Re: poly: Infinite calculations in an open universe?

From: Hal Finney <hal@rain.org>
Date: Tue Dec 09 1997 - 14:58:37 PST

Perry E. Metzger, <perry@piermont.com>, writes:
> Lets say there is an ever expanding set of problems that we want to
> calculate, and the set of problems we wish to calculate expands faster
> than the amount of computation we have available per unit time.
>
> As someone with a mathematical background and a background in C.S.,
> I find a lot of the arguments unconvincing because of just such
> difficulties. The problem is never having unlimited computation. In
> theory, if the universe lasts infinitely and we have enough
> electricity my laptop will perform an infinite amount of
> computation -- but it can't do enough per unit time to be of interest.

This is one of the paradoxes of infinity. It is like the hotel with
an infinite number of rooms, all occupied. Now a new guest comes along.
The manager makes space for him by moving the person from room 1 into
room 2, the one from room 2 into room 3, etc. This leaves room 1 empty
and the new guest can stay.

Next, an infinite number of guests come along. Again the manager finds
space, this time moving the person from room 1 to room 2, from room 2 to
room 4, from room 3 to room 6, ... from room n to room 2n. This frees
up all the odd numbered rooms, of which there are an infinite number,
making enough rooms for the infinity of new guests.

In the case of the computation, suppose every time you solve one problem
you generate 10 new ones. You still are able to get to every problem
eventually. You generated problems 0-9 at time 0, problems 10-19 at time
1, problems 20-29 at time 2, and in general problems 10n - 10n+9 at time n.
You are still able to retire problem x at time x, so every problem gets
dealt with in its turn and there are no problems which never get solved.

This reasoning obviously does not apply to any finite number, only to
infinity, which is why it is so important to have an infinite amount of
computation and not just a very large but finite amount.

Hal
Received on Tue Dec 9 23:06:48 1997

This archive was generated by hypermail 2.1.8 : Tue Mar 07 2006 - 14:45:29 PST