Meeting Minutes
Wendy Reid (she/her): That's good.
All right. Why don't we get started?
the first thing that I had on the agenda this week was, a call was put out
For,
breakout sessions at TPAC. So, typically at TPAC, the Wednesday is set aside for breakout day, and breakout sessions are essentially just.
One hour, slots,
Where people can come and present on various topics.
it's kind of like a fun, kind of fun break in the middle of the week.
from the standards work to talk about it's all standards-related, but it's often a chance for groups to maybe present on an idea that they have, or for people to maybe get consensus around something that could.
Be a group or be brought to a group as a new feature.
I got the reminder email about breakout sessions, the other day, and I was like, I it hit me that it might be worth considering if we would like to do a breakout session.
it can be a good way, too, to kind of introduce work to, the broader community, because Breakout Day is kind of one of those days that everyone.
It doesn't have stuff scheduled except breakouts, so it's easier to get people to attend.
but I wanted to run it by everyone to see what they thought about us proposing a breakout session, and if we did, what might it be about?
Jutta Treviranus: I think that'd be a great idea if we did that. Certainly to make people aware of what we're doing and what's at stake.
the I'm going to, unfortunately, be, presenting I'm doing a talk at the opening plenary of the ISOSC42 at that time, but,
Wendy Reid (she/her): https://github.com/w3c/tpac2026-breakouts/issues
Wendy Reid (she/her): I just posted the link. You can kind of get an idea of what the topics are just by.
seeing what other sessions have been submitted. None of these have been accepted yet, notably, so it's not
Looking at what's been proposed is not a guarantee that.
This is what's accepted. Notably, there's a lot of it, like, Agentic Web.
Submissions.
But,
so I'm noticing it's interesting, only a few years ago, there was, I think, one session about AI, in the whole day, and now it's probably gonna be, like.
50%, I don't know, maybe more.
Ned Zimmerman: It looks like there might be a little bit of overlap between issue 12 and what we've been doing, which is the web accessibility evaluation using AI agents.
Wendy Reid (she/her): Yeah, I think if I'm thinking about topics for what ours would be, it's probably, like, kind of, like.
Maybe it's like concept of like reintroducing ATAG, but talking about like ATAG in the new era.
And, like, I mean, maybe we can give it a fun title, like, Reimagining ATAG or something, and it's
Talking about our
about our work.
And maybe gathering some use cases from the community that we didn't think about.
Shivaji?
Shivaji Kumar: So we can collaborate with others to do the session, or it has to be independent of others?
Wendy Reid (she/her): It doesn't have to be independent of others. Typically what tends to happen is that people
see it on the schedule, and they see if it's a topic that interests them, they attend. And usually, I've seen most sessions go this way, where, like.
No more than the first half is like a presentation of sorts. Like, here's our presentation on this new thing that we want to do or this new idea that we have. And then the rest of the time is dedicated to discussion, you know, answering questions or raising ideas.
We can the fun thing with breakouts is we can structure it any way we want. We could do the
presentation, we could
You know, keep the, like
explainer part of it quite short and say, hey, this is actually we want to have an open discussion about what people want to see in ATAG or opportunities. We can structure it however we want.
Jutta Treviranus: I mean, one other thing that we could do is to pull in more people to explore the idea of greater adaptability, sort of a re-look at compliance, or what can be achieved in terms of.
diversification of accessibility in and with AI.
there. The evaluation seems to be largely, you know. Let's use AI to test the same things as before. But if we go, if we open up the topic of.
How do we make WCAG more adaptable or the notion of accessibility more adaptable to encompass more?
but that may be too thorny an issue to or too controversial an issue to bring up at at a breakout, but, I I think it would
attract discussion.
Wendy Reid (she/her): I think it would attract discussion. I also have a feeling it might get interpreted as us trying.
to step on some toes that I don't think we're intending to step on.
Jutta Treviranus: Right, yeah.
Wendy Reid (she/her): Conformance is a very popular topic.
Jutta Treviranus: I know.
Shivaji Kumar: Yeah, so
And we could
Just, you know, going back to your initial thought about.
it titled something like, Reimagining ATAG in the New Era.
might not be all that offensive to anybody, because we're just exploring it, right? And so is everyone.
So, you know, it is naive to some extent, but also at the same time, we're trying to explore new territory, and if you want to stick with something like that.
