DALIInterface/TAPRegExt evolution

[Markus Demleitner]

Dear DAL,

When we tried to work out how to deal with authentication in the
Registry, it turned out that our conventional way of registering
complex, multi-endpoint services had a few rather severe problems,
which I discussed in the Caproles Note
(http://ivoa.net/documents/caproles/).  Among our current standards,
this mainly concerns TAP, and more concretely, its rules for the
Registry, TAPRegExt.

As a proposal for how to evolve TAPRegExt to address Caproles, I've
just created PR#5 against TAPRegExt, which you're welcome to review;
a pre-built rendering is at http://docs.g-vo.org/TAPRegExt.pdf.

As I'm writing in the PR comment, I'd say we should iron out any very
rough edges in the PR but then try for a WD even if people don't feel
entirely comfortable with the proposal in order to keep at least a
bit of that discussion on the mailing lists (as opposed to github
discussions).  

That would mean that at this point hard-core TAP enthusiasts would be
invited to review and comment.  The wider community probably can wait
until a WD is out; the whole thing can, I expect, still be shot down
then.

        -- Markus