Research Software Engineers as Maintainers: Keeping code alive

Abstract In recent years, scientific research groups, especially at universities, have desired the expertise of the human equivalent of a Swiss Army knife for software: the research software engineer (RSE). RSEs are known by many job titles, like data scientist and research support staff, and likewise have a plethora of duties from writing software for HPCs to building models to visualizing results in plots. As RSEs prominence rises, a wave of software sustainability efforts has blossomed, centering software as an increasingly vital component in research. At the same time, funders of software and infrastructure projects stress reusable components to streamline development efforts across institutions. In addition, the shift from closed systems to ones using open source software components has led to the question: How can we view RSEs over time? In this position paper, I argue that RSEs should be viewed as maintainers of software. Software cannot be sustainable or reusable without a maintainer performing the work. In university spaces where graduate students come and go, and PIs are focused on research results, RSEs emerge as key through lines of continuity in long-term projects, fonts of historical knowledge, and navigators of the increasingly thorny scientific software stacks