Personal Tech

Beauty and Magnets 

Why don't robots ever move in a tentative manner?
Why doesn't your computer purr to your touch?
Why must technology always be useful?
One former science student says it's time to set things right.


when he discovered electromagnets. But not until he was in high school did it strike him like a poem that each winding of a wire coil increases the strength of its magnetic field. At that point, he remembers, he started winding electromagnets until his wrists got tired. Once, when a coil grew to the size of an acorn squash and the weight of a bowling ball, he lugged it over to a wall socket and plugged it in. By some miracle the fuses held, but lights throughout the house grew dim. "My parents asked me not to plug it in in the evenings when they were trying to read," he recalls.

In 1977, Durlach went to Princeton and tried studying physics, math, and electrical engineering, but he was unhappy. Academic science was too dry for his tastes, and he had little in common with students in traditional liberal arts courses. After four years, he left without taking a degree, moved back to his native Massachusetts, and spent several years in a succession of odd jobs--volleyball coach, math tutor, stand-up comedian--but otherwise rarely left his apartment. Soon he was running out of the small inheritance he had been living on. He needed an income, he needed more human contact, and he had a vague sense that he could turn his hobby--electromagnets--into something he could present to the outside world. He joined a support group for entrepreneurs.

"We spent half the time giving each other therapy and the other half business advice," he says. It was then, at the age of 26, that his tinkering with electromagnets took a serious turn. One day he brought the group a tray of iron filings, with a few small electromagnets wired underneath. He had also rigged up a computer to control the amount of electricity in each coil--and thus the strength of its magnetic field--and had programmed the computer to make patterns form and shift in the filings. The support group loved it.

Thus emboldened, he tried out his creation on a more discerning crowd. He took it to an open house at an artists' cooperative, where some 3,000 to 5,000 people saw his dancing iron dust (he had replaced the filings). He watched their faces. "People looked like they look when they see a child starting to walk," he says. "There was this warmth and support. It was just delight. And people laughing. And interesting conversations. It would go from 'What chips did you use?' to 'This reminds me of Tina Turner.' There were hard-core techies and people who were dancers." It finally occurred to him why he loved physics so much: not because it was useful, but because it was beautiful.

Yet those same people whose faces lit up in delight turned and asked: What's it for? And that kind of thing upset him. "If it were a puppet show, with the same audience impact exactly, no one would ask 'What's it for?'" says Durlach. "But because it's high tech, people think it must have been developed for some 'practical' purpose."

Despite the awesome application of science and technology in solving practical problems, Durlach realized, virtually nothing had been done to harness the same tools in the service of emotions. And though he doesn't remember a moment of epiphany--or a voice in the back of his mind saying "My destiny!"--this realization had a deeply personal significance. It gave meaning to the years of discontent. Durlach began to feel passionately that only half the potential of technology to enrich people's lives had been tapped. "Medical researchers think nothing of spending billions of research dollars on dialysis machines and heart replacement machines to keep people physically alive, but they don't build virtual reality systems so that old people isolated in intensive care wards can imagine they are having dinner with their family. Now, it would cost a lot, but we could do it equally well."

Durlach no longer has time to coach volleyball or sit around his apartment. He now runs a three-room lab in a modest single-story building in Somerville, complete with a motley staff. Several students from nearby Harvard and MIT work there, as well as an opera singer and a fortyish man by the name of Wes Keyes, a jack-of-all-trades equally adept at designing machine parts on a computer and working the lathe. Amid the clutter of books and boxes, manuals and machines, Durlach, now a stocky, balding 38-year-old, flits about in a state of nervous excitement, showing one of his young helpers how best to cover a container of oily liquid with Plexiglas to keep it from evaporating, or negotiating with another what hours he will work between classes. Durlach keeps several projects going at once--ideas he's been cultivating for years but that he hasn't had the time or money to try to realize. Things got a bit easier for him last fall when the National Science Foundation, as part of its program to encourage small, innovative, science-related businesses, awarded him a grant of $300,000 over two years.

