next up previous
Next: Implications of this Up: No Title Previous: What we can

Initial demonstrations

Amorphous computing will surely precipitate startling applications in distributed sensing and active materials. Realizing these applications will also require further progress in fabricating microsensors and microactuators. Our present plan, however, is to choose initial applications that focus on computational issues, and do not depend on ongoing developments in micromechanical-electrical systems.

One area we plan to explore is solving partial differential equations. As described above, large collections of amorphous computing particles may be effective platforms for adaptive gridding techniques. At a minimum, we should be able to deal effectively with Laplace's equation, which is interesting in its own right, and provides the basis for exploring some of the more speculative ideas describe above, such as organizing systems by solving reaction-diffusion equations.

It is also important, however, to explore an application in which the particles interact directly with the physical world. Here we expect to choose a purely sensing application, so as to avoid the need to deal with low-power actuators. One plausible application would be to include a photosensor in each particle and program a lawn of particles to recognize and interpret shadows passing over it.

Gerald Jay Sussman
Thu Jun 27 16:56:19 EDT 1996