Ambiguity and constrained polymorphism.

Abstract

This paper considers the problem of ambiguity in Haskell-like languages. Overloading resolution is characterized in the context of constrained polymorphism by the presence of unreachable variables in constraints on the type of the expression. A new definition of ambiguity is presented, where existence of more than one instance for the constraints on an expression type is considered only after overloading resolution. This introduces a clear distinction between ambiguity and overloading resolution, makes ambiguity more intuitive and independent from extra concepts, such as functional dependencies, and enables more programs to type-check as fewer ambiguities arise. The paper presents a type system and a type inference algorithm that includes: a constraint-set satisfiability function, that determines whether a given set of constraints is entailed or not in a given context, focusing on issues related to decidability, a constraint-set improvement function, for filtering out constraints for which overloading has been resolved, and a context-reduction function, for reducing constraint sets according to matching instances. A standard dictionary-style semantics for core Haskell is also presented.