Re: AOM with APA at TPAC?

On Mon, Sep 9, 2019 at 3:22 AM Léonie Watson <lw@tetralogical.com> wrote:

>
> On 08/09/2019 11:07, Janina Sajka wrote:
> > 1.) is existing AOM and API architecture suitable for immersive Web or
> > should AOM just be restricted to Web Components?
>
> The AOM isn't intended to be restricted to Web Components. As far as
> it's been implemented, it can already be used anywhere ARIA can be applied.
>
> One of the future phases of the AOM intends to introduce the ability to
> add virtual nodes to the accessibility tree in the browser. These
> virtual nodes need not have a corresponding node in the DOM tree.
>
> One of the use cases for this phase, is the ability to create a virtual
> branch of the accessibility tree for the purpose of providing
> accessibility semantics for things like the <canvas> element, which is
> where the AOM may well be particularly useful to the WebXR space.
>
> Léonie.
>

Everything Léonie said!

My main concern would be that the existing ARIA vocabulary/existing AT
interaction patterns would be too limiting for UX designed for an immersive
environment (orthogonal to the AOM API design), which Janina touches on
below.

> 2.) Do we need bi-directionality for good XR support? Semantics can be
> > consumed by user agents but may be modified in an imersive environment
> > and change as interactions are happening. Like React is data driven, XR
> > semantics may be interaction or results driven.
> >
> > 3.) What would be the ideal architecture to support XR accessibility? We
> > seem to be currently aiming at patching XR with current and even legacy
> > AT, so that architecture may be temporary, or move away from browser and
> > API interactions towards AT being embedded in an immersive environment.
> > What does "good" look like in this situation?
>

These are really interesting and important questions - I don't know enough
about XR to start answering them.

> 4) Are Object Oriented approaches to accessible XR preferable to
> > declarative or author applied semantics?
>
> > Please confirm whether 11:00 Thursday works.
>

It's open for me, although if the topic is primarily going to be XR I'm
unclear why this would be a separate session from the proposed plenary
session <https://www.w3.org/wiki/TPAC/2019/SessionIdeas#XR_Accessibility>
on Wednesday.


