REDMOND, Wash., June 21, 1999 — When Nathan Myhrvold first approached him to start a research laboratory at Microsoft, Rick Rashid was skeptical. For one thing, he felt extreme loyalty to Carnegie Mellon University, where he had taught for the past dozen years. He was passionate about his research, he enjoyed the academic work environment there, and he was next in line to become dean of the university’s computer science department. Moreover, Microsoft was still a relatively small, young software company. How could he know if the company was really committed to research, or if it would even still be around in five years?
But Myhrvold persisted. He visited Rashid at Carnegie Mellon University (CMU). He even persuaded Rashid to fly to Redmond, Wash. to talk to Microsoft’s top executives. After three months, Rashid gradually warmed up to the offer. “It was one of those things where I had a lot of loyalty to the organization I was in, and I wanted to continue and support it,” Rashid said. “But I ultimately decided after spending the summer thinking about it that going to Microsoft would be just a huge amount of fun.”
Eight years after agreeing to start Microsoft Research, Rashid has built one of the world’s premiere computer research organizations. He is credited with hiring a cadre of top researchers from an array of technical backgrounds. He has created an open, academic-style environment in which researchers are free to pursue their interests. And he has successfully worked to incorporate his group’s research into nearly every Microsoft product.
“Microsoft has built the strongest computing research organization in the world in a period of eight years,” said Ed Lazowska, chairman of the University of Washington’s Department of Computer Science and Engineering. “Rick has been in charge. That’s a pretty significant contribution.”
Sitting in a tidy, second-floor office overlooking a small stand of fir trees, Rashid, 47, explains his role as vice president of Microsoft Research. While Myhrvold, as Microsoft’s chief technology officer, provides broad direction for the research group, Rashid’s job is to manage the day-to-day operations of the organization. It’s a role that will not change with Myhrvold’s decision to take a yearlong sabbatical beginning July 1.
“The group has been reporting to me, and the group will continue to report to me,” Rashid said. “The only difference now is that if I have a critical issue to deal with, I’ll ask Bill Gates directly whereas I used to ask Nathan. Nathan will be on e-mail, and I do plan to talk to him. It’s just that he’s not going to be here on a daily basis.”
From Literature Major to Operating System Developer
Clean-cut and casual, Rashid immediately puts people at ease. He is both soft-spoken and personable, and punctuates his conversations with dry wit and a warm laugh. “It seems like everything I’ve done, I’ve done by accident, but nothing bad has happened to me yet,” he joked.
The first accident for Rashid was getting into computer science. A comparative literature and mathematics major at Stanford University, Rashid was first introduced to computers by fellow classmate Dan Ling in the early-1970s. Ling, now a director of Microsoft Research, convinced Rashid to take a graduate-level computer programming course. The course turned out to be disappointing, but Rashid went ahead and took a second computer science course — this time in computer architecture, which got him hooked. “The great part about that course was that you were given access to a couple of Hewlett-Packard mini-computers,” Rashid recalled. “And I can remember the evening where I got my class-project program working. It was just a rush knowing that my brain had animated this piece of metal. I floated on my way back home to my dorm room.”
Despite his newfound interest in computers, Rashid decided to study graduate-level mathematics and was accepted at the University of California at Berkeley. But before he could start the program, a professor at Stanford invited him to join the new computer science department he was forming at the University of Rochester. “I really had great fun with computer science, so I thought I’ll go ahead and try it, and if it doesn’t work out, I can always go back to graduate school in mathematics,” Rashid said.
Upon arriving at Rochester, Rashid was the first of eight students and three faculty members to show up on campus. The others eventually joined him, and they became a tight-knit group that frequently headed to lunch in the same station wagon and spent hours working together on computers. Fascinated by the ability of software to control the underlying hardware of a computer, Rashid spent most of his time developing operating systems. At the same time, he built hardware and earned a Ph.D. in “computer vision,” the analysis of computerized moving images. “That’s been the hallmark of my career,” he said. “I always like to do research in many different areas and work on many different things simultaneously.”
Rashid’s diverse interests continued after he joined CMU as a faculty member in 1979. He published several papers in the areas of computer vision, operating systems, programming languages for distributed computer systems, computer network protocols and network security. He also developed one of the earliest networked games for the Xerox Alto computer called “Alto Trek.” Most notably, Rashid headed the effort to develop the Mach operating system. “That project produced a lot of good ideas, lots of good students and a highly influential system,” Lazowska said.
The Mach architecture formed the basis for the Open Software Foundation’s version of Unix operating systems and will be incorporated into the next version of the Macintosh operating system. “It’s been very influential in the way people go about building operating systems,” Rashid said. “People derived a lot of the concepts that are used today from that system.”
