State-based formal methods in scientific computation

Abstract

Control systems, protocols, and hardware design are among the most common applications of state-based formal methods, and yet the types of modeling and analysis they enable are also well-suited to problems in scientific computation, where quality, reproducibility, and productivity are growing concerns. We survey the challenges faced by developers of scientific software, characterize the nature of the programs they write, and offer some perspective on the role that state-based methods can play in scientific domains.