The proof of correctness could be given in holographic form, after which a human could check the first eighty bits or so and for all practical purposes it would be correct. See http://www.cs.bu.edu/fac/lnd/expo/holo.htm On 3/3/06, Pacher Christoph <Christoph.Pacher@arcs.ac.at> wrote:

...

Such a computer program would be as intricate as the proofs it is expected to verify, and is subject to the same human error. There cannot be certainty of correctness of a proof without certainty of the correctness of the proof-checking software. Could the program check itself?

this reminds me to the halting-problem, which is known to be not solveable...

--Chris

_______________________________________________ math-fun mailing list math-fun@mailman.xmission.com http://mailman.xmission.com/cgi-bin/mailman/listinfo/math-fun

-- Mike Stay metaweta@gmail.com http://math.ucr.edu/~mike