is going to do with this money, it is first necessary to dwell for a moment longer on his magnum opus, the dancing iron dust, because all of his newer work is in some way related to it. He is eager to demonstrate its charm. In a 15-by-15-inch tray, the iron dust has arranged itself into nine rows, each composed of nine little humps. Below the tray, directly underneath each hump, a static, permanent magnet holds the dust in formation. The humps look like walnut-size porcupines, the dust forming spikes as it follows the direction of the magnetic field outward. Durlach pops a compact disc into his computer and hits a few keys. The dust springs into action, shifting and swirling to the strains of Flight of the Bumblebee; one moment it's like a line of synchronized, kicking Rockettes, the next like a troupe of modern dancers racing around a stage. The humps are always there, but the shapes move and change with uncanny swiftness and fluidity. They bend this way and that, their spikes reach for the sky or fall humbly toward the floor. At some moments the performance is touching, at others comical.

This five-minute display is the result of hundreds of hours of programming, not to mention the computer language Durlach had to create to specify the details of choreography. In this language, Durlach not only shifts dust through the three spatial dimensions, he also plays with time, changing the rate at which it appears to flow, moving it forward and backward, to achieve especially fluid motion. "One of the interesting things is that you can make time go forward and backward like a sine wave," he says. "It's as if you took the reels of a movie, and instead of rotating them continuously forward you moved them back and forth, woonk-woonk, woonk-woonk. That's a very cool effect." Durlach goes on to enumerate the different ways you can play with time, and the effects they generate. He knows intuitively how the 16 electromagnets underneath the tray act to create these effects, but he doesn't understand it in a rigorously mathematical way, and he doesn't particularly care. Although he is engaged in research of a sort, he is first and foremost an artist, which is to say he is more concerned with figuring out how to create an effect for his audience than in coming to an academic understanding of the physics involved.

He is taking a similarly pragmatic approach in his efforts to write a computer program that can listen to music and, as the music unfolds, create a choreography for his iron dust. At present, his choreography is laboriously handcrafted to each piece of music. Although some widely available devices already respond to music in what computer scientists call real time, they are primitive and not particularly interesting. So-called light organs, which give you a moving graphic representation of the volume of the sound at each particular band of frequency, are easy to build but deliver no particular emotional impact. At the other extreme, artificial-intelligence researchers are working on analyzing music through supercomputers.

For obvious reasons, Durlach eschews the supercomputer approach. Instead he likes to conceive an idea and play with it until it yields the effects he desires. He hit on one idea by thinking about the problems he faces in choreographing his iron dust--an especially difficult task, precisely because the dust can be molded by the magnets into almost any shape. From the standpoint of choreography, it is better to have some constraints, such as those imposed on human dancers by bodies that can run only so fast, bend into only so many positions. Since Durlach cannot actually impose physical constraints on his dust, he is writing a computer simulation that treats the dust as though it were not millions of unconstrained particles but rather a thick liquid able to slosh around the tray only so fast. Then he'll use data extracted from the music to create wave patterns in this virtual liquid. The computer will translate these wave patterns to the dust. He may also try changing the properties of the liquid to certain cues in the music--when the flutes come in, say, the liquid might get thinner, or perhaps get thicker when the tempo slows. "We want to give the iron dust a personality that is coherent and separate from the music," he says. "This may or may not work. It's a research question."

appreciated the need to keep his solutions simple. Eight years ago, desperate for funding to develop his iron dust technology, he accepted an offer from the owner of a chain of shopping centers to create a display that could tour the centers. The only catch was that it had to be rugged enough to be placed outdoors. Durlach had to build a metal casing to protect the sensitive electronics from salty ocean spray, from graffiti, saliva, rain, snow, and sudden swings in temperature--the machining bill alone came to more than $10,000. To keep the electronics at the proper operating temperature, he installed a $2,000 air-conditioning and heating unit that had been designed for radar on military ships. "I learned all kinds of things that I didn't want to know," he says. "It was nuts, absolutely nuts. The manual alone took me two and a half months to write." He spent two years on the project and was paid $60,000 for expenses and labor.