Wendy Reid (she/her): Yeah, I like the approach. I've always liked, with breakouts, trying to invite conversation because I do think it's a rare opportunity at TPAC to get
people in the room that
May not ever attend a meeting.
they might be interested, but it's just, like, a time commitment they can't make, or you know, they're vaguely interested, and they might be interested once we publish, but they're not interested enough to, like, join the work, or anything like that. But it is a good chance to get, you know, some, like, voices in a room that you may not hear from otherwise. Like, I've definitely
I tend to attend breakout sessions that are either, like, stuff that I'm interested in, but also sometimes just stuff that, like.
I find interesting, but I maybe don't have, like, any stakes in. And it's always been, like, really cool to see what
Is going on.
And also sometimes bringing new perspectives. Like, I've occasionally attended ones for, like, CSE, where
they're presenting some, like, really interesting concepts, but they clearly never thought about accessibility. And so it's like, I get to be the person who's like, hey, have you considered keyboard navigation? Like
So it's a fun day. I will say it's a really enjoyable time.
Jutta Treviranus: And would our goal be to go beyond, like, the Way community to the other, yeah, W3C members and
Wendy Reid (she/her): Yeah.
So the breakout sessions are available to literally everyone attending. So you do see a lot of cross-pollination.
Depending on topics, you can expect some people, but
Jutta Treviranus: But, like, would we aim
Yeah, but we could specifically aim to gather, like, to garner more interest from a broader group.
Wendy Reid (she/her): Yeah. And I think I would try and frame the description of the session in a way that
maybe leans a little more on the, like, AI side, just to to pique the interest of
Wendy Reid (she/her): Maybe people who don't consider themselves accessibility interested, but
Jutta Treviranus: Yeah, even a primer on, you know, what are the implications of AI for disability as part of it might be of interest to the larger group.
Wayne D: what about something like imagining,
Authoring tools.
Modern authoring tool.
Accessible authoring tools.
I mean, not even presented as
Jutta Treviranus: Yep.
Wayne D: Taking on a guideline at the moment, but
envisioning what
We can do with AI authoring. I mean, it's.
Anyway.
Lisa Liskovoi: First of all, I'm gonna plus one Wayne's idea. I think that would be a really great topic of conversation. I also love the idea of having it really hands-on, like, to keep our presentation sort of to a minimum and dedicate the majority of the session to discussion.
And I was actually wondering if we want to go.
And that just in terms of framing it.
more so something like imagining the future of authoring, rather than having like, I'm almost thinking there are so many sessions with AI in the title.
we're going to be talking about AI realistically, no matter what session, probably in all of them. So I wonder if we want to kind of differentiate ourselves by actually not mentioning AI explicitly, but still talking about it.
Wendy Reid (she/her): All that kind of fun stuff. Stuff about policy, which will be — I have a feeling I'll be attending that one if I have the time.
Any other ideas for possible breakout session?
I mean, I think we have quite a bit, so I'll try and draft, like, a proposal, and I'll share it with everyone. the deadline to submit, I think, is October 8th, so we have a little bit of time. not a ton, but, you know, enough to
I'd have to draft something and get feedback on it, so
Hidde de Vries: Can I ask, do we what do we want to get out of people, in the session? Because sometimes breakouts are really good to, like, get feedback from people that maybe don't have time to normally join the meetings, but they happen to be walking around on the hallways in Dublin.
Wendy Reid (she/her): I will end.
Hidde de Vries: we got them into the room, and they suddenly gave us great, great ideas. Are there, like, specific things that were
looking for, I was thinking, like, definitions of authoring tools, but we've talked about that quite a bit, right? The different types, but maybe there are other things that we could use.
Mike Gifford (CivicActions): I mean, I do wonder if the,
since there's been a change with the how, WCAG 3 is being what the scope of WCAG 3 is.
in the latest draft, and this is a smaller footprint or smaller expectations than what we were, than what people were hoping for. I mean, and ATAG was one of the first things to get dropped from that WCAG 3.0 release, and maybe there's an opportunity to just sort of talk about.
how how important it is, and how strategic it is to go off and to think about the authoring tools simply to go off and to say, if we want this to be successful, we have to be able to like. It all starts with the authors like.
