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.