[mercury-users] Idea for a currying library

[Ralph Becket]

> Peter Schachte's preferred notation shares one advantage with lambda
> notation which the callperm stuff lacks:  the form that is abstracted
> can be compound.  On the other hand, anything more than a single goal
> is likely to merit a comment or two, so putting it on more than one
> line is going to pay off in readability anyway and we might as well
> use lambda-notation.

I'm inclined to agree, although I find the need to supply mode and
determinism declarations for lambdas in Mercury makes code look
clumsy and cumbersome.  I would like some sort of facility for
turning on mode inference on a selective basis, e.g.

:- pragma infer_modes(foo/3).

foo(X, Y, Z) :- <<something containing an unmoded lambda>>.

I don't mind supplying top-level modes for foo/3, but it does grate
having to write them for lambdas where it should be bleeding obvious
what is meant in most cases.

Ralph
--------------------------------------------------------------------------
mercury-users mailing list
post:  mercury-users at cs.mu.oz.au
administrative address: owner-mercury-users at cs.mu.oz.au
unsubscribe: Address: mercury-users-request at cs.mu.oz.au Message: unsubscribe
subscribe:   Address: mercury-users-request at cs.mu.oz.au Message: subscribe
--------------------------------------------------------------------------