Date of this Version
Without software, modern research would not be possible. Understandably, people tend to marvel at results rather than the tools used in their discovery, which means the fundamental role of software in research has been largely overlooked. But whether it is widely recognised or not, research is inexorably connected to the software that is used to generate results, and if we continue to overlook software we put at risk the reliability and reproducibility of the research itself. The adoption of software is accompanied by new risks - many of which are unknown to the majority of the research community. The practices of software sustainability minimise these risks and help to increase trust in research results, increase the rate of discovery, increase return on investment and ensure that research data remains readable and usable for longer. Funders are well aligned with the goals of software sustainability: both seek to support reliable, trusted research. This means that funders are well placed to play a pivotal role in advocating software sustainability. Funders can help raise awareness, and can make some simple, low-cost changes, such as encouraging the adoption of software management plans, that could lead to significant improvements in the software used in research. Improving software sustainability requires a number of changes: some technical and others societal, some small and others significant. We must start by raising awareness of researchers’ reliance on software. This goal will become easier if we recognise the valuable contribution that software makes to research - and reward those people who invest their time into developing reliable and reproducible software. We must educate the research community on the issues raised by software adoption, and provide training in the software engineering skills that are needed to overcome them. We cannot rely on researchers to adopt all of the skills needed for software sustainability, we must also allow research groups to recruit software experts, and we must create organisations that develop and disseminate expertise in software sustainability. The adoption of software has led to significant advances in research. But if we do not change our research practices, the continued rise in software use will be accompanied by a rise in retractions. Ultimately, anyone who is concerned about the reliability and reproducibility of research should be concerned about software sustainability. This argument alone may rely too much on the stick and not enough on the carrot. To that end, we must also show that software sustainability promises to identify and make available the software that is most likely to advance research.