| Home | Back to List of Articles |


A scared conversation with Bob O'Kane

by Volker Grassmuck
for InterCommunication Magazine # 14, Autumn 95

I think, and there is light. I think different thoughts and the illumination dims, the little creatures in front of me slow down and finally stop. Mind controls matter. Is this a dream? Is it magic? Have I suddenly turned into a god shaping the world with a mere idea? Or maybe into one of these body-suit robots that multiply the pilot's power beyond any human range? It's neither, and a little of all of them. It's art. "Terrain" is one of two installations of the Germany-based studio "Otherspace" of Ulrike Gabriel and Bob O'Kane, recently shown at Tokyo's P3 gallery.


Multimedia, by nature, crosses borders, and requires versatile talents. While Bob O'Kane was studying electrical engineering at the University of Buffalo, he started to work for the closed-circuit TV station on campus. There his abilities were discovered, and a professor invited him into the department of media studies, which belongs to the faculty of arts and letters. As such its curriculum focusses on critical theory, analyzing film, video and digital arts, the use and the content of media. It also provides production facilities so students can learn in a hands-on way how much the technology shapes the media. With its rather rare interdisciplinary design it attracts theoreticians as well as students who want to become a director or cameraman. After work in electronic music, computer productions, computer generated video, animation, and film, O'Kane graduated with a Masters is arts and humanities, an interdisciplinary degree between the media studies department and the music department. His hobby is indeed playing music, and building his own amplifiers, fixing his mixing boards or his base he taught himself electronics.

After graduation, his professor at Buffalo, Peter Weibel, invited him to the newly founded Institute for New Media in Frankfurt where he became systems' manager, was teaching computer graphics and started getting into interface design. Here, another hobby of his came in, video games. "I'm a game boy," says O'Kane, "Games are interaction and response, an environment that you have to work in. This rubbed off into my interface designs a lot. An interface should be much more than just a button or a joystick. It can be many things."

He has many things in common with the other well-known artist- engineer Gideon May. Both are an interface, a "linker" between the technology and the artist. While O'Kane's strength is more the interface and hardware aspect, May is more geared toward information and software, and the pieces they worked on reflect that. Both worked with quite a number of artists, May doing work at the ZKM in Karlsruhe with Jeffrey Shaw and Agnes Hegedus, O'Kane at the Institute for New Media with Peter Weibel, Constanze Ruhm, Michael Saup, Ulrike Gabriel and others.

One of the biggest criticisms of computer art in the 1980s was that technicians were doing it, while artists lacked access and also a deeper understanding of the machine. "If interactive electronic art is going to go farther," says O'Kane, "this literacy will be needed. The people making the concepts should be much closer to and more aware of the technology and its language and syntax."

Recently there are more artists with a traditional art background doing studies in computer science and engineering. There is a limitation to what e.g. an animation artist using off-the shelf programs can do. If she wants to get a certain color, shape, or movement, it might be impossible if the author of the software didn't put that option into it. But more and more people are learning that you can do it yourself. "The machines are open," says O'Kane. "Someone who can program and has the conceptual point can always push the border of each."

After Weibel's institute underwent reorganization, he moved on in 1993, and together with Ulrike Gabriel set up the media art studio Otherspace in Offenbach, just outside of Frankfurt. It is a rather unique cooperative relationship between the two of them, with both in a Renaissance kind of way sharing a background in both art and technology. Still they focus on their respective field of strength. O'Kane takes care of the more technical and detail work for specific machines, while Gabriel is in charge of the contents and the concepts, although she as well is quite capable of programming herself and making her own interfaces. O'Kane thinks of himself as a "linker," linking the technology with the art. The concept of Gabriel or other artists he has worked with exists beforehand. But in order to implement it correctly there still needs to be a valid translation. The installations the two create are always work in progress, with the exhibited pieces only snap-shots in time. It's neverending, these links are always growing.


Ulrike Gabriel invited O'Kane to work on the robots for "Terrain" while he was still at the Institute for New Media. It was first shown at Ars Electronica 1993, which is generally the same version now shown at P3. The installation consists of a large steel disk on which a pack of insectoids are running around. The robots, about the size of a hand, are fitted with solar cells that are fed by an array of lights above the disk. One visitor a time interacts with them through a brain-wave sensor. The brain waves are analyzed for alpha and beta components that are emitted when a person is in a state of relaxation. These are then linked in a bio-feedback loop to the lights. Dependent on how relaxed a person is, more light falls on the Terrain, and the robots get more active.

Each of the robots is autonomous. They do not communicate, just search for more light and try not to bump into each other. O'Kane speaks about the robots as being "alive." He qualifies that this is not to be mistaken for Artificial Life, but rather to be understood in a general sense of autonomous functioning. But, as with all systems that display behavior, the analogy to biology lends itself, and especially to their creator. "They look for light," says O'Kane, "so they understand that it belongs to their survival." In order to perpetuate the survival it turned out they need an act like panic. When the robot has enough light, but cannot find any more, it goes into a panic mode and starts running away from where it is. When it finds more light, it continues normally.

One of the ideas of Terrain is the dynamic structure of cellular automata, to create patterns of population and distribution in space, recognizable groups that are synergistically related.

Technically speaking, the robots are analog, infinitely accurate state machines. There is no CPU, no general purpose computer inside, and they are not programmed. An analog system is designed by selecting devices and setting range and threshold values, but where exactly a transition in its state will take place can only be estimated, and to optimize it is very difficult. You cannot pick it up, open it, turn a resistor, and put it down again to see whether that was it, because as soon as you pick it up from the Terrain to adjust it, it's not in its state anymore. To adjust it constantly while it's moving in the space is very difficult.

