Redmond, Wash. - When Microsoft Research Vice President Rick Rashid called Dan Ling to offer him a job, Ling hesitated. After all, Ling knew Microsoft as a proprietary, quick-to-market company that barely contributed to the academic community.
But still Ling left IBM's prestigious T.J. Watson Research Center where he worked on the RISC processor to join Microsoft's fledgling research effort.
That was 1992, a year after Microsoft Research was started. Seven years later, as director of Microsoft Research, Ling sports a smile that says he made the right choice.
The biggest reason Ling left IBM is because Big Blue was afraid that research efforts would obsolete the company's existing products. "Microsoft doesn't have this fear," he says. In fact, the company encourages research teams to work closely with product groups, and if need be, toss out old features in favor of what research is developing.
The main backer of this attitude is Nathan Myhrvold, Microsoft's chief technology officer. Myhrvold, who cowrote "The Road Ahead" with Bill Gates, has a unique combination of futuristic dreams and creative energy. A physicist by trade, Myhrvold studied under cosmologist Stephen Hawking. Meanwhile, he relishes gourmet cooking and has a certificate in French cooking. It is this convergence of unlimited thought and real-life trial and error that guides Microsoft Research.
"Even though we do not force our researchers to do short-term [research], when they do come up with something great, we're able to productize it," Myhrvold says.
Rick Rashid, who helped Myhrvold found Microsoft Research, says the drafting of technology into products as it is developed is key to the company's success. "Today, virtually all of Microsoft's product groups use technology created in our research organization," he says.
Myhrvold says Microsoft's research efforts are different from other corporations. "We do broad, curiosity-driven research. This is the same model that Bell Labs and IBM Research among others have [followed] in the past, but more recently, those places have tended to focus on applied research... and near-term problems," he says. Many projects at Microsoft are at least five years from completion.
He says another difference is that Microsoft is a software company doing software research.
"Most software research in the past has been done by companies in very different lines of business - telephone companies (Bell Labs), telephone equipment companies (Lucent), computer hardware companies (IBM and DEC), or office equipment vendors (Xerox)," he says. "The reason these companies have a hard time with technology transfer is simply that they are doing research in areas outside their business."
Software is a wide playing field and the three top executives agree that there is room for far-reaching projects. "A number of our projects are 'out there' in my opinion," Myhrvold says. "Some include work on quantum computing, new approaches to theoretical computer science and a couple of projects looking to revive the dream of artificial intelligence."
But even the most far-out projects are expected to contribute to today's product line. For instance, the intelligent interface group works with the Encarta team to improve its search engine and with the Office team to enhance the applications' understanding of foreign languages.
Inside these hallowed halls Microsoft says a key to its success has been to preserve an academic environment for researchers.
"We operate much like a university computer science department," Rashid says. "Individual researchers make their own decisions as to the directions of their work and are validated by their research results and by the normal academic peer review processes."
More than half the staff has Ph.D.s, according to Ling, who is working to expand the staff by next year. But Ling says it wasn't always easy to attract researchers, let alone Ph.D.s, to the Redmond campus. Now Microsoft has facilities in Cambridge, England, Beijing and San Francisco.
"In the beginning, it was difficult to hire people," Ling says. "It required charm and persuasion. Academics doubted whether we were really going to be open and contribute to the research community." To prove that Microsoft was serious, company researchers started publishing their findings in journals, appearing at conferences and co-authoring standards. "That's when we started having credibility in the academic community."
"For us the most important question is 'what is the impact of our research on the computer science research community and how does Microsoft Research improve Microsoft as a company?" Rashid says.
And as other companies slash their research budgets, Microsoft continues to beef its up. Rashid says the fact that so many of the technologies have gone into products, helps prove the investment is worth it.
"We consider our investments in basic research to be critical to the long-term success of the company and don't foresee any cutbacks or stagnation," he says.
