Only IT can save the world

Once upon a time, I was a member of the Green Party. I went as far as standing for a local council seat in East London, in a part of town where the only green thing was the meat in the dodgier kebab shops. My experience of polling day was surprisingly similar to eating one of those kebabs — it taught me a lot about expectation and reality, and the consequences of trusting appearance without questioning content.

I get much the same visceral reaction when I read many of the press releases that pollute my inbox. Companies big and small are only too happy to cloak themselves in earth-friendly camouflage, for all sorts of reasons. None of it's worth a single sheet of recyclable toilet paper without evidence that they've understood and bought into the real philosophy of environmental thinking — so expect a lot more scepticism from this website.

The real irony is that IT has the potential to transform itself into the most environmentally sound industry on the planet. More than in any other area of human activity, the science that lies behind our technology is capable of driving revolutionary changes in the way we work, without impacting on the effectiveness of what we do. The creation, distribution and consumption of material goods will be forever in thrall to Newton's laws of matter and energy: information, however, is different.

It may seem a given that computers, like any electrical gizmos, will consume power. We've come a long way from ENIAC's 160 kilowatts for 5000 calculations a second in 1945 to a Core Duo running at around 20 billion calculations a second for around sixty watts — ten billion times more efficiently in sixty years. But according to some readings of information theory, that's not even scratching the surface.

For a long time, information scientists considered that it was impossible to do useful processing without losing energy as heat. Every time information was destroyed — a bit erased, say, or two numbers combined — a tiny amount of heat had to be generated. The limits for that are very small, set by an equation called the von Neumann-Landauer Limit, but inescapable. There is a way out, though: reversible computing. This is a theoretical yet widely studied idea that says if you don't destroy information but maintain it in a state where it can be reconstituted, you can get away with virtually no energy lost at all. Like travelling at the speed of light, this may be unachievable in practice, but as the semiconductor industry has shown, once you have a target you can get an awfully long way even if you never quite get there.

And everything else in IT has similarly astonishing potential. At 3GSM, the first portable devices using e-ink were on show. Electronic ink uses power only when the display changes; when it's static, it takes as little energy as a gravestone. Communication is also defined by friendly physics: if you send a bit of information by throwing a photon down a fibre, then the energy you get out at the end when the photon excites an electron is the same as you put in when the photon was spat out. The rest is practicalities.

Of course, this ignores the way that IT lives in material goods that have to be created, distributed and consumed like everything else. Well, not quite like everything else: once again, the fundamentals of information theory offer an enticing vision. This time, Alan Turing is the eco-warrior of choice. He invented the Turing Machine, the basic genie of data processing that's at the theoretical heart of everything we use to compute. One of its intriguing attributes is that any Turing Machine can emulate any other. Get one that's big and fast enough, and you can innovate to your heart's content within it. Combine that with efficient communications, and you end up with a planetary matrix of superb hardware where innovation exists purely as software — and one day, further mathematics predicts, there'll be nowhere else to go. In 2000, MIT professor Seth Lloyd combined a number of fundamental physical ideas in a paper in Nature to show that you'll never be able to do more than around 1 x 10 to the 50 calculations per second per kilogram of matter.

There's a long way to go before even Intel hits that particular limit to Moore's Law (in around 600 years, says Lloyd, we'll be able to simulate the entire universe), but its entirely conceivable that with good enough networking, advances in hardware can be made in very efficient factories and then just hooked up in situ for software innovations to use, with no transportation required and raw materials recycled from the previous generation.

The picture I've painted above may seem like science fiction. To some extent, it is. But in every case, even when the ultimate zero-energy, zero-impact machine is doomed to remain a theoretical creature of the imagination, there is a great deal to be won just by knowing what could be done and deciding to aim for it. And vision is the key component to any worthwhile change — vision, backed with a sense of the possible. IT is the only technology that can dare to think like this: bear that in mind next time you wince at some heavy-handed greenish propaganda, and believe.