Inference-Based Overloading Resolution for ADA

We present an incremental type inference and overloading resolution algorithm for ADA. The algorithm is used within a language-specific ADA editor; it is based on inference rules and order-sorted unification. The method is founded on the notion of context relations, a calculus for the incremental analysis of incomplete program fragments. This calculus has recently been extended for the analysis of user-defined polymorphic or overloaded objects, and the semantic analysis for ADA described in this paper was the first field test for the new extensions. Although overloading resolution in incomplete fragments in general is NP-complete, we have obtained an editor with reasonable response time behaviour for up to medium-sized programs.