'How do you plug in an atom-sized computer?'

Well?

You haven't a clue.

That makes two of us.

Maybe three, if there's someone else reading this.

But I digress.

Financial Times senior technology correspondent Alan Cane's above-headlined essay in today's paper addressed this topic, which may become relevant much sooner than anyone dared dream even ten years ago.

Here's his excellent piece.

How do you plug in an atom-sized computer?

Everybody knows that the computers of tomorrow will be smaller, faster and much more powerful than those of today.

What is only just becoming apparent, however, as news of the latest research developments emerges from laboratories around the world, is just how much smaller, faster and more powerful these devices could be.

Science fiction is becoming science fact more rapidly than anyone could have predicted. It raises serious questions about the implications of the availability of virtually unlimited computing power and data storage.

IBM, for example, earlier this year published some breathtaking research from its San Jose and Zurich laboratories which brought the prospect of computers no bigger than a speck of dust closer to reality.

In one piece of work, the researchers used a scanning tunneling microscope - essentially a sharp point capable of probing the fine structure of tiny objects — to arrange individual iron atoms on a copper surface. They were then able to measure, for the first time, what is known as the "magnetic anisotropy" of the individual atoms, the specific orientation of the atom which can be used to define it as either a 0 or a 1 of binary code.

IBM thinks it might be possible to create small clusters of atoms or even individual atoms able to store information: it estimates that in such a system nearly 30,000 feature length movies could be stored on a device the size of an iPod — that is, about 1,000 trillion bits.

In a second development, the IBM researchers were able to demonstrate "switching", turning a flow of electrons on and off, in a single organic molecule, a compound called naphthalocyanine.

Switching, the basis of computer logic, is conventionally carried out on processor chips by arrays of transistors. These are fabricated by a complex process in which miniaturisation is limited by the laws of physics and we are already close to the limit. Molecular switches could be used to construct supercomputers little bigger than a pinhead.

While practical results from IBM's research is some way off, spectacular progress is being made in another approach to superfast computing — the quantum computer.

These replace ordinary binary digits, the 0s and 1s, of conventional computing with quantum bits or qubits. Electrons, photons, atoms and molecules are all possible candidates for physical qubits. These particles obey the laws of quantum mechanics and have the remarkable property of existing both as 0 and 1 at the same time, a phenomenon called "superposition". This makes it possible for a quantum computer to perform many calculations at once — perhaps 1m at a time.

In a recent article, the Oxford University scientist Prof David Deutsch (who developed the first theoretical description of a quantum computer) speculated that a general purpose quantum computer might be only a decade off.

Indeed, Canadian company D-Wave has constructed a quantum computer, which, the company says, can solve Suduko puzzles. Some experts are dubious about D-Wave's more ambitious claims but experimental quantum computers capable of solving simple problems have been constructed and there seems no reason why more advanced machines should not be possible.

So if within a decade or two it may be possible to construct molecular and/or quantum computers, there remains the question of how to manage computing power and storage capabilities on an unimaginably large scale.

There is, of course, the difficulty of making input/output connections to a computer no bigger than a few molecules and, indeed, how to provide the power for computation. Research into this latter problem is already in progress. A group at the University of Illinois at Urbana-Champaign is proposing that energy could be provided by the mechanical deformation of piezoelectric material no larger than a molecular computer itself. This material generates an electric current when deformed.

But just as increases in computing and communications capability have opened new commercial and social possibilities, so the availability of unlimited computing power will open doors that do not exist now.

In the 1990s, predictions of how technology would change our lives included the liberating effects of the video cassette recorder and the mobile phone, telecommuting and telemedicine.

These have proved generally accurate. Writing effective computer code to control even today's fairly slow computers (compared to a quantum computer, that is) is not a simple matter. There are plenty of examples, from the failure of big private and public projects to the regularity with which home computers "crash".

With unlimited computing power, however, a new methodology might obtain. It might be possible simply to define inputs and outputs and allow the computer heuristically to arrive at the best solution. The possibility of the right software every time would be revolutionary, but could prove a Pandora's Box.

John Diebold, the "father of automation", wrote in 1987 that technological advance revises how mankind perceives itself.

To be surrounded by superfast computing machinery with apparently limitless ability to recover from error might prove a chastening experience for the human race — one with which evolution could be ill-equipped to deal.

....................

Somwhere, Richard Feynman is smiling.