Home > News > Ken Hinckley introduces the "Ubiquitous sensors"

Ken Hinckley introduces the "Ubiquitous sensors"

From The Wall Street Journal
June 25, 2001

Technology (A Special Report) --- In The Lab: Window Into the Future --- At Microsoft's research division, scientists spend their days studying technology that may one day lead to new products; Then again, it may not -- By Rebecca Buckman

REDMOND, WASH. -- Want to play poker with Bill Gates and Steve Ballmer?

Since Mr. Gates, Microsoft's chairman, and Mr. Ballmer, the chief executive, agreed to have their faces digitally photographed with a video camera for the project, the images look pretty realistic. But neither man's avatar has hair. And Mr. Colburn hopes to improve the game's lip-synching, so it looks as though the players are actually saying "deal" or "fold" at the appropriate times. Right now, the Gates and Ballmer figures "just flap their lips when . . . talking," Mr. Colburn explains.

As frivolous as the poker game may seem, it is all part of Microsoft's commitment to research and development. Unlike some newer technology companies without huge cash reserves, Microsoft spends large sums on basic computer-science research that might or might not find its way into new products, such as games or new software. It does that through a quasi-academic pure-research division -- Microsoft Research, where scientists like Mr. Colburn get to devise the technology of the future.

"You should try some [risky] things like that if you're a company our size," Mr. Ballmer said in an interview earlier this year, after Microsoft's research group trotted out many of its showiest projects at a technology conference in San Jose, Calif. "Let's say 70% of [research] sees the light of day . . . that's a good payback."

This year, projects at Microsoft Research will receive about $250 million of the $4 billion Microsoft plans to spend on research-and-development activities. Last year, the company spent about $3.8 billion, or more than 16% of its total revenue, on research and development.

Microsoft, founded by Mr. Gates and Paul Allen in 1975, is trying to emulate the success of storied research labs at much older companies, such as International Business Machines Corp., Xerox Corp. and AT&T Corp.'s Bell Labs, now part of Lucent Technologies Inc. In some ways, Microsoft doesn't yet measure up: Those companies boast Nobel Prize winners and huge, famous inventions like the transistor and the touch-tone telephone. They also do basic research in areas including physics. (Microsoft notes that it does have three winners of the Turing Award, which the company says is comparable to a Nobel Prize for computer science.)

But even with those impressive accomplishments -- and as overall corporate R&D spending in the U.S. increases -- some big companies are spending a smaller percentage of their revenue on research. IBM spent 5.8% of its revenue on R&D in its last fiscal year, down slightly from 6% in 1999 and 6.2% in 1998. A spokesman notes that IBM remains committed to research, and has hired 100 new researchers in the past two years. The company is simply trying "to invest our R&D dollars more intelligently," he says.

Equally worrisome to some academics is that many information-technology companies are devoting more and more of their research dollars to pure product development. It's a trend some fear could stunt true scientific innovation, particularly as federal support for technology research has remained flat or even declined over the past decade.

"For most companies, R&D is D," says Edward Lazowska, chairman of the computer-science department at the University of Washington in Seattle, who also sits on a technical advisory board for Microsoft's research group.

Indeed, five years ago, a National Research Council report found that about 90% of the R&D budgets at technology firms went toward product development, rather than to more forward-thinking research. That could be because many companies have come under increasing pressure to manage costs and boost earnings for shareholders. The council also pointed out that the industry has shifted to churning out products with lower profit margins, so as a result companies don't have the kinds of profit they once did to plow back into research.

But some companies are bucking the trend. IBM still does "tons of basic, exploratory research," says spokesman Tim Blair, such as its recently reported efforts to develop carbon nanotubes as a possible replacement for silicon chips or other silicon-powered devices.

Chip maker Intel Corp. -- while still linking its research efforts very tightly to product-development teams -- has announced several advances in chip making involving super-small transistors. Eventually, they could be used in inexpensive computers that deliver the same amount of computing power as much bigger mainframe machines that cost millions of dollars.

Microsoft is focused also on developing way-out products, including some neat gizmos for consumers. It might seem surprising, since the company has long had a reputation for skillfully marketing others' ideas -- and not coming up with innovative ideas itself. Indeed, when he first joined the company in 1991, Senior Vice President Rick Rashid, who now oversees the research group, notes that "most people thought Microsoft Research was an oxymoron."

