Detecting object-oriented code smells through social network analysis

Abstract In software engineering, maintaining high code quality is essential for ensuring the software's maintainability, scalability, and overall performance. A major obstacle to attaining high-quality software is the presence of code bad smells, which are indicators of potential weaknesses and design flaws within the code. This paper investigates the potential of Social Network Analysis (SNA) as a new method for identifying and examining code smells. By applying SNA techniques, we aim to provide a novel perspective on identifying and analyzing code smells, which are indicative of potential design flaws and vulnerabilities within software systems. Our methodology involves modeling software components and their interactions as a network, allowing us to apply SNA metrics to uncover hidden patterns and relationships. In this paper, we model four Java open-source projects (JUnit 5.10.2, Spring 6.1.4, Apache Commons BCEL 6.8.2, and Guava 33.0) as social networks. Using SNA metrics, we then formulate and identify code bad smells.