Microsoft's Rick Rashid on Building a Corporate Research Giant

The winner of the 2008 IEEE Piore Award considers Microsoft's impact on astronomy, biomedicine, Yahoo Research, and HP Labs

6 min read

On 10 June, Rick Rashid received the IEEE Emanuel R. Piore Award ”for contributions to the design of modern operating systems, and for innovation and leadership in industrial research.” Rashid is a senior vice president at Microsoft and founder of Microsoft Research, the world's largest software research and development organization. Senior Associate Editor Steven Cherry interviewed Rashid by phone shortly before the awards ceremony about his own software-development days and his vision for Microsoft Research.

IEEE Spectrum: As a researcher, you're best known for the Mach operating system. The Piore Award announcement notes its enormous impact on operating systems to this day.

Rick Rashid: That was something I did when I was at Carnegie Mellon University [CMU] back in the early 1980s, and it has had a pretty significant impact on a lot of different systems. It's the basis for the modern Mac OS. In fact, it was used directly by Steve Jobs for the NeXT OS.

Spectrum: NeXT was the visionary company Jobs founded between his two stints at Apple. Where else was it used?

Rashid: It was used by the Open Software Foundation for OSF/1, which became Digital Equipment Corporation's UNIX, Compaq UNIX, and HP Tru64 UNIX. Mach has really had a significant impact on Windows as well. The early work that led to the creation of the Windows NT project was based on some of the operating system work I did at CMU, which is how [Microsoft's then-chief technology officer] Nathan Myhrvold knew about me. So when Microsoft came looking for a director for a new research lab they wanted to start, I think that's why they had my name on the top of the list.

Spectrum: Besides operating systems, you did some applications research.

Rashid: I'm particularly proud of the fact that I did one of the earliest network computer games. It was something called Alto Trek, back in the days of the Xerox Alto. Aside from being a lot of fun, it influenced subsequent systems. It really explored the notion of what eventually became multicast on the Ethernet. So that was fun. Then later in the mid to late '90s, I adapted that same code base into a game that Microsoft released in 2000 called The Legion. That was Microsoft's first online-only game, and a few years ago we released it as free source code for anyone who wanted to do anything with it. It's still around in various forms--people are still playing it on the Web today.

Spectrum: Now your focus is on managing Microsoft Research?

Rashid: On the research management side of things, the nature of Microsoft Research is the thing I'm most proud of. It's something I created, and I think most people now regard us as the top research organization in the world. At a lot of top conferences, you'll see that anywhere from 10 percent to 20 percent of the papers are from Microsoft.

There's a huge impact on so many different fields. Look at SIGGRAPH [ACM's annual SIGGRAPH Conference on Computer Graphics and Interactive Techniques]; over the past 10 or 12 years we've been the dominant publisher. We've had a huge impact on a lot of areas of graphics research. If you look at the work we've done in computer vision, again, that's been very significant. In programming languages...we've really pioneered modern techniques for proving properties of large bodies of code. They've not only been adapted into our own internal development processes but they're available as part of the software developer kits for Windows Vista. So any hardware manufacturer that wants to use that tool can now prove over 100 properties of their programs just from their device drivers, directly from their source code.

Spectrum: Most of the previous honorees of the Piore award are in academia. In fact, you can go back 15 years and there's only one other industry recipient--Leslie Lamport, a senior researcher at Microsoft. Do you think that says something about the company?

Rashid: A lot of people look at us as an innovator in industrial research. The approach I took was to create the equivalent of a Carnegie Mellon computer science department--that type of environment--in the context of a software company. In many respects we behave much more like a university. Our people publish freely; there are no lawyers who read over their papers before they're published. We don't have any approval process like that.

We're a very open environment. We have more than 1000 Ph.D. students who work with us during the year as interns. We have as many or more members of the National Academy of Engineering as the entire University of Washington. I have more Ph.D.s working for me than Brown University. In China, we actually have the right to grant degrees in computer science; we run a joint Ph.D. program with Asia Pacific University, and we also run a large postdoctoral degree program over there.

Spectrum: There must be quite a few things that couldn't have come out of a less university-style organization. Is there one thing that comes immediately to mind?

