Path: utzoo!utgpu!attcan!uunet!cs.utexas.edu!uwm.edu!rpi!rpics!spencert
From: spencert@rpics (Thomas Spencer)
Newsgroups: comp.arch
Subject: Re: LEGOs -- computationally complete?
Message-ID: <1989Oct1.023735.10234@rpi.edu>
Date: 1 Oct 89 02:37:35 GMT
References: <3300071@m.cs.uiuc.edu> <1801@brazos.Rice.edu> <218@visix.UUCP>
Reply-To: spencert@turing.cs.rpi.edu (Thomas Spencer)
Organization: RPI CS Dept.
Lines: 27

In article <218@visix.UUCP> jeff@visix.UUCP (Jeff Barr) writes:
>In article <1801@brazos.Rice.edu>, preston@titan.rice.edu (Preston Briggs) writes:
>> In article <3300071@m.cs.uiuc.edu> nelson@m.cs.uiuc.edu writes:
>> >We are interested in building something (possibly a Turing Machine) out of
>> >  LEGO blocks.  Various ideas have been popped around, but there seem to be

[All comments but the original querry deleted.]

I'm not sure what constitutes a Lego computer, but the following results
from automata theory might be useful:

1. a two stack PDA can simulate a Turing Machine.  Thus, one can imagine
a computer whose main storage consisted of two stacks of Legos.  Of course,
it is necessary that there be two possible kinds of blocks in each stack
and the finite control needs to be represented somehow.  The latter is a
small matter of engineering.

2. A four counter machine can simulate a Turing Machine.  Thus, if you are
willing to provide 4 stacks of legos, the stacks can contain all the same
kind of block.

I hope that this helps.

			-Tom Spencer
			 spencert@turing.cs.rpi.edu
			 uunet!steinmetz!itsgw!spencert
"First figure out what you are trying to do."  -Me.