Hello,
Muldis D provides infix operators like the following for testing the membership
of a value in a collection:
in | ∈ |
!in | ∉ | not-in
has | ∋ |
!has | ∌ | not-has
There currently exist 6 distinct type-specific functions for each the 4 infix
operators, whose parameter types are as follows:
value | collection
----------+-----------
Tuple | Relation
Universal | Set
Universal | Array
Universal | Bag
Universal | SPInterval
Universal | MPInterval
Where you see "Universal", the typical usage is a scalar value, but it can in
fact be any value at all, since collections may be nested in collections.
Now, generally speaking I intend to overload the 4 operators so that you can use
the same syntax for all container types.
However, due to reasons of overlapping types plus differing semantics for the
same inputs to different functions, the Tuple/Relation pair must have distinct
syntax, though the other 5 can all have a shared syntax.
My first question for you is, should the Tuple/Relation option get the shorter
syntax given above and a longer/alternate syntax be invented for the other 5 to
share, or should the 5 share the above syntax and T/R use the invented?
My thought is for the 5 to share the above syntax and T/R get something like
this instead:
within | %∈@ |
!within | %∉@ | not-within
holds | @∋% |
!holds | @∌% | not-holds
My second question is whether you have any proposals for other in/has synonyms
to use instead of within/holds, or the Unicode syntax, which read better.
Finally, I assume people would be using the 5 more often in code but would you
agree or might you see people using the tuple membership test more often instead?
Note that SQL's "in" I believe corresponds to the T/R version if I'm not
mistaken.
-- Darren Duncan
_______________________________________________
muldis-db-users mailing list
muldis-db-users@mm.darrenduncan.net
http://mm.darrenduncan.net/mailman/listinfo/muldis-db-users
