REC for ASCII MOCs

[Markus Demleitner]

Hi Pat,

On Thu, Oct 25, 2018 at 12:58:37PM -0700, Patrick Dowler wrote:
> cell in an interoperable way. DALI could be the place to define the syntax
> or it could refer to a MOC standard directly if a syntax was normative.

Do you have a preference?  Me, I'd now say "let's keep the MOC things
together and have ASCII MOC in the MOC Rec".

> You don't really need a MOC function (constructor) -- would someone query
> with a single MOC or enter one by hand? -- but for symmetry and
> completeness it could be a stretch goal.

Well -- I'd say it's a use case, at least in connection with
cut-and-paste.  And if your database supports MOCs, it's trivial to
make such a constructor function.  The interesting question would be
what its argument is: just a string literal or would we allow an
expression, too?  In that latter case it's the database that would
have to understand ASCII MOCs, and while pgsphere does, that's a lot
more to ask for.

But I'd say let's postpone the issue of what's desirable in ADQL and
first get ASCII MOCs properly defined.

I like Grégory's proposal, including, I have to say, the PEG grammar.
So... Who's volunteering for taking MOC to 1.1?

       -- Markus