Graphics ARIA Roadmap

Hello all,

Joanie asked me to review the graphics-related bits of the ARIA roadmap
<https://www.w3.org/WAI/ARIA/roadmap> & deliverables timeline
<https://www.w3.org/2018/03/draft-aria-charter#timeline> (for the new
charter) that were put together at last week's face-to-face meeting.

I've already sent the following comments to her & the other chairs. I'm
sharing now with everyone in anticipation of a discussion on the next ARIA
WG call (which I plan to join):

For the roadmap: "Add new roles/attributes to achieve parity with SVG
elements"

   - The graphics ARIA (Ievel 1) roles mostly achieve this. The exception
   are the text elements, which don't have directly mapping roles. However,
   those text roles would directly map to roles equivalent to HTML <p> and
   <span>, so it may not be necessary to track this requirement separately for
   SVG.

   - The only other potential SVG-specific feature might be <view>. SVG-AAM
   currently recommends that it should be exposed only through the effects it
   has on a target <svg> element, but we do have a large open issue
   <https://w3c.github.io/svg-aam/#issue-8> about it.  If a different
   approach is chosen, a new role may be required.

For the timeline deliverables:

   - I do not currently expect there to be a "Graphics ARIA 1.1" (or 1.2)
   spec. There is no planned incremental enhancement to the three foundational
   roles in the current spec. As mentioned above, the roles that are required
   for SVG parity would be shared with HTML and therefore included in the
   update to the main ARIA spec.  (This could change depending on the <view>
   issues, but that's very hypothetical currently.)

   - There has in the past been extensive discussion about establishing a
   complete graphics taxonomy of roles for charts, diagrams, and maps (axes,
   legends, data points, data series, connectors, regions, etc. and so on).
   This was always discussed as a full "Level 2" Graphics ARIA Roles spec.
   The numbering may not be that important, but it should be clear that this
   is not a minor undertaking.  It would involve creating roles & attributes
   that do not correspond to anything that is currently exposed to or consumed
   by assistive tech through standard Accessibility APIs.

   - I don't know whether it is possible to set any timelines for
   deliverables of the next level Graphics Role spec until we have some
   commitments from people to edit it & implement it. As it currently stands,
   it would definitely not be reasonable to expect it to get to full
   recommendation status within a year and a half.  Q4 2020 might be a
   feasible deadline, but it would remain aspirational until there is a team
   in place working on the plan.

   - We would need active people from multiple implementation teams, at all
   levels of implementation (authoring tools, browsers, OS accessibility APIs,
   and assistive tech tools of different types), to get this from theoretical
   proposal to a practical addition to the web platform.

A couple weeks ago, it was agreed to close down the SVG Accessibility Task
Force for lack of activity. It or something similar could be revived if
there was a group of people eager and able to work on more advanced
approaches to accessible graphics.  But I truly believe that it's not
enough just to have spec editors, we need implementers, too.

On a somewhat-related note, SVG-AAM <https://w3c.github.io/svg-aam/> should
be republished as an updated working draft this week (after various delays
with the publishing process).  We'd very much appreciate feedback on the
remaining open issues.  The next step is the test suite, and we're trying
to organize a small team to work on it. Please let me know if you're able
to contribute to that.

Best,
Amelia Bellamy-Royds