From: hal@finney.org
Date: Sat May 06 2000 - 12:24:44 MDT
Billy writes:
> hal@finney.org wrote:
> > Robin has a paper on the topic of robots who try to use reversible
> > computing. My understanding was that there are unavoidable costs
> > due to the non-reversible nature of the world,
>
> Yes.
>
> > as the robot senses and
> > interacts with it.
>
> No. The big issue is the non-reversibility of the physical mechanism of the
> computer's internal hardware. It doesn't really matter what the data in the
> computer represents.
For a robot to make a measurement in a reversible and non-dissipative
manner is possible. For it to reverse that measurement is also possible.
However for the reversal to work I believe it is necessary that the
outside world must be in the same state that it was in at the end of
the measurement. This was one of the points Robin made in his paper.
If the outside world is reversed when the robot is, as is possible if
the robot is interacting with a VR, then this condition will be met.
But if the outside world is the real world, then in general this condition
cannot be met. Things will have changed in the interim and when you do
the reversal, the state of the measured system will not be the same as
it was when the initial measurement was completed.
Therefore measurements can be reversed in the virtual world but not in the
real world, and you thus have additional opportunities to save energy when
interacting with a virtual world over interacting with a physical one.
Hal
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 15:28:28 MST