Distributed computing

I thought the concept behind distrubuted computing was quite straightforward: have several computers do what a single one can't in a reasonable amount of time.
I thought that calculating the digits of Pi using an BBP algorithm was a fitting, if trivial, example of task suited for distributed computing : instead of one system calculating N digits of Pi, you'd have several, the first one calculating only a handful, the second one the following digits and so on. But I've been told that it's not the best example, that it just shows that distributed computing involves distributed (computing) power.
So I'm at a bit of loss here. I'm not sure anymore what would easily illustrate how DC works. Supposedly the missing key is concurrency: the force of DC is also how the system as a whole distribute sub-tasks on the fly, collects them and constructs the whole result. While that does sound nifty and all, I have no idea of what kind of computation does that.

Care to enlighten me ?

Folding at Home

Give 1000 computers the task of folding a protein 1000 different ways, then collect statistics on the results of the simulated folding.

And say, when one computer reports a successful folding, will the project as a whole direct its attention to possible variations, using the previous result to try and see how to improve it ?

Well, out of 1000 different general folding pathways, you could select the few pathways that would seem the most probable, and then do 1000 subtle variations on those most plausible pathways, and keep repeating that in an iterative manner, until you reached a consensus folding pathway for that protein.  At each iteration , the central processor reviews the data from all the foldings carried out by users' PCs, selects the best, and then distributes a list of new foldings.

So I guess that's pretty much the same as your BPP example, in that you don't actually save any computing time by distributing the workload.  I don't see how any algorithm would cut down the total amount of computation time by distributing 1 task to n computers, as opposed to just running n tasks on one computer.  But then again I just study biochem so lol wut do i know.

oh, and while we're on the subject

peer.gomez.com

If you have some spare computers, this could be profitable.

look at grope looking for  e.t."s

I have a warehouse with 1000 386s running as a cluster