Scientific Research Software Ecosystems

In the past 70 years, science research methods have expanded from test tubes and beakers to include simulations in software. Scientific software is produced in a larger environment of collaboration that includes not only software developers, but also scientists who both use and extend the software for their research endeavors. The objective of this research is to understand the factors that promote and inhibit ecosystem formation and sustainment specific to scientific research software development. We have identified a set of relevant problems inhibiting sustainability in the scientific software context, including development, evolution and maintenance, as well as funding and leadership. We have collected data and used existing business, software, and innovation analysis techniques to examine the artifacts from several scientific research projects. We interviewed significant actors in select scientific ecosystems. We describe the differences between scientific software ecosystems and commercial software ecosystems, our initial results from interviews and data analysis, and we provide a set of recommendations and approaches towards handling software sustainability. Data suggests an ecosystem strategy can contribute to the survivability and sustainability of research software; however, such a strategy, which may distract from sustainability, requires thought and effort be given to establishing and nurturing the ecosystem, which in turn distract from the research mission. At this stage these initial results point us to future potentially fruitful investigations.