The best solution is self-optimizing which is what neural networks do. Technically, right now the robots are fuzzy logic systems, whereas in a neural network the result is fed back into the system, which allows it to base its decisions on its previous learning history.

O'Kane talks about the current version as a "brain stem." In the next step towards a neural network he plans to add another level, a "cortex" that would take over when the survival of the robot is ensured, when it has "free time." Then it could work on its own, maybe cooperate with others, search each other out, and create new group patterns.

Using the example of Terrain, O'Kane describes the cooperation process between artist and linker: "The resulting robots are somehow what Ulrike described to me, for example with the limitations of size or power. How we actually made them may not be exactly the way she thought about it. The philosophy might come from a different direction, like the fact that they are analog machines. If they were digital machines, they possibly would be a little closer to the actual description of the circuit. I specialized on the circuit board. The whole philosophy of light and dark is hers. Just from this difference in philosophy or perspective you'll get different things from each of us about how it works. She is not an electrical engineer. She shouldn't have to deal with that aspect of it. If you concentrate full-time on art, you can't concentrate on anything else. Or if you concentrate full-time on electrical engineering you can't really concentrate on the art aspect. Both of us agree, anyway, that the art is the important aspect. She concentrates on the whole view of the installation, not just the robots, but the lights, the grid, the brain sensor, the light machine, the whole. I concentrate on the detail work. Not one works without the other. That's the evolution of Otherspace. It functions as a whole this way, it's not divided."

The piece works at many conceptual levels. The user is the link in the system, but he is not completely in control. A point Gabriel likes to stress is this being in control of something by not being in control, by relinquishing the control. Experiences with showing Terrain in the West and in Japan, taught them about interesting differences in the cultural attitudes. In the West, people sit down and want to be in control of the installation. With the Asian ideas of meditation and relaxation, observes O'Kane, the willingness to be removed from a situation seems much stronger here. It is generally easier for users here in Tokyo to make the system work, rather than in Frankfurt. There, people were really bothered by the idea that they are responsible for making it move. The viewers stand around the user, looking at him expecting him to "hurry up and relax!" At the Ars Electronica, someone would sit down and it just wouldn't work. They would close their eyes, so they remove themselves from the situation. They relax, the robots start moving, they open their eyes to watch them, and then the lights would go off again, because they started to think about it, and the brain sensors picked up that activity. This would happen several times, because they wanted to be in control. The ultimate goal is to be able to watch the robots while they move, to "bear it with open eyes," as Gabriel describes it, to be rewarded, to observe a function of what you're doing.

The prototype of the next robot is already sitting on O'Kane's desk in Frankfurt. In the next version they want to go beyond the robot's characteristics of feeding on light, speed, panic and avoidance. They are planning to control these characteristics externally, to be able to turn on and off the panic or the avoidance, or install it, no matter what state they're in. They want to learn how to communicate with the robot, and give it information it should respond to, e.g. by using different colored lights in order to trigger certain behavior.

Another potential concept for the future is to have two people connected to the Terrain with brain sensors at the same time. They would influence different groups of lights, and if one person is more relaxed the robots will come to him.


"Perceptual Arena" was the first version of this piece that came out of the cooperation with Canon Artlab, shown in 1993. It is a classic VR installation using data-glove and data-helmet, and high-end graphic computers. The environment tries to detect how people react when they perceive something interesting. The environment is modified based on this detection, and also the representation of the environment is constantly being updated. The system doesn't learn from it, yet, but Gabriel and O'Kane do, and they evolve the system through reprogramming.

The first version used a database the user can enter and out of which he can pull objects towards himself. In the next version, "Arena Life," the graphic objects, or, as Gabriel calls them, "tribes," are self-generating according to Artificial Life principles. Groups of objects interconnect and grow into complex shapes. The trick is to give the polygons more awareness, more ability to structure themselves based on their own history, their parent, or their tribe. The user will still be able to interact with them, but their range of reactions will be expanded.

A different way to extend Arena is planned for the Internet. "As you perceive something on the net," explains O'Kane, "you might like the 3D representation of this information. The environment would enhance the growth or the importance of this object. They would become more active and grow, and maybe get more information that would make it more interesting. Or maybe it gets too big and parts of it die off, because you haven't accessed that part, but it's still generally something interesting. Other types of information maybe boring, and you don't look at them, so they get very small and wither away. Or maybe they go to a new location, travel to a new server, for example, and someone there thinks they are interesting, so they would flourish in that environment."

The user interface could be anything from a data helmet plugged into a VRML (Virtual Reality Modelling Language) page, to just a window with mouse or arrow keys. But no matter what the implementation is, there would always be some kind of graphical representation, preferably 3D. Much of the development on tracker information and drawing routines has been done for Arena, and its is rather simple to translate VRML to the database. Arena is spatially but also time-based, which is the biggest limitation at this point. There are, as of yet, no animation tags in VRML. But this is expected to change soon. There should be instructions of how objects move, whether they are attracted to other objects. Eventually genetic programming and artificial life principles will be implemented. They are also thinking about agents, but not in the sense of a dog that you send out to fetch the newspaper - more like an autonomous object that survives and grows in its element because it has interesting things around it or because someone thinks its interesting.

Maybe someday, it could even become a general browser for the network. But they are not working on a tool. Otherspace's work is, after all, still art. It should be a representative, aesthetically pleasing, dynamic system, based on the dynamics of the Internet.