Triaging incoming change requests: bug or commit history, or code authorship?

Abstract

Combining information retrieval and source code authorship information as a novel approach to recommend expert developers to assist with software change request is presented. This approach is based on the presence of code authorship information in source code files and does not require software repository mining or training for past bugs/issues. Our approach is compared with two other representative approaches 1) using machine learning on past bug reports and 2) based on commit logs, over three open source systems ArgoUML, Jedit and MuCommander. Without the need of any historical change information our approach provides recommendation accuracies equivalent or better than the two compared approaches.