Clothes Make the Network

Continued from page 1

By Howard Rheingold

December 4, 2002

By and large, these efforts have involved participants who know one another, but Kortuem's hypothetical communities could be made up of strangers. "I want to make it possible for people who don't already know each other to interact productively," he says. "The biggest effect of mobile wireless computing devices will become visible only after large numbers of people start using the technology to engage each other."

While other early pioneers of wearable computing, such as Thad Starner, director of the Contextual Computing Group at Georgia Institute of Technology, and Steve Mann (see Cyborg Seeks Community, TR, May/June 1999), focus on developing systems to create communities, Kortuem attempts to show how communities of computer-wearing users might cooperate if these systems enabled them to find their own connections.

Kortuem, who came to the University of Oregon from Germany in 1993 on a Fulbright grant, witnessed an early example of this sort of spontaneous interaction while studying for his master's degree in computer science. During the summer, he took an internship at Apple, where he worked on software for the Newton PDA, an early handheld with an infrared communications link. He became fascinated by the way Newton users beamed business cards from one unit to another. Familiar with the emerging power of desktop-bound Internet communications, he realized that networks of wearable computer users would hold the potential for an entirely different medium. As much as the PC was changing the world, its impact would be severely limited by its stationary character. Portable devices would be networked wirelessly within a decade, and the killer apps would be social.

The University of Oregon was the right place to pursue such ideas. The same year Kortuem arrived, Zary Segall, formerly of Carnegie Mellon University, one of the birthplaces of wearable computer design, cofounded the Wearable Computing Lab with University of Oregon Computer Science Professor Steve Fickas.

Kortuem and Jay Schneider, another graduate student, began assembling hardware/software packages capable of automatic wireless interactions. In 1999 the team completed its first ad hoc community application called Walid, a program that negotiates chore sharing among parties with complementary tasks on their to-do lists. The software agents representing each party might determine, for instance, that one person could pick up some blank CDs at the store if the other would return a book to the library. The playlist-sharing application, known as Pirat, came next, followed by mBazaar, which mediates swaps of CDs, books, bikes, furniture, and electronics.

It soon became apparent that ad hoc community applications relied on a common set of functions, having to do mostly with detecting nearby parties, querying and comparing information, and keeping track of contacts. Kortuem realized that putting these functions in a common code library would speed up development. Completed in mid-2001, the Proem peer-to-peer platform consists of 135 Java commands optimized for spontaneous social organization. "You can have a message that's sitting in your buffer delivered whenever you meet someone," Kortuem explains. "You can say, 'This message is only for people nearby,' or you can send it to nearby devices and have them route the message further."