Cyclomatic Complexity and Maintainability in Modern Software: A Systematic Review

Abstract

Cyclomatic Complexity (CC) has emerged as a pivotal metric in contemporary software development, exhibiting robust correlations with testing initiatives and maintenance challenges. This systematic literature review analyzes five high-quality studies (2020-2024) covering more than 200 projects to evaluate the predictive value of CC in structural testing and maintenance contexts. The findings indicated substantial correlations between CC values and several key metrics, including: test coverage requirements (r=0.82), an augmented development time of 35-45% for CC>8, and a diminished code comprehension (r=-0.74). Research has demonstrated that CC values above 10-15 consistently predict test complexity and maintenance difficulty. Concurrently, the utilization of CC in CI/CD pipelines has reached 85% effectiveness as a quality gate, signifying its growing integration into software development processes. Nevertheless, challenges persist in the standardization of measurement and the necessity for a domain-specific threshold framework. The present study proposes the implementation of adaptive threshold-based CC, in conjunction with machine learning, to enhance the predictive quality analysis.