Science 2.0 has gradually been colloquially Smurf'ed into meaning many different things but at its core we have always intended to reboot science for the 21st century by modernizing publication, participation, communication and collaboration.   Collaboration was always known to be the tricky part and little progress has been made but groups like Mendeley are at least making it possible to better organize research.  Thankfully they violate the trademarks of last.fm and iTunes rather than us doing so and it's unclear if their system is better than PubMed, for example, but I am generally convinced the private sector will do better than the public sector so they have our support.   

If I could figure out a way to make a secure data-sharing tool that goes behind organizing papers I'd certainly raise the money and do it but it is tough going.

However, because it is always on my mind, I correspond occasionally with folks at the Leukippos Institute for Synthetic Biology and they recently wrote a guest post at Genome Engineering on how the logical structure of data-driven research and deducing patterns from a data pool will make it possible to improve the scientific method itself.

Will it work?  Can feedback from the world community boost the speed of science discovery?   Sure, that is the idea, but between the idea and code is where the empires are - and no one is able to cover that gap yet, including me.  But if you have ideas, I am listening.