Programming Dangerously! Can Formal Methods and Empirical Software Engineering Come to the Rescue?

Summary form only given. This award prompts me to reflect on how my team at Microsoft Research was able to bring research in formal methods and empirical software engineering into practice. How did I select the topics for my teams and my research? How did we conduct the research and which methods did we choose? How did we show progress, and equally important when did we decide to stop? I will also discuss the importance of knowing your customers, i.e., who, how and why a customer will use your research contributions. Of equal importance is to understand the constraints of a product group that productizes your research, and how your role changes when a product team adopts your ideas. Three years ago, I changed roles and founded a product group, called Tools for Software Engineers, which builds developer tools and service at scale. I will explain what new research challenges I encountered in building software services. I will argue that a combination of both, formal methods and empirical software engineering is the most promising approach to impact existing practice.