jwz - Turing Trains [entries|archive|friends|userinfo]
jwz

  www.jwz.org
  userinfo
  archive
  rss

Links
[»| DNA (Log) (iCal) WebCollage (LJ) Mixtapes ]

Turing Trains [Tue, 10-Feb-2004 2:35 PM]
Previous Entry Add to Memories Tell a Friend Next Entry
[Tags|, , ]
[music |Meat Beat Manifesto -- Mass Producing Hate]

These folks have constructed a Turing Machine out of a model railroad.

Allowing ourselves to fleetingly believe in an earlier historical miscalculation that "... Computers in the future may have only 1,000 vacuum tubes and perhaps weigh 1 1/2 tons." (Popular Mechanics, March 1949), we decided to put some hundred tons of scaled steel together in order to build these calculating protozoa. The operating system of this reckoning worm is the ultimate universal calculator, the Turingmachine, and is able to calculate whatever is capable of being calculated.

I, for one, welcome our new locomotive overlords.

linkReply

Comments:
[User Picture]From: [info]deadprogrammer
Tue, 10-Feb-2004 2:40 PM (UTC)

(Link)

But will you be helpful in rounding up other humans to toil in underground fluorescent plastic mines?
[User Picture]From: [info]abates
Tue, 10-Feb-2004 2:47 PM (UTC)

(Link)

That brings a whole new meaning to the term "train of thought".

Not to mention the term "system crash".
[User Picture]From: [info]baconmonkey
Tue, 10-Feb-2004 2:54 PM (UTC)

Re:

(Link)

and think of the havoc that could be wroght by a bug like a grasshopper or locust!
[User Picture]From: [info]ivan_ghandhi
Tue, 10-Feb-2004 3:21 PM (UTC)

(Link)

As far as I remember the definition, a train system cannot represent a Turing machine, because a Turing machine has an infinite tape.

Also, there is no generic Turing machine. A Turing machine, afaik, is any finite automaton with an attached infinite tape.

The guys could probably build a finite-state machine out of railroad blocks, that's easy. But not a Turing machine!
From: [info]premchai21
Tue, 10-Feb-2004 4:30 PM (UTC)

Re:

(Link)

So just build the automaton-emulator automaton—the universal Turing machine. Though I gather that's not what they did... ?

[User Picture]From: [info]quercus
Tue, 10-Feb-2004 5:07 PM (UTC)

(Link)

A finite TM can certainly "represent" an infinite TM - you just have to have a tape long enough to represent all the calculations of interest to you.

Calculating just how big this minimum tape length is then becomes an interesting problem, a bit like the "busy beaver" problem - how long can a FSM (finite state machine) run for, for a given number of states.
[User Picture]From: [info]omnifarious
Wed, 11-Feb-2004 1:06 PM (UTC)

Re:

(Link)

That makes the term 'Turing machine' a pretty useless definition of a computing system than since it can never actually be applied to any real computing systems, only theoretical ones.

In order that the term be made more useful, I propose that it be accepted as a description of real systems with the obvious implied caveat that it's not actually a real turing machine by the most strict definition as it does not have infinite storage space.

That should satisfy all the pedants who object to such an implied definition without having it stated anymore, no matter how obvious it is.

[User Picture]From: [info]omnifarious
Tue, 10-Feb-2004 3:49 PM (UTC)

Many welcomed overlords, duking it out for control of my life

(Link)

You seem to be of the mindset:

Anybody who wants to be my overlord is welcomed. They're welcome, even if they haven't actually stated they want to be my overlord but might possibly, in the far distant future, contemplate wanting the position. The more overlords, the better, because then they'll be too busy fighting to worry about me.

or something similar.

It's a fine mindset to have, but I just thought I'd note it in passing. :-)

[User Picture]From: [info]ivan_ghandhi
Tue, 10-Feb-2004 4:25 PM (UTC)

(Link)

Do I understand you correctly that you suggest to redefine Turing machine to suit the purpose of the publication? Have you ever tried to apply this approach to solving other mathematical problems? E.g. prove that x^n + y^n = z^n does not have solutions for integer x, y, z and natural n > 2? Proof: let's redefine ^ operator to be multiplication. Voilá. Now you can go ahead and crack RSA.

[User Picture]From: [info]omnifarious
Tue, 10-Feb-2004 6:36 PM (UTC)

Re:

(Link)

I wasn't proposing that anything be redefined. I was just noting [info]jwz's desire to submit to it should it ever become an overlord.

[User Picture]From: [info]ivan_ghandhi
Wed, 11-Feb-2004 12:42 PM (UTC)

(Link)

Sorry, I misunderstood you. :)
[User Picture]From: [info]myasma
Tue, 10-Feb-2004 4:43 PM (UTC)

(Link)

Your posts are some of the most interesting on my friends list. Even the ones I don't understand.
[User Picture]From: [info]jfedor
Tue, 10-Feb-2004 5:01 PM (UTC)

(Link)

[User Picture]From: [info]blowtorch_betty
Tue, 10-Feb-2004 5:55 PM (UTC)

(Link)

I have no idea what any of you are talking about but I like you anyway.
[User Picture]From: [info]spendocrat
Thu, 12-Feb-2004 11:43 AM (UTC)

(Link)

*I* think you posted this for the matching colour scheme.