Eavesdropping on aliens

By CHRIS DANIELS

The Oxford English Dictionary, claimed to be the greatest book ever published, was not written by a single person.

No one individual could have finished this monumental task in a lifetime - defining every single word in the English language, then finding different examples in published work, of how it had been used, and when it was used first.

The answer lay in the division of labour. Like many other human endeavours, it was split up into thousands of discrete parts and shipped out to the workers.

It did, however, rely on communication, a reliable postal service ferrying the vital information to and fro.

Today's communication network, the internet, is taking this concept to a new level of sophistication.

For the Oxford Dictionary, thousands of volunteers were solicited, then sent a copy of the rules they needed to follow - look through books, pick out odd words or common words used in an unusual way, write them down then post the results back to those assembling the dictionary.

The first edition was finished 11 volumes and more than 40 years later, all thanks to thousands of volunteers and the postal service.

Now the internet can connect brains in an instant, using the untapped calculating potential of that box sitting in many homes, the humble PC.

By dividing up a task that would require booking a massive supercomputer for years, it is possible to get a big question answered by millions of people for minimal cost.

Known as "distributed" or "grid" computing, its potential for accomplishing huge tasks has been shown by the Search for Extra Terrestrial Intelligence at Home: "SETI@home".

Run from the University of California, Berkeley, it is not just a search for ETs. Its designers describe it as a gigantic computer science experiment, with more than 3.5 million users globally taking part.

Becoming part of the experiment is simple. You go to setiathome.berkeley.edu The system send you some information (not much, it takes only a few minutes to download), then you're off.

A screensaver is loaded and whenever you are not working, the computer is chugging its way through a block of data, sucked in from the outer reaches of space by a huge radio telescope in Puerto Rico.

And it looks pretty too, with colourful graphics moving down the screen, convincing this scientific amateur that some impressive analysis is going on.

The experiment got so huge recently that almost 40 per cent of all the data streaming into, and out of, Berkeley, was part of the SETI@home project being run from its Space Science Laboratory.

What are they after?

Well, your computer is basically looking for something scientists call a "Gaussian". These, they say, are the "bell-shaped power curves that are typical of continuous signals coming from space". Intelligent aliens communicating, in other words.

A special detector on the Arecibo telescope sucks in signals from space, while other scientists are using it for their own projects. The information is divided up into small blocks, sent out to all the millions of members who have downloaded the SETI@home software.

When each block is finished (for me it took 76 hours and 45 minutes) it is sent back to the Berkeley boffins, who try to see if your little PC at home found the crucial contact.

They say that of the suitable 1.25 million Gaussians found, only 1397 qualify as being likely cases of multiple detections, another criterion they use to whittle down all their possible ETs.

A single computer would have taken 400,000 years to process what has been done so far.

Since SETI was such a massive success, able to convince legions of sci-fi enthused PC users to donate computing time, the idea of distributed computing took off like ET on a BMX.

But alas, the power of distributed computing has not proved the answer to the world's problems some had hoped.

Vijay S. Pande, an assistant professor of chemistry at Stanford University, described the problem facing distributed computing enthusiasts in an interview with the ABC.

"Just giving someone 100,000 computers doesn't solve the problem," he said. "It's like giving someone 100,000 secretaries. What you need is a way to organise these guys and come up with ways that you could actually use all the secretaries. Otherwise, you end up wasting them."

The skill in designing such a system is to get a task very carefully split up into the tiny little pieces, ready to send out to all the willing participants.

SETI@home is a task custom-made for distributed computing, because, said Pande, it was similar to giving his 100,000 secretaries one simple page of dictation to type.

Many other problems need one part solved before the next stage can be processed - which would take forever using a distributed network. Other problems that need sheer computing grunt also fail when attempted by thousands of small computers, such as weather forecasting.

Where one part of the equation has an effect on all the rest, all those thousands of screen savers whirring through the coffee breaks just cannot match the power of a super-computer.

The future of distributed computing may then remain with pretty screensavers simply crunching numbers in the search for ET rather than looking for the more important, but far more complex, cure for Aids or cancer.

Links

* distributed.net - A non-profit organisation that uses all its members' computing power to solve encryption problems. They try to win prizes for breaking fantastic codes by sheer brute force.

* Entropia - A firm that develops the software needed for distributed computing systems to work properly.

* Folding@home - This site, along with the genome@home, is part of the attempt by Stanford scientists to solve some of the greatest riddles of biophysics.

* Popular Power - A defunct company set up to make a buck out of convincing companies to lease out their unused computing power. Check out their original site to see the high hopes.

* The Golem Project - Another example of the power of distributed computing coming up a bit short. These guys offered screensavers that would help to evolve robots. But things all got a bit too hard for the 30,000 home computers involved and the robots were not getting any smarter. The spectre of Terminator walking out of the computer laboratory of Brandeis University in Massachusetts has thankfully been averted.

* www.naic.edu - Home of the National Astronomy and Ionosphere Centre, which runs the huge radio-telescope at Arecibo, Puerto Rico. This is where the data being analysed by home PC users was collected. James Bond fans will also recognise the dish as being used for nefarious purposes in the 1995 film Goldeneye. For a more "realistic" depiction of its capabilities, check it out in the 1997 film Contact.

* Mersenne.org - Otherwise known as the GIMPS project (Great Internet Mersenne Prime Search). Bit of a mystery to the layperson, but it is all about finding great big (4 million digit) prime numbers. Download this software which, running in the background, can find a Mersenne prime number. It takes about a month to process all the calculations and even then you have just one chance in 60,000 of finding one of these things. But imagine being able to tell all your friends!

* Compute Against Cancer - Part of a company called Parabon, which is also trying to get into the business of organising unused computing power for charitable purposes.