Times have changed. These days, hundreds of geeky Ph.D.s as well as psychologists, anthropologists and sociologists troll the hallways of the research group's home. Researchers work in a hands-off, university-like atmosphere. The goal is publishing intellectual research papers and getting them discussed at industry conferences -- not shipping products out the door, Mr. Rashid says.

Research teams often present their ideas to Microsoft engineers who are developing specific products, and researchers occasionally do temporary stints working inside product groups. Microsoft researcher Jerome Lengyel, for instance, parlayed his edgy work in computer graphics into a temporary assignment with the team building Microsoft's Xbox video-game console.

But products don't drive research, Mr. Rashid insists. What's more, "we don't have budgets for projects," he says, since fixed costs can place limits on how well people work together and share resources. Instead, researchers often just work on what interests them, whether the subject is 3-D movie graphics or complicated computer-database configurations. Researchers are Microsoft employees and subject to regular employee reviews, not questions about tenure. They are in many ways unfettered in their work; Microsoft doesn't edit or even review papers before they're sent to conferences, for instance.

Ken Hinckley , who joined Microsoft four years ago, had an interest in what he calls "ubiquitous sensors." That's the idea that electronic sensors, which can detect things like touch and proximity, could be used more widely in consumer electronics. So he plunged in. Now, he likes to show off a prototype handheld organizer (a PocketPC machine made by Casio Computer Co.) that he's doctored up with various sensors, including some attached to the machine with white and blue tape. "I did most of the soldering myself," he says proudly.

His prototype, outfitted with "tilt" and "touch" sensors, can automatically sense when someone is picking it up. That means the device could be programmed to turn itself on when it receives those signals. And when Mr. Hinckley picks up the machine and tilts it away from himself, the text on the screen begins to scroll. That enables him to read a document or e-mail without ever using a stylus.

Another possibility: Using the sensors to generate commands for voice recording. In other words, the machine could recognize when a user is picking up the machine and putting it to his or her ear, ready to talk into it. A handheld organizer or a cell phone could be programmed to automatically turn on its recording mechanism at this time. That would allow people driving cars, for instance, to talk on their phones without having to punch buttons, and possibly to pay more attention to the road.

The day before a recent demonstration of his souped-up organizer, Mr. Hinckley met with the Microsoft team in charge of making software for PocketPC devices to see if they'd be interested in using some of his technology. Their reaction? "It's still in the research stage, but it looks promising," reports Mr. Hinckley, who came straight to Microsoft after receiving his Ph.D. in computer science from the University of Virginia.

Mr. Colburn, of poker-game fame, has also dabbled in other graphics work that could find its way into computer games. At the San Jose conference, sponsored by the Association for Computing Machinery, Mr. Colburn's computer display of "real-time fur" drew plenty of attention from visiting schoolchildren. The technique is simply a way to draw realistic-looking fur on things like animals and even fuzzy dice. It's "eye candy for games," Mr. Colburn says.

Other researchers are working on projects that could help disabled people. Steven Shafer leads the "Easy Living" team, which is developing systems using wireless technology and video cameras that allow people to log onto a computer simply by sitting in front of a keyboard. In one of his team's labs, which has been decorated to look like a living room in a home, test subjects plop down on a tan leather couch and are suddenly greeted by a computer-generated voice: "Hello, Person 140," it says. Then, sensing that Person 140 is sitting in front of a computer keyboard and might want to log on, the image of a Windows computer screen appears like a slide projected onto the living-room wall. When the person gets up, the screen goes away.

Already, some of that research has been incorporated into Microsoft's Outlook Mobile Manager product. That software helps figure out which e-mail messages are important enough to send to people when they're on the go, and how to send them, such as through their cell phones or pagers.

Overall, pioneering work done in labs such as Microsoft's and IBM's is becoming more important as newer Internet companies fail; most of them didn't have funds for basic research, and simply relied on using their once-highflying stocks as currency to buy companies with new technology they needed. Basic research requires a big investment and a long-term view, says Mr. Rashid, who came to Microsoft after working as a professor at Carnegie-Mellon University in Pittsburgh.

"We've decided it's important for us," he says. "We think that's what's going to drive the long-term future."

Ms. Buckman is a staff reporter in The Wall Street Journal's San Francisco bureau.

Journal Link: What products and developments do you think will be key to Microsoft's future? Join a discussion in the online Journal at WSJ.com/JournalLinks

Original Article | Local Copy

Return To List