[mercury-users] Idea for a currying library

[Ralph Becket]

A while back there was a discussion about how
best to deal with the issue of currying, in
particular how one might deal with the fact
that often one has to permute arguments in
higher order to code to make things work.
Richard O'Keefe, I think, made the suggestion
that some extra syntax be introduced, e.g.

	P = foo(X, #2, Y, #1)

would translate into

	P = (func(A, B) = foo(X, B, Y, A))

or somesuch.  I, for one, would find such a
facility very useful.

If adding more syntax is undesirable/too much
trouble for now, it occurs to me that we could
get some way towards solving the problem with
a suitable library.  The idea would be to
provide currying functions for all the non-
standard argument permutations for all functions
up to a given arity, say 5.  This library would
look something like the following:

	% cy12 omitted as unnecessary.
	:- func cy21(func(T1, T2) = T3, T2, T1) = T3.
	cy21(F, A, B) = F(B, A).

	% cy123 omitted as unnecessary.

	:- func cy132(func(T1, T2, T3) = T4, T1, T3, T2) = T4.
	cy132(F, A, B, C) = F(A, C, B).

	:- func cy213(func(T1, T2, T3) = T4, T2, T1, T3) = T4.
	cy132(F, A, B, C) = F(B, A, C).

and so on for cy231, cy312, cy321 and on up to 
arity 5 or whatever.

Of course, we'd almost certainly have to limit
the library to (in, in, ..., in) = out modes to
avoid an explosion in the size of the library.

Mind you, generalising to the case where only a
single output is allowed might not be too bad.

Any comments?

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
--------------------------------------------------------------------------