More recently, I was bombarded with email and letters when I suggested that today's secure coding and encryption systems might be short-lived. I proposed this on the basis of them inevitably being overtaken by computing power that will unlock their secrets. But it appears we have also developed a microprocessor-dominant mindset. This is despite the fact that Colossus, the Second World War Bletchley Park code-cracking computer, is still far faster than a high-end PC at cracking its specific class of codes. An architecturally task-specific machine always wins against a general machine running task-orientated software. But almost all of the 12 billion-odd microprocessors on the planet fall into the latter category. Only recently has the cost of cutting silicon rivalled that of cutting code, thereby allowing significant performance enhancements.

Suppose we had a technology that could examine all the options. Well, it now looks as though a quantum computer is feasible. If we could use the states of elemental particles, such as electrons and protons, as analogues of today's transistors, we could get all the world's computing power today in a jam jar of fluid. But the biggest stride forward would be the ability to examine all paths, states and solutions simultaneously. This would be possible because particles have a quantum nature, with states defined by spin or dispersed location. However, quantum leaps - infinitesimal, almost unmeasureable changes - present a challenge for quantum computing: detecting the changes of state looks to be fundamentally tricky. Recently, though, a quantum computer calculated 1 + 1 = 2; not very impressive, but important. This was the transition to reality of a dream spelt out in 1985 by Richard Feynman - just 12 years ago. Predictions are that a practicable quantum computer will be with us in 10 years or so, and perhaps on our desks 10 years after that. If this turns out to be true, it will give us undreamed of computing abilities.

Almost all of the outstanding classical mathematical and physics problems could be solved in minutes by such machines. No doubt, we could also enhance our ability to predict the weather, or business activity. But perhaps the key contribution such power might afford would be the ability to model people, society and change. Instead of being subject to the chaotic change invoked by technology, perhaps we could get ahead of the game to exercise some control. We might even be able to educate people for the future, as opposed to educating them for the past - as we do today.

Peter Cochrane holds the Collier Chair for the Public Understanding of Science & Technology at the University of Bristol. His home page is:
http://cochrane.org.uk