RFC: Use of xml-roles object attribute for MathML without ARIA

Hi all.

As you know, there is no ARIA MathML module (at least not yet; I'm
starting to think perhaps we should have one). There is, however, a need
to improve exposure of MathML elements and their roles. One proposed way
to do this is via xml-roles. See this Mozilla bug:
https://bugzilla.mozilla.org/show_bug.cgi?id=1177765.

I haven't yet decided how I myself feel about this. On the one hand, why
create a functionally-equivalent new object attribute? On the other
hand, at least in my mind, xml-roles was the place to find the ARIA
role, and the possible values for xml-roles was defined by the ARIA
specification. What is proposed in the Mozilla bug perhaps is setting a
precedent that it's ok to expand the possible values into items not in
any specification.

Thoughts?

--joanie

Received on Tuesday, 14 July 2015 08:43:26 UTC