More than one-fifth of academics who develop their own software for use in research have had no training in software development, research has revealed.
Some 92 per cent of scholars surveyed say they use research software, and 69 per cent say that their work would not be practical without it.
However, 21 per cent of the 56 per cent of academics who develop their own programs have received no training in software development, according to research undertaken by the Software Sustainability Institute, which advises on producing better research software.
“Look around any lab and you will see software – both standard and bespoke – being used in all disciplines and by all seniorities of researchers,” said Simon Hettrick, the Software Sustainability Institute’s deputy director.
“Software is clearly fundamental to research, and I find it terrifying that we don’t understand quite how much we rely on it,” he added.
For his survey, Dr Hettrick questioned 417 researchers selected at random from 15 Russell Group universities.
He discovered that 20 per cent of the respondents who were responsible for writing grant bids had not included software development costs in their proposal, even though they knew that software would be needed.
“Anecdotally, we had heard that researchers tend not to include software development costs…because they think that it will weaken their bid,” Dr Hettrick said.
There was also evidence of a gender divide – with 70 per cent of men saying they had developed their own research software, compared with 30 per cent of women.
The use of research software has caused problems in the past. Phil Jones – one of the University of East Anglia’s scientists at the centre of the “Climategate” scandal of 2009 – was criticised by the Commons Science and Technology Committee for not revealing details of the computer code that had been used in his research. This code had been created specifically for the project on which Professor Jones was working.
Some working in the sector have called on universities to employ dedicated teams of software engineers to work with researchers when required – in order to safeguard against the poor use of computer programs.
“Software use is not going to decline,” Dr Hettrick said. “We need to be aware of our reliance on software so that we can ensure that the [academic] community have the tools and the skills they need to stay at the forefront of research.”