From: m5@vail.tivoli.com (Mike McNally)
Date: Sat, 23 Jul 94 06:29:32 PDT
To: rjc@gnu.ai.mit.edu (Ray)
Subject: Re: GUT and P=NP
In-Reply-To: <9407230956.AA28103@geech.gnu.ai.mit.edu>
Message-ID: <9407231321.AA00766@vail.tivoli.com>
MIME-Version: 1.0
Content-Type: text/plain


>    We are not talking about physical computers, we are talking about
> turing machines. If there is some *finite* deterministic process to
> get from the initial data to the final result, no matter how long it
> takes, it is an algorithm.

I don't see the need for determinism; it depends on the underlying 
computational model.

>   I can't think of a single thing which is non-algorithmic
> except true randomness or non-determinism. 

The "essence" of nondeterminism may not be algorithmic, but I don't
see why that's important.  If nondeterminism can be sufficiently
characterized that I can express an algorithmic process involving
it (and of course we can; that's how NP problems are expressed) then
my boat floats.