Thinking About a Computing Community

By Hope Green

Special to the Gazette

Think back for a minute to those early digital computers, circa 1948. Each of the contraptions filled a room. One math problem at a time, they helped researchers to calculate the trajectory of a ballistic missile, say, or crack the language of a secret military code.

Despite everything we know today about what computers can do, says Bunting Fellow Lynn Andrea Stein, at a subconscious level our culture has not shaken the image of that clumsy, oversized hardware. Innovations ranging from VCRs to the Internet redefine what computation means in everyday life. Yet in many classrooms, even among computer scientists, the outdated notion persists. According to Stein, we still have not erased the image of computers as "what-do-I-do-next machines."

In a recent lecture, "Why Your Computer Is Not An Abacus: Deconstructing the Central Myth of Computer Science," Stein suggested that a change must occur in our mental construct about computers, and that academia has a part to play in fostering this change.

"As we start acknowledging the language and metaphor of interactive community - I call it 'preaching what we practice' - we will start solving problems in different ways," she said during her presentation at the Radcliffe-based Mary Ingraham Bunting Institute.

Stein '86 is an associate professor of computer science at the Massachusetts Institute of Technology. She is on sabbatical as a 1997-98 Bunting Science Scholar, funded by the Office of Naval Research. Her work with artificial intelligence has helped to convince her that a new framework is needed for thinking about the role of computers in the world. Think of them, she tells her students, not merely as a bunch of processors performing sequential tasks, but as a set of interactive, concurrently functioning entities. In other words, don't think of "computation" as just calculation. Think of computation as a community.

Traditional computer science courses, Stein said, tend to narrowly define their discussion of computer programming. Typically, they will use a step-by-step metaphor such as the creation of a peanut butter and jelly sandwich.

A really simple computer, students are told, is trained to "Do this, and then do that" until the operator determines when the "sandwich" is done. More complex programs are explained as patterns of clustered steps, so in the sandwich model, the commands to spread peanut butter on one slice, then spread jelly on another, and then fold the slices together, are performed in a single stroke.

While the sandwich-making metaphor does illustrate the way "how-to" knowledge is encoded in a computer, Stein said, it is only part of the story of computation. "It's a nice, linear, sequential path with a well-defined end. The problem is, I don't think that's what a computer does!"

Instead, Stein sees the computer as "essentially a living being," because it interacts with its human operator and also interacts with other computers, even while the operator is asleep or doing other things.

A better metaphor, she said, is the operation of a restaurant. Human "agents" including the customer, waiter, and kitchen staff make requests of each other for information and for food. These agents cooperate in a network, with a mutual understanding that their requests will be carried out. Rather than evaluate results, "you evaluate the system by whether you get the appropriate ongoing behavior."

Likewise, she said, "our computers are communicating with each other and not just producing a result at the end. They keep going and keep spinning out results over time."

Stein predicts people will see the limitations of the traditional sequential model for computers, because the facts of our everyday world disprove it. "As people start to use their intuition more, I think you will see them solving problems that they thought were too difficult to solve . . . that people didn't know how to talk about in a traditional framework."

Stein adds that it is not her mission to speculate on how a new paradigm in computer science might change the way we use the Internet or influence other specific technologies. She prefers to focus on the larger issue: "In the long term, computers will become as much a part of everything as motors are. We don't talk about motors anymore! There's a motor in my camera, my projector, my food processor. Computers will become just as ubiquitous and will be used in ways that interact with other things . . . Computers will become part of the background, but a more consistent part of the background, and everything will be networked together."