When Everything Computes

When I was in college, most computers were segregated away from their human users behind closed doors in air-conditioned rooms with false floors (under which the masses of cables ran). Users interacted with them through remote, character-only terminals and punched cards. But despite this immediate reality, it was quite obvious to us where the future led: small, affordable computers that anyone could own, with graphics displays, all networked together into what we called the "world net." It's taken thirty years for that prediction to become real, and many of the details are different than we might have suggested at the time, but the broad structure of things is just as we thought it would be.

Recently, I asked myself if I could make a similar prediction for the future, or is that sort of clear and confident prediction only available to people who've just arrive on the scene. I'm happy to report that I had very little trouble covering both the next iteration (call it thirty years, since that's what the last iteration took), and also the iteration following that.

The main theme for the next phase is ubiquity. We already have processors in toasters and watches and cell phones and MP3 players. But this is really just the first step. Except for embedded microcontrollers, which are integrated into specific physical artifacts, I suspect that the eventual prevalent form of ubiquitous computing will be paint-on computers. A little more work in the MEMS (Micro Electro Mechanical Systems) line of research and we should be able to manufacture paint consisting of millions of small processors. Paint it on any surface you like. It will come with graphic and aural outputs, and sound, touch, and video inputs, and be instantly and wirelessly linked into the Internet. (Power is perhaps the hardest issue here, which I'm going to gloss over, using the excuse that I can't be expected to predict everything.)

Once paint-on computing is a reality, user interface as we know it becomes freed from virtually every limitation. You don't have to sit at a computer screen any more. You don't have to use a mouse. You don't need a keyboard unless you decide it's the best way to input data. There's enough computing power to do anything you want in terms of graphics and animation. In short, the UI is limited only by your imagination and human psychology. You can paint the computer onto a piece of 8.5x11 paper, onto a business card, onto a wall, onto a desktop, or onto your own skin. All those separate pieces are connected to be one interwoven, interactive, real-time computer system.

It's worth stopping and thinking about what you would do with that sort of capability. Even if you're (temporarily) stuck with today's technology, it can help you understand the difference between user interface design and the limitations and peculiarities of our current computer implementations.

But that's only a step to the eventual goal: computing all the way down. Eventually, we'll be able to incorporate computers and computer-driven sensors and actuators into the very substances we use to build artifacts. Not something painted onto a wall, but the very material the wall itself is made of. Eventually, every man-made object and material will be full of computers, literally composed of computers. This will require completely new ways to manipulate and manufacture materials, what we now call nanotechnology.

I tried to think of some of the "obvious" applications for this in the chapter I wrote for the book Nanotechnology: Molecular Speculations on Global Abundance. But by far the more interesting aspect of both this and the earlier stage are the applications we can't think of because we haven't lived with the technology long enough or deep enough.

One example of the possibilities comes from the MEMS work at U.C. Berkeley and elsewhere: Using MEMS, it is possible to construct a display with active pixels, such that each pixel performs a raster scan projection. That is, each pixel shines modulated light out sequentially in each possible direction in turn. Because the light is modulated on a per-raster basis, it can be different for each angle from which the display is viewed. Someone looking at it from the right side would see something entirely different than if they looked at it from the left side. This, in turn, means that potentially every viewer of the display might see something completely different. And the display could be fully 3D with no special glasses required, since it would project different images to each eye of the viewer. Those scenarios require sophisticated viewer and/or eye tracking as well as the basic display mechanism. But this is not something that is expected in thirty or sixty years, but something that people are working on in the lab right now.