Structure in Theorem-proving Programs

We describe how clauses in resolution programs can be represented and used Without applying substitutions or cons-ing lists of literals. The amount of space required by our representation of a clause is independent of the number of literals in the clause and the depth of function nesting. We introduce the concept of the value of an expression in a binding environment which we use to standardize clauses apart and share the structure of parents in representing the resolvent. We present unification and resolution algorithms for our representation. Some data comparing our representation to more conventional ones is given.