Your viewpoint is very agreeable, but how do you propose to do it otherwise? Given limited resources (which continues to be the case even today), you have to allocate them somehow. Can you think of a simple and objective measure to distribute the available resources between research projects?

It's all right to decry financialization of research, but it is a very difficult problem to solve.

I would borrow from the "20% free for personal research" deal Google offers software developers: "spend 20% of your time on anything you want." Gmail, for example, was born as a 20%-spare-time project. (The creator of Gmail, Paul Buchheit, is now at Y Combinator.)

People who have dedicated their life to science should get more than 20%. Say, 50%: "Spend 50% of money/time/people on the funded project, 50% however you want."

I think my 50/50 idea would work quite well. Scientists still would have to demonstrate and justify the value of near-term projects to get funding, but would also get the resources they need to fund long-term projects with impossible-to-estimate present value.

Round robin or following a uniform distribution?