A Novel Framework to Estimate Truck Factor in Software Repositories Using Dynamic Knowledge Decay Model
Arghya Bhattacharya, Samba Narayana Duggirala, Aditya Ramdas, K. Muthukumaran · Frontiers in artificial intelligence and applications · 2025
The concentration of technical knowledge among a limited group of contributors can undermine the sustainability and stability of critical software systems. In software engineering, the truck factor highlights the risk related to knowledge concentration. The truck factor refers to the minimum number of key contributors whose sudden departure would jeopardize the project’s development. A higher truck factor indicates a more robust project that is less dependent on a small subset of individuals, making it a useful metric to evaluate the reliability of external libraries in mission-critical applications. In this work, we propose a novel algorithm for estimating the truck factor that addresses limitations of previous studies. We employ a knowledge-based and time-based framework that assesses the distribution of expertise across the repository. Notably, the framework recognizes that multiple contributors may possess partial familiarity with specific segments of code, even if they have not directly authored them. Furthermore, we propose a knowledge decay model, inspired by psychological principles, which adjusts the expertise metrics over time. This model aims to account for the diminishing relevance of historical contributions, ensuring that the truck factor estimation accurately reflects the current resilience of the project.