exactly what functional needs
they need to correct for and prioritize right away.
… A need from original research showed that there is a need for functional analysis.
wendyreid: another consideration with conformancy-things is audiences, and the different audiences that will receive the reports.
… As a11y lead of a company, have different hats. Is procurement of product going to meet the needs of the people i'm trying to get the product for.
… I don't care about your score, I care that you are meeting the needs. But when reporting to the government, need to be able to meet the metrics the government is asking for.
I'm not very comfortable with us providing the mechanisms for prioritising some groups over others.
wendyreid: When putting together reports for customers, I need to be clear about when I'm meeting their needs.
<kevin> +1 for recognising different audience needs
wendyreid: Don't think we will ever be able to write a model that covers all the needs entirely at the same time, but we still need to be thinking about these and encapulating the ways people are thinkign about this so that we can communicate.
Rachael: Hearing earlier today and yesterday, need to talk more about how we are talking about the prerequisite baseline.
… Baseline enhanced, not baseline.
… Need to talk about the different levels of harm, and how we are thinking about task completion.
… Instead of thinking about levels, can we think about how we could categorize groups of success criteria?
… How to distinguish between these needs that will create a more effective way to understand severity.
Chuck: really do like the prerequisite level. Just required. Full. Stop. Observing that there are a number of different interpretations of what might be in there, though. Can we set on what is in prerequisite.
Rachael: Some people define harm as obvious physical harm. Flashing, seizures, motion that causes other vestibular reactions.
… Others incorporate emotional harm, sensory harm. If you make it too hard, everything can cause harm.
… How do we encapsulate what is most important?
alastairc: Example of how it might be used in practice. Sharing screen with decision tree for focus appearance. Is the user-agent default agent used (pre-requisite)? Firefox and safari, indicator not very apparent. Baseline, however, is if the default indicator meets contrast across backgrounds.
<jeanne3> Research from Bentley University UX department
<Zakim> jeanne, you wanted to say foundational outcomes
alastairc: In practice, shouldn't need to move our provisions based on user agents improving if we write the decision trees properly.
jeanne: Found comment in research done from university ux department, would like to see us do more foundational outcomes. Basic things that are difficult to remediate, like structure.
… Things organizations should be aware of when designing, that are hard to remediate because you have to redo the entire site.
Jeanne - my point was that if we write the methods properly, we won't need to move provisions.
jeanne: Patrick_H_Lauke: Alastair's email -- who decides and documents the decision tree? Otherwise authors will still be on their own.
… Is that the responsibility of the working group? To maintain a yearly list of the things that you can leave up to the environments?
<Zakim> Rachael, you wanted to respond
Rachael: conversation we are having in the working group, will follow up with Patrick and integrate thoughts into that conversation.
Patrick - see w3c/
Chuck: liked seeing us in conference room and misses us!
Rachael: categories to discuss include,
… safety and harm
… task completion
… If we have categories independent of the functions, then we can talk about how those fit different categories in a more meaningful way.
Chuck: Rachael will start and share a document for us to use
Rachael: welcomes others to suggest categories, or how to better define the ones we've talked about
Chuck: personal definition of harm, when in doctor's office and asked about pain threshold. Imagine a train track. Can I leave the track? If not, it's a 10. Anything less than that is a lesser level.
… if on a site that was causing migraines, could I leave the train track?
Ben: Different levels willing to endure based on the task itself. E.g., renewing a passport, there is no alternative. So would be able to cope with a higher level of pain than on a news website, could go to a different website for news.
wendyreid: Want to think about it in an inverse way. not so much that the website is causing harm or difficult, but rather the risk of me not completing the task.
… If I'm put at great discomfort paying power bill, there is a high level of risk if I don't pay my power bill. But if I'm trying to buy shoes, it doesn't really matter if I don't finish the task.
wendyreid: Shouldn't have to be put at risk of great harm or safety by doing the tasks that have to be done.
tiffanyburtin: often the pain and harm is cumulative or additive. Cognitive load to think through all the different things and tasks that you need to do. Stress of switching between ATs because a site isn't fully compatible, or different extensions based on compatibility and what need to accomplish.
<Ben_Tillyer> spoons!
shadi: Different types of content or websites, wondering if that is the place for us to be going? Or is that more about the regulatory bodies?
<tiffanyburtin> Exactly, impact on spoons and also ticket theory!
<Zakim> Ben_Tillyer, you wanted to react to shadi
shadi: Websites are often treated differently if they are commerce or government, etc. Shouldn't be part of the ruler.
