IIAG: a data-driven and theory-inspired approach for advising how to interact with new remote collaborators in OSS teams

摘要

Open source software development (OSS) team members often need to figure out how to initiate a collaboration with a new remote collaborator. An inappropriate strategy could lead to failures in developing cooperation. In this paper, we propose an approach and corresponding intelligent system called IIAG (Initial Interaction Assistant based on Game theory analytics), which identifies and advises its users about strategies for initial interactions with new remote collaborators. IIAG integrates game theory, decision models, and social factors with the collaborative traces mined from empirical project data to achieve this goal. When a user seeks IIAG’s advice, it simulates an individual’s decision processes to find the strategies that yield the best outcomes. Thus, it can advise proper strategies for users. IIAG is evaluated extensively. We design and perform virtual experiments to evaluate IIAG with empirical data collected from three large open source projects. The results show that IIAG can identify the payoff-optimal strategy with over 80% accuracy. We also conduct a lightweight user study to evaluate the IIAG’s usefulness from the potential users’ perspective. The results are also promising. Thus, IIAG can help OSS team members in making informed decisions about interacting with new remote collaborators.