|
|
            |
Super-computers Nature, according to modern physics, is a dynamical system - everything in it affects everything else. And it's constantly changing. Which makes it difficult to put your finger on one bit of nature and say "this is it", because the next moment it's changed form or disappeared completely. Artists have known this intuitively, but scientists have tended to study nature idealistically, as if, like a car engine, it could be broken down into a fixed number of parts. Part of the problem was that scientists lacked the tools they needed to keep abreast of nature's moods. There were just too many variables to contain in any notebook. Using supercomputers it is now possible to study nature as a changing set of relationships, rather than a set of fixed phenomena. In Touch talks to Jeremy Main, Scientific officer at the University of Cape Town (UCT): "Supercomputing is giving us a chance to run the whole lot of scenarios and see might what happen if certain key aspects of the climate system or the earth system were to change, particularly the ones that humans has some control of, like gas emissions and pollution," says Jeremy.
"The Supercomputers are really giving us a tool, an ability to have a model of the planet earth - a model at least in atmospheric terms - and allow us to tinker with it," says Jeremy. By building models of natural phenomena, supercomputers broaden our vision. But they can also act as virtual microscopes. The National Bio-informatics Institute is using a CRAY supercomputer to construct a virtual human genome - a vast database that contains over a million genetic sequences.
From the smallest levels to the largest, supercomputers are making sense of our universe. And as they push out the frontiers of knowledge, they're also transforming the academic community. Says Debbie Hudson Lecturer at UCT :"Well, I wouldn't be able to run the type of models I'm running without the supercomputers. The reason is that these climate models are really big, there are only a few of them world wide and they were developed by a team of physicists and meteorologists."
For all their processing power, supercomputers are not miracle workers. The quantum physicist Richard Feynman complains: "It takes a computing machine an infinite number of logical operations to figure out what goes on in no matter how tiny a region of space." Why should it take an infinite amount of logic to figure out what one tiny piece of space/time is going to do?" This is something that weathermen are only too aware of. No matter how much data they collect to feed into their supercomputers, their predictions quickly go awry - in a world of infinite variables, they inevitably miss a few.
Is there a form of knowledge beyond mere number crunching? Perhaps intuition is still our greatest gift. CONTACTS
Jeremy Main, Scientific Officer
Dr Win Hide, Director, SA National Bioinformatics Institute (SANBI)
Debbie Hudson, Lecturer
David Saphire, Professor of Philosophy |
|
Created and maintained by Intekom works Copyright © 1998 Intekom |
|
Contents and images © 1998 All rights reserved |
Scientists have always built models of nature. Using supercomputers they're getting a lot closer to the real thing. Dynamic systems such as the atmosphere can now be studied as they change over time, rather than by taking isolated snapshots.
Dr Win Hide at the SA Nat. Bioformatics Institute, explains : "Hiroshima is the reason that the human genome project is currently being sequenced and what really happened is that after Hiroshima, the Americans in the Department of Energy decided they would like to know what the effect on human DNA of radiation is, so they decided that it would be a great idea to look at the actual sequence of each of our genes and see if radiation have had some effect. That was a long time ago and a lot of money was put aside to sequence every single one of the genes in the body. To get there, first of all, we had to make a map. The map took many, many years. The job that we have at the Institute is to take that information, which is raw, it has no meaning, it is just digits - and turn it into knowledge such as if you have this particular gene in this particular form you have a very strong chance of gaining cancer at the age of 45."
By empowering scholars from every walk of life, these unassuming boxes are contributing to the gradual democratisation of knowledge. As databases grow and converge, there is a growing recognition among scientists that all areas of knowledge overlap. Using machines, it may be possible, even necessary, to study physical phenomena holistically, rather than in separate categories.
"These data-bases that we generate are very important because they are made up of work that comes all the way from medicine right through to computer science and that means that computer scientists have to work with doctors and Doctors have to work with insurers and insurers have to work with lawyers and those sort of things," says Dr Hide. " At a closer level - in terms of disciplines in science - we are working with microbiologists, biochemists, health scientists, geneticists, etc. Everyone has a particular area that they need expertise for and by working with this centralised, unified kind of database concept that we are generating uniquely in South Africa, we're bringing together disciplines that have never before interacted."
"Perfect prediction in all cases is unattainable and in that sense must be obsolete. From quantum theory we know you can only hope to predict things with probability," says David Sapire, Professor of Philosophy at Wits University. "Chaos theory shows us that we can never have sufficiently accurate information about the way a system starts off to accurately predict where it will end up. As the old examples goes - a butterfly flaps its wings on one side of the world and this evolves in such a way as to end up causing almost anything - a vast storm perhaps."