The four elements of achieving research software sustainability for long tail projects

Abstract

At deRSE19 we presented Hexatomic, a project to investigate what small research software projects need to make their software more sustainable. This talk reports our results. Developing annotation software for multi-layer linguistic corpora, we found that - there is a minimal infrastructure that needs to be in place to activate a potential for sustainability in the first place, and - that four elements play key roles in reducing the risk of software collapse. These four elements are: a clearly defined, resourced maintainer role; multi-modal documentation to guide maintenance and use; automated tests; code review and triage processes supported by static code analysis. We have tested this by performing changes in maintainership and observing success in reviving development and maintenance activities. Our results furthermore clearly point to the necessity for RSEs to be involved in software projects as part of research projects. In this talk, we describe our research project, the minimal infrastructure as well as development, maintenance, and release and publication workflows we implemented, and how the four elements helped maintainers take up their roles.