Ten years since the Software Sustainability Institute was established, we look forward to the next decade: what can we expect to see?
As Richard Gunn, Head of e-Infrastructure at UKRI-EPSRC points out, “it’s probably foolish to make confident predictions during times of crisis” like the coronavirus lockdown in the UK, but “current events are exposing the downsides of efficiency at the cost of resilience, and I’d expect this to be one of the significant themes, post pandemic.”
“Certainly, there is an increased, and uncomfortable awareness of our reliance on many aspects of the modern world. This is extremely relevant to the work of the Institute in promoting software sustainability,” Gunn says.
“There will be increasing interest and debate in the social values encoded in software – e.g. the current debate over contact-tracing apps and surveillance, use of facial recognition, etc. This relates to one of the great successes of the first 10 years of the Institute: its commitment to promoting an understanding of how the process of software development intersects with research.
“The ability to interface between increasingly complex and powerful technologies and research disciplines will become an increasingly important and valuable skill. As software becomes more powerful and integral to research methods, the SSI stands to play a key role in the debate over reproducibility and accuracy. This will be increasingly consequential (e.g. the recent debate over modelling inputs and the impact/spread of COVID-19),” says Gunn.
Kirstie Whitaker is Programme Lead for Tools, Systems and Practices at the Alan Turing Institute. Looking ahead, she plans to continue to work with the Software Sustainability Intistute on the Turing Way and in increasing collaboration across academia.
“No one knows how to collaborate! There’s a lot of work needed to teach people to work together, to work on Github, to run collaborative events… so we’re trying to curate a lot of the content that’s on the Institute’s blog and make it more discoverable.”
“I also want to see the Institute fit further into a network of other institutes around the world that are aligned with its values – its decentralised approach, and equitable and inclusive practices.
“In future, I’d just say keep doing what they’re doing! And increase their friends and allies to amplify the message. The way they run events, the way they select Fellows: it’s always been ‘person first’ and that, I think, is the future”.
“My take is that the SSI has always had a psychological space of curiosity, of wanting to learn and to connect with passionate people. That will stay the same, but the things they connect with will change. There will always be new problems to tackle, but their focus on people is – and always has been – the answer,” Whitaker says.
Sally Price is Professor of Physical Chemistry at University College London (UCL). She has been involved with the Institute from the beginning, writing a letter of support for its first funding application and providing one of the first projects for the team: creating a stable version of the NEIGHCRYS/DMACRYS software.
While ten years may seem a long time to some in the computer modelling world, she says, “it is actually quite a short time in the development and use of large computational chemistry codes”.
“The landscape has changed in that computer modelling is being used far more widely as a complement to experimental work, as the speed of the computers is allowing electronic structure and atomistic modelling (DMACRYS performs atomistic modelling of the crystal structures of organic molecules using anisotropic atom-atom intermolecular potentials derived from the electronic structure of the molecule) to be applied to more practical problems with worthwhile results. Thus, we find ourselves reusing and adding to crystal structure prediction calculations performed many years ago, when additional experimental results appear.
“To be able to add the corresponding calculations we have needed to dig out old versions of the code from the SVN that was set up by the Institute. We have also distributed the code to various groups, both academic and industrial, who can use the test-suite that the Institute developed to check that they have set up the code correctly on their machine,” Price says.
“I hope that SSI will continue to flourish for decades to come, and be able to integrate sustainability into software, so that less software falls into disuse from lack of people being able to support the continued use of the codes. There are many codes that still rely on support from the original developers who are, or are approaching retirement,” she says.
The Institute has had many successes over the past ten years: please read our article from yesterday on the impact it has had over the decade, and our Q & A with Neil Chue Hong and Simon Hettrick. With the success of the Research Software Engineer movement, the establishment of an influential network of Fellows and its great community and events schedule, the Institute sees a promising future for research software.