Research software at the University of Southampton

Abstract

Most research would not be possible without software. It is the enabling technology behind major advances, from decoding the human genome to the discovery of the Higgs boson, it lies at the heart of strategically important technologies, such as artificial intelligence, and it is a tool of constant use and fundamental importance in day-today research. Despite its pivotal role, support for software and software skills within academia has not kept pace with the rapid adoption of software as a research tool. This is a major barrier that inhibits research and puts results at risk.

A 2014 study [1] found that 92% of UK researchers use software and 69% report that it is fundamental to their research. It is not possible to disentangle the reliability of the software used to generate results from the reliability of the research itself. Unreliable software leads to unreliable results that cannot be trusted. Trust cannot be established in poorly engineered software, because it is too difficult to understand whether the software is reliable. Without trust, researchers will not build on existing software and will instead recreate it themselves: an avoidable duplication of effort and a waste of resources.

In 2019, the Southampton Research Software Group (SRSG) [2] conducted a survey of research staff to identify the relationship between software and research at the University of Southampton. The SRSG works across faculties to provide software engineering expertise and training to researchers. The survey found that software use is near ubiquitous across all but one faculty, researchers report that software is vital to their work, a third of researchers develop their own software, and around a third who develop software are interested in commercialisation. Worryingly, around two thirds of staff who develop software do not believe they have had sufficient training to develop reliable software.

Software has an almost unparalleled power to advance research and enterprise. Much of this potential is likely to remain unrealised at the University, unless researchers are given access to the necessary software skills and expertise.

More information

Identifiers

ORCID for Simon Hettrick: orcid.org/0000-0002-6809-5195

ORCID for Stephen Crouch: orcid.org/0000-0001-8985-6814

ORCID for Philip J Grylls: orcid.org/0000-0001-9677-5852

Catalogue record

Export record