– DRAFT –
Accessible Platform Architectures Working Group Teleconference

SAZ: Thanks e'one, we had intro meeting two weeks ago

Looking at research and organising symposia etc

Looking at more indepth analysis

Interest was strong.

To make use of interest and expertise

The question is how to use the opportunities of the symposia, round table discussions etc

This is my understanding - the question of if we use this for specific topics, or do it more broadly

to frame questions or look at bigger gaps etc

Sorta meta research, or looking into what is needed

To allow for follow on work and contribs

We can also do a combination of that

First two could be broad, and then more specific

or create an exemplar

SAZ: I'd like to get a sense today of what we should do

And what the first symposium should be

There has also been work done on keywords in conference proceeding etc

SAZ: We can also look at what was not looked at

<CarlosD> https://drive.google.com/drive/folders/1uLkgAvLGFr3UApAaOtrG4guPZRo64JqK

Carlos and Leticia from FCID

JS: Should we start with something specific and end up with more meta topics?

SAZ: We could

CD: We could combine them - do them other way around

Could be done at the same time.

We have possibilities

JS: I have a suggestion..

We could frame this both ways...

The most immediate issue in APA, is from Google and Vispero

Both stating a problem.. relating to pronunciation

How do we design this in a way that works via an AMM or something else into UI A11y or iAccessible2, various APIs

The spec will fail if we can't solve this

JS: There are also XR meetings relating to AOM, that we've been invited to.

People are talking about APIs, their limitations, what do they need to do etc?

Could be framed as a research question.

JW: Jason and Josh

JW: System level interfaces is a good topic

JW: Based on recent comments in APA work, this could be turning into a more urgent issue

JS: It is for pronunciation - but it is in general an emergent issue

<janina> Joshue108: Agrees on framing

<shadi> JoC: short-term need may be driver to look into this issue

<janina> Joshue108: Suggests the short term need provides a way to possibly structure a broader conversation around these issues

<janina> Joshue108: we do have a broader sense that aapi issues are arising

<janina> Joshue108: also notes that groups outside W3C are also looking at this

JB: This direction is good

SAZ: Other reactions?

JW: If you want a topic then this is a candidate/

CD: My reaction is that this fits quite well, in terms of stuff in our keyword search

If the research community is not aware of this, then we can do that.

CD: I do have some project concerns, deep analysis may be hard if there are no publications

JB: We ran into issues in the past with lighter weight approaches?

SAZ: We have no required format - can be as we want.

There is the issue that with the less publications there are, it can be tricky

CD: There are enough people here to come up with good ideas

We would need to connect this with what the community is doing.

CD: I wonder if the first symposium should be helping to build bridges

<janina> Joshue108: Should look for greater coop between W3C and other communities; this is an opportunity for that

<janina> Joshue108: Especially when they propose tech solutions in areas we care about, such as XR, WoT, etc; where it's hard to find a locus of control

<janina> Joshue108: Not sure we should go all the way into an area where there's no research--but maybe we do?

SAZ: I'd like to switch a little..

Can we derive some criteria for topic selection?

There may be urgent topics, that can help an existing issue

another angle is impact in accessibility

There could be criteria so we say - as Jason indicated there could be other topics

We dont need to set those now, but we can have candidates.

Across certain criteria

This gives a paper trail etc

JW: Seems that topics that affect multiple aspects of a11y would be useful

There are both sides about prior publications, or could mean if there are few, then the reserach community is not looking at an area.

I don't think just having recent publications should be a deciding factor

We can discuss short and long term

We should have envisioned outcomes also

What is the benefit of looking at x or y?

SAZ: Thats helpful

I agree but are recent publications a sign of something?

Could show gaps etc

We should be thinking about this either way w other criteria

JW: It connects with my larger concern around architecural issues around how a11y is implemented and in general dont attract the research community in a substantial way

Focus on empirical HCI stuff

If we could stimulate the architectural discussions that would be good.

