14:51:21 RRSAgent has joined #ag 14:51:25 logging to https://www.w3.org/2026/03/24-ag-irc 14:51:25 RRSAgent, make logs Public 14:51:26 Meeting: AGWG Teleconference 14:52:18 agenda+ Charter next steps 14:52:31 agenda+ Sub-group check-in 14:52:42 agenda+ Accessibility Support Sets https://github.com/w3c/wcag3/discussions/621 14:52:55 agenda+ Provision survey results https://www.w3.org/wbs/35422/wcag3-provision-survey-01/ 14:56:59 shawn has joined #ag 14:57:55 present+ 14:57:59 chair: alastairc 14:58:19 Adam_Page has joined #ag 14:58:21 mehm8128 has joined #ag 14:58:24 regrets: Azlan 14:58:53 Francis_Storr has joined #ag 14:59:03 shadi has joined #ag 14:59:15 present+ 14:59:31 present+ 15:00:05 present+ 15:00:10 Heather has joined #ag 15:00:24 present+ 15:00:29 filippo-zorzi has joined #ag 15:00:42 Patrick_H_Lauke has joined #ag 15:00:45 ginader has joined #AG 15:00:51 Charles has joined #ag 15:00:53 giacomo-petri has joined #ag 15:00:58 janina has joined #ag 15:00:58 present+ 15:00:59 present+ 15:01:01 present+ 15:01:03 present+ 15:01:05 present+ 15:01:15 ShawnT has joined #ag 15:01:18 present+ 15:01:24 Ben_Tillyer has joined #ag 15:01:38 bbailey has joined #ag 15:01:44 present+ 15:01:48 present+ 15:01:57 TOPIC: Introductions and Annoucements 15:02:02 q+ before we get going fully, a small out-of-schedule reminder about wcag 2 TF issues 15:02:15 present+ 15:02:29 present+ 15:02:36 jtoles has joined #ag 15:02:37 q- 15:02:41 present+ 15:02:42 AWK has joined #ag 15:02:43 q+ Patrick_H_Lauke to say something 15:02:45 kirkwood has joined #ag 15:02:49 q+ to tell Patrick something 15:02:51 present+ 15:02:51 q- 15:03:09 q+ to make a good point, please 15:03:11 q+ to dance arround 15:03:11 ack me 15:03:12 hdv, you wanted to make a good point, please 15:03:14 q- 15:03:14 CClaire has joined #ag 15:03:15 q+ to gumble about the weather 15:03:17 Jennie_Delisi has joined #ag 15:03:20 GreggVan has joined #ag 15:03:20 q- 15:03:21 present+ 15:03:23 present+ 15:03:24 present+ 15:03:24 present+ 15:03:27 scribe+ 15:03:31 present+ 15:03:31 scott has joined #ag 15:03:33 present+ 15:03:34 present+ 15:03:40 present+ 15:03:44 q? 15:03:48 alastairc: opportunity for introductions; no takers. 15:03:53 Frankie has joined #ag 15:03:56 present+ 15:03:57 ack Patrick_H_Lauke 15:03:57 Patrick_H_Lauke, you wanted to say something 15:03:59 alastairc: Announcements 15:04:07 Detlev has joined #ag 15:04:07 BrianE has joined #ag 15:04:13 https://lists.w3.org/Archives/Public/w3c-wai-gl/2026JanMar/0172.html 15:04:13 chrisg has joined #ag 15:04:15 present+ 15:04:24 present+ 15:04:32 present+ 15:04:36 Patrick_H_Lauke: Email sent yesterday listing the latest proposed pull request from the WCAG 2 Backlog Task Force. There's been a number of questions since sending the email. 15:04:58 q+ Rachael on a new github discussion 15:05:00 julierawe has joined #ag 15:05:01 q+ to say WCAG-EM review period ending this week 15:05:16 present+ 15:05:23 Makoto_U has joined #ag 15:05:31 present+ 15:05:57 ... This pull requests that represent potential errata, and that caused a little confusion. This is only a very early look at these. Bruce and Patrice will be be meeting early next week. The CSC will be for potential addition to the next version, which happens twice a year. 15:06:12 Wilco has joined #ag 15:06:24 ... If at that point of the CFC, these proposed PRs are deemed too fundamental, then they get put into a new version of WCAG to. This is has been an ongoing topic. 15:06:30 q? 15:06:34 Is "errata publication" a republication of WCAG or of an errata document? 15:06:46 q+ kevin on W3C announcements 15:07:25 laura has joined #ag 15:07:29 ... Look at do we have enough, why do we need to do this. And maybe look at the WCAG 2 wider group. Unintended misinterpretations and it's causing confusion. 15:07:37 present+ Laura_Carlson 15:08:26 ... Second, In the set of pull requests that were included in the patch emailed yesterday, wanted to draw the groups attention to one in particular because it was hefty. It makes a lot of very big changes to understanding documents, and it's the one that tries to disambiguate pointer gestures and dragging movement. 15:08:32 AWK - two streams, 1st is errata, 2nd is potential future publication. 15:08:40 thx 15:09:03 Disambiguation of path-based gesture versus dragging movement #4843 15:09:03 https://github.com/w3c/wcag/pull/4843 15:09:14 q+ to ask about rendered page diffs for PRs. Can we still do this? 15:09:39 ack awk 15:09:39 AWK, you wanted to ask about rendered page diffs for PRs. Can we still do this? 15:09:46 ... Welcome feedback on those. It tries to clarify how auditors should approach determination of if this is something that falls under this one success criterion or the other when its a failure. 15:10:19 AWK: The challenges of trying to understand what's actually being changed sometimes and whether we can see a page that shows the diffs. 