Difference between X\=Y and dif(X,Y)

Question

What is the difference between this:

X \= Y

and this piece of code:

dif(X, Y)

I thought that they should behave the same, but they do not. Here's the example:

n_puta(L, N, X) :- nputa(L, N, 0, X).
nputa([], N, C, _) :- N = C.
nputa([G|R], N, C, X) :- G = X, nputa(R, N, Y, X), C is Y - 1.
nputa([G|R], N, C, X) :- dif(G,X), nputa(R, N, C, X).

And here are some calls:

?- n_puta([a,a,b,b,b], 2, X).
X = a ;
false.

?- n_puta([a,a,b,a,b,b], 3, X).
X = a ;
X = b ;
false.

X should be the atom that occurs exactly N times in the list L. If I replace dif(G, X) with G \= X, I don't get the expected result. Can someone tell me what is the difference between these two operators? Can I use anything else except dif(G, X)?

This example works prefectly in SWI-Prolog, but doesn't work in Amzi! Prolog.

Answer 1

dif/2 and (\=)/2 are the same as long as their arguments are ground. But only dif/2 is a pure relation that works correctly also with variables and can be used in all directions. Your example clearly shows that you should use dif/2 in this case, because you use your predicate not only to test, but also to generate solutions. The most widely used Prolog systems all provide dif/2.