> I was just about to write something similar - the reason nasa projects fail is because they refuse to allow for iteration.

If that's why they fail, do they regularly succeed in spite of it or also because of it?