A World-Class Research Organization in Eight Years
Rashid’s first challenge at Microsoft was recruiting other top researchers to join the company. Gates and Myhrvold wanted to turn Microsoft into a company that invented its own futuristic technology, and Rashid’s job was to hire the people and create the atmosphere to make it happen. At first, convincing researchers to come to Microsoft was a tall order. “It was very hard to hire in the first year because there was nothing to show,” Rashid said. “It was just my smiling face and Nathan’s smiling face. I had to convince other people that this really was going to be successful.”
With a little persuasiveness and a lot of luck, Rashid said he was able to hire several good people during the first two years of Microsoft Research. And once a few good people had joined the organization and began obtaining results, recruiting gradually became easier. “It’s a little like gravity,” he said. “The more you have, the more you get.”
Today, Microsoft Research employs 420 top computer scientists and engineers in Redmond, San Francisco, Cambridge and Beijing. Its scientists conduct research in nearly 30 disciplines ranging from computer architecture to artificial intelligence. Some are striving to achieve Gates’ vision of an intelligent computer that can see, listen, speak and learn. Others are working to make information securely available over a variety of computerized devices and home appliances. All of the group’s research is focused on simplifying and enhancing people’s experience with technology.
During the past eight years, Rashid has tried to emulate the academic-style environment he learned at CMU. He strives to create an open atmosphere free of bureaucracy so employees can focus on research without the distractions of administrative tasks or office politics. Within each discipline, employees must decide for themselves where to focus their research. They’re free to publish papers without the approval of Rashid or anyone else at Microsoft. There is no set budget for any area of research, so researchers can request the money they need to complete projects, but no more. “At CMU, there was this notion of ‘the reasonable person principle,'” Rashid said. “There weren’t a lot of rules. There were just intelligent people making intelligent choices, and that’s the approach we’re trying to take here.”
In addition to maintaining an open atmosphere, Rashid strives to create a collaborative environment in which people are rewarded for sharing ideas — not just developing them. This atmosphere has been responsible for innovations like ClearType, a technology that dramatically improves the clarity of type on computer screens. “That work has brought in people from the hardware and prototyping group, our computer graphics group, our signal processing group and our theory and algorithms group,” Rashid said. “It wouldn’t have come about if all those people hadn’t worked together.”
The environment at Microsoft Research is widely regarded as a model for corporate research organizations worldwide. It is also praised by Microsoft researchers, who credit Rashid with building a highly productive work environment from the ground up.
“Rick delegates both responsibility and authority,” said Jim Gray, a Microsoft researcher who recently won the prestigious A.M. Turing award. “He does not direct the research group; he enables it and helps it when it gets stuck.”
“Rick has been a tremendous research manager,” Ling said. “He hires outstanding people and gives them the freedom to be entrepreneurial in defining research directions, starting new activities and working on new ways to interact with product groups.”
“Rick provides a steady hand for Microsoft Research,” said Linda Stone, director of the organization’s Virtual Worlds Group . “He is thoughtful, practical, patient. He’s excellent operationally and great at growing people as well as thinking creatively about how technical projects can get pushed along.”
Transforming Innovations into Products
Under Rashid’s leadership, Microsoft Research has successfully incorporated numerous research innovations into Microsoft products. “Microsoft Research has a tremendous track record in impacting the company’s products,” Lazowska said. “Very advanced work from Microsoft Research in areas such as natural language processing, statistical decision theory and computer graphics has been embedded into Microsoft products in ways that users take for granted, but that really change the functionality and usability of those products.”
Asked to name some of the innovations that have found their way into Microsoft products, Rashid rattles off a long list. The grammar checker that identifies grammatical errors in Microsoft Office. The natural language query that allows users of the Encarta encyclopedia software to search for information by typing in questions in plain English. Microsoft audio technology that enables music lovers to store a greater number of audio clips on their personal computer. The index tuning wizard that automatically figures out the best way to lay out a database for SQL Server users. Technology in Microsoft Commerce Server that enables Web site owners to offer customized information to customers. The optimization technology that has made every Microsoft software program take up half the memory it used to.
“You can pretty much name any activity that we have going on in Microsoft Research, and there’s probably a particular result that’s made it into Microsoft products or a collaboration between our group and a product group that’s ongoing,” Rashid said.