Ben_Tillyer: suggest instead of leaving it up to regulatory bodies to determine harm, should be up to the user. Provision that if we know if the user has reported harm, that has impact on conformance.
<DJ> +1 Patrick_H_Lauke
Patrick_H_Lauke: distinction into the three categories remind of priority levels that are used by companies. Such as P1 is may cause seizure and makes a critical task impossible, while P2 is makes it significantly more difficult, and P3 categorized as would improve user experience for certain types of users.
… Has a sense of hierarchy in it, even though we aren't intending hierarchy.
Glenda: In work done they have broken down into 4 levels, with 4 at high risk. When had the 4 levels, found that if everything is a priority, nothing is a priority. Had an overwhelming percentage of 4s. Did a reassessment of default impact factor for common issues, and added a top category of a "blocker" and de-escalated a lot of things out of
level 4. But stuck with 3 or higher as "you have to fix it or you are at risk" and 4-5 as brutal.
… Causes physical harm, causes an unsafe position that is super serious, are examples.
… For example, if I'm driving and suddenly have a blinding headache.
… Not saying the other terrible experiences aren't terrible. Maslow's hierarchy of need, "threatens your existence"
<Zakim> alastairc, you wanted to ask would it help to separate "will cause harm" compared to "could cause harm" and the outcome level, then cumulative harm at the conformance level?
alastairc: Might be useful to separate will and could cause harm, and then address cumulative harm.
<JenStrickland> I can't be there right now, as I'm in APA-Equity meeting. Regarding "will cause harm" - I would really like to speak to this, because the impact of harm on individuals can be devastating. Will there be another opportunity to discuss this?
alastairc: Often severity of harm is about context. Has to be generalized because we have very little predictive power.
… Agree that if everything is "red" people can't prioritize
GreggVan: when you talk about p1, p2, p3 and injury, these are safety rather than accessibility standards.
… What if it blocks all AT?
… What is blocks today, it may not block tomorrow, because AT is evolving quickly and AI will add confusion.
… Makes it easier to use is a concerning concept, because if you look at cognitive, language, and learning, those are the things that make it easier to use, as well.
… Two kinds of priorities. Which things to do we tackle first. Vs. which ones are important. Order of attack, such as low hanging fruit. If it's something that's really serious and we have no idea how to solve it, we might save it and address the immediately fixable stuff first.
<julierawe> +1 to Gregg comments on the harm of poor navigation and findability
<jeanne3> +1 Gregg
GreggVan: Safety, also thinking about the immediate harm, but what about people not being able to use the page causing a family to go hungry? Overall today, not being able to use the internet is a harm. To daily living, wellbeing, integration.
<kevin> +1 to Gregg's comments
GreggVan: Worried about sorting creating shortsightedness where we are losing awareness of the people that aren't top of mind.
To me, tackling that kind of issue (needing access to food) is a regulator thing where they should have higher standard/level to meet.
<tiffanyburtin> +1 Gregg comment
<Patrick_H_Lauke> trying to work out how WCAG could address the problems that Ben is talking about...
<tiffanyburtin> https://
Ben: considering the complicated interactions when addition or mental health are in consideration, being able to audit or identify risks of things like ads and data aggregation, etc.
<Zakim> DJ, you wanted to things moving over time
<Patrick_H_Lauke> are we trying to solve, through technology/normative guidelines, lots of problems that would fall outside of what authors/auditors CAN address?
DJ: Gregg mentioned criteria moving as browsers change. Not sure how that is a problem. If things need to change over time, can we let them change?
shadi: Not being able to access food for your children is one thing. Not being able to buy your shoes is another. Different categories. Come to type of content and category of website, but this is outside of the scope of WCAG. Up to the policy makers to decide.
… What is something that will stop someone being able to carry out the task. What the task is and what that means for someones life is a consideration for a policy maker.
<Patrick_H_Lauke> riffing on what shadi said, that's exactly what EAA does, defining different requirements for different industries
tiffanyburtin: Impact and daily harm. Shoe store example, taking away someone's autonomy if they are unable to go out to a shoe store, isn't fair either. So what are the levels of impact to daily life. We are an internet society now. It's giving more freedom to those in the disable community.
wendyreid: Tiffany read my mind. There is a valid use case of buying shoes online being an absolute essential function to daily living. The degree of impact is part of regulation. We cannot as WCAG cannot measure that.
… One person may get through this fine, another may not be able to complete it. Impossible for us to be able to quantify or predict that. But a regulator would be able to because they have the context.
