Meeting minutes
<wendyreid> date: 2025-12-19
<Saif3> testing
wendyreid: *gets everyone up to speed with IRC*
wendyreid: you can queue via IRC or Zoom, will keep an eye on both
Draft document
wendyreid: I have a pull request open for the draft document
wendyreid: ATAG 2.0 is published at W3C in the “TR” space (technical report). I thought it'd be helpful for us to have a copy of our own, eg to make changes or adjust
wendyreid: most new W3C documents use a tool to make the document, usually respec or bikeshed. ATAG 2 is old enough to not use either
<mgifford2> I'm back with irccloud.com (maybe it will be less of a pain.
wendyreid: so I converted the document to make it work with respec. Some links are still broken
wendyreid: *shares screen*. this is what the doc looks like. In the right top corner there's errors, currently 95 of them, they are mismatches for links / terms
<mgifford2> Think this is the right link to respec https://
wendyreid: I fixed some of them already, for some I'm not sure if the correct term is being attached
<Zakim> hdv, you wanted to offer help
hdv: Happy to help find some of the errors and fix them
wendyreid: not pressing, but hoped to get it up there so it's easier to talk about
Deliverables / areas of work
wendyreid: this group was formed with the goal of updating ATAG
wendyreid: one of the biggest drivers of this will be AI, and what it means for authoring tool accessibility
wendyreid: finding out what we want to deliver is going to help us put together the Charter, which we'll need to do for the group
wendyreid: we're a Community Group… we can write and produce documentation, but we can't update Recommendation Track documents. What we can do is incubate ideas and lay a framework for how we want to do the updates
wendyreid: at that point we can hopefully take it to a Working Group that can actually do the update
wendyreid: but the first part of it is incubation. It's the fun part, high in the sky big ideas.
<wendyreid> w3c-cg/
mgifford2: Is that Respec code up on GitHub yet?
wendyreid: yes
<Zakim> Charles, you wanted to consider version method before deliverables
Charles: I'd say there are some homework tasks before coming up with deliverables, like big picture scope. Does ATAG need editorial updates, modest updates, new criteria, an version update, like 2.x or are we making a new document?
wendyreid: with some of the discussions we had with shawn and others in WAI, we hope to align with WCAG 3. That's a big chance that will add a lot of modernisation to accessibility standards in general.
wendyreid: we spoke to the chairs of the Accessibility Guidelines Working Group… we all want to coordinate with them, not work in a silo
mgifford2: Can we also do a minor release to support WCAG 2.x? WCAG 3 will take some time and implementation and requirement by governments will take longer
wendyreid: I'd say that's a good idea. I don't think there's a group that “owns” ATAG?
shawn: my recollection is that the current AG charter says that it can maintain ATAG
<Charles> same q and concern as mgifford2. time is a big factor for staying aligned.
shawn: the current version says 2.0, it might be possible to do 2.x releases. I need to double check that
<itmaybejj> I wonder if we could make the existing one pretty and add techniques and test cases without revising the requirements.
wendyreid: current charter says 'maintain errata'
shawn: the AGWG is going through rechartering right now, it's not too late to suggest adding to the current charter
shawn: we did talk about next generation authoring guidelines going into a separate document vs same document
shawn: what we perhaps want to propose… to have the next charter enable minor updates to ATAG 2, if we wanted to decide that as an interim step
<wendyreid> hdv: There might still be time, they haven't discussed rechartering as much, there is room to discuss
shawn: depending on timing… can add it as an option rather than absolute requirement
wendyreid: I can also look at the Process to see how much of the change would fall under 'errata'
wendyreid: eg maybe changing references to WCAG 2.0 to WCAG 2 may just be errata
<Charles> note on AG re-chartering, the current charter ends 30 April 2026
itmaybejj: I wonder if it's worth looking at how much we can do updating techniques rather than the actual requirements
wendyreid: can be challenging to update non normative requirements without normative requirements too
shawn: I'll have to double check how this would apply… the WCAG understanding docs aren't in TR space anymore so there's more flexibility… a community group could possibly publish understanding and techniques for ATAG
shawn: and they could be published on the W3C website, if approved. That's what the ARRM group does, a community group that mgifford2 is involved in
<mgifford2> ARRM Link https://
<itmaybejj> Interesting.
<mgifford2> Our work in GitHub to push this forward w3c/
wendyreid: they would be Community Group reports
shawn: yes
wendyreid: I also don't want us all to get too caught up in the details of how the publishing workflows work… people like shawn on W3C staff are always super helpful to help us make it happen
wendyreid: there will be ways, so feel free to share ideas you have about what to publish
wendyreid: there's possibly also work in improving the structure of ATAG
wendyreid: eg the first requirement of ATAG is to meet all of WCAG… that could freak people out as it is a very large requirement
wendyreid: possibly we need to consider a different conformance structure, as the current one is fairly complicated
wendyreid: and then there's different tools too
wendyreid: and considering how we can look at the role of AI
<Zakim> shawn, you wanted to note Implementing ATAG https://
shawn: there's also a document called 'Implementing ATAG', a bit like Understanding and Techniques
shawn: would that be a useful starting place, or not?
