Modern object-oriented languages support higher-order implementations through function objects such as delegates in C#, agents in Eiffel, or function objects in Scala. Function objects bring a new level of abstraction to the object-oriented programming model, and require a comparable extension to specification and verification techniques. We introduce a verification methodology that equips each function object with side-effect free (pure) methods for its pre- and postcondition, respectively. These pure methods can be used to specify client code relatively to the contract of the function object. We demonstrate the expressiveness of our approach through several non-trivial examples. It can be combined with any verification technique that supports pure methods, as illustrated by our experiments with Spec#.

Abstract. We present a logical system suited for specification and verification of imperative ML programs. The specification language combines dynamic logic (DL), explicit state updates and second-order functional arithmetic. Its proof system is based on a Gentzen-style sequent calculus (adapted to modal logic) with facilities for symbolic evaluation. We illustrate the system with some example, and give a full Kripke-style semantics in order to prove its correctness.

c t i v it y e p o r t 2009 Table of contents