Your computer sits utterly idle much of its time, even when you're working. It works faster than you do, so it gets things done quickly and waits like a panting sheep dog for you to issue the next order.
If you're in the mood to be generous, you
can donate those "spare cycles," as they're called, to help other people do research. It's called "distributed computing,"
where a program sits on your computer and uses whatever time the computer isn't doing anything constructive to perform calculations.
It's a safe and interesting way to participate in the worlds of science, art, language, mathematics, finance and more. Scientists
have turned to distributed computing as an alternative to having huge "supercomputers" do the work. Supercomputers are extremely
expensive. Getting thousands of individuals to donate computer time is cheap.
Typically, you'll download a "work unit" and then upload the results when it's done. Most projects are highly technical in fields like cosmology and fluid dynamics. They typically offer information about what they're doing, so you can feel more like a part of the entire effort, which is rather nice. But if you're more generous than geeky, you can just nod, smile, and let the software run without needing to know what it's doing behind the curtain.
Distributed computing has been around a long time. I first wrote about it in January 2003. Then, there were only a few projects you could be part of. Now, there are dozens. Most of them use a single downloadable piece of software from the University of California, Berkeley. It's called "BOINC" (boinc.berkeley.edu/download.php).
Once you've downloaded it, you can choose one or more research projects, and off you go. In most cases, you won't even know you're crunching somebody else's numbers. It'll happen when you're not using the machine, which for many of us is actually most of the time. Some people become obsessed with distributed computing and construct whole racks of computers just to run the experimental data. Most people have it on only one little machine. And that's fine.
There's a long list of projects on the site DistributedComputing (distributedcomputing.info/index.html). The oldest of the distributed computing projects is SETI@home, which processes the hisses and crackles from space in an effort to detect alien signals. It's been going since 1999, and has logged more than 2 billion results since then. So far, no clear signal. But researchers intend to keep trying. You get a nifty screen saver that shows what's happening during the processing, but you have to know a good deal of signal theory to fully appreciate it.
In other science projects, ClimatePrediction.net wants to predict what the earth's climate will be in 50 years. The organization runs more than one experiment at a time, and you can choose which one you want to host. AfricanClimate@Homeis trying to do the same thing, but using a very different method. AfricanClimate@Homeis part of World Community Grid, a bunch of projects bundled together. MilkyWay@Homehas set itself the task of modeling the evolution of our entire galaxy.
A bit closer to home, Orbit@Hometracks near-earth asteroids (NEAs) to help detect possible asteroid impacts on the planet. The Clean Energy project is using big computing power to design better organic solar cells. Compute Against Cancer has several projects still ongoing, all of them somehow tied to cancer research. FightAIDS@Homedoes similar work in the field of AIDS research.
Not all projects are scientific. ChessBrain uses the power of distributed computing to create one huge chess-playing computer. In 2004, more than 2,000 computers on the network played a game of chess against a human grand master, and achieved a draw. Sudoku finds the minimum number of clues that have to be given for a Sudoku puzzle to be created with a unique solution.
In the art world, a program called BURP (Big and Ugly Rendering Project) makes 3-D animations, an application that's notorious for needing massive computing time. The same sort of work is being done by IMP, the Internet Movie Project. MoneyBee tries to predict how the stock market will move by using something called "neural networks." In a similar vein, GStock tests investment strategies worldwide to find the best ones. There are even projects to devise better search engines than Google, such as Grub and Boitho.
There are also projects that aren't computer-automated. If you want to participate yourself in a good work, you can sign up for Project Gutenberg, which is scanning vast numbers of public-domain works and needs proofreaders. Twenty Questions is based on the simple game by the same name, but is used for artificial-intelligence training.
All these projects and many more can be found at DistributedComputing. There's something for everybody.
Altom is an independent local technology consultant. His column appears every other week. Listen to his column via podcast at www.ibj.com. He can be reached at firstname.lastname@example.org. Find his blog at usabilitynome.blogspot.com.