<Saif3> There is a new thing in AI world they are called "Generative UI" which we can recommend including in an AI section.
wendyreid: question for people here… why are you here, what are you hoping to get out of your participation in this group?
mgifford2: something I had wanted to work on… can we use AI to make a complicated process like ATAG simpler, so that more authoring tools can complete the process?
mgifford2: there's a huge barrier to entry to begin with this
Miriam4: for WCAG, we have a lot of materials focused on specific topics like altenative text and subtitles… when I tried to implement ATAG at my place of work, content management system and education tools, it was hard to tackle it 'as a whole thing', there are not enough specialised resources and tutorials to look at
Miriam4: there's a lot more content for WCAG. And ATAG is seen as something to implement 'whole or nothing'
Miriam4: if it can be broken down in more parts in specific areas it could be easier to start implementing it and easier to make a business case
<hdv> +1 Miriam4
<mgifford2> Progress over perfection
Miriam4: and makes it easier to tackle one area at the time
<shawn> [ shawn notes to self : ATAG missing the supporting material that WCAG has, e.g., tutorials, ... ]
<wendyreid> hdv: Wanted to +1, great point, ties in to what I was saying earlier about meeting all of WCAG, doing that requires a lot of planning, breaking it down would help
wendyreid: earlier we were talking about different types of authoring tools and use cases… I work on an ecommerce site, the requirements are different from those for a government site. Same is true for different authoring tools
wendyreid: it could be a venn diagram, perhaps with a lot of overlap, but the focus is pretty different between different types of websites.
<wendyreid> hdv: Totally, I remember now we discussed this during ATAG promotion, looking at different sectors and making examples for each
<wendyreid> ... alt text in education, the text can't give the answer away
<wendyreid> ... on a non-educational site, it's completely different
Miriam4: also think it's a good approach to make very specific use cases… or set specific borders on what a company is working on… eg if they make a CMS, do they tackle the developer side, or future customers, or the output…
<shawn> [ these resources are promotional, rather than implementation support: https://
Miriam4: we can show which parts of ATAG apply to which part of what they're working on, making the use case more specific.
shivaji: shawn, did you suggest we can make an implementation document or is there one already in existence?
shawn: it already exists
shawn: the best place to start is the ATAG overview
<shawn> ATAG Overview https://
shivaji: my other question: I had taken a task to do some research on AI and how it impacts different tools… what I found was that focusing on different areas, like Miriam4 and others suggested, would really help
itmaybejj: was also thinking of the examples the ARIA WG published, it was successful and everybody started using it. Presenting things as examples, I could see us have a lot of impact. Most developers are in the weeds, rather than teaching them a lot, giving them 'here's what the HTML needs to look like' might be more appropriate
<Zakim> Charles, you wanted to clarify this breakdown of applicable criteria
Charles: we're still talking about deliverables, figuring out what's next. We have a lot of ideas.
Charles: a way to make ATAG easier to assume by breaking it apart in what's applicable and what's not, I think doing that by sector, is probably something we can't decide yet as we'd need to define what all those sectors are. Much like the market sectors in the original implementation report
Charles: perhaps the list of deliverables could consist in coming up with list of ideas for simplifying ATAG
<mgifford2> It woudl be good to have a discussion about ideas for simplification here https://
wendyreid: possibly… I think use cases are really important for this
mgifford2: we do have a discussion space in GitHub where we can keep track of them
wendyreid: yes we do have a thread for use cases
<wendyreid> https://
<itmaybejj> You asked why we're here -- I think I'd summarize for me as "plain language and examples"
Saif3: is it things we expect or where things are heading, or more what we have right now?
<itmaybejj> That's where my head is. I tend to be much less interested in the spec details than the implementation docs.
wendyreid: standards can take a while and it can be hard to predict the future, but we know there's trends. So there's an element of flexibility we want to build in
shivaji: re use cases… will we have use cases from tools like CoPilot or code generation, not just content creation?
wendyreid: yes, with use cases, not thinking about specific products but specific needs for a specific user group, eg 'as a user, …', 'as a teacher, …'. Have any tool in mind, but want to think about the goal and desired action
<Charles> idea: a path to simplification and to include AI as a tool might be to look at ATAG by agency – things the tool does and has agency for; and things the human does and has agency for. this way the type of tool matters less. at least from an information architecture perspective.
wendyreid: I set up our meeting cadence for every 2 weeks…but have cancelled January 2nd. We'll meet again January 16th
<Saif3> saif3: Is there a similar group that recommends accessibility guidelines for software developers (non-training tools)?
wendyreid: this gives us lots of time to think about use cases… feel free to contribute to the open discussion threads or open a new one if there's something you want to discuss
wendyreid: Saif3, that's WCAG
<Charles> note: i do not know github well enough to create a new discussion thread
wendyreid: WCAG 2 is focused on web content… there's no spec or standard for how to do everything in every language or framework, but there's like framework specific documentation that provides that.
Saif3: just thinking of scope, we don't want to go into software development if other groups are going into that
wendyreid: in general WCAG takes the overall, we're more narrowly targeted at authoring tools and their output
wendyreid: thanks all!
<mgifford2> So what should go here, vs Slack?
<wendyreid> Just meeting minutes :)
<wendyreid> Since most people (myself included) don't have perpetual access to IRC.