This is a TED talk about open science. It starts with a description of a new math problem which is offered on a blog, and which eventually attracts enough mathematicians working on it to solve, not just the original problem, but a more difficult version of it. It was enough easier than the usual way of doing math that it was described as being like driving a car instead of pushing it.
Then the speaker talks about more ambitious projects-- like a wiki about quantum computing-- which get started, but no one is actually willing to do the work, so that the wiki lies all but vacant.
He suggests that public science isn't what scientists get paid for nor what builds their careers, and has some ideas for pushing the standards of science to change. There's been at least one success involving publishing genomes.
Perhaps the reason the math project succeeded was because the problem was small enough that success was both well-defined and possible, not to mention that working on it was probably more fun than figuring out how to do tolerable and sensibly-linked wiki articles.
There may be a way to get publicly funded science to be open source. We're already got proof of concept for solving math problems if they're interesting enough, so I suggest going public if you've got a math problem people might like to work on.