I was thinking exactly the same thing. Only the motion component
along one of two dimensions has any effect in that system. The other
component need not exist at all. Heck, if you can impart momentum
to a ball without disturbing the rest of the system, why not apply
that technique to the vehicle itself and forget about the balls? I
think it's fairly clear that the Dean Drive (centrifugal) aspect of
the mechanism is not, in itself, the solution to the reaction mass
problem, since making a Dean Drive work still requires that the
reaction mass problem be solved by some other means. So why is the
centrifugal arrangement there at all? Just a smoke screen?
---Peace, William Kitchen
bill@iglobal.net