In the News - New Center Will Help Software Development "Grow Up"

U ic Basili knows the meaning of the phrase "there's no profit in prophecy." Basili, professor of software engineering at the University of Mary-land, has long championed empirical research and experimentation in the development of large software systems, but his message fell on deaf ears. "For a long time in the early days, people didn't even know software had to be measured, Basili says. "In the late '60s, software development was about building programs. And then it shifted to building systems-and it didn't scale u p right." Not "scaling up right" can cost a project thousands of hours and millions of dollars t o correct. That's where the Center for Empirically Based Software Engineering comes in. Basili and his colleague at the University of Southern California, Barry Boehm, have created CeBASE and are now forging a network of colleagues in industry, the academic community, and research positions t o establish CeBASE as an indispensable resource. CeBASE's resources are intended to help software engineers build on existing empirical data t o reduce costs of developing large projects. The center went public in October, announcing it had received a two-year, $2.4 million grant from the US National Science Foundation. "CeBASE is a recognition that experimentation and the building of empirical models are important," Basili says. "I wouldn't say that's a majority opinion yet-but it's a large minority opinion." CeBASE's near-term task, says Basili, is t o demonstrate that this large minority opinion can be translated into tangible resources, such as a strong Web presence, conferences, books, and a network of researchers willing to share their experimental and empirical software development data. If that can be done in a timely way, he is confident the grant will be renewed, and CeBASE will be well on its way. "First of all, I think we'll show there's a stronger interest in this issue," he says. "Then. we have to show that we will have some basis of information on that network. We'll have t o show some models that are shared among the community. So the key is that we attract that community." Software quality guru Watts Humphrey, whose Personal Software Process and Team