Re: "All random number generators eventually exhibit periodicity"?????

From: Douglas A. Gwyn (DAGwyn_at_null.net)
Date: 07/28/04


Date: Tue, 27 Jul 2004 23:38:48 -0400

Paul Rubin wrote:
> I think PRNG is supposed to mean about the same thing as a PRF from
> concrete security, that is, a function from N to {1,0} that can't be
> distinguished from random by any P-time computation.

No, for one thing that would make some ?RNGs not known
as to whether they are PRNGs or not. Traditionally, a
PRNG is *any* deterministic sequence generator whose
output is intended to have *some* degree of resemblance
to a random sequence. Note that even a true RNG isn't
necessarily uniform. If one wants to narrow down the
meaning to just generators with specific properties,
more must be stipulated than merely "(P)RNG".



Relevant Pages

  • Re: Random Number Generation -----> Hardware or Software?
    ... Is a Pseudo-Random Number Generator's output functionally ... equivalent to a random sequence? ... A: PRNG: The next number in the sequence is knowable with sufficient ... netcom. ...
    (comp.arch.embedded)
  • Re: new /dev/random
    ... >computationally undistinguishable from a true RNG. ... Research papers ... >these days tend to just say PRNG for this, but CSPRNG makes it clear ...
    (sci.crypt)
  • Re: new /dev/random
    ... >>turn it into a PRNG. ... > series of posts make one suspect that you don't know the ... Get off this crap about "you don't know the difference between a true RNG ... The problem with this tread is even after I post "the latest patch has ...
    (sci.crypt)
  • Re: new /dev/random
    ... >>computationally undistinguishable from a true RNG. ... Research papers ... >>these days tend to just say PRNG for this, but CSPRNG makes it clear ...
    (sci.crypt)
  • Re: OTP4U (new cipher proposal)
    ... > Nice hobby cipher but it requires a large constant output from a PRNG. ... Actually, true RNG. ... > I could only get about 20,000 bytes a second at 99% randomness from my ...
    (sci.crypt)