Poster: Discovering Code Dependencies by Harnessing Developer's Activity

Abstract

Monitoring software developer's interactions in an integrated development environment is sought for revealing new information about developers and developed software. In this paper we present an approach for identifying potential source code dependencies solely from interaction data. We identify three kinds of potential dependencies and additionally assign them to developer's activity as well, to reveal detailed task-related connections in the source code. Interaction data as a source allow us to identify these candidates for dependencies even for dynamically typed programming languages, or across multiple languages in the source code. After first evaluations and positive results we continue with collecting data in professional environment of Web developers, and evaluating our approach.