Poof! Goodbye computer...

New project aims to make computers like air

The film clip from a security camera caught it all: A harried executive's tensed face as his desktop computer seems to mock his attempts to use it.

First, he pounds the keyboard with his fist, then rips it from the computer, and finally heaves the screen over his cubicle wall.

It's an extreme form of a frustration that Michael Dertouzos shares.

"For 40 years, people have served computers," says the director of the Laboratory for Computer Science at the Massachusetts Institute of Technology in Cambridge, citing his own horror stories of balky laptops and online delays. "We're going after a unique and powerful goal: to make machines human-centered."

On June 21, Dr. Dertouzos, along with colleagues from MIT's Artificial Intelligence Laboratory and six major computer and telecommunications companies, announced an alliance to revolutionize the way people and computers interact. Dubbed Project Oxygen, the effort aims to develop technologies and standards that will make computers as pervasive as oxygen and as effortless to use.

In your pocket: a voice-activated, digital equivalent of the Swiss Army Knife. It wraps a personal digital assistant, cell-phone, wireless Web browser, e-mail "post office," GPS receiver, video camera, and other functions into one small case.

In your home or office: systems that sense your presence, anticipate your information needs, and seamlessly take the handoff from your pocket device when you walk through the front door.

Nor is MIT alone. University computer labs and high-tech companies worldwide are pursuing similar goals.

On June 22, for example, Microsoft announced Microsoft.Net, a version of pervasive computing that Microsoft chairman Bill Gates calls a "bet the company" shift from providing boxes of software to providing "interactive services" where software resides online.

With its approach, Microsoft hopes to enable "knowledge workers to create, browse, edit, and share information ... no matter where they are or what device they're using," according to company president Steve Ballmer.

For home users, it could mean the end of trying to be your own information-technology manager. Tasks that drive many home users to distraction - upgrading software, adding new functions, or backing up files - "will happen automatically and transparently," Mr. Ballmer says.

A number of factors have converged to give these efforts some hope of technological success, analysts say. One is the federal government's willingness to put up seed money for research into pervasive computing. Many of the coordinated projects at US universities began last year after the Defense Advanced Research Projects Agency opened its purse strings (DARPA).

While Oxygen's corporate partners are putting up more than half the $50 million for the five-year project, the balance comes from DARPA, Dertouzos says.

DARPA money also is supporting projects such as "Endeavour," a pervasive-computing research program at the University of California at Berkeley that embraces the use of micromechanical devices and robotics in its vision of computing's future.

Another factor combines a shared vision of the future with maturing technologies. "In general, there is a widespread community vision that computers will continue to get smaller, cheaper, more tightly networked, and more pervasive," says Kevin Werbach, editor of "Release 1.0," an industry newsletter.

In addition, he says, the growth of the Internet has built worldwide links between once-disparate computers with different operating systems, while the growth of digital wireless networks is cutting the umbilical cords that tie computers and their users to phone jacks.

"Any trend individually is not enough," he says. Instead, their convergence has helped pave the way for growth in the number of pervasive-computing projects.

For all the George Jetson-like uses researchers envision, a variety of enabling technologies still have a way to go. For example, vision technologies are good at recognizing faces, but not small objects, Dertouzos says; engineers must develop longer-lasting batteries and circuits; speech-recognition software, where a computer acts on commands issued in idiomatic language, needs work.

"That will play a large role in making devices and computing seem much more natural," says Steve McClure, who specializes in collaborative computing technologies at International Data Corporation (IDC), based in Framingham, Mass.

Speech processing - a device's ability to recognize speech and act on it - "is almost there," he says. But "natural language" approaches are more difficult. "You're dealing with phrases like, 'Gee, that's rad,' or colloquialisms like 'off the cuff,' " he explains. If devices are to eventually anticipate a user's needs, they will need to distinguish between dialects, ages, and genders, he says.

With pervasive computing come concerns about privacy, since files are likely to live on networks instead of within individual devices. Moreover, systems that monitor a person's environment could include TV cameras and infrared detectors to help systems anticipate movement and location, as well as information needs.

Dertouzos is confident that with legislative safeguards, privacy is manageable. "We have all the technology we need to ensure all the privacy we need," he says.

IDC's Dan Kusnetzky, vice president of systems software research, is less sanguine. "I do have concerns about security and about what all this will cost," he says.

In addition, some industry analysts suggest that software companies geared to lining store shelves with programs could face tough times if programs reside on network servers instead of home and office PCs.

Yet it's clear "the future of computing is not in the desktop," says Mr. McClure.

The race is on to define standards and technologies that will be that future.

(c) Copyright 2000. The Christian Science Publishing Society