Timo Borst

According to the Open Science paradigm, (open) research software is expected to play a constitutive role besides the other two pillars, Open Access and Open Data. While scientific infrastructure providers like computer centers, research data centers or libraries may have a clearer picture on how to support the other two aspects, the perception of research software and the corresponding activity of developing and maintaining it has not been defined or fixed yet. In that sense, the talk will first introduce into the concept of research software from the point of view of a scientific (digital) library mainly responsible for managing and providing ‘traditional’ research output. The approach towards research software may be conceived as a mix of activities like standardizing, maintaining, versioning, archiving, monitoring and hosting of research software, with the overall goal to identify and to support those pieces of software code, which are most likely to be (re-)used for scientific purposes, while others may be ephemeral, proprietary, or just too specific to become reusable in other scenarios.
In its second part, the talk will have a closer look in particular at statistical packages, which is a prominent, if not dominant category of software in Social Sciences and Economics. In the light of a question like "What kind of software is commonly used in research, in a certain community or discipline?", we examined the acknowledgement, citing or operational in-depth-use of statistical packages in different academic contexts like job announcements, academic papers, search engines for academic publications, and data repositories as indicators for their scientific significance and value. By this study, we identified packages which might be more relevant to be hosted, maintained and preserved in the light of processing and reproducing research results, than others. The talk concludes with options for infrastructure providers responsible for managing research software.