A sketch of the programmer's coach: making programmers more effective

Abstract

As programmers work on source code, they ask an array of questions that are difficult to answer manually. To help answer these questions, programmers often employ software tools; often in attempting to use these tools, the programmers encounter many obstacles which frustrate their efforts and lead to less than optimal tool utilization. Possibly worse, programmers often intentionally under utilize available tools as they prefer to answer questions only with tools they have used before. We hypothesize that we can coach programmers towards a more systematic use of appropriate software tools that would enable the programmers to be more productive in the completion of their work. We propose to use activity logs collected automatically to deduce the questions a given programmer asks a frequently and then to coach the programmer automatically on appropriate, possibly unfamiliar, tools to answer those questions more effectively. By using activity logs to inform coaching decisions, our approach is based on an objective cost metric. We envision an environment that enables a programmer to learn how to use appropriate tools systematically.