Sustaining Research Software: Why and How

Abstract: Humanity seeks new knowledge for its own purpose as well as for its potential solution to problems, situations, and crises, which at large scale, often relies on HPC systems. We want to be able to verify (or disprove) such knowledge (reproducibility), then build on it (reuse), as simply and as cost-effectively as possible. In this talk, I will focus on knowledge captured in research software, which can be both read, executed, and extended. However, software, unlike data, requires ongoing human activity to fix bugs and to adapt to frequent changes in the software and hardware environment on which it depends, as well as changing user needs. Software sustainability is the process of assembling resources and using them to do this work, including through a mix of open source communities, industrial or government support, and commercialization. This leads to a number of overlapping challenges and corresponding efforts, including making research software FAIR (findable, accessible, interoperable, and reusable), publishable, and citable, as well as considering how to support the human effort needed to maintain and sustain the software, including incentives and career paths.  This talk will highlight some recent activities in these areas, including FAIR for research software principles, software citation, the Journal of Open Source Software, and software career paths.