Being an astrophysicist (and not american), I'm entirely pro-science, and would support spending more on NASA vs say on war. But for some reason, the video by Tyson make the case that spending on science (and particularly big PR projects like flying to Mars) is the solution to all problems. I don't think it is. I think spending a good chunk of GDP on science is very productive way to incurage innovation etc., but it is not a panacea. Furthermore, I'm a bit skeptical about projects like flying to Mars, which are good PR, probably very good for engineering and technology, but not that exciting from scientific prospective.

It is probably more effective to have many instruments instead of duplicating each one. Because if you have 10 instruments and 1 doesn't work, you still have 9 left, but if you have 5 pairs of instruments, and only one fails, what's the point of having 4 pairs of identical working instruments ? I think redundancy only make sense when you are talking about instruments which are absolutely crucial for the mission success. Otherwise it is better to just have more different instruments

My guess is that there are two main reasons for the decline from linux users -- one is that the old projects are already in distrib's repositories, while new projects don't really go to sourceforge, because of its insfrastructure. For the project admins code.google.com, github and etc. are way easier to manage comparing to sourceforge (I'm speaking as owner of a few projects on sf.net, code.google and github).