SETI@Home is a blast, and effortlessly cracking encrypted messages for fun and profit is a neat idea, but do these distributed computing projects really help society or businesses do “real work” more quickly or efficiently? Probably not.
Outside the realm of these projects to build a global supercomputer, researchers and computer scientists are continuing to improve the Internet’s ability to enable all of us to benefit from distributed computing through better remote communication tools and easier real-time interaction.
From enabling colleagues in opposite corners of the office — or planet — to edit a white paper simultaneously to allowing a customer and a service representative to work together in real time to complete an order over the Internet, the “human” side of distributed computing remains a high priority for workflow experts.
Stephen C. Hayne, associate professor of information systems at Arizona State University, researches pragmatic distributed computing problems, including collaboration across a network. As it turns out, what seems like the most basic of tasks can actually be a substantial challenge for the careful developer, Hayne said. Even something that seems as simple as allowing two people to edit a document at once presents a number of hassles.
The problem lies in deciding who gets to make changes — and who gets to make their changes stick. “You can try everybody wins,’ but people don’t like that because I don’t want to try to make something italic, and then you come along and make it bold, and we wind up with bold-italic,” Hayne said.
Part of the solution might lie in old-fashioned person-to-person communication, such as using live chat and cursor gesturing to announce intentions back and forth. Even more important will be intermediary software that can anticipate and resolve disputes transparently. Right now, however, nobody is quite sure what form that software will take, or where it will reside — on one or more client computers or on a centralized distributed-computing server.
“I happen to be in the camp that says the operating systems that manage the PC are not up to managing these problems in an uncontrolled environment like the Internet,” Hayne said. He is working on building systems with an independent intermediary that resolves conflicts in collaborative distributed projects.
Of course, the march of technology isn’t the only barrier to effective collaborative computing. Sometimes interactive distributed computing dreams are realized only to demonstrate that perhaps they weren’t such great ideas in the first place.
Paul Dourish, a member of the famed research staff at Xerox’s Palo Alto Research Center, was a pioneering developer of audio-video conferencing technology over the Internet. “We envisioned a world where everybody would have a camera on their desk and would be videoconferencing with everybody else,” he said.
“It was a very strange way of working,” Dourish said.