If we overlook the authors and just look at the output, we're missing the value chain in fixing these issues.
Jutta Treviranus: Yeah, I agree with Mike.
Wayne D: Yeah, I agree that I think that's really the bottleneck is authoring.
And.
So, and I think that authoring accessibly.
is so tractable these days.
As opposed to when ATAG was written.
Jutta Treviranus: Yeah, and a focus on strategy as opposed to specifics, I think would be good, sort of looking at the systemic.
approach to, achieving the goals of WAI.
Hidde de Vries: Yeah, I agree with that, too. I do think we shouldn't make a session that's against WCAG 3, necessarily, saying as a you know, someone in that group as well, they're they're trying hard. and I don't know if
Authoring was
explicitly dropped for all purposes, like, I think
there are still benefits in the WCAG 3 approach that also benefit authoring, so
yeah, maybe focus on the strategic, kind of, where does authoring fit?
in achieving your accessibility goals, without mentioning WCAG 3, it'd probably be a good way to go.
Mike Gifford (CivicActions): WCAG 3 is an epic journey. Definitely don’t want to diss anything they are doing.
Jutta Treviranus: And, a number of you weren't here, but, the, the other thought I had is to look at how we can expand the, the sort of adaptability, flexibility of.
what is accessible. So, and I don't think we need to do it in an anti-WCAG 3 way or an anti-WCAG way or any sort of controversial way, but is there an opportunity here?
like a generational opportunity to rethink the bounds of accessibility or how accessibility is defined.
Meaning more differentiation, more a greater nod to the heterogeneity of needs.
greater flexibility in what the end user can expect or see.
Wendy Reid (she/her): Yeah, I think
I was struck yesterday by someone, something someone shared at work, and it was about this new, model.
They were working with, and this model is,
Much, much smaller and more focused on.
Like, browser testing.
and at a much lower cost than other LLMs.
And one of the things that it's very clearly highly dependent on is being able to traverse.
the tree, like the accessibility tree and the DOM tree of a page.
and they were showing a demo of it completing a, task on Google Flights, like, requesting a like, putting in, like, the information to, like, see the prices for a flight.
And it was able to do that in 7 seconds, which is really fast.
And I was watching it and watching it go.
I'm reading their explanation, and it was so clear that.
It only worked because
That page has relatively good semantics and accessibility.
You know, it knew that, oh, that's a button, that's a button, link, combo box, like, and it knew what to do.
And my first thought was, I would love to test this on a page.
That is a web app that uses.
Divs with click handlers.
for everything just to see if it can figure it out or fall to just absolute pieces because it doesn't know if that's a button or a link or even if it's interactive.
And that's the thing that has been striking me so much about the AI push is the pages have to be accessible. You have to use semantics correctly.
Jutta Treviranus: Of course, yeah.
Wendy Reid (she/her): You have to use ARIA.
Jutta Treviranus: Yeah, and that I mean, our UI options had the same sort of experience. It was very much dependent on.
Wendy Reid (she/her): Yeah, everything is dependent. Okay, I have some ideas floating around in my head of what we could do. So I'll try to put together, maybe put a couple variations on a possible breakout. I mean, arguably, we could attempt to submit multiple. It might be a bit of a push.
But we could do it with no guarantee that.
one or any of them get accepted, of course. So there's still that element to it.
But good brainstorming. I like this.
Let's pivot now to our
other topic for today, which was, I'm going to put the link in the chat.
Wendy Reid (she/her): Getting back to our use cases and requirements.
Document.
And, as I said last week, what I think I wanted to kind of push us back towards looking at was
the
Getting back to the use cases and requirements, right now we have
and just
I'm on the third tab, which is the cleaned-up version of the use cases and requirements document.
Jutta Treviranus: Some interesting thoughts: https://theopeneyedman.com/mainstream-commercial-interest-aligning-with-accessibility-needs-double-hurrah-with-extra-chocolate-on-top/
Lisa Liskovoi:❤️
Wendy Reid (she/her): We have right now,
a lot of use cases, which is amazing. They are very highly specific, which is also really good.
I do think we probably want to take what is now, like, a
Very, very, very big and expansive list and turn it into some more core.
Use cases. we already have, like, these
Kind of third level header headings.
that we put together, like, users must be users must be able actually, I'll share my screen, so I'm not
Some people can follow, but I'll make sure I'm saying the headings as well.
