Object Propositions

The presence of aliasing makes modular veri cation of objectoriented code di cult. If multiple clients depend on the properties of an object, one client may break a property that others depend on. We have developed a modular veri cation approach based on the novel abstraction of object propositions, which combine predicates and information about object aliasing. In our methodology, even if shared data is modi ed, we know that an object invariant speci ed by a client holds. Our permission system allows veri cation using a mixture of linear and nonlinear reasoning. We thus o er an alternative to separation logic veri cation approaches. Object propositions can be more modular in some cases than separation logic because they can more e ectively hide the exact aliasing relationships within a module. We validate the practicality of our approach by verifying an instance of the composite pattern. We implement our methodology in the intermediate veri cation language Boogie (of Microsoft Research), for the composite pattern example.