15:10:27 could we link to diff link example here? 15:10:33 ljoakley has joined #ag 15:10:50 Patrick_H_Lauke: There are diff links; the diff doesn't include images, which part of the change that is being made as well. Links being provided in IRC. 15:10:53 present+ 15:11:01 Conforming alternate versions & equivalents https://github.com/w3c/wcag3/discussions/623 15:11:01 ack Rachael 15:11:01 Rachael, you wanted to comment on a new github discussion 15:11:16 Essential Exception https://github.com/w3c/wcag3/discussions/624 15:11:38 ack hdv 15:11:38 hdv, you wanted to say WCAG-EM review period ending this week 15:11:41 SydneyColeman has joined #ag 15:11:45 Rachael: Announcement: Two new discussions in GitHub. First is conforming alternative versions and equivalents. it's a follow up from last week. Second, is the essential exception. Jump in there and start conversations. 15:11:47 present+ 15:11:49 Glenda has joined #ag 15:11:56 awk: diff for pointer gestures https://services.w3.org/htmldiff?doc1=https%3A%2F%2Fwww.w3.org%2FWAI%2FWCAG22%2FUnderstanding%2Fpointer-gestures&doc2=https%3A%2F%2Fdeploy-preview-4843--wcag2.netlify.app%2Funderstanding%2Fpointer-gestures and for dragging movement https://services.w3.org/htmldiff?doc1=https%3A%2F%2Fwww.w3.org%2FWAI%2FWCAG22%2FUnderstanding%2Fdragging-movements&doc2=https%3A%2F%2Fdeploy-preview-4843--wcag2.netlify.app%[CUT] 15:11:56 anding%2Fdragging-movements (those two diff links are in the description - first comment - of https://github.com/w3c/wcag/pull/4843) 15:12:12 q+ 15:12:16 Link for WCAG EM https://www.w3.org/WAI/news/2026-02-05/wcag-em-2/ 15:12:18 cdv can you post the link again? 15:12:21 present+ 15:12:22 q- later 15:12:56 ack Charles 15:12:57 hdv: WCAG Evaluation methodology opened, and we'd love to hear from you. Getting comments already, encourages others to participate in GitHub or reach out via email. 15:13:27 Charles: Question about Rachael's topic if it is open to the public. 15:13:28 Thanks Patrick_H_Lauke - I was missing the small "diff" link and was clicking on the 'deploy-preview" link. 15:13:31 s/WCAG Evaluation methodology opened,/Review period for WCAG-EM 2 is still open until Thursday this week/ 15:13:35 alastairc: Yes, that is open to the public. 15:13:40 ack kevin 15:13:40 kevin, you wanted to comment on W3C announcements 15:13:58 -> https://www.w3.org/calendar/breakouts-day-2026/ Breakouts 15:14:24 LenB has joined #ag 15:14:28 pointer gestures could be fun ;) 15:14:30 presnet+ 15:14:34 present+ 15:14:53 Kevin: First announcement: W3C has breakout sessions running tomorrow and the day after. Link provided in IRC. Two session on each of the days, with overlapping talks. One talk is from COGA and their research, and Pointer Gesture talk. 15:15:18 woo! congrats, Adam :) 15:15:23 Congratulations, Adam! 15:15:24 👏👏👏 15:15:32 ... Second announcement: Work of the chairs. When Chuck retired, we've considered onboarding another co-chair. This will be Adam Page. 15:15:33 Thank you Adam! :) 15:15:48 please send mentioned github links 15:15:53 🎉 15:15:54 Congrats Adam! 15:15:58 Thank you, Adam! 15:15:58 Congratulations! 15:15:59 Yay! Thank you, Adam! 15:16:00 Thank you Adam! 15:16:06 congrats adam 15:16:12 +1 adam and his technicolor background 15:16:31 ... Adam brings a huge amount of experience, is active in WAIARIA and WCAG 2 for the past 3-4 years. He's on the West Coast US. Great perspective from a large organization, and he has extensive UX and technical expertise. 15:17:19 Thank you Adam! Nice meeting you in person at CSUN! 15:17:23 👏 15:17:25 Adam_Page: I've gotten to know quite a few of you, honor to be asked to fill this role, excited to join Alistair and Rachael. WCAG 3 is vital work, and is eager to do as much as he can to help make this group the best it can be. 15:17:50 zakim, take up next item 15:17:50 agendum 1 -- Charter next steps -- taken up [from alastairc] 15:18:40 Kevin: Charter Next Steps. Introduced new charter a few months back, a survey intakes comments, and are now in a position to move forward. Current charter runs through April. 15:18:54 -> https://w3c.github.io/charter-drafts/2025/ag-wg.html New draft charter 15:18:55 ... Worked on a number of comments and an updated draft has been posted in IRC above. 15:19:23 ... the charter is a way to communicate to the broader membership what work is happening within the working group. 15:19:28 I already found a flaw in it. Missing a chair ;) 15:19:53 Gez has joined #ag 15:19:57 ... hoping to get this submitted through to AC next week. Asking for feedback or comments (Wilco mentioned there is a missing chair). 15:20:01 present+ 15:20:38 ... process is to submit to AC reps for review, which includes possibility of formal objection. Aware of a potential formal objection, and has prepared for that possibility. 15:21:27 detlev has joined #ag 15:21:43 ... intention of the charter is to progress the work. The charter also covers the AG does, including WCAG 3, WCAG 2 maintenance, Mobile accessibility task force, COGA, APA Working Group, WCAG to ICT, and ACT, and others not specifically mentioned here. 15:21:49 q+ to ask about the duration of the review period 15:21:50 q+ 15:21:54 q+ 15:21:58 ack shadi 15:21:58 shadi, you wanted to ask about the duration of the review period 15:21:58 ... Any last minute adjustments, please make aware so we can progress that forward 15:22:09 shadi: Asks about the review period for charter review. 