JS: I'd like to see TTS engines work in a multi-lingual manner

There is no consistency today, across browsers etc

<Gives overview of various library content and browser experiments and lack of language switching and lack of support for phonems>

JS: The push for the standard is coming out of necessity - people are not looking at the standards for pronunciation and they fail

We can do better - no one is going there.

may not be traditional a11y research

SAZ: My question.. Is it more impactful for a11y, to look at this problem, the lack of support etc..

Or is it better to shed light in the research community and the fact they are not looking at it.

We do need to look at the reviewers and what they want etc

We will need to justify it

I'm just questioning - its interesting to hear the research community is not looking at this.

JS: Google may not be doing this as they want market share etc

This may not be via co-operation, and the development of standards

But we can point to easy things that will help for the general population etc

I'm hoping the EU may care more - with language diversity

<Zakim> Judy, you wanted to make a case for a higher-level approach

JB: Reflects on the first conversation/meeting...

We can have framing conversations to raise visibility

What about doing a workshop on a topic that frames a space?

Like AI etc and how we could plug that into testing and remediation?

To try to structure the space, so we could work on these

JW: Seems some topic that would engage the architectural issues of a11y, and intergration with APIs etc

That would satisfy the criteria of earlier suggestions

Seems like a candidate

Regarding machine learning etc on the web.. we can see that on the authoring and testing side, also meets that criteria

<janina> Agree we have two good candidates

Promising on initial analysis

Framing it on a more generic level could be appropriate

Important hard problems

<janina> Joshue108: agree with jason; have preference for arch at the moment; like the notion of doing the hard work identifying gaps in but also pulling players together who can meet the needs

<janina> Joshue108: Suggest we need to lead in this space

<janina> Joshue108: AI/ML is less of a preference; but do think these are related

<janina> Joshue108: believe pronunciation is a massive issue

<janina> josh add xr and we have yet more

SAZ: +1 to Josh, we have more and more pieces

CD: Just want to echo that this is an interesting conversation..

Most research in a11y is done by HCI related groups

This is an opportunity also to bring other groups in..

Solve the hard problems that we should be working on.

CD: Could pronunction be framed in the language domain?

And Natural Language Processing?

This is a good topic to justify to the commission, many EU languages

<Judy> +1 to broadening the framing of the pronunciation to a broader language issue

SAZ: Yes, there are interesting intersections

I'd like us to have some tangible outcomes.

SAZ: Would like to hear from Michael also.

SAZ: Here is a concrete idea, Jason referred to criteria but we need to agree and document those.

Various topics etc, all important.

Its important to see that there are gaps in various communities etc

So i propose for our first topic, we ask what are the gaps, topics that the W3C feel are urgent?

Can they described in a set of criteria that we agree on?

Our first symposium could be around good practices in a11y research, and what the community needs?

Could set out our stall in broader circles.

Set parameters in the field, foundation and we can pick specific things

AI I still imagine as a vertical

Could be a model that others can follow.

How does this sound?

JS: What comes to mind, while it has value, who would come to that kind of symposium?

It would be hard to get industry.

We've never gotten AT vendors to participate at W3C.

SAZ: Good point but devils advocate, wouldn't it be great to address them, we could build a network.

JW: There could be value in a wider discussion - in what are the hard problems.

JW: Could be good - when thinking of hard issues that need to solved. Would need to be clear.

We'd need strong underlying criteria

<janina> Joshue108: not thrilled about too general a framing

<janina> Joshue108: having spent last two years on emerging tech am concerned that we know a lot of what needs focus that isn't being covered

<janina> Joshue108: concerned to resource more known issues

CD: I would need to think about this more.

I would be concerned that we just have three. There is a value bringing these challenges to the research community

I see a place where various stakeholders could present their issues and challenges to the research community

So think we already have good topics.

But there could be value in the more general approach but could we achieve more value going direct to the topics that we floated.

SAZ: My worry is that we jump into firefighting.

Josh is saying we know what some of the issues are but I don't think we have fully formulated this

Could be really good for the community based on data

SAZ: Am thinking about road mapping.