”The approach I took was to create the equivalent of a Carnegie Mellon computer science department in the context of a software company”

Rashid: Many. We worked with the SQL team and developed all the data-mining tools that came out with SQL Server 2000 [a database-management product]. All the natural language and speech technologies, which are an important part of Vista and Office and which a really large number of people use, those all came out of research. I started the first e-commerce group in the company. DirectX, our 3-D graphics technology, came out of the research group, and in fact, I ran the DirectX team in its early days, kind of as a side thing. Going back to the early '90s, we developed the technologies that were then interactive TV and streaming video and audio over the Internet, and then we spun that out as the digital media division.

On the academic side, we're also moving in areas that are sort of off the beaten path but very much in line with where computer science technology is going. Some of our researchers are working on things like AIDS vaccine candidate analysis, looking at conditions like malaria, hepatitis C. With things like the WorldWide Telescope and the SkyServer--we just released WorldWide Telescope a few weeks ago--we've had a huge impact on the astronomy community. It's really changed the way they think about how they do their science.

Spectrum: Few companies organize research the way you have. One of the few in Silicon Valley is Yahoo. Not long ago they bulked up their pure research area.

Rashid: Yes. There have also been some changes going on at HP Labs. I see some of the changes that people are making as a kind of nod to what we're doing. If you look at some of the changes in direction at Yahoo and at HP Labs, I think it is to be a bit more like us, because I think people see us as a very successful model.

Spectrum: If the Microsoft-Yahoo merger had gone through, do you think it would have been fairly easy to integrate the research organizations?

Rashid: Well, one never knows. I know the people at Yahoo Research quite well. Many of them are friends of mine. It's like the academic community.

Spectrum: You mentioned former chief technology officer Nathan Myhrvold before. After Microsoft, he started Innovation Ventures, which, in a nutshell, tries to harness the wisdom of crowds and applies it not just to prediction markets and answering questions about the future but to innovation itself. Do you have any thoughts about applying the wisdom of crowds to research?

Rashid: It's a really big subject; there are lots of different angles and things to look at. First off, the wisdom of crowds doesn't quite apply to research. There just aren't that many people. If you look at the number of people who even hold Ph.D.s in computer science, it's not a large number, and the number of people capable of working at the state of the art in the field and at the highest level is just not very big. In any given research area...[there are] a few hundred to a few thousand total in the world. Those are not big numbers.

Spectrum: But it's not just computer science Ph.D.s. You pointed out that some of the most interesting areas involve the intersection of computer science and something that's completely different--biochemistry, astronomy, and so on.

Rashid: When you build a basic research organization at a company like Microsoft, or a basic research infrastructure in a country like the United States, it's an investment in being able to handle change and survive when things go badly. You get a lot of great technology, you get a lot of great people--all those things are by-products of having a basic research organization. If you go back to the writings of Vannevar Bush and the origin of the National Science Foundation, he looked at the experience the country had during the two previous wars and said, ”Look, we need to invest in having the capacity to change, the capacity to deal with situations that might be wars or famine or disease.”

For a corporation, it might be a new competitor, a new corporation, a new business model. In some sense, I think those investments are survival investments. The greatest testament to Microsoft Research is the fact that Microsoft is still here. Most of our peer organizations back in 1991 are gone, and hardly any of them invested in anything that resembled research. We were the only ones I can think of that made a significant investment in basic research. I think that's a testament to Microsoft's view about the future and a willingness to invest in that long-term view of the future and the long-term success of the company.

This article is for IEEE members only. Join IEEE to access our full archive.

Join the world’s largest professional organization devoted to engineering and applied sciences and get access to all of Spectrum’s articles, podcasts, and special reports. Learn more →

If you're already an IEEE member, please sign in to continue reading.

Membership includes:

  • Get unlimited access to IEEE Spectrum content
  • Follow your favorite topics to create a personalized feed of IEEE Spectrum content
  • Save Spectrum articles to read later
  • Network with other technology professionals
  • Establish a professional profile
  • Create a group to share and collaborate on projects
  • Discover IEEE events and activities
  • Join and participate in discussions