Towards Evidence-Based Software Quality Practices for Reproducibility: Preliminary Results and Research Directions

In the computational science and engineering (CSE) community, there is a prevailing belief that adopting better software development practices and investing in software quality will directly lead to more robust, reproducible software. There is, however, relatively little evidence to indicate what specific aspects of software quality influence reproducibility or which practices lead to more reproducible software. To better inform this discussion, we present preliminary findings from an ongoing study of how software quality practices among CSE teams affect the reproducibility of the software they create.