Rashid on the Value of Research

Posted by Rob Knies

 

Once again, for the second year in a row, Rick Rashid delivered a bravura performance during his opening keynote address of Microsoft Research Asia’s 15th Computing in the 21st Century Conference.

Last year, during the same event, he closed his presentation by demonstrating to an awed crowd in Tianjin, China, a working example of speech-to-speech translation. Declaring that “personally, I believe this is going to make for a better world,” he left the stage to deafening applause. Some members of the audience admitted to being brought to tears.

His talk on Nov. 1 at the Hefei (China) Grand Theatre included no comparable razzle-dazzle. It was titled The Role of Basic Research in Innovation, and it simply, engagingly delivered as advertised.

But the keynote also represented so much more. Rashid, who founded Microsoft Research and spent 22 years as head of the organization, recently chose to return to his roots in operating systems, the realm in which he first staked his claim as one of the great computer-science minds of his era.

Given that context, when he examined the scope of what an investment in basic research can provide for a company and for an industry, he was delivering a valedictory of sorts. What the gathering of 1,500 Chinese students and professors heard was nothing less than a survey of a remarkable career that includes almost a quarter-century at the helm of the foremost computing-research organization in the world today.

Hard to Predict

Some would argue that the previous sentence need not include the word “computing.” But Rashid’s talk made no grandiose claims, no attempts to carve out a legacy. From a research perspective, that work already has been achieved.

Instead, he followed a simple script: take a cautionary look at the perils of predicting the future, examine his own involvement with an evolving computer ecosystem, and end on a positive note. It didn’t take much imagination, though, to connect the dots.

Noting that the conference’s theme was Decoding the Future, Rashid wryly observed that “the future isn’t really what it used to be,” that history is littered with failed, sometimes ludicrous, attempts to divine how technology would develop.

“It’s very charming how people in the past thought about the future,” he noted, but he also observed that a few have proved prescient, notably noted science-fiction writer Arthur C. Clarke, who described a communications satellite that, decades later, became reality, and Star Trek, with its vision of a communicator, something eerily similar to today’s ubiquitous mobile phone.

In the realm of predicting the future, though, those examples are hardly the norm. Such notables as Thomas Edison, Albert Einstein, and mathematician John von Neumann have uttered maxims that sound laughable in hindsight.

Rashid quoted von Neumann as saying, “You have to be careful with the predictions that you make, because you can look very foolish in five years.”

The problem, Rashid explained, is not just that you can’t predict when a technology will make a mark, nor the impact it might have, but even how it could be used. And the technology might be right, but the timing wrong.

“Technology does not move in a straight line,” Rashid said. “It’s hard to predict how cultures and individuals will react to technology.”

A Singular Career

With that, he pivoted to a discussion about his career has shaped—and been shaped—by the technological advances of the past four decades. While technology might not proceed in a straight line, for some people, the path of success certainly does.

Born in 1951—the same year the first UNIVAC computer was sold—in a “small town in a poor area of rural America,” Rashid began an odyssey that took him to undergraduate studies at Stanford University, post-graduate work at the University of Rochester, and, eventually, to a professorship at Carnegie Mellon University.

Along the way, he got an opportunity to work with a Xerox Alto, the first personal computer. It was a different era, though. Today’s computers operate at speeds of billions of operations per second; the Alto clocked in at only a hundred thousand. Storage was similarly minimal. In fact, Rashid noted as evidence of how much things have changed that even in the mid-1980s, the entire Earth had less computer storage capacity than today’s garden-variety memory stick.

Still, his experience with the Alto helped set Rashid on his life’s path. The first notable achievement of that journey came in 1977, while he was a student at the University of Rochester, with Alto Trek, a game he developed that helped define the first multicast standard for the modern Ethernet. In 2000, well into his leadership of Microsoft Research, he developed the game Allegiance based on the same code used in Alto Trek. Allegiance, received with plaudits by the gaming community, retains a cadre of adherents to this day.

In the ’80s, Rashid made an even bigger splash, with Mach, an operating-system kernel created in conjunction with his graduate students at Carnegie Mellon, that moved computing into a new world of multiprocessors.

“It didn’t just work in a network environment,” he recalled, “but worked across both small-scale and very large-scale multiprocessor systems.

“Mach went from being a system built in a university environment,” Rashid told the assembled academics, “to having an enormous impact, even on you, on the things that you do.” In fact, due to its flexibility and extensibility, the OS became a basis for Mac OS, iOS, UNIX, a version of Linux, and, indirectly, Windows NT.

Turning Point

In 1991, he began discussions that eventually led to his founding of Microsoft Research.

“If you go back to the investments we made at Carnegie Mellon in basic research,” Rashid said, “those investments paid off handsomely in the sense of having a huge impact on products, on industry, even though we were not working with a company. My idea was to go to Microsoft to create that kind of research environment, where we could create the type of technology that could have a huge impact.”

One of his first actions upon arrival was to write a mission statement.

“When I got to Microsoft, I had a philosophy with three key elements,” Rashid said. “These were the basis for the mission statement that we use even today within Microsoft Research. Those elements are a statement about the things you need to do to make the investments that do allow you to create the future.”

Those enduring principles are pretty simple:

• Extend the state of the art in the areas in which we do research.
• Contribute technologies to Microsoft products.
• Ensure that Microsoft has a future.

The efficacy of this approach can be measured, in part, by the way Microsoft Research has flourished. From one lab, based in Redmond, Wash., in 1991, the organization has grown to 13 research facilities, positioned around the globe.

In lockstep with that growth has been Microsoft Research’s continual contributions to Microsoft products. As Rashid stated proudly: “Every product has something from Microsoft Research. That’s really a testament to the opportunities that exist in a company like Microsoft—and the fact that basic research really can make those kinds of contributions.”

For Rashid, that’s a badge of honor.

“The real value in investing in basic research,” he said, “is the ability to survive, for Microsoft to be able to survive, for society to be able to survive. When you need a great idea, when you need a technology, it might be there if you’ve made those investments.”

He then offered a few of the firsts to come from Microsoft Research: the first terabyte database on the web, TerraServer, which eventually led to the WorldWide Telescope and then Power Map; the first large-scale use of program-proof technology, in SLAM, a software-verification engine used in the Static Driver Verifier; the first practical use of 3-D computer vision as an accessory, in Kinect; the first major use of program synthesis, in Flash Fill—and the speech-to-speech technology debuted a year earlier in China.

“For me,” Rashid said, “that was the culmination, not just of a long career in computer science, but also of a dream about the kinds of things that we would be able to do with the technology we’ve created.”

As for what the future holds, Rashid wasn’t hazarding a prediction, he was stating a fact.

“I think the key thing to realize,” Rashid concluded, “is, as scientists, it’s not that we create the future. We create the raw materials, we create the pieces, from which a future can be created.”

Opens in a new tab