15:22:22 Kevin: Thinks about a month. needs to double-check in the process. 15:22:37 Wilco: Can you summarize the changes? How much time do we get to approve it. 15:23:16 Kevin: The working group does not get to approve it, they only contribute to its contents. My responsibilities is to ensure there is a charter, and it's maintained and reviewed by AC. 15:25:22 LenB has joined #ag 15:25:38 ... Change summary: Expanded out the scope with details, included section 2.2 on WCAG interim publications allowing for release, expanded the policy support section, introduced and included more info on the support for non-Latin scripts (including impact to WCAG 2), deliverables have maintenance of normative specifications and non-normative 15:25:38 deliverables, introduced definition of done for clarity; the rest basically remains unchanged. 15:25:39 q? 15:25:41 jkatherman has joined #ag 15:25:42 ack Wilco 15:25:44 ack AWK 15:25:48 present+ 15:26:04 AWK: Asks if there is a diff version relative to last time. 15:26:25 maryjom has joined #ag 15:27:07 ... Specifications talks about maintenance of WCAG 2.2, but it says that no normative changes will be undertaken through the chanter under WCAG 2 maintenance. Wants to understand what's being said there. Saying we're going to identify what changes to the normative content may be warranted and valuable related to discussions. What are the edges to 15:27:07 that? 15:27:30 q+ 15:28:05 Poornima has joined #ag 15:28:11 Rachael: Feedback on this point was not having a scope of what we would want to change. We stepped back from that, and we can use this two-year period to continue the task force work, and at the end of the 2 year period, a clear outline of exactly what would change if we wanted to publish WCAG 2.3. 15:29:12 q+ to respond to questions 15:29:14 ack Wilco 15:29:32 alastairc: adding to last comment, the way it could pan out is that we have two streams, if we agree that things are, and for those things that aren't. We can store up the items that could be the next version or iterate on a newer version. 15:30:07 ack kevin 15:30:07 kevin, you wanted to respond to questions 15:30:11 Wilco: Asks if he is expected that there is an objection. Only seeing this for the first time. Not sure what's going on. 15:30:57 OK I did not get anything. Lets put that down to comm issue then. Lets get in touch off call 15:30:58 -> https://www.w3.org/policies/process/#CharterReview Charter review 15:31:27 -> https://services.w3.org/htmldiff?doc1=https://www.w3.org/2023/11/ag-charter&doc2=https://w3c.github.io/charter-drafts/2025/ag-wg Diff from previous charter 15:31:43 -> https://github.com/w3c/charter-drafts/commits/gh-pages/2025/ag-wg.html Changes to charter 15:32:00 Great, I'll look there. Thanks 15:32:01 Kevin: This shouldn't be the first time you've seen it. Wilco has submitted issues and those were responded to, and initial change have been made. There were others in the communication that have been sent. The charter review period stipulates a minimum of 28 days. A diff can be done from the current charter and the new charter. View GitHub 15:32:01 history for the Charter Drafts. 15:32:02 q? 15:32:29 zakim, take up next item 15:32:29 agendum 2 -- Sub-group check-in -- taken up [from alastairc] 15:32:53 Jen_G has joined #ag 15:33:09 Present+ 15:33:23 alastairc: Quick round robin to each subgroup lead/delegate. How are you doing? Do you have any provisions for review soon? One survey out, looking to get the next survey out. Do you have any blockers? 15:33:26 q+ 15:33:30 q+ 15:33:32 q+ 15:33:34 ack hdv 15:33:38 q+ 15:33:43 q+ 15:33:45 q+ for user-control 15:33:54 q+ for single sense 15:34:01 q+ for SIgn Language 15:34:19 Charu has joined #ag 15:34:32 present+ 15:34:45 ack giacomo-petri 15:34:45 hdv: Subgroup Errors is doing great, have a new liaison. We have potentially one rule ready for review in the next round, and will be sharing soon. Asks for more people if anyone is interested. Meet on Monday afternoon in Europe (10 or 11 EST) 15:35:06 Patrick_H_Lauke has left #ag 15:35:22 GN015 has joined #ag 15:35:59 ack Makoto_U 15:36:00 2.1 Images and media https://www.w3.org/TR/2026/WD-wcag-3.0-20260303/#images-and-media 15:36:03 giacomo-petri: Subgroup is User orientation, structure. Lots of discussions, especially around structure, including 1.3.1. Reminder that we are aiming to progress and iteration and not perfection. We should be ready with one provision for Survey which is Page View title. We ask for more participation. We meet on Wednesday 9am EST. 15:36:46 Jon_Avila has joined #ag 15:37:27 Makoto_U: Image and Media Alternative Subgroup has been working on the reorganization of subjections and are almost done. We want to bring it to survey for the larger group. It is difficult to understand what should be done when people take a look at the current subsections, and there's a 2.1 image and media section. We want to make a proposed on 15:37:27 organizing sections and provisions, so people can more easily understand provisions. After this, we'll pick up the pace. We meet at 10AM EST on Tuesday. 