The more serious problem with the dancing iron dust, however, is the expense of making the display any bigger than it already is. The cost of additional dust, magnets, and the electricity to run them quickly gets out of hand when you start increasing the dimensions of the tray. Nevertheless, Durlach has found a market among manufacturers who want eye-catching exhibits for their trade shows. And he's done a version of the dust for Ford in which permanent magnets spell out the company's name.

For some years, however, he has had in mind another project, one that would be far less costly to scale up. Like the dancing dust, Durlach says, it would demonstrate "interactive physics" and also work as a catchy sign. He calls this creation his Tower of Triangles, and at the far end of his lab, near the milling machine, sits a prototype. The triangles are like those you'd make while building a house of cards: each consists of three square faces joined to one another along two edges. They revolve around an axis, like the letters that Vanna White flips in Wheel of Fortune. Although at the moment he has only a few triangles mounted vertically, one on top of the other, he envisions making long chains of them. Durlach flips the top one with his hand and sends it wobbling back and forth. The other two wobble as well, though they lag behind each other as though they were abdominal segments of a loose-hipped snake. "The idea is there'll be a lot more segments," he says, "and when you move one end, you'll see a wave propagate through the entire thing." With motors operating the triangles at either end, you could set up standing waves, in which the hills and valleys of the wave alternate with each other, while the points in between--the nodes--appear to stay in the same spot.

"Remember how cool a barber shop pole looks?" he says. "It's just a turning thing. You'll be able to wind this thing up and turn it, like a barber shop pole, but you'll also be able to superimpose a standing wave on the twist, creating a sort of trill." Durlach continues rattling off descriptions of other patterns he'll be able to generate with the tower. His ideas fall one after the next like fruit toppling from a stand. With two towers next to each other, you could set waves propagating down one and going up the other. Then you could set the contraption up so that each time the wave got to the end of a pole, it bounced back, growing bigger each time. And so on. "Having two towers would be wonderful," he says. "Imagine one winds, and the other winds, and they wind in synchrony. There are so many relationships of dance between two of them. I think that kind of thing in the aggregate will be just totally cool. And, of course, no matter how big you make it, you only need two motors at either end, so the cost stays under control."

Art collectors, Durlach has found, do not generally invest in high-tech art, so he has looked instead to science museums and corporations to sponsor his work. In addition to the Ford logo, he has made a logo for the Dexter Corporation's magnetic materials division in Billerica, Massachusetts, out of its own magnets. For the Clippard Instruments Laboratory, a pneumatic-valve manufacturer in Cincinnati, he is working on a sign made out of bubbles: small valves at the bottom of a tank of water or some other fluid would release bubbles in such a way that they would form the company's name as they rise to the surface. "I'm very interested in using the physics of a corporation's own products to do promotion for them," Durlach says. "There's also a whole market in retail stores and trade shows that is right for serious research into displays."

he is doing for Clippard, the physical challenges are formidable. For one thing, when a bubble passes through water it leaves behind a slight bit of turbulence, which subtly and unpredictably affects the way a second bubble following in its wake will behave. For another, he must worry about what happens if the power fails when the valves are open and the water floods back into the electronics. And then there's the expense: for the display to work visually, Durlach figures it has to be about eight feet long, six feet high, and two feet deep. He once priced an aquarium this size, without the valves or electronics, at $30,000. For these reasons he's put the bubble sign on hold, though he hasn't stopped dreaming about it. "It could be a very beautiful artwork," he says. "It could also be a really cool clock, where it releases the time in bubbles. You could also do three-dimensional imaging, like releasing a DNA double helix from a circle of valves at the bottom. Or an interactive display where you move your finger over a touch pad and trace something out, and you see the same pattern in the bubbles. You could put it in a hotel lobby where guests come to register, and when you put their names into the computer the bubbles say, 'Welcome Mr. and Mrs. Schnitlau.'"

