For obvious efficiency reasons, it is good to avoid dynamic type-checking in a programming language. The Prolog type system proposed by Mycroft and O'keefe has some good properties which avoid such a runtime overhead. One of this property called type preserving or transparency is a restriction of the polymorphism. This restriction is not compatible with some programming techniques, for example with higher-order predicates.

Higher-order logic languages seem particularly suitable to be implemented on parallel architectures for at lest two reasons. On one hand, they offer more opportunities for parallelization than first-order logic languages. On the other hand, the greater computational costs of higher-order unification of typed -terms should allow the exploitation of parallelism with a larger granularity, possibly leading to higher efficiency. In spite of their potential interest, there are a few implementations of higher-order logic languages, and, to our knowledge, no experience has been made about their parallelization, up to now. We report here a preliminary study about parallelization of an interpreter for the higher-order logic language Prolog. Starting from the available experience in parallelizing first-order logic languages, the possible sources of parallelism in execution of Prolog programs are presented, and a characterization for their actual exploitation on parallel hardware given....