Lack of support for software engineering is holding back efforts to improve the reproducibility and openness of scientific research in the UK, a panel of MPs heard yesterday.
With data analysis at the heart of many cutting-edge scientific fields, the need for support from software engineers has never been greater, but it is badly needed, according to two researchers speaking to the Commission parliamentarian for science and technology.
[Universities] think, ‘Oh, you’re a computer guy. We have a note for that. How many people do you manage? Any? In this case it’s £42,000 a year’…and it just doesn’t work
Dr Jessica Butler, head of analysis and research fellow at the University of Aberdeen, said universities were not designed to support people like software developers or statisticians, who have no ambition to be published in a “fancy” academic journal.
“They want accurate stats. They want to write code that works. They could easily work for a company and get paid six times as much and get promoted,” she said of the engineers.
Dr Ben Goldacre, head of the Nuffield Department of Primary Care Health Sciences at the University of Oxford, told MPs his team employed software engineers who could get ‘very good low six-figure salaries’ outside of academia.
“They’re willing to halve their salary to work with me if it’s a worthwhile project. They’re not willing to cut it to £30,000 or £40,000,” he said. It’s $40,000 or $53,000.
The problem was that university pay levels didn’t recognize the skills of software engineers, Goldacre added.
“A full-featured, commercial-grade software developer has a different skill set than the person who fixes your printer when it breaks down. [Universities] think, ‘Oh, you’re a computer guy. We have a note for that. How many people do you manage? Any? In this case it’s £42,000 a year’…and it just doesn’t work,” he said.
Goldacre said his team managed to “hack” the system by using research money to employ its own developer and create software tools that other academics can use, but the practice was not common.
“There is almost no open competitive funding available anywhere for anyone who wants to build this type of middleman [software] tools. They’re intellectually and academically creative, and incredibly powerful and high impact in terms of the good they do, but there’s almost no competitive funding open to do this kind of work, so this kind of work doesn’t is not done,” he said.
Software engineering in these teams was usually done as an “amateur activity” by scientists whose primary roles lay elsewhere. While capable, they were no substitute for experienced, skilled software engineers, he said.
“It’s research software engineers who are at the heart of this new way of working. You need people who are really good at writing software, not people who are researchers who have dabbled a little here and there. and a legitimately funded activity.”
The problem of recruiting and retaining software engineers goes back to the so-called reproducibility crisis in science.
In its written evidence to the committee, the joint academic group Software Sustainability Institute said that about 69% of research is produced with specialized software, which can be anything from short scripts solving a specific problem and worksheets complex calculations analyzing the collected data, to the millions of lines of code behind the Large Hadron Collider and the Square Kilometer Array.
“With many studies, research published without the underlying software used to produce the results is unverifiable,” the submission states.
Goldacre told the committee today: “The true and complete story of how you produce your results is the code you wrote, attached to the data you ran it against. But unfortunately, our release model scientific discoveries was built in the 19th century.” ®