But so we have, like, users must be able to author accessible content.
And then we have many, many — this is one of the examples where we have a lot within just that first section.
And I'd love to be able to kind of take it.
To, like, more canonical.
Instead of, say, authoring actions.
Authoring use cases, and then do that for each.
So we can start getting this closer to, like, something that can be
Maybe our very first publication.
What I'm gonna do, though, so we don't lose this kind of original
Because I'm going to duplicate this.
so there's gonna be a fourth tab, just FYI.
We're editing.
ATAG future.
Okay.
I also realize when I do this view that I can't see anything like the chat anymore.
Nice.
Will you let me see you?
So Zoom's — oh, there it is. It's behind my other window, of course.
Zoom's interface is a bit wacky.
That is a really cool — I will check out that link later.
Okay, so if we start with the very first, users must be able to author accessible content. We have a bunch of examples in here.
Jutta Treviranus: The reemergence of decoupling the UI from the underlying function and direct accessibility for Web and general UI is having a moment.
Wendy Reid (she/her): from different perspectives, obviously, like, as a student, I must be able to conform to the guidelines of my university.
as a marketer, I must be able to create and assemble packages of content.
As a scientist, I must be able to add and work with math formulas and scientific notation.
How do we
I'd love to be able to distill a lot of these into like
Jutta Treviranus: This harks back to Sun Microsystems work and early work in OS2 by IBM and Jim Thatcher
Wendy Reid (she/her): Some essential use cases.
Jennifer Bullard: so, and I'm, I'm new here. I think that I'm just looking over this, list and trying to wrap my head around it, and it seems like it would be useful to group it by user type, because the usage scenarios do tend to be, you know.
somewhat, like, there are a lot of usage scenarios for students here, there are a lot of usage scenarios for teachers, and so, like, as I'm reading it, it seems like I'm reading as a student, I'm switching to a different user, and then I'm back to a student.
Jennifer Bullard: and then, I don't know if you ever do this with your use cases, but it's kind of nice to have a follow-on statement. So, like, as a student, I must be able to blah blah blah
you know, what is what is my goal, right? but then it's kind of nice to, like, break it down. Okay, to achieve this, I need XYZ, where the XYZ here could be, like.
What does this mean in terms of like?
an accessible experience? Like, what is specific here to
Accessibility. Those are sort of my thoughts.
Mike Gifford (CivicActions): I don’t know if Jim’s related work from that time. He did have such a big impact.
Wendy Reid (she/her): Yeah.
Do you like the idea of like
We do have some
duplication.
Evelyn Wightman: I also see themes of type of content.
Like across both student and marketer, they're talking about.
creating videos or multimedia things. And those are probably
Closely related actions.
Wendy Reid (she/her): Yeah, that's exactly what I'm
what I think
at the end, even though there's, like, very different roles, or they I guess, correction, they look like very different roles, but there's a lot of overlap between what, say, a student wants to do and what a marketing professional wants to do.
They may be different tools, but like the thing that they're trying to, the action they are attempting to perform is the same.
With very possibly very like what Jennifer was getting at, like similar outcomes.
Wendy Reid (she/her): Or similar expected outcomes.
But in a very different context. So it might feel different, but it's actually like a very similar task.
So if we
2, 8
I think you had a, I really liked Dr.'s idea of like the outcome.
The sound. So I have, as a user, I want to be able to create multimedia content, including text, images, video, and or audio, that is accessible to the standards of my organization in order to ensure my content is perceivable to anyone who will use my content.
Jennifer Bullard: I think if we're focusing on this one specifically, there's also a,
a nod needed to, like, where that content is going to be posted, or how that content is going to be shared, because the accessibility has to travel with the content. Like, it can't be
app specific.
Necessarily, like, if you author in one place and post in another place?
Wendy Reid (she/her): I'm trying to remember, I think we have a
created these kind of like higher level.
headings.
And I'm trying to remember if we had one about.
This is maybe the closest this one might be the closest one. Users must be able to save or export content in multiple accessible formats and with any accessibility properties preserved.
Shivaji Kumar: Well, so I'm thinking
So when you say that, you write out the last thing, and I'm
I've been thinking about that. so when the content is being converted into multiple.
formats, and those formats should be accessible? Does it mean that the authoring tools should have.
