Search All of the Math Forum:
Views expressed in these public forums are not endorsed by
NCTM or The Math Forum.



Re: Wolfram's 2,3 Turing Machine Is Universal!
Posted:
Oct 25, 2007 5:57 PM


gjedwards@gmail.com wrote: > Wolfram seems, to me at least, to believe his work is hugely important > (even more important than James Harris' ? ). I'm not qualified to > judge whether or not he's a crackpot, though clearly his success in > other areas would mark him as if very unusual crackpot if that were > the case.
His work is not important. I'd tend to classify him as a crank. He's arrogant (as you noted) and ignorant (of the prior art). I liked this review of the book:
http://www.cscs.umich.edu/~crshalizi/reviews/wolfram/
> I genuinely want to know if anything significant has been > discovered due to ANKOS that otherwise might not have been.
Not in the sense of ANKOS contributing materially to the discovery, since it didn't contribute anything to the field of computer science. Someone might have read it and been inspired to do research they wouldn't otherwise have done, and discovered something that wouldn't otherwise have been discovered. I don't know of any example of that happening, and it could just as well happen with The Da Vinci Code.
ANKOS seems to be a lot like memetics. My impression (which may be totally wrong) is that they've both attracted a lot of people who want to do research but aren't part of the academic system and probably couldn't get hired in it. They set up their own journals and conferences and do research as a hobby. They basically have the field to themselves because it's not really a new field of study; they're duplicating work the real academics have already done. I don't know if this has a name ("indie science"?) or if it's been studied (by, er, real sociologists), or if it's all in my mind.
In any event, I'm sure this proof is real and that it's a new result. It's not a very important result, since lots of simple computationally universal systems are already known. This one just happens to be a cellular automaton of the kind that Wolfram is keenly interested in for some reason. I don't think it would have attracted much interest among serious researchers if not for the prize money.
My favorite computationally universal automaton is SMETANA:
http://catseye.tc/projects/smetana/
Someone explicitly constructed a Turing machine in it a few years ago. You need an infinite (algorithmically generated) program for this, though, since a finite program can only be in finitely many states. Malbolge has also been shown universal with caveats:
http://www.lscheffer.com/malbolge.shtml
And Conway's Life was proven universal ages ago; I think all of those are more interesting than this new result. By the way, if you haven't been following the progress in Life technology over the last couple decades, download Golly (golly.sourceforge.net) and play with the included examples. Now /that's/ cool indie research.
 Ben