> > Looking forward to seeing you both in Fukuoka,
> >
> > Janina
> >
> >> Alice Boxhall writes:
> >>> On Wed, Aug 28, 2019 at 6:02 AM James Craig <jcraig@apple.com> wrote:
> >>>
> >>>>
> >>>>
> >>>>> On Aug 27, 2019, at 10:31 AM, Janina Sajka <janina@rednote.net>
> wrote:
> >>>>>
> >>>>> Dear Alice, James:
> >>>>>
> >>>>> I'm sending this note to ask for an hour or so of your time at TPAC
> for
> >>>>> a meeting with APA regarding AOM. I'm aware you're also meeting with
> >>>>> ARIA earlier in the week, but I'd like to take our conversation in a
> >>>>> somewhat different direction, and hope you would be available for
> that.
> >>>>>
> >>>>> I would propose 11:00-Noon on Thursday. Let me know whether that
> works,
> >>>>> or some alternate suggestions.
> >>>>>
> >>>>> One of our top concerns this year is the emergence of XR. Our
> Research
> >>>>> Questions Task Force has been exploring[1]1 some of the
> opportunities and
> >>>>> challenges afforded by augmented and virtual reality technology, and
> >>>>> we'd like to understand how AOM might fit into the XR stack.
> >>>>
> >>>> I’m open to meeting with a clearly defined agenda. Generalities of XR
> >>>> seems too broad for a productive APA discussion IMO. If there are
> specific
> >>>> proposals for changes or API, that’s something to discuss.
> >>>>
> >>>
> >>> We could talk about use cases for virtual accessibility nodes?
> >>>
> >>>
> >>>>> Another topic is the resurgent interest in developing a CSS-AAM.
> >>>>
> >>>> CSS-AAM seems proposed as a solution to an ill-defined problem. I’m
> still
> >>>> uncertain that mapping document is the right solution for CSS.
> >>>>
> >>>
> >>> I would be more interested in fleshing out a shared understanding of
> >>> the relationship
> >>> between the DOM tree, the block model and the accessibility tree
> >>> <https://github.com/WICG/aom/blob/gh-pages/accessibility-tree.md>.
> There
> >>> was a lot of interest in this idea last year, but then I dropped the
> ball
> >>> on pushing it forward.
> >>>
> >>> I think that would go a long way towards helping spec authors
> understand
> >>> how their proposals will affect assistive technologies, and helping
> page
> >>> authors understand how the code they write will be translated into the
> >>> accessibility tree.
> >>>
> >>> As I mentioned when I withdrew from APA, I had much more success
> working on
> >>>> CSS accessibility directly from inside the CSS WG than from the
> external
> >>>> APA group. I can attest that Alice has also had a lot of success
> getting
> >>>> accessibility features added to the specs by working directly within
> the
> >>>> scope of those working groups.
> >>>>
> >>>
> >>> Right; for better or worse, the CSS group really has its own way of
> doing
> >>> things, which doesn't easily lend itself to casual participation.
> >>>
> >>>
> >>>> For the APA’s issues with CSS, I'd suggest developing a concrete list
> of
> >>>> problems and filing them as individual issues to the CSS tracker. For
> >>>> example, there is already language in flexbox for how CSS should
> affect AT
> >>>> order. It’s flawed, but possible to address with follow-up issues.
> >>>>
> >>>> https://drafts.csswg.org/css-flexbox/#order-accessibility
> >>>
> >>>
> >>> Unfortunately this seems like a counter-example; folks have long been
> >>> trying to persuade CSSWG to allow flex order to apply to focus and
> >>> accessibility order.
> >>>
> >>> There are also some existing issues that relate to accessibility in
> some
> >>>> way.
> >>>>
> >>>> Sequential navigation for example:
> >>>> https://github.com/w3c/csswg-drafts/issues/3377
> >>>>
> >>>> Good luck,
> >>>> James
> >>>>
> >>>>
> >>>>> What that might cover is not yet defined in a document, but it seems
> >>>>> reasonable to me it could be most powerful if it could be tied to OS
> >>>>> services through the AOM. I think this would be a good time to
> explore
> >>>>> what might be reasonably possible.
> >>>>>
> >>>>> I expect there will be a few additions, so hope you can be available
> for
> >>>>> a joint conversation. It would definitely benefit APA to have a
> >>>>> conversation.
> >>>>>
> >>>>> Best,
> >>>>>
> >>>>> Janina
> >>>>>
> >>>>> [1]
> >>>>
> https://www.w3.org/WAI/APA/task-forces/research-questions/wiki/Main_Page#XR_.28Augmented.2FVirtual_Reality.29_Accessibility_and_Associated_Issues
> >>>>>
> >>>>>
> >>>>> --
> >>>>>
> >>>>> Janina Sajka
> >>>>>
> >>>>> Linux Foundation Fellow
> >>>>> Executive Chair, Accessibility Workgroup:     http://a11y.org
> >>>>>
> >>>>> The World Wide Web Consortium (W3C), Web Accessibility Initiative
> (WAI)
> >>>>> Chair, Accessible Platform Architectures
> http://www.w3.org/wai/apa
> >>>>>
> >>>>
> >>>>
> >>
> >> --
> >>
> >> Janina Sajka
> >>
> >> Linux Foundation Fellow
> >> Executive Chair, Accessibility Workgroup:    http://a11y.org
> >>
> >> The World Wide Web Consortium (W3C), Web Accessibility Initiative (WAI)
> >> Chair, Accessible Platform Architectures     http://www.w3.org/wai/apa
> >>
> >
>
> --
> Director @TetraLogical TetraLogical.com
>