Capability to turn text into, let's say
an audio, produced, speech-based thing.
You know, people can hear it instead of reading it.
Jutta Treviranus: Replying to "I don’t know if Jim’s related work from that time.":
He proposed an alternative to the Model View Controller, direct access to functions by the UI, not interpreting the visual or the keyboard/mouse control but a completely adaptable link to the underlying functions.
Wendy Reid (she/her): That's an interesting one. Yes, I can hear you. I was just thinking.
Wayne D: I think that what our goal is, is that the authoring tool automatically produces semantically accessible structure.
And that
that a semantically accessible structure can construct.
text-to-speech, braille.
Large print with word wrapping. Important accessibility presentations. I think that that's what we really hope for. I mean, once you
I mean, once you have that format, you can use it, you know, as a person who wants to read a document, who wants to navigate a web page. If it's semantic, you know, you're gonna have you're gonna be able to find a tool to search it.
Shivaji Kumar: Got it. So, I what really got me thinking about this is that I,
So there is a text which is available for me to read with screen readers, but
If I want to read it on a braille display, it has to be in a particular format, BRF format, right?
So,
That does not
exist today universally. So BRF format is very specific.
and it needs some work in order to make it accessible on displays, braille displays.
Shivaji Kumar: Right? So
Yeah, that's where my thinking started.
Israel Cefrin: Hello, everyone. Good morning. Just on that point for the accessibility property preserves, my question is, these accessibility properties are the feature itself, like headings, image description, or —.
All the expertise of the content or do you need to be more specifically say about accessibility metadata as well? So all the properties include the features of the accessible content like headings, including metadata or are these separate properties?
Just for modification of my anger.
Wendy Reid (she/her): I think I was taking properties to be inclusive of metadata.
Israel Cefrin: Okay.
Wendy Reid (she/her): but we can also be
We can be prescriptive.
Israel Cefrin: Okay, thank you.
Yeah, so just a background here. I was just reading the European Accessibility Act. And they say that the accessibility data is not an optional so you need to declare even though the code is not accessible. So it's more like a regulation.
from the European and everything, every part of part that works with them.
should be aware that we need to say, okay, is it accessible? Yes or no. So this silence is not an option anymore.
Wendy Reid (she/her): Yeah, for sure.
I do wonder what the
Shivaji, like the case that you brought up of like.
Converting to, say, like, a
Actually, like, the examples that I know the most about are, like, accessible EPUB to
to Braille.
Or eBraille. Umm.
Shivaji Kumar: Yep.
Wendy Reid (she/her): And how right now, like, I think there is this kind of implicit
assumption that if it's an accessible EPUB with accessible HTML, that
It will just work.
on a Braille display.
Without some kind of conversion, but I also.
I haven't looked into this in a while. There was, I think, some work going around around, like, essentially a standard for eBraille, because there wasn't, like, a consistent one before this, and so that might have also been a blocker in the sense of, like, without a consistent standard for eBraille.
it's really hard to make it, like, a conversion tool, because you don't know what you're
what you're converting to, I guess.
But that's
It's a good thing to keep in mind.
Maybe, like, another way to look at that actually is
I'm trying to think of how to frame it, because it may it may be less that we expect that every tool should magically be able to or should be expected to be capable of conversion.
but that
As long as the source is accessible in its format.
that other tools can take that format and convert them.
It's a very cotton.
I'm clumsily trying to describe this.
Because it may be, it may be, like, an unfair expectation to say, like.
You know.
one tool that creates EPUBs must also be able to convert those EPUB files into eBraille, into PDF, into a Word document, into
LaTeX into all these other formats.
Because there's complication with all of those.
Shivaji Kumar: And also into text-to-speech.
Jutta Treviranus: Yeah, well, this gets back to our question of the fuzzy boundaries between the user agent, the authoring tool, and the content, right?
So how do, what are the boundaries of.
The authoring tool. I think a lot of those,
boundaries that we've created are being redrawn with much of this.
Wendy Reid (she/her): I'm going to put this in our research question section. What are the new boundaries between.
Answering.
Cool. The fine tones.
User agent at present input.
Jutta Treviranus: Yeah, and, and of course there, the, the assistive technology, or the user agent, or the, the agent, comes into play as well. There's, there's a few new.
roles or functions.
