RISC - Linz Research Institute for Symbolic Computation

interpretation Jacques Carette and Stephen Forrest Computing and Software Department, McMaster University Hamilton, Ontario, Canada {carette,forressa}@mcmaster.ca Abstract. We continue our investigations of what exactly is in the code base of a large, general purpose, dynamically-typed computer algebra system (Maple). In this paper, we apply and adapt formal techniques from program analysis to automatically infer various core properties of Maple code as well as of Maple values. Our main tools for this task are abstract interpretation, systems of constraints, and a very modular design for the inference engine. As per previous work, our main test case is the entire Maple library, from which we provide some sample results. We continue our investigations of what exactly is in the code base of a large, general purpose, dynamically-typed computer algebra system (Maple). In this paper, we apply and adapt formal techniques from program analysis to automatically infer various core properties of Maple code as well as of Maple values. Our main tools for this task are abstract interpretation, systems of constraints, and a very modular design for the inference engine. As per previous work, our main test case is the entire Maple library, from which we provide some sample results.