Define computedrole for the SVG roles table

[cookiecrook]

Based on the outcome of this ARIA issue…

• Need a decision on what the computedrole should return on extension specs aria#1887

And this graphics-aam issue/pull:

• Define computedrole for graphics-* roles graphics-aam#7
• Add computedrole column graphics-aam#9

And this html-aam issue/pull:

• Add computedrole column for HTML-AAM html-aam#464
• Addition: computed role mappings for html elements html-aam#465

SVG-AAM (or svg-aria) may need an update similar to this issue/pull:

• Which ARIA synonym role should computedrole return? core-aam#166
• Add computed role section and mappings core-aam#167

However, since SVG-AAM primarily points to the role mappings of core-aam, html-aam, and graphics-aam, it's possible this isn't needed in svg-aam... To avoid duplication. Not sure.

[cookiecrook]

Reminder to self on this approach discussed in the ARIA F2F Meeting, May 2023 in Seattle.

• Need a decision on what the computedrole should return on extension specs aria#1887 (comment)

Quoting:

Consensus from the ARIA F2F was that I would draft a PR for SVG-AAM adding a computedrole column to each of the elements... Many would be "not mapped" but the ones that use the graphics-aria superclass roles would be return svg-<tagname> such as svg-polygon for the <polygon> element. The role instance would still be graphics-image, which is a subclass of the image Core-AAM role... Though the returned computedrole svg-polygon is just a string.

[cookiecrook]

Until there is a resolution on:

• Should concrete shape/object roles like circle have a more specific role description? #30

…I'm not certain there is a need to add the host-language defined computedrole values, and I don't want to write spec changes or tests for either assertion (the graphics-aria roles or these new svg-specific computedrole return values) until there's group consensus on the other issue.

…so I think the default-role-for-element portions of web-platform-tests/interop-accessibility#69 may be blocked.