Hidde de Vries: possibly relevant, TAG finding on agents https://www.w3.org/TR/web-user-agents/
Wendy Reid (she/her):👀
Jutta Treviranus: But one thing that becomes clear, the labels, the markup is really critical here, or it becomes even more critical.
Wendy Reid (she/her): Yeah, that's it.
Yeah, I was just thinking about with.
with what Shivaji was mentioning.
Text-to-speech, for instance, is a function of the user agent. It's not actually a function of the content itself.
It's a it's a
I know way too much about this from my years in publishing, but text-to-speech is a feature that must be built by the user agent that consumes the content and displays it.
And
All the content can really do.
is be formatted appropriately to facilitate that text-to-speech interaction.
But it doesn't actually like, you don't write a text-to-speech version of
a document.
Shivaji Kumar: Yes. Yep.
Jutta Treviranus: Yeah, and this harks back to SGML and the XML,
Write once, present multiple times, and the flexibility that that allowed.
Evelyn Wightman: The “text” part of “text to speech”
Hidde de Vries: I think in a way it can also be seen as an extension of the headless CMS kind of thing where people are making content for whatever, like the content isn't for this specific website.
But it's, like, for whatever wants to consume it.
I think that started with, like, doing things in XML, and it is now all the rage, I think, around CMS vendors that are
Smart about things, like, you know, make your content once, put it in your website, and your app, and your
And your agent, like, they're just different, things that consume the same structure in a way.
Wendy Reid (she/her): Great. So just added this new one.
And it might even be a replacement for the one above it. As a user, if I have an accessible source document, I expect a tool that can perform conversion of that document to other formats to be able to preserve the accessibility properties and metadata of the source so that accessibility is preserved in the new format.
Mike Gifford (CivicActions): Tricky with copy/paste but important.
Hidde de Vries: It looks good to me. I wonder if there are formats where some of the metadata is different, or needs to be different, or something?
Wendy Reid (she/her): I was thinking about that.
Hidde de Vries: Yeah.
Evelyn Wightman: I was gonna circle back up to
It seems like we've got these high-level headings that are
maybe sort of the list that you were trying to make of the compacted list of high-level use cases. Is there a way to just, like, look at just those? Oh, you've got it there, expanded, cool, yes, thank you.
Wendy Reid (she/her): Here, let me Unfortunately, some of them are so long that I think they, like
Truncate.
Israel Cefrin: Yeah, so just about the question from Mike about the metadata. I'm just placing the chat a link to a possible schema for exhibited feature, and they provide some examples for metadata for different formats, RDF, JSON, or that's all.
Israel Cefrin: https://schema.org/accessibilityFeature
Israel Cefrin: we can
Maybe recommend I use like the marriage schema or international standard.
or just mention something like that. So I expect to different formats provide some kind of an idea in different formats as well.
It's not a single format.
But possible ways to represent the metadata and external metadata in different formats of content.
Israel Cefrin: I'm sorry if there's any noise behind me. I'm in a coffee shop here in my car, but I hope that's fine.
Wendy Reid (she/her): There we go. This guy. Yeah, I —.
Actually, maybe I can
Oh, my gosh. I'm trying to find the link for this. It's always
Publishing accessibility metadata. So I think almost all of the schema metadata is
derived from
The publishing.
I think we're the ones that submitted it to Schema.
I don't think this — this is the display techniques document, though. Where's the original? Is it here?
Yes.
And that's the package meditating. There's so many documents.
This is maybe like it's the most human readable.
Why does it switch to dark mode when I do this?
They have a weird bug in our
Some of our documents.
I don't know why it's in dark mode.
Wendy Reid (she/her): https://w3c-cg.github.io/publ-a11y/metadata-display-guide/guidelines/
Wendy Reid (she/her): Yeah, Evelyn, I guess go back to what you were saying.
So we have these, like, high-level, I think, headings. I do think there's still even with these ones, there's still, like
specific behave like, use cases within them that are
I think, like, just to get to, like, a good example.
So, like, users must be able to author accessible content, but there's, like
Higher order access, like
Use cases within being able to author, I guess is the way I would.
If that makes sense.
And I think there it's once you like if you look through the very, very specific.
Use cases, you can see the
The patterns of the
I guess it's like taking hyper-specific and turning them into general.
use cases.