<Zakim> alastairc, you wanted to ask if we could identify outcomes which can be cumulative? Also, we can't answer these questions at the outcome level.
wendyreid: This isn't just about accessibility and the web, it's also about privacy. Like the privacy working group, there is accessibility impact to the failures of other parts of the web to do what they are mean to do. May not be for WCAG, but may be for a privacy or ads standard that we haven't explored.
<Ben_Tillyer> rssagent, pointer?
alastairc: Curious if we can identify which outcomes could be cumulative? Maybe we can then build that into a conformance model?
<Patrick_H_Lauke> +1 to wendy, having some of the issues covered by other groups like privacy. doesn't all have to fall under AGWG just because "harm"
<tiffanyburtin> Willing to join the conversation on cumulative issues
alastairc: Curious if we can identify which kinds of sites might be held to a higher standard than others? For example, EAA picked on banking, commerce, travel and media.
<shadi> +1 to Alastair
alastairc: Enable them to have multiple bars. If we have certain sites that need to go above baseline, is there a hook that we can provide to regulators that they can use?
Rachael: We've captured a couple of points. Fundamental question, each of these categories sits on a continuum. We can place many outcomes on that continuum somewhere. How do we decide where we cut off the line?
… Safety and harm are highly talked about and also has a lot of emotional ties.
… AT support is also a continuum. What are some examples of that and the likelyhood of that being changed?
alastairc: Wonder if it's less of an organizational exercise where we pick levels, but rather a thing that each subgroup makes as they go. Easier to do with the decision tree, to help build it up and ask questions together, rather than arbitrarily looking at the flat list.
Rachael: Task based prioritization is something that came up a lot, and we explored it in the first public working draft. If we define a priority around task based, how do we do that?
How to prioritize priority (level) based on tasks
Ben_Tillyer: When the useage of a set of webpages is completely different to author's intention, wouldn't want to leave it soley on the authors to decide what tasks exist. Don't think they would get the full list.
alastairc: Slightly disagree. Yes, not really a way to say here are all the tasks. But could allow for people making conformance statements to choose tasks.
… May be some standards within certain industries, such as learning management systems.
<Patrick_H_Lauke> 1) prevents task altogether; 2) makes task difficult to complete for x users
alastairc: But then how do you assess the impact of an instance of an issue. Different instances will have different severities based on context. Up to the conformance statement to say that you would be able to complete a task based on certain expectations.
Glenda: favorite way to figure out what the core tasks and functionalities are is to think about disaster planning information security. If the whole system went down, what are the systems they'll bring back up first? Those are the core task flows. For example, log in is down. Or if you are pizza joint, it is order and delivery.
<Patrick_H_Lauke> similar with EAA
Glenda: Also reminding of the airline act where there were 7 core tasks, such as review flight availability, check schedule, contact information, call for help, etc.
… So could use this as a relatively straightforward way to get these core tasks.
… And if I'm with a client in front of a judge, would I be able to share that it is not accessible and justify it?
Chuck: acknowleding that we just went through that here as a reference. When hotel lost power, first brought up the generator. Then the lights and elevators.
Rachael: agree this is useful as a regulator, but harder as a customer. Is someone going to be able to interpret this when they are trying to assess this?
… And how do we know how important it is to an individual?
… Difficult to come up with those kinds of tasks.
… Is browsing by itself a task, or not a task. Trying to find important health information is a task. So how do we figure out what a task is?
tiffanyburtin: Thinking if I'm going to a site where i have to order a specific medication that is specialty, and it is on a PDF that is an image, it won't be part of their core tasks. Not a core task for a large portion of clientele, it's that I need this rare medication.
Patrick_H_Lauke: can't preempt every possible user need. There will always be things that go outside of what the site owner finds as the primary purpose of the site. At some point or other, as an auditor, I have to be guided by what the site owner tells me are the primary journeys
<Zakim> alastairc, you wanted to comment on how to use tasks in conformance, e.g. doesn't replace view, but provides prioritisation.
<kevin> +1 to Patrick's point about the difficulty of identify every possible task
alastairc: If we use tasks in a conformance model, not replacing the standard page based or view based aggregate thing we do. Tiffany's example should be caught if done correctly. What about assistive technology testing that is very focused on the specific test? You only catch the things in that specific task rather than a more whole understanding.
<tiffanyburtin> I fully understand prioritization but it tends to be certain personas that are considered later than others, and it stinks to continually being underserved.
Rachael: One of our goals is to move the conformance model forward in a useful way. Looked at conformance models that we thought had the most promise. Where is everyone on those? When we look back at the models we were talking about originally - slide 37 - talked about baseline + and prerequisit +
<Zakim> Chuck, you wanted to change scribe and chair
Rachael: Do people feel good about these to explore further, or do we need a full pivot?
