Knowledge-Oriented Models Based on Developer-Artifact and Developer-Developer Interactions

Abstract

INTRODUCTION: Software development is organized around developers working collaboratively promoting two types of interactions for knowledge sharing. Developer-Artifact interactions indicate developers define or access pieces of information within artifacts. Developer-Developer interactions indicate the exchange of information among developers using a collaboration platform to clarify an issue, promote an idea, or expose any thoughtful comment. PROBLEM: The number of such interactions grows over time and makes it difficult to capture and assess the evolution of the developers' knowledge about specific software project artifacts and tasks. Further, this knowledge decreases over time due to the natural limitations of human cognition that restrict our capabilities to cope with information overload. Besides, who has more knowledge about specific project elements are important to promote collaboration. AIMS: The $K_{a}$ , $K_{s}$ , $K_{c}$ , and $K_{p}$ models capture the evolution of the developers' knowledge about software project elements such as artifacts, tasks, similar tasks, and the whole software project. These models represent not only the knowledge developers have about these elements but also capture how this knowledge decreases over time based on forgetting and relearning functions. EVALUATION: An experimental study analyzed some developers' interactions on artifacts for the purpose of predicting the evolution of developers' knowledge in six software projects. The results show that the developers' rankings by performed tasks and by our models have 72% or more of similarity. CONCLUSION: Our models can capture and assess the evolution of the developers' knowledge and help to identify which developers have more knowledge about specific elements of software projects.