Evelyn Wightman: So for that, does taking away the specific, like, marketer, student, scientist, and just
Sort of collapsing that into.
As an author, as a
authoring tool.
creator as a
consumer of
Stuff that has been authored.
Wendy Reid (she/her): A little bit, but I think it's also thinking about, I guess the example that spawns the first one was the student example was, as a student, I must be able to conform to the guidelines of my university.
and ensure readability of my documents, and as a marketer, I must be able to create.
and assemble packages of multimedia content.
And, of course, the core
They sound kind of different, but the core thing is I want to be able to create content.
That is that meets organizational standards and.
is accessible to anyone who will use it.
Wayne D: We also have a whole category of authors that aren't people, like content management systems, learning management systems.
version control systems.
And they're very adept at just throwing in inaccessible stuff into a document that you've already.
attempted to create excessively.
Hidde de Vries: Without any users?
Wayne D: Well, you don't have a lot of control over the database that you're pulling information from.
Or the software that's creating the.
User interface.
that you're doing unless you're in an organization that supports that.
And so, I mean, there are really non-human authors, a lot of them.
Wendy Reid (she/her): Does that mean that some of our user needs are not about?
Humans, but they're about.
Systems? Or agents? Or
Wayne D: Yeah.
Yeah. Yes.
Jutta Treviranus: And this was previously the category of what was authored by default, right, by the authoring tool, that it, that needed to be accessible. But this expands it quite significantly.
Evelyn Wightman: I'm very hesitant to
treat machine needs as user needs.
It could get convoluted, but I think you can still frame
The needs or the requirements for tool interactions.
as user requirements, like, as a human author.
I need
This is like the database I use to preserve or not inject whatever stuff such that.
This thing, this
accessibility that I'm building is preserved.
I don't know. I'd want to at least try to phrase it from a human perspective before.
framing use cases for machines.
Hidde de Vries: Yeah, I agree with you, Evelyn. I think
Yeah, if you're describing user needs, it makes sense to center them.
around humans, I think we can
Frame a lot of the.
things that Wayne was talking about as things that humans want.
and I think philosophically speaking, like, having
attributing a need or a want to a system, doesn't doesn't make a whole lot of sense, I think.
Those things fundamentally need humanity, like, to have a desire, or to have, like, a
something that you want to be the case. so I also think we're making it much easier for ourselves if we phrase it in terms of humans rather than.
then systems. and the third thing I think is,
Ultimately, most content will be the responsibility of humans. So even if you use AI or agents to do stuff for you.
You are the person publishing it.
We just had the Dutch Prime Minister being called out for being using AI for his speeches, and he was definitely personally blamed.
for the speeches, and, he couldn't put the blame onto the system, right? so, yeah, I think that's culturally also important to,
center these things around humans rather than systems.
Jutta Treviranus: Yeah, those functions are an authoring tool according to our definition. So they are authoring something and therefore they should fall under the requirements of an authoring tool. I think we can use the fact that these.
other authoring actors or whatever require the structure as a way of making it clear that these are important things that go beyond accessibility, but I think
The way that we've looked at what is an authoring tool is anything that's creating, and so those are tools, and they would be a part of what is required of a tool.
Wayne D: And there are also the administrators. I mean, there's the learning management system administrator.
And that's a person or a group.
And so they have the responsibility of.
Maintaining
that system in a form that it produces accessible content.
And the the learning management system.
Developer
Should.
produce a product so that people can upload stuff that is accessible and remains accessible.
when they put it.
put it into the system.
I mean, I'm not saying this from inexperience, you know, I've seen people do perfectly good stuff, run it through the learning management system and have it not be accessible.
Jutta Treviranus: It is about human choice offered. If the choice is restricted, the default should be accessible.
Wayne D: It took well, I mean, it took us a lot of time to get
Blackboard to be accessible as it is as accessible as it is.
Like, we had to shut off their access to the California State University system.
for a year before they came around.
Wendy Reid (she/her): Yeah.
We're at time. This was a really good discussion.
I'm gonna try and clean up this document to group things up to make.
This a little bit easier.
but I think we made some really good progress and had some really good discussion. I'll also prep some, like.
Example. Breakout proposals. So for people to check out.
but
Thanks, everyone, for another good meeting, and we'll talk again next week.