RE: starship-design: Casimir-Foreward balloon

[Steve VanDevender <stevev@efn.org>]

L. Parker writes:
 > Well, that is the problem. The total momentum (and energy) must be defined
 > in the terms of the system, which as you point out is rather large in this
 > case. I seem to remember that it WAS possible to violate conservation
 > locally as long as you didn't violate it in terms of the larger system. The
 > physics for this are beyond me I'm afraid.

So far as anyone knows momentum and energy are always conserved in
relativistic physics.  If you take a group of particles and keep track
of them through all their interactions with each other, their total
momentum and energy remains the same.  If they interact with other
particles you didn't initially include, you have to go back and include
those other particles in your entire analysis until you have a closed
system -- one that does not interact with any parts other than its own.

The way this balloon thing or my simpler one-sided mirror example work
is that the system is defined to be the balloon or mirror plus all the
photons that interact with it considered over a period of time.
Starting out you have the balloon or mirror sitting "at rest" (by
definition, for our purposes), and a bunch of photons initially
uniformly distributed in direction and with no net momentum.  At the end
you have a moving balloon or mirror, and a now non-uniformly-distributed
bunch of photons whose net momentum is now opposite to that of the
balloon or mirror.

The physics are really quite simple -- and when you get to chapters 7
and 8 of _Spacetime Physics_ you'll see how.