Would anyone object if I replaced this syntax with a new pragma named suppress_tail_recursion(name/arity)? It would make documenting it significantly easier. (The internal representation already corresponds to the proposed new syntax, because it is clearer.) Zoltan.