Why compute with electrons? Why not use something a bit less exotic, like, say, water? Sixty years ago, strangely enough, a New Zealander invented a computer that did exactly that. Next time you are in Wellington, pop into the Reserve Bank Museum at #2 The Terrace and take a look at the MONIAC:
The Monetary National Income Analogue Computer computes using the flow of water through tubes and pipes. It was built in 1949 by Bill Phillips, an economist regarded by many as New Zealand’s greatest for his discovery of the link between inflation and unemployment. The MONIAC was designed to simulate a national economy: interest rates and taxes can be set by adjusting pipes and valves, and the flow of water should stabilise at a level that, with any luck, represents a real economy in equilibrium.
The particular model in the Reserve Bank Museum (above) is set up to simulate New Zealand’s economy. Sure enough, when it was last switched on, it sprung a leak (on Ruth Richardson, I am led to believe), presumably from the bucket that Phillips used to model our multifactor productivity.
In 1949, however, people were also playing with electronic computers (such as ENIAC), and it is descendents of these, rather than those of MONIAC, that dominate our lives today. In fact, the electronic computer won out over the fluidic computer, in part, because of the way power is dissipated as electrons flow.
Electrons in a metal wire behave more like a gas than a liquid. This means that if you shrink the diameter of a wire by a factor of 10, the power required to drive the same amount of current goes up by 100. If you shrink the diameter of a pipe by 10, the power needed to drive the same flow of water goes up by a factor of 10,000. This makes it much more difficult to miniaturise flows of water than flows of electrons.
So while electronics has reached the nanometre scale, mechanical fluidics has been stuck at the millimetre scale for some time. Not a problem if you just want to read a newspaper on your iPad, but not such a good thing if you are interested in manipulating individual molecules dissolved in water, such as DNA.
Nature, naturally, has solved these problems. To regulate the flow of water across cell membranes, many of our cells have proteins called aquaporins, which contain pores so small that water molecules flow through in single file. Despite their size, these pores allow water to through with ease, but block the passage of dissolved ions like sodium or chloride. If you live in a major Australian city, chances are that you’ll be drinking water that has been desalinated by nanopores like these in the near future.
It’s also not out of the question that devices could be built to do computations with pores like these. In particular, our cells can switch aquaporins on and off, depending on whether they have an over or under supply of water. While it’s unlikely we would want to build a nano-MONIAC to simulate the economy, a simple nanopore-based logic could be very useful in small devices that manipulated individual molecules.
I will be talking about this next week with Bryan Crump on Radio New Zealand Nights (Thursday, 22nd April, 8.40pm).