15:37:31 ack Francis_Storr 15:38:42 ack julierawe 15:38:46 Francis_Storr: Input and Focus Interactive components subgroup. We meet Wednesday at 12 P EST. We have one provision in a survey already, and another ready for the next survey. We are working on a third, and are aiming to get that done at the end of the week. Some people in the group have another meeting on Friday to keep the progress moving. 15:38:46 Invites others to join 15:38:50 q+ to clarify what is a provision 15:39:33 A provision is a requirement or an assertion. Provision is the umbrella term. 15:39:46 https://www.w3.org/TR/wcag3/#types-of-provision 15:40:03 thanks 15:40:05 ack alastairc 15:40:05 alastairc, you wanted to discuss user-control 15:40:09 julierawe: Text detectable in the most recent survey. Another is ready for text-to-speec provisions for the next survey. Other provisions have drafted ACT rules, and have had great conversation. We have 3 provisions on clear language. Meetings are Wednesday 10 AM EST. 15:40:15 q- 15:40:58 ack Rachael 15:40:58 Rachael, you wanted to discuss single sense 15:41:00 alastairc: User Control subgroup - last few weeks have been working on animation and diving into pseudo-motion. We meet on Monday at 10 AM EST. Note the time zone. 15:41:26 s/subsections, and there's a 2.1/subsections under the 2.1 15:41:53 ack kevin 15:41:53 kevin, you wanted to discuss SIgn Language 15:41:55 Rachael: Single Sense - 2 provisions may be ready for the next survey. About to kick off the color contrast research. Meet on Friday at 8 AM EST. Contrast research group will be a different time. Invites others to join, but other areas need it more. 15:42:45 Kevin: Same Language subgroup. Another participant joining (now has 4, which seems sufficient for the work to be done). Have a solid set of requirements, a few ACT rules, and is progressing relatively well. 15:42:54 agenda? 15:42:58 s/Same Language/Sign Language/ 15:43:00 alastairc: Asks for any questions for the subgroup work. 15:43:02 zakim, take up next item 15:43:02 agendum 3 -- Accessibility Support Sets https://github.com/w3c/wcag3/discussions/621 -- taken up [from alastairc] 15:43:04 ... none. 15:44:02 alastairc: Next topic - continuing conversation from last week. link is in the agenda items and is sharing screen. Only one comment provided from last week. Will treat as an introduction and continuation of the topic, and may not get into a lot of detail. 15:44:10 https://docs.google.com/presentation/d/1Vov5k9XXlQSy7AtVuBVQvhSMDIOPA0kj69xJoq71BSQ/edit?slide=id.g3d044950233_0_0#slide=id.g3d044950233_0_0 15:44:17 ... link to presentation above. 15:44:39 ... summary: Define what an accessibility support set includes without naming products. 15:45:23 ... We can go into more detail about we use and what our assumptions have been. We don't see this as a per-claim, it's more a case of organizations or different regions that may need emphasis of different technologies. 15:45:56 ... these give us a little bit more abstraction in-between the sort of normative W3C content. Reliable data is hard to come by. 15:46:31 ... Main question is: How to describe int he normative text, what an accessibility support set needs to include. 15:46:58 ... Various factors such as cost, availability, usage, and may vary by region. Main reasons from trying to take this approach is that there's a lot of examples. 15:47:44 ... alastairc: reviewing the slide content, including an example. 15:48:23 ... comment from GreggVan to split into two parts, such as browser support and assistive technology support, and alastairc started a reply. 15:48:30 q+ 15:48:32 q+ 15:48:46 ack GreggVan 15:48:48 ... comment to define platform accessibility support and define AT accessibility support. 15:49:39 For internal situations - I don't think we can require the most common AT. 15:49:49 GreggVan: In the past, we've only talked about Web content. If we are trying to make this apply to web and non-web, we need to account for the platform features, and the fact that the browser would be missing. 15:49:50 q+ on how much overlap there is? And internal situations 15:50:08 Software should not depend on what AT does, but AT should used APIs agreed upon, like proper ARIA support. 15:51:11 q+ also on responsibility of user to choose best AT 15:51:19 is plain language technology included in the “accessibility support set”? 15:51:43 It would need to be available to folks - provided without more than nominal cost. 15:51:45 +1 to Gregg’s question. had same ? 15:52:23 q? 15:52:27 q+ 15:52:43 The return of "AT-push" from WCAG 2.0 days! 15:52:43 ack Charles 15:52:43 GreggVan: I don't think you can say that you can make a web page accessibility by just assuming that there's an AT that can make something accessible. or rely what is in the platform. We might want to define what we means for web, and what is meant for non-web. Anytime you wander off the web, it gets more complicated. We did not account for "it's 15:52:43 ok as long as AT can do it." Are we going to assume that all users have all AT? All users may not have that. It is not simple anymore. There are lots of types of AT, and we've never took into account cost, etc. 15:53:37 q+ to say more platform accessibility support complications 15:53:45 Charles: Physophical question as part of the example in the slide deck. The language "In at least one" triggers the thought of "how can this be abused?" Should we also be recommending that they're proportionate? 