Abstract

We give a semantics for a typed object calculus, an extension of System F with object subsumption and method override. We interpret the calculus in a per model, proving the soundness of both typing and equational rules. This semantics suggests a syntactic translation from our calculus into a simpler calculus with neither subtyping nor objects.