<Glenda> Air Carrier Access Act (ACAA) first prioritized the 7 core tasks of airline websites. The 7 core tasks were defined as: 1) Book a Flight, 2) Check In, 3) View Personal Travel Itinerary, 4) View Flight Status, 5) View Flight Schedule, 6) Get Info to be able to Contact Airline, 7) Access Frequent Flyer Account
Chuck: Don't think we need a full pivot.
Next steps on Conformance Model
GreggVan: Everyone is doing Level A and AA. Eliminating the distinction would eliminate people comparing them to WCAG 2.
<Zakim> alastairc, you wanted to say there is a question about using % for baseline.
alastairc: Differences between Baseline and Prerequisite.
julierawe: Questions on how to meet Levels 2 and 3 for both models.
… Language about how to fail and "Not applicable" is unclear.
Rachael: Explore the concept of having a baseline. Fuzzy line between baseline and prerequisite.
<Rain> +1 to Wendy
wendyreid: Prerequisite felt less like a level to achieve
… and more like: Am I even at a place to do this?
Baseline and prerequisite help people do the basic before doing more complex work.
<Patrick_H_Lauke> ... AI ...
GreggVan: 1) You can't do any sorting based on AT.
… In 10-15 years, AT will be able to access even the most inaccessible webpages.
2) Provisions need to be added to models in order to see how the models will work or not.
GreggVan: 2) Provisions need to be added to models in order to see how the models will work or not.
<Zakim> alastairc, you wanted to comment on how AT support would progress.
alastairc: To Gregg: I don't think we need to move provisions based on AT support changing. If we write the decision tree properly.
<Patrick_H_Lauke> (e.g. have something like "the structure of this document is exposed to AT users". once AT gets better - thanks to the hovercar/AI coming any minute now - it will be satisfied without author needing to do anything)
alastairc: If browser improve focus to be dual-color, then we would default to browser-defined focus indicator style
GreggVan: I understand Alastair's statement that writing provisions properly will self fulfill as AT get better
alastairc: 1) Two things have been raised. When we talk about "Not applicable". One thing was counting fails rather than passes.
… But if you exclude "Not applicables", in the Prerequisite+, you'd need to adjust the result to "Less than X%" fails.
<kevin> +1 to my own idea
alastairc: I agree that it helps to count fails.
<Patrick_H_Lauke> "I'm Kevin, and I approve this message..."
alastairc: In the enhanced level we could categorise into functional requirements
<julierawe> +1 to being tired :)
Rachael: Everyone is tired :)
alastair: But that leads to a spreadsheet exercices rather than discussion.
<Patrick_H_Lauke> cheesecake factory is a prerequisite
<MelanieP> stay with us!
<Patrick_H_Lauke> the fun with covid is that...you may not know if you already have it or not / whether you should be at the covid table ;)
<JenStrickland> <3
Patrick_H_Lauke - like sunday night.
GreggVan: 1) This is the first time I've seen pass fails. That means we have testable items.
Adn I don't understand it, and we can talk about it later.
… Conformance was being conflated with enforcement.
… We don't have to tell them what to do, but we just want to make sure they're aware. We should capture them all.
… We have a lot of wisdom to pass on.
… We can help them understand the consequences.
Rachael: Kevin, could you speak to this?
GreggVan: To be clear: We should tell them what to keep in mind.
JenStrickland: In enhanced, video provides additional content.
… I can't fully follow the slides and would appreciate talking to someone about them.
<JenStrickland> Here's what I was trying to speak to, trying to make sense. An Enhanced scenario is if video provides alternative content for content on screen.
<JenStrickland> Where as if there is video only and there isn't text on the screen that is baseline (also the reverse of if text only and no video).
<JenStrickland> Do we in the conformance flow, first ask if X (forms / multimedia) exists first?
<JenStrickland> Then, if it exists, is it accessible?
<JenStrickland> That could set the stage to open a set of conformance standards if X exists.
<JenStrickland> The ultimate score output would be the same, but if X exists then an additional bank of possible points is considered into the total.
<JenStrickland> In Baseline a total is 100 for 100%. In Enhanced the total might be 125 for 100%.
<JenStrickland> [Baseline Set of 50 items]
<JenStrickland> [Enhanced Set of 35 items]
kevin: Yes, there is scope to explore guidance for policymakers. It's a sensile thing to explore. We don't write regulations and we don't tell regulators what to write
… We are in position to provide guidance to regulators