Code Maintainability and DataStructures in Energy-AwareSoftware: An Empirical GitHubStudy

Sammanfattning

As software becomes an increasingly larger part of everyday life, writing maintainable, energy aware code is important. Saving energy is good both for operation costs as well as the environment. In this thesis, I analyze software commits from public GitHub repositories. The work was framed around a reproduction study of ”Mining Energy-Aware Commits” by Moura et al. The commits were mined using the tool World of Code. Then using thematic analysis, grouped into themes which describe what effect they have on energy usage. The results show the steps developers take to save energy through scaling the CPU frequency, offering the users options to customize their experience, or implementing more energy-efficient data structures. We could also see that performance and stability are often prioritized over energy saving and that it can affect maintainability characteristics.