# Fw: Is the universe computable?

From: Frank <logical.domain.name.hidden>
Date: Tue, 4 Nov 2003 13:40:35 -0800

I submit this link to Shmidhuber's second paper, which discusses various
probability distributions on the set of computable Universes.
ftp://ftp.idsia.ch/pub/juergen/toesv2.pdf

Sorry if this has been already covered. I'm not a mathematician, and I'm
not
entirely "into" hardcore computer science.

This other site contains the links to Shmidhuber's other works.
http://www.idsia.ch/~juergen/computeruniverse.html

Peace

----- Original Message -----
> From: "David Barrett-Lennard" <dbl.domain.name.hidden>
> To: "'Hal Finney'" <hal.domain.name.hidden>; <everything-list.domain.name.hidden.com>
> Sent: Monday, November 03, 2003 11:52 PM
> Subject: RE: Is the universe computable?
>
>
> > An interesting idea.
> >
> > Where can I read a more comprehensive justification of this
> > distribution?
> >
> > If a number of programs are isomorphic the inhabitants naturally won't
> > know the difference. As to whether we call this one program or lots of
> > programs seems to be a question of taste and IMO shows that "probability
> > calculations" are only relative to how one wants to define equivalence
> > classes of programs.
> >
> > I would expect that the probability distribution will depend on the way
> > in which we choose to express, and enumerate our programs. Eg with one
> > instruction set, infinite loops or early exits may occur often - so that
> > there is a tendency for simplistic programs. On the other hand, an
> > alternative instruction set and enumeration strategy may lead to a
> > distribution favoring much longer and more complex programs. Perhaps it
> > tends to complicate programs with long sequences of conditional
> > assignment instructions to manipulate the program state, without risking
> > early exit. Importantly such "tampering" doesn't yield a program that is
> > isomorphic to a simple one. We seem to have a vast number of
> > complicated programs that aren't reducible to simpler versions. This
> > seems to be at odds with the premise (of bits that are never executed)
> > behind the Universal Distribution.
> >
> > - David
> >
Received on Wed Nov 05 2003 - 00:03:42 PST

This archive was generated by hypermail 2.3.0 : Fri Feb 16 2018 - 13:20:08 PST