In the meantime, Durlach is beginning to explore the use of ferrofluids--liquids that contain iron particles, which make them magnetically attractive. It's already late in the day and time is short, but Durlach decides to open a bottle of ferrofluid to demonstrate what it can do. It has the color and consistency of motor oil. "This is the first time we've done anything with this stuff on this scale," he warns. "I don't know what it's going to do." His assistant, Anne Harley--the opera singer--has finished mounting little metal dowel-shaped magnets into a circular tray. When you look down on it, you see that the dowels have arranged themselves into the numerals of a clock.

Durlach begins to pour the liquid, but he pours so cautiously that the liquid dribbles back up along the side of the bottle. He gets flustered, and for good reason: the quart bottle of liquid in his hand retails for $4,500, though the manufacturer donated it in exchange for access to Durlach's research. He and Harley scour the office for a funnel, but instead Durlach proposes using an ice cream stick as a channel. Meanwhile, the teaspoon or so of liquid that made it into the tray is crawling eerily up one of the dowels. It perches on top, in the shape of a beetle with odd ridges along its back. "Wow! Cool!" Durlach and Harley say almost in unison. Durlach continues pouring. More liquid runs onto the tray and creeps up onto other numerals. "David, we should videotape this," says Harley. "You're right," says Durlach, and once again he interrupts his pouring. He scurries around setting up lights and a video camera on a tripod. "We want to document everything we do," he says. "Otherwise it would be impossible to go back and re-create it later."

Eventually, in the glare of a floodlight, Durlach finishes emptying the liquid into the tray. Then he activates the electromagnets to the same choreography he programmed for the iron dust. The liquid runs from one end of the tray to the other, forming little vortices and currents. Now and then it makes a plop or a gurgle. Durlach and Harley emit squeals of laughter. Of course, Durlach explains, the programming needs to be tailored to the physical characteristics of the new medium, but this primitive demonstration gives an idea of what it will be like. Just then a few drops of liquid squirt straight into the air, as if in agreement.

Perhaps the oddest thing about Durlach is that he has trained himself to look at technology in a way that is almost completely foreign to most people, except perhaps a handful in the entertainment industry. "I often have to build what I envision before other people can see it," he says. "So it might take two years before a lot of people even understand what I'm talking about. But I'm able to do this because I understand how people react to kinetic things, and I know enough physics about what to build to sort of do that."

Durlach has come to the conclusion that some fields of science are handicapped because researchers don't appreciate the emotional point of view. For instance, computers charged with searching quickly through reams of data would do better, he says, if they had the ability to ignore some information simply because they didn't care about it. Likewise, robots are built to lift heavy loads or assemble products, but few people make robots that can, say, move gracefully or with tentativeness. "There are plenty of chess programs that are good at winning, but none that pout if they don't get a chance to play," Durlach says. "The point is, this is in fact relevant. It may not be possible to build artificial intelligence that's functional without having agendas of the kind that are emotionally based, not cognitively based."

But since emotion is left out of the language of science, discussing it invariably seems silly. These days, however, Durlach has mustered enough confidence to venture a few silly ideas. "One of the things you could imagine doing is having your whole computer and everything around it be sensitive to touch, so that there's no way you could touch it without it reacting in some way. That would make it so much more human."

Durlach recalls once seeing a computer program that could engage people in conversations. It wasn't really intelligent, he says, but it nevertheless seemed more human than most programs he had encountered. "The reason it seemed human is that it really didn't listen to you very carefully and sort of wound the conversation back to the few things it cared about. Now, not listening is not what most researchers would think of as a research topic. But it is! It's what makes things human. We don't have the skills in this lab to do formal research in artificial intelligence. But we don't necessarily need to understand this stuff. We need to dance. Dancers don't understand the physics of what they do."