Computer scientist defends corporate research

by Sylvain Comeau

Does corporate involvement in research mean less freedom? Not according to Yuri Gurevich, senior researcher and manager of Microsoft's Foundations of Software Engineering group.

"I have more scholarly freedom at Microsoft than I ever did in university," Gurevich said at a lecture here on April 27. "At Microsoft, they basically told us, 'Just do whatever you think is right.'"

In an interview following his lecture, Gurevich explained that financial constraints can still act as a curb on research options.

"In university, your research activities are limited by your funding, especially in computer science. You can only pursue your project if a funding body or a corporate partner also shares your vision."

Gurevich's team is developing algorithms to speed up debugging of computer programs in the design stages. This process can be very time-consuming and expensive, as bugs easily creep into the millions of lines of code required in new software. Ironically, Gurevich feels that his team would not have accomplished as much had they been given strict, more results-oriented guidelines or deadlines.

"Research, especially fundamental, basic research, requires a lot of creativity," he explained. "When we're allowed to explore our ideas and intuition, good things happen. And if a team is given few constraints, that is a fun atmosphere in which to work. If people enjoy what they're doing, they do their best work, and practical applications will eventually follow from their work."

Gurevich says that Microsoft set a precedent by establishing a group that was free to follow its own inspiration. He points out that other high-tech companies are following suit by establishing their own freewheeling research teams; Gurevich himself briefly worked with IBM's Theoretical Computer Science Group. He feels that the corporate world should show the same foresight when funding university work because neither side wins when there are too many strings attached.

"Companies should not dictate. I think they should be very careful in finding the right people and the right groups. If they find a group who are sufficiently self-motivated, they are likely to come up with something the company can use. That is the best strategy."

Gurevich teaches electrical engineering and computer science at the University of Michigan. His lecture, which was titled "Formalware Engineering," was sponsored by the Faculty of Engineering and Computer Science.

Photo: Seen before his lecture in the DeSève Cinema on April 27, computer scientist Yuri Gurevich (above, second from right) paused for a photo in the Faculty of Engineering and Computer Science offices with Louise Quesnel, the Faculty's Advisor on External Affairs, Professor Gosta Grahne, Associate Chair of the Department of Computer Science, who played an active role in organizing the lecture, and Professor Rama Bhat, Associate Dean, Graduate Programs and Research.