Committing to reproducibility and explainability: using Git as a research journal

Abstract

AbstractTraditional critical editions in print offer a model for creating a culture of reproducible and explainable research in Digital Humanities. In addition to explaining their editorial methods and practices in a preface, editors provide a critical apparatus to allow readers to evaluate the evidence and decide for themselves whether the editor’s judgment is sound. This paper will argue for viewing distributed, version-controlled data repositories (e.g., Git) in the same way, to empower users to follow the choices and methods of DH researchers and to reproduce their work. Using examples from a current project, this paper will propose some guidelines for using Git to promote reproducibility and explainability in the publication of DH research of all kinds.