At best true for toy problems and if you're satisfied with O(n!) performance. There's quite a difference between:
randomsort(L1, L2) :-
permutation(L1, L2), ordered(L2).
quicksort([X], [X]) :- !.
quicksort([H|T], S) :-
partition(H, T, L1, H2),
append(S1, [H | S2], S).
I like Prolog a lot, but the "describe the problem instead of the solution" claim is only partially true. Using it as the only reason to say you love Prolog mostly indicates that you should look for better arguments.
BTW: "Zen-like"? I think SLD-resolution is more responsible...