No action needed. Just discussion

The accessibility api was the solution to meet screen reading needs after
the change from text based i/o to the GUI.  It is a brilliant model, but
today's discussion of programmatic determination seems to consider the job
done once an accessibility api is produced.

I do not believe that the accessibility api approach is a viable model for
many disabilities. Any disability that needs an adjusted visual interface
(visual linearization, enlargement or any spacial adjustment) could
potentially be served by this model, but it is unclear whether the current
vision of the accessibility tree is sufficient to support visual
restructuring.  Mobile technologies bypass a accessibility API. So do many
low vision ATs (some do use the, API some don't)

There are problematic ARIA parameters like aria-hidden that make a
judgement that no person with a disability needs to know that a body of
code exists. I have usually seen it used as tool for screen readers without
consideration of other users with  disabilities.

One thing that is certain is that the API serves screen readers, but what
else does it reliably serve.  Think about this carefully.  Are essential
needs for all disabilities being met by the accessibility APIs? And, more
important is the accessibility API the place where we should focus our
attention on the role of programmatic determinism.

Wayne