SEATTLE, Wash., July 16, 1999 — Watch carefully: Now you see them; now you don’t. If some of the nation’s leading computer scientists are correct, personal computers are about to become invisible. No, they’re not actually going away — quite the contrary — but computers will become less obtrusive as they become simpler and more natural to use. In many cases, computers literally will become invisible, being hidden within pens, shoes, wallets, refrigerators and other common objects, where they’ll operate on our behalf, often without any conscious intervention on our part.
A computer-in-a-pen might send copies of your handwritten notes or letters to your desktop PC without you having to synch-up the two devices. A computer-in-a-shoe might track your aerobic activity or emotional state based on your speed, gait, body temperature and galvanic skin response. And future computer displays might be printed as a part of the wallpaper covering the walls in your home or office, activated by a minor electrical charge so that the displays are visible only when needed, disappearing when not in use.
Such invisibility is a far cry from today’s trend in consumer computing, in which PCs take on stylish designs and colors that call more, not less, attention to themselves. But it may not be so far-fetched, or very far away, according to Dr. Turner Whitted of Microsoft Research and Prof. Gaetano Borriello of the University of Washington’s Department of Computer Science and Engineering. On behalf of their respective institutions, these two scientists are co-chairing the “University of Washington/Microsoft Research Summer Institute: Technologies of Invisible Computing,” to be held July 19-23 in Seattle.
Forty of the nation’s top computer scientists and engineers will attend the Institute, representing the A-list of American technology institutions, including Hewlett-Packard, IBM, Intel, Microsoft and Xerox, as well as Carnegie Mellon, MIT, the University of California at Berkeley, Stanford, and the University of Washington. Their mission: to chart new directions in invisible computing over the next decade, to identify the challenges, and to spark innovative collaborations among the participants. The sessions will cover everything from new user interfaces and sensors to network-based services and network infrastructures.
“The individuals coming to the Institute rarely, if ever, meet in the usual, discipline-oriented forums,” notes Borriello, a veteran researcher from Xerox and Berkeley who is now participating in the University of Washington’s prestigious research for DARPA, the government agency that helped create the Internet. “It’s our hope that this unique collection of people in an interdisciplinary forum will provide innovative insights into the real-world problems encountered in the design of new ubiquitous and invisible products.”
Two big topics for the Institute — and two of the biggest challenges that stand in the way of invisible computing, according to Borriello and Whitted — are battery power and communications.
“There’s no Moore’s Law of Battery Life,” says Whitted, a former Bell Labs researcher who started the Hardware Devices Group at Microsoft Research. “While computing power may double every 18 months, battery life only doubles every five to 10 years. Since many of the new, invisible computing devices will be mobile, ensuring that they have a satisfactory power supply will be a major challenge.”
Much of the solution will come from innovations in power conservation, says Borriello. Invisible devices will save power by handling just a few functions or applications themselves, off-loading most others to computers that are connected to a continuous power supply, such as a desktop computer connected to a wall outlet. How computing functions and power requirements get partitioned between mobile and fixed computers is a key question for Institute participants.
Heat dissipation is another power-related constraint on mobile, invisible computers, adds Whitted. “People are going to wear many of these computers — in their belts, rings and watches, in jewelry, around their necks — and we can’t allow them to generate so much heat that they become uncomfortable,” he says. “While most power for mobile, invisible computing devices likely will remain chemical-based, we need to explore alternatives. For example, work at MIT shows you can harvest enough power from heel compression while walking to power a hand-held PDA (personal digital assistant). And third-world countries without electricity are powering cell-phones via hand cranks. We have to be creative about the possibilities.”
With mobile computers likely to share the computing and power loads with their stationary cousins, effective communications between those devices is part of the solution — as well as another challenge to be addressed, according to Whitted.
“Connections among devices is the toughest task we face,” he says. “The interplay of bandwidth, power, and mobility presents us with challenges that don’t arise in fixed computing. The problem is that bandwidth isn’t free. It’s like real estate: no one’s making any more of it. So getting the most out of available bandwidth is important.”
Whitted and Borriello concede that the emerging Buetooth specification for short-range, low-power wireless communications may be part of the answer. Whitted says that fine-tuning radio frequencies for specific functions — an innovative project on which Microsoft Research is already at work — is another. That invisible accelerometer in a shoe, for example, might send its data at a lower, and less power-hungry, radio frequency than voice communications require.
“UHF radio waves aren’t the only way to transmit a signal,” adds Whitted. “Infrared, visible light and ultrasound may also become part of the communications solution. “I’m not willing to discount even far-fetched alternatives, especially if they are cheap and require little power. They may not offer much bandwidth, but then again we may not need much bandwidth for some applications.”
“Many communications needs for invisible devices may only extend a matter of feet, from a computer in your ring or watch to the cell phone or pager in your backpack or purse,” says Borriello. “We might create invisible computers that can monitor your existing communications devices and take advantage of them to send information when those devices are not otherwise being used.”
Innovative small-scale networks would minimize both the communications and power challenges facing invisible devices, as well as help address another key issue: security. Invisible computers will likely collect far more data about their owners than computers have ever collected before. How that information is used, and who has access to it, are questions with major privacy and other public policy implications. Making sure that information isn’t transmitted more than a few feet is at least one part of the security solution.
The Summer Institute’s sessions on invisible computing will focus on commerce as well as art — because making invisible computing a reality will require new business models and new technologies, according to Borriello and Whitted. There’ll be a need for businesses to provide communications services, possibly including subscriptions to content and applications. Much as cellular companies and even some Internet providers are giving away phones and, now, PCs, some of the new invisible computing devices might be given away by companies eager to acquire subscription customers for services using those devices.
The July 19-23 session is the third Summer Institute program convened by the University of Washington and Microsoft Research. Previous sessions covered biological models for computing and data mining. The fourth session follows on August 2-6, when the two institutions will host a Summer Institute on Technologies to Improve Software Development. This program will focus on tools and technologies to support designing, constructing, evolving, analyzing, debugging and testing very large software systems.
“A lot of the hard problems in writing software still haven’t been solved, and that’s why software is often too fragile,” says James Larus, senior researcher at Microsoft Research and an organizer of the program, along with Senior Researcher Daniel Weise of Microsoft Research and Prof. David Notkin of the University of Washington. “We want to expand the field of software engineering research by building a community of outstanding researchers to identify technical problems and solutions. At Microsoft Research, we’ve been working on tools to help programmers understand large systems and how their software fits into them. That’s one approach we want to share with others throughout the field.”
For both Summer Institute programs, the weeklong sessions will mark only the next steps in a very long journey. Consumers may not actually see many of the invisible computers that Borriello and Whitted contemplate but — if those two experts are correct — consumers likely will be seeing the benefits for years to come.