15:54:00 q+ 15:54:49 acl alastairc 15:54:52 ack alastairc 15:54:52 alastairc, you wanted to comment on how much overlap there is? And internal situations and to comment on responsibility of user to choose best AT 15:55:03 alastairc: Potentially. One way to think about it is... as a standard we are trying to get a certain level of coverage, it doesn't necessarily ned to be perfect coverage. It is also going to be difficult to get the kind of data, and to say how many are in which category. Think of diminishing returns on the bigger sample. 15:55:55 clarifying. my concern is proportions over numbers. a set should not highly favor one need. 15:57:07 alastairc: Got on queue to respond to Gregg's comment. The things we test with the most are things like web, native mobile, etc. There's a lot of overlap. Having that user agent somewhat replicated on the native platforms (text sizing, contrast adjustments, etc). Repsonsiblity is on the author to utilize those in a fairly similar way to the web, 15:57:07 but there are differences. Platforms may not have any assistive technologies features. There is a lot of overlap. 15:58:38 q? 15:58:43 ack GN 15:58:49 ... Most difficult question to answer IMO, is the assumption that the user is going to select the best assistive technology for them. An expensive AT that does everything for you, and leaves the author with nothing to do. But even with free AT, there are complaints. Can we include such a thing where we meet certain requirements and make an 15:58:49 assumption that it's not universally available. People may not have it, but can we assume that if they need it, they should have it. 15:59:16 the legal answer is: burden 15:59:49 q+ on user-agent bugs 15:59:53 ack maryjom 15:59:53 maryjom, you wanted to say more platform accessibility support complications 16:00:18 GN015: Is the accessibility support set used for testing? If we give such an accessibility support set to an author, and the author can't do anything to make an AT work with it because the AT itself doesn't work with the Accessibility API? it seems to push the authors in the wrong direction. 16:01:05 maryjom: Would want to tie the accessibility support set to the platform, you would name the specific operative system. 16:01:40 scribe- 16:02:03 zakim, pick a scribe 16:02:03 Not knowing who is chairing or who scribed recently, I propose Glenda 16:02:13 Detlev has joined #ag 16:02:20 scribe+ 16:02:24 q? 16:02:26 present+ 16:02:32 +1 to gundala and mary jo's comments 16:03:01 ack GreggVan 16:03:11 Gregg: Has to be a11y support on all of the platforms that the application runs. For web, might be the top 3 browsers that are free. 16:03:35 q+ to ask about AI data use with AT 16:04:12 Gregg: only works on 1 screen reader. What if it only works on one really obscure browser. I don’t think that works. I think it has to at least be on the native screen reader. 16:04:29 q+ to comment about naming specific ATs, platforms, and user agents 16:05:23 Gregg: just saying any 1 AT isn’t good enough…because it might be an AT that hardly anyone uses. 16:05:38 I still recommend to rather have authors fulfill well-defined APIs. Otherwise the are pushed to run after errors in platforms, browsers, and assistive tools. 16:05:50 ack alastairc 16:05:50 alastairc, you wanted to comment on user-agent bugs 16:06:00 Frankie has joined #ag 16:06:10 For some system like a SmartTV you are limited to use the ones that are made available on that platform if you can't install any third party technology. 16:07:22 ack Jennie_Delisi 16:07:22 Jennie_Delisi, you wanted to ask about AI data use with AT 16:07:32 Alastair: Picking up Gundula’s thread. it does seem wrong to punish authors for long standing bugs in AT. Like Dragon. It is a user agent bug. But it also hurts real users. 16:08:37 we should also be thinking mobile first btw (due to popularity) 16:08:42 ack Heather 16:08:42 Heather, you wanted to comment about naming specific ATs, platforms, and user agents 16:08:50 +1 to CPU/GPU and privacy and security as characteristics of the AT 16:09:15 q+ on how support-sets interact with methods 16:09:17 q+ 16:09:28 q+ 16:09:40 ack alastairc 16:09:40 alastairc, you wanted to comment on how support-sets interact with methods 16:09:40 q+ 16:09:54 Jennie: in old days, we used to talk about issues related to bandwidth and having the ability to run AT at the same time. Wonder if now we also need to have consideration related to AI with could be use by AT, may store data in different places. related to the end user’s privacy and ownership of their own data. I wouldn’t want to require someone ot use an AT that stored data in a way that wouldn’t be comfotable for them or in alliance wit[CUT] 16:09:55 work requirements. 