home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.theory
- Path: sparky!uunet!stanford.edu!CSD-NewsHost.Stanford.EDU!Sunburn.Stanford.EDU!pratt
- From: pratt@Sunburn.Stanford.EDU (Vaughan R. Pratt)
- Subject: Re: Cryptography and P=NP
- Message-ID: <1992Nov23.033725.22456@CSD-NewsHost.Stanford.EDU>
- Sender: news@CSD-NewsHost.Stanford.EDU
- Organization: Computer Science Department, Stanford University.
- References: <1992Nov17.043640.5525@CSD-NewsHost.Stanford.EDU> <1992Nov22.171847.27106@athena.mit.edu> <102829@bu.edu>
- Date: Mon, 23 Nov 1992 03:37:25 GMT
- Lines: 42
-
- In article <1992Nov22.171847.27106@athena.mit.edu> patl@athena.mit.edu (Patrick J. LoPresti) writes:
- >
- >Isn't it true that, with respect to a random oracle, P!=NP with
- >probability 1?
-
- The same argument with "set of Gods" in place of "oracle" proves the
- existence of a God. There are infinitely many possible sets of Gods,
- only one of which is the empty set. Hence with probability 1 there is
- a God.
-
- See you in church. :-)
-
-
- With regard to the analogy with Goldbach's conjecture, I believe the
- number of representations of 2n by a sum of two primes is close to that
- expected statistically, certainly for as far out as this test has been
- conducted (anyone have concrete figures on this?). Since this hit rate
- grows at something like n/log^2 n, the probability of getting 0 hits
- for some value of 2n instead of the expected n/log^2 n hits, out beyond
- where the conjecture is currently known to hold, can be calculated
- quite explicitly to be a certain *very* miniscule yet definitely nonzero
- number.
-
- The corresponding statistics for random oracles are rendered
- meaningless by the arbitrariness of the measure chosen for their
- distribution. For Goldbach there is no arbitrariness in the measure,
- the set of primes below 2n consists of around 2n/log n primes with a
- quite well (though not perfectly) understood distribution.
-
-
- In article <102829@bu.edu> gene@cs.bu.edu (Gene Itkis) writes:
- >There has been a number of fairly recent results contradicting the random
- >oracle conjecture. Basically, these results indicate that the "statistical
- >evidence" from the random oracle is not very reliable. The results include
- >IP=PSPACE, MIP=NEXPTIME and others (these theorems are false with respect to
- >random oracle with prob 1, but true in the absence of oracles).
-
- I overheard at FOCS that Juris Hartmanis has recently been marshalling
- further arguments of this type against the significance of
- relativization results to P=NP, which I'm looking forward to seeing.
- --
- Vaughan Pratt A fallacy is worth a thousand steps.
-