Subject:
|
Re: Building a computer from Lego's
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Thu, 1 Feb 2001 01:52:08 GMT
|
Original-From:
|
Harley Myler <(h.myler@myler.org)AvoidSpam()>
|
Viewed:
|
1380 times
|
| |
|
|
> I took a look at the Dewdney article, and I agree, it is a quasi-digital
> device, but not a general-purpose computer at all.
Nonsense. I quote:
"No general-purpose computer is complete without a memory. The memory of the Apraphulian computer consisted of hundreds of special storage elements we would call flip-flops. Here again the remarkable simplicity of the Apraphulian mind is immediately evident."
and
"The Apraphulian computer is believed to have been programmable. If it was, part of its vast memory would have been used to store the program. Program instructions would also have been merely patterns of 0's and 1's retrieved by the same mechanism outlined earlier. Those patterns would in due course have been sent to an instruction register for interpretation by the computer's logic unit."
...but the elephants needed to run the thing. Boggles the mind.
> I have been working on a "universal Turing machine" made from Lego technics
> parts, but I am cheating and using an RCX for the controller. The idea is
> that Lego parts are used for the cells on the 'tape'. (I am using the 3x3
> L-shaped lift arms in an array for the cells). This has been done before, I
> know. However, my idea is that the actual Turing program is read into the
> RCX from a specific pattern of bricks that represents the states, 0/1 input
> and output states, and left/right moves. The machine is initialized by
> 'reading' the bricks that represent the program. This program is then
> created in RCX memory, and then it is executed using the L-arm array as the
> tape. The advantage is that many different kinds of programs can be run
> without reprogramming the RCX. I think I have figured out a pattern of
> 'programming' bricks that can be reliably read for up to an 8-state Turing
> program. My 'tape' has only 32 cells, which makes the automaton a linear
> bounded automaton, rather than a real Turing machine, which needs an infinite
> tape.
>
> -Bill Leue
> wleue1@nycap.rr.com
I like it. How do you "write" a brick? Why not do something on the order of an assembly line (that would get you away from the LBA) and have the machine read a brick, place a brick (for write) or remove a brick (for erase). Your RCX is then the FSM that runs the TM. You could do some very nifty things with this machine.
BTW, can you add a pair of arbitrarily large binary numbers with an FSM?
Can you multiply a pair of arbitrarily large binary numbers with an FSM?
Of course, you can do both with a TM.
Harley Myler
http://macmyler.engr.ucf.edu
|
|
Message has 1 Reply:
 | | RCX's and Lego Trains
|
| Can you "drive" a Lego train with an RCX? The Lego 5300 has a power connector on top, or perhaps just power the rails? (24 years ago, 1-Feb-01, to lugnet.robotics)
|
Message is in Reply To:
| | Re: Building a computer from Lego's
|
| I took a look at the Dewdney article, and I agree, it is a quasi-digital device, but not a general-purpose computer at all. I have been working on a "universal Turing machine" made from Lego technics parts, but I am cheating and using an RCX for the (...) (24 years ago, 1-Feb-01, to lugnet.robotics)
|
23 Messages in This Thread:
 
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
This Message and its Replies on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