16:10:26 InaT has joined #ag 16:10:48 q+ 16:11:23 ack hdv 16:12:21 Heather: Suggested including at least one assistive technology (AT) from each major category when defining an accessibility support set (e.g., four types of AT). Reiterated that testing should focus on meeting the standard, regardless of technique. Acknowledged that AT-based testing is one valid technique among many. Emphasized the need for thorough testing while balancing efficiency. Noted diminishing returns when repeatedly testing the same requ[CUT] 16:12:22 using multiple techniques. Concluded that the priority is verifying conformance to the standard, not mandating specific testing methods. 16:12:24 q+ 16:12:27 ack maryjom 16:12:50 qq+ 16:13:07 ack alastairc 16:13:07 alastairc, you wanted to react to maryjom 16:13:58 ack Jon_Avila 16:14:06 Alastair: clarified the relationship between requirements, methods, and accessibility support sets. Requirements remain normative and are assessed on a pass/fail basis, while multiple methods or techniques may be used to meet them. The proposal introduces a defined default accessibility support set to make underlying assumptions more explicit, with AGWG methods validated against this set to ensure consistency. It was acknowledged that different r[CUT] 16:14:07 contexts, such as intranets, may require alternative support sets, and while the default is expected to reflect commonly used assistive technologies (likely centered on US and European contexts), flexibility is necessary. The intent is not to require testing across all assistive technologies, but to ensure that methods are reliably tested against a defined support set. For those using AGWG methods, this testing is already accounted for, while for 16:14:07 alternative methods, the framework provides a way to demonstrate adequate testing. 16:15:44 Hidde: supports not explicitly including AI-based browsers in the definition, noting concerns around privacy, security, and the current economic models of AI platforms, which may not be stable or representative of true costs. Also points out potential risks associated with AI-based systems and agreed with keeping the language more general rather than naming AI browsers specifically. 16:16:27 ack giacomo-petri 16:17:20 MaryJo: raised a concern that requiring assistive technology (AT) for testing may conflict with existing approaches, such as the U.S. Trusted Tester program, which validates accessibility without using AT by focusing on programmatic determinability. She questioned whether this would imply such approaches are insufficient. 16:17:49 Alastair: responded that WCAG and ACT rules already rely on implicit assumptions about user agents and testing contexts to determine pass/fail outcomes. He explained that the proposed approach makes these assumptions more explicit by defining a test set, while still allowing others to establish their own. The goal is to clarify which environments methods are validated against, rather than mandate specific testing techniques. 16:17:58 q+ 16:18:22 q+ to say that the lists thing depends on the requirement, is it done visually? 16:19:51 ack GreggVan 16:20:03 we cannot predict the future 16:20:16 Safari with VoiceOver is an example of where if you hide the list style type the semantics of the list are hidden. 16:21:04 Jon Avila: highlighted a gap between conformance requirements and how they’re actually applied, warning that people might claim conformance by meeting success criteria while ignoring broader conformance expectations. He cautioned against creating loopholes—especially when relying on platform features (like OCR on iOS)—since those features can be fragmented and incomplete, making the end-to-end user experience unreliable. 16:21:17 ack alastairc 16:21:17 alastairc, you wanted to say that the lists thing depends on the requirement, is it done visually? 16:21:29 Giacomo: emphasized that we shouldn’t avoid recommending semantic HTML methods (like lists) just because some supported technologies don’t fully meet requirements. He pointed out that lists can still be semantically correct even if their visual presentation changes (e.g., bullets removed). However, he raised a concern: if a supported user agent strips list semantics (rendering UL/OL as generic elements), should that affect whether we recommend[CUT] 16:21:29 semantic list elements at all. 16:23:00 q? 16:23:11 Gregg: G1) Testing without assistive technology can be valuable if it ensures information is exposed in a standardized way, but it requires clearly defined testing setups. 2) We should plan for a future where AI is local and more deterministic, while recognizing current variability (like alt text differences) and access gaps on older devices. 3) Platform features should only be relied on when they truly work—and work consistently across major 16:23:12 platforms—so conformance depends on real, dependable functionality. 