A Quantitative Analysis of Open Source Software Code Quality: Insights from Metric Distributions

Code quality is a crucial construct in open-source software (OSS) with three dimensions: maintainability, reliability, and functionality. To accurately measure them, we divide 20 distinct metrics into two types: 1) threshold-type metrics that influence code quality in a monotonic manner; 2) non-threshold-type metrics that lack a monotonic relationship to evaluate. We propose a distribution-based method to provide scores for metrics, which demonstrates great explainability on OSS adoption. Our empirical analysis includes more than 36,460 OSS projects and their raw metrics from SonarQube and CK. Our work contributes to the understanding of the multi-dimensional construct of code quality and its metric measurements.