Formal Systems Laboratory

Evolution-Aware Monitoring-Oriented Programming

Abstract. Monitoring-Oriented Programming (MOP) helps develop more reliable software by means of monitoring against formal specifications. While MOP showed promising results, all prior research has focused on checking a single version of a target software application. We propose to extend MOP to support multiple software versions and thus be more relevant in the context of rapid software evolution. Our approach, called eMOP, is inspired by regression test selection---a well studied, evolution-centered technique. The key idea in eMOP is to monitor only the parts of code that changed between versions. We illustrate eMOP by means of a running example, and show the results of preliminary experiments. eMOP opens up a new line of research on MOP---it can significantly improve usability and performance when applied across multiple versions of an application, and is complementary to algorithmic MOP advances on single versions.