16:23:21 To Giacomo's concern: I'd rather say the specific browser does no longer qualify to be part of the accessibility support set. 16:23:52 q? 16:23:57 zakim, take up next item 16:23:57 agendum 4 -- Provision survey results https://www.w3.org/wbs/35422/wcag3-provision-survey-01/ -- taken up [from alastairc] 16:24:36 Alastair: previous topic Alastair noted that assistive technologies like VoiceOver may still interpret lists correctly even when visual list styles are removed, due to built-in heuristics, which complicates how we evaluate semantics. He suggested that conformance may need to be assessed on a per-requirement basis—clarifying whether we judge based on what is programmatically provided by the author or what actually works in user agents. He reinfor[CUT] 16:24:37 WCAG 2 sometimes overlooked accessibility support, meaning something can appear to pass success criteria but still fail in real-world use, raising the need for clearer expectations. 16:25:44 Rachael: Brian made comment about ACT rule. Alastair resolved it. 16:26:02 q+ on trigger warnings 16:26:03 q+ 16:26:35 Rachael: Scott you had a question. Scott confirmed he had added it as an issue. 16:27:01 ack Charles 16:27:47 +1 to Charles, good point 16:27:55 +1 to that point 16:27:57 ack alastairc 16:27:57 alastairc, you wanted to comment on trigger warnings 16:28:01 Rachael: Bruce’s example of YouTube videos that warn about flashing and include timestamps, and asked whether that approach is sufficient to meet the core “no flashing” requirement. Let’s discuss it briefly to help the subgroup determine if such warnings can count as acceptable compliance. 16:28:04 +1 to Charles 16:28:18 I agree that it's too easy to miss the warning. Would the warning need to e audio described as well? 16:28:31 q+ 16:28:31 q+ 16:28:36 ack Heather 16:28:54 q+ 16:29:11 q+ 16:29:11 q+ 16:29:11 ack giacomo-petri 16:29:12 Alastair: noted that while warnings about flashing are important, the foundational requirement still allows some essential flashing, and it’s unclear how warnings and the core requirement should work together—suggesting this needs further clarification beyond the current requirement. Charles added that the effectiveness of a warning (like on YouTube) depends on how content is accessed; if users jump directly to a timestamp after the warning, t[CUT] 16:29:13 never see it, making the warning insufficient. 16:29:14 q+ 16:29:28 q+ 16:29:47 ack gn 16:29:55 Alastair: we will need to take this good point into account. 16:30:14 Heather: I thinkn this sort of thinking is dangers. Setting a precedent for other standards. 16:30:50 ack GreggVan 16:30:50 ack GreggVan 16:31:04 Reminds me of the law in the UK around having to display what products contain what allergens from a list of 14... "All products may contain traces of: nuts, shellfish, celery, sulph....." 16:31:16 Giacomo; To add to what charles said, “this video may contain flashing lights” you don’t know if it wil be ther or not. If you add a warning to fix it…it may preclude people from watching (but there might not be flashing) 16:31:50 s/Giacomo; /Giacomo: 16:31:54 ack julierawe 16:32:19 Gudula: This leads to making assumptions what the user will perceive. They may be ditraged when the warning appears. Do they have coga issues. Might take time to process it, or forget to process it. 16:32:55 ack maryjom 16:32:57 +1, what if an operating system/browser update has 'reset' your preferences 16:33:01 q? 16:33:18 Gregg: Gregg emphasized that warnings about flashing are necessary but not sufficient—while lacking a warning makes content inaccessible, simply adding one does not make the content accessible. Because flashing poses a serious health risk, the underlying issue must be addressed, not just mitigated with warnings. 16:33:20 +1 you also may be at a place where you can't set such a preference 16:33:21 +q 16:33:22 q+ to ask whether having defined "flash" makes sense 16:33:40 +1 to Mary Jo. In my survey responses, I recommended renaming this short name. 16:33:42 ack Jon_Avila 16:33:55 +1 No flash over threshold 16:34:04 Julie: if you respect a user pref for reduce motion, you could pass, but what if you are looking at someone elses device. This seems like a special case to me. 16:34:27 q+ to say No flash over threshold 16:34:57 MaryJo: Confused about the short name of this core requirement. No Flashing. But we do allow some minimal flashing. Short name should be Minimal flashing, or under threshold 16:36:00 ack jkatherman 16:36:36 Jon Avila: Jon Avila noted that flashing can be unpredictable in real-world scenarios (like games), so future solutions may need automated detection or filtering. He also pointed out that current testing methods (like viewing distance) don’t account well for different user experiences, especially for people with low vision. Also, not knowing where someone is going to be viewing this..it could be on a TV in a doctor’s office. Relying on user pref 16:36:36 isn’t enoug. 16:36:37 ack alastairc 16:36:37 alastairc, you wanted to ask whether having defined "flash" makes sense 16:37:34 Jeremy noted that simply warning users about inaccessible or flashing content does not make it conformant, but those warnings still have value and should be encouraged—just not treated as meeting WCAG conformance. 