Scientists at universities perform much of the world’s cutting-edge scientific research—often while relying on shaky, homemade computer software written by students and postdocs. Schmidt Futures, a philanthropic initiative founded by former Google CEO Eric Schmidt and Wendy Schmidt, his spouse, hopes to remedy that situation by investing $40 million over the next 5 years to establish a Virtual Institute for Scientific Software, the organization announced today. The institute will help scientists obtain more robust, flexible, and scalable “open-source” software that can be easily shared.
The institute will include centers at the Georgia Institute of Technology (Georgia Tech), Johns Hopkins University, the University of Cambridge, and the University of Washington (UW). Each university will hire software engineers who will help meet the needs of scientists, explains Eric Braverman, CEO of Schmidt Futures. “We believe that a network of people developing software will be essential to the onward development of so many areas in the scientific enterprise,” he says.
“When I heard about this initiative, I was like, ‘Oh, this is gonna be great!’ because I can easily see the need in my lab,” says Nancy Allbritton, a bioengineer and dean of engineering at UW Seattle. Allbritton, who develops microdevices that incorporate living tissue, credits Schmidt Futures for addressing a crucial need. “Someone was very smart and thinking, ‘How could I invest money for the biggest payback?’”
These days, researchers depend on computers for everything from running their equipment to collecting and analyzing their data. As computing power continues to grow, scientists face new challenges. They have to make sure their software can scale up to handle the massive data sets much research now produces, says David Beck, a chemical engineer at UW Seattle. “You get a graduate student who works out a really nice solution on their laptop for 1/100th of the data, but they don’t have the skills to necessarily get them to the full petabyte data set,” he says.
Similarly, as computer hardware changes, researchers may struggle to make software that ran on one machine work on the next, especially if the programs require high-performance computers, says Alessandro Orso, a software engineer at Georgia Tech. “Basically, researchers are faced with having to build this system to handle massive amounts of data on a shifting platform,” he says.
Software engineers can handle just such problems. However, compared with government and private labs, universities often struggle to hire these professionals, who often receive high salaries and other compensation in the private sector. “Universities offer terrible stock options,” says Stuart Feldman, a computer scientist and chief scientist at Schmidt Futures.
A grant from the National Science Foundation (NSF) or National Institutes of Health (NIH) typically only pays for a fraction of a full-time software engineer, says Andrew Connolly, an astronomer at UW Seattle. “You can’t get people to come in and work on your project when you can’t provide them with some kind of long-term career development,” he says.
With $2 million per year, each center in the new virtual institute will hire a team of professional software engineers who will provide their services to the entire university. Georgia Tech envisions hiring a half-dozen software engineers, including a lead engineer, Orso says. UW plans to hire five, Beck says. Connolly says those engineers might, for example, help UW astronomers deal with the enormous data sets that will come from the recently completed Vera C. Rubin Observatory in Chile.
Each center will seek to pay salaries that are slightly below private sector levels and rely on the allure of the science to attract candidates, Feldman says. “Instead of ad optimization, you get to explain the expansion of the universe or the evolution of the climate 50 million years ago,” he says. “There are both psychic and social rewards.” After 5 years, Schmidt Futures will evaluate the centers’ performance and reconsider the need before making more funding decisions, he says. Allbritton hopes federal funders might take note of the initiative. “One might imagine,” she says, “that NIH and NSF could pick up on something like this and really amplify it.”