Industry giants put millions of pounds into university research, but what do they expect in return? The THES spotlights three of the big spenders. How does the computing giant decide who to back and how does it treat them? Kam Patel reports
Microsoft Corporation does not call its headquarters in Washington state the "campus" for nothing. It aims to recreate the feel of a United States campus university - it is even organised along similar lines.
This is especially true for its seven-year-old research arm, Microsoft Research (MSR). There, says Dennis Adler, director of MSR's university relations department, 300 researchers work in divisions on areas ranging from operating systems and graphics to man-machine interfacing. "These are top-notch researchers, and the divisions are organised along similar lines to university departments," he says.
Last year MSR spent $2.3 billion on research and development - 18 per cent of Microsoft's turnover. Included in this figure is part of the Pounds 50 million being invested over five years to set up a research laboratory at Cambridge University. This is Microsoft's single biggest academic R&D expenditure outside the US. The Cambridge team conducts basic computer science research in fields such as data security, information retrieval, operating systems and networking (see Research, June 19).
Microsoft's "internal" expenditure is complemented by other spending on higher education that totals about $7.3 million a year. This outlay is in the form of cash, software, endowments and gifts and includes some funding for basic research. The funding is not broken down in detail by project areas: "We have not tracked it enough," admits Mr Adler, who adds that company is tightening its coordination and monitoring. It all seems a fairly laid-back approach, a reflection perhaps of the firm's vast resources.
Leaving aside the content of its collaborations with academic researchers, Microsoft has a well-developed philosophy for forging links with academics. Mr Adler explains that at conferences, MSR researchers often meet academics whose work they find exciting. "If MSR researchers feel strongly enough about the research an academic is doing, they will come back and tell us that we should be funding his work."
Mr Adler stresses that the work has to mesh with MSR's short to medium-term research interests. "Often with our academic links we are looking to the long term. Our product cycle is six to 18 months, and we continually look to integrate advances into new products. But it is difficult, for example, to work on integrating vision and speech on that sort of timescale. Our job is to look beyond that horizon, to back interesting work in universities. We help universities push out research, it helps us to see what the possibilities are, and eventually the consumer will benefit."
MSR has a group of academic advisers who review the company's research efforts. It includes Ed Lazowska, professor at the department of computer science and engineering at Washington University, and Raj Reddy, dean of the school of computer science at Carnegie Mellon University.
Microsoft's policy on publishing results of academic and internal research is relatively relaxed, Mr Adler says. "We do not really put any restrictions over publication of research in universities or within the company. Researchers decide what to publish, when to seek patent protection and so on. Nobody says to researchers that they cannot go to a conference and talk about this or that. We see putting our ideas out in the world as a form of peer review."
Mr Adler hesitates to pinpoint key research themes for the development of the computer systems: "It is really difficult to look ten to 20 years ahead in this industry. We tend to have a three-year planning cycle, but even within that things move extremely fast." Microsoft is working hard to develop vision and speech interfacing. This work, now in its infancy, aims to extend to the development of video-conferencing.
As for future university collaborations, Microsoft "has its hands full" with the Cambridge initiative, Mr Adler says. By 2000 the Cambridge MS research laboratory will boast 40 researchers.
"At Microsoft we like to have a rich interaction between product teams and researchers. I think that if our work with researchers (in Cambridge) is as successful as at our Seattle laboratories, there could be potential for other university collaborations."