16:37:55 When I was at HSBC a few years ago, they worked with Zoom to add the feature on that platform, the technology is there and will become more widespread : https://www.epilepsy.org.uk/news/zoom-and-hsbc-develop-epilepsy-safety-feature 16:37:57 +1 to Jeremy's statement 16:38:03 Alastair: Yes, we do have a separate req for providing a warning. How those two interact is a good question. 16:38:25 q? 16:38:37 ack GreggVan 16:38:37 GreggVan, you wanted to say No flash over threshold 16:38:39 ack GreggVan 16:39:52 Alastair: noted that while tools to automatically detect or reduce flashing may emerge, they aren’t reliable yet. He raised the tension between banning all flashing versus allowing it with warnings and user choice, and clarified that WCAG doesn’t prohibit all flashing—only flashing above defined thresholds based on size, contrast, and frequency. 16:40:07 warning of the existence of something harmful does not prevent harm. placing that thing behind an explicit user action where they have full agency seems like the only way to prevent harm. 16:40:19 q? 16:41:26 Gregg: Gregg said that while future tools (he is working on an open source option) should be able to detect or reduce flashing, they aren’t ready or reliable enough to depend on—since they’d need to be universal, fast, and on by default, which is unlikely. So they can be a helpful fallback, but not something to rely on for accessibility. 16:42:05 Rachael: Charu if you have any outstanding issues can you specify in survey or here. 16:42:35 q? 16:44:16 q+ 16:44:16 q+ 16:44:22 ack julierawe 16:44:59 q+ on one of the questions "Should we replace it with a “Text selectable” provision that is more specifically geared to help text-to-speech tool read all visible text?" 16:45:18 ack Charles 16:45:27 q+ 16:45:47 It's not limited to that. 16:46:37 ack alastairc 16:46:37 alastairc, you wanted to comment on one of the questions "Should we replace it with a “Text selectable” provision that is more specifically geared to help text-to-speech tool 16:46:40 ... read all visible text?" 16:46:53 Julie: the “Open Questions” section in the doc was intended to spark broader discussion, but it seems like participants may not have engaged with it or seen it, and she’s unsure it generated meaningful feedback. 16:47:46 Charles: doed “text-detectable” includes OS-level text recognition (like detecting text in images). 16:47:51 q+ 16:47:55 ack Charu 16:48:17 I agree selectable and detectable are related but different needs. 16:48:23 Rachael: Great confersation point for this group, because it crosses subgroups. 16:48:25 ack GN 16:49:21 Alastair: uggested focusing on “programmatically determinable” text rather than “text selectable,” noting that visible text can sometimes be hidden from assistive tech. He emphasized that if text is presented visually, it should be programmatically available, and that visibility, programmatic determinability, and selectability are distinct concepts. 16:49:25 agenda? 16:50:08 present + 16:50:11 present+ 16:50:12 present+ 16:50:12 Charu: concern was that key open questions—like those Julie mentioned—had not been addressed, which is why they raised them in the survey, emphasizing that these issues need to be resolved before moving forward. 16:50:14 present+ 16:50:17 present+ 16:50:34 present+ 16:50:34 Thanks all 16:50:44 present+ 16:50:50 present+ 16:52:40 Gundula: asked whether certain text can be visible but not accessible to screen readers or selectable, and clarified it’s not image-based text. 16:53:03 unfortunately have fought this battle 16:53:06 Alastair: confirmed this can happen (e.g., with aria-hidden), noting it’s a flaw but not something that would currently fail conformance. 16:53:23 and lossed every time so far 16:54:09 Alastair: wrapped up the meeting, thanked participants, and outlined next steps: subgroups will review survey results, make updates where possible, and log unresolved items as issues. All issues should be addressed before the next draft, with another survey planned soon. 16:54:17 jkatherman has left #ag 16:54:20 zakim, end meeting 16:54:20 As of this point the attendees have been alastairc, Francis_Storr, shadi, shawn, Heather, Patrick_H_Lauke, filippo-zorzi, janina, giacomo-petri, Charles, Adam_Page, bbailey, kevin, 16:54:23 ... ginader, hdv, JeroenH, AWK, Jennie_Delisi, Ben_Tillyer, Rachael, CClaire, kirkwood, jtoles, GreggVan, scott, Frankie, Detlev, BrianE, chrisg, julierawe, Makoto_U, 16:54:23 ... Laura_Carlson, ljoakley, SydneyColeman, Glenda, LenB, Gez, jkatherman, Jen_G, Charu, Jon_Avila, maryjom, GN, ShawnT, InaT 16:54:23 RRSAgent, please draft minutes v2 16:54:24 I have made the request to generate https://www.w3.org/2026/03/24-ag-minutes.html Zakim 16:54:31 I am happy to have been of service, alastairc; please remember to excuse RRSAgent. Goodbye 16:54:31 Zakim has left #ag 16:57:12 Glenda has left #ag 17:58:26 Glenda has joined #ag 18:53:13 laura has left #ag 21:43:45 janina has left #ag