Rashid even has a few of his own innovations to add to the list. He developed the original software for Windows Media Technology that enables users to broadcast videos over the Internet. He created some of the advances in high-speed networking that have led to faster Internet connections. He even used his spare time to develop an online space-action computer game called “Allegiance,” which Microsoft plans to launch this winter. “It’s amusing to see Microsoft Research as a game developer, but what the heck,” Rashid said.
Another measure of the group’s success has been the large number of research papers it has authored at industry conferences, Rashid said. For example, Microsoft was responsible for more than 25 percent of the research papers delivered at last month’s Programming Languages Design and Implementation Conference. And Microsoft authored 20 percent of the papers delivered at the 1996 SIGGRAPH computer graphics conference. “These are statements about the quantity and quality of research that’s being done by our people here and of the level of innovation taking place,” Rashid said.
In addition to accelerating the pace of technological development, Microsoft Research has helped the computer industry by reviving the notion that corporate research is a good investment, Rashid said. “When we first started out, everybody was cutting back — IBM, AT & T, Xerox — everyone,” he said. “Everybody was saying that they weren’t getting a benefit from this. We’ve shown that we can create a first-rate research organization. We’ve shown that being open and being involved in the academic community is important and valuable, and I think that’s having an influence on how other labs are being run or are starting to be run.”
Building a Global Think Tank
To keep the pace of research in line with Microsoft’s overall growth, the company two years ago decided to expand the size of its research organization to 600 people by July 2000. “It was Bill’s feeling and Nathan’s feeling that research needed to grow faster to keep pace with the company,” Rashid said. “The company had been growing faster than Microsoft Research had, and there was a sense that increasing our efforts was one of the best investments the company could make toward its future.”
But while Microsoft Research is growing, Rashid said complaints that Microsoft is raiding the academic world for researchers are unfounded. Fewer than 50 of the 420 employees hired by Microsoft Research have been faculty members at universities, he said. And that’s a small number compared to the more than 3,000 computer science faculty members that are out there. “Over the past eight years, we’ve probably hired on average six people from teaching positions in a year,” Rashid said. “That’s a blip. That’s unnoticeable in statistics.”
Microsoft Research has actually strengthened university research because its scientists and engineers are active participants in the academic world, Rashid said. The group shares its research with the academic community at industry conferences. It also encourages university professors to give talks at Microsoft and to gain a solid understanding of the real-world concerns and pressures facing Microsoft software developers.
While growth is an investment, it also presents challenges. One challenge is managing research facilities that lie half way around the world from Redmond, Rashid said. Another is maintaining effective communication, both within Microsoft Research and between the research group and the company’s product teams. “Certainly as the company has grown, and as the size of Microsoft Research has grown, it’s been more challenging to keep those contacts,” Rashid said. “There are 13,000 developers now, which is far more than the company had when I first got started.”
Yet another challenge is keeping up with the rapid pace of technology. From 1998 to 2001, the quality of computer graphics will improve by a factor of 100, Rashid said. And within two years, the storage size of disk drives will increase by a factor of seven. “Those are enormous changes, and it’s challenging to make sure the technologies we’re working on are relevant and state-of-the-art,” Rashid said.
Despite the breakneck speed at which technology is changing, Rashid said he often feels like technological advances are occurring in slow motion. “In some sense for me, the change doesn’t happen fast enough,” he said. “I look back at the last 20 years, and I think, wow, so much has changed. And I also think, wow, it doesn’t seem as if we’ve done as much as we could have. There’s so much that remains to be done.”
Bringing Star Trek Down to Earth
So what changes does Rashid expect to see within the next few years? For one, computer users will be able to store the entire contents of the Internet on their personal hard drives. “That’s going to be a huge opportunity because when I do research, I’ll be able to analyze information in ways I’ve never been able to do before,” he said.
People will also be able to view graphic images on their personal computers that offer the same quality as Toy Story, the movie hailed as a breakthrough for computer animation. “That’s going to create tremendous opportunity, not just in entertainment, but in medicine, user interfaces and so forth,” Rashid said.
And like the Star Trek communicator, a communications device that Rashid has admired since childhood, people will be able to communicate instantly with anybody, anywhere. “The technology to do that exists today,” Rashid said. “Deploying it and creating an environment for it will take a little bit longer, but we’re not talking significant science fiction.”
As for him, Rashid will continue to bring more top-notch researchers to Microsoft Research and to guide the organization as it expands globally. At the same time, he will continue to roll up his sleeves and pursue the many areas of research that have always fascinated him. “I’ve been told by a lot of people that I have the best job in the world,” he said. “I’m not sure if that’s the case, because I don’t know what everybody else’s job is like. But there’s no question in my mind that this is the best job I’ve ever had. It’s just a huge amount of fun.”