On Tue, 3 Oct 2000, Kevin Elliott wrote:
>A
>cryptographically strong PRNG would then be a PRNG with a very large
>period and some way of reinjecting randomness to guarantee the device
>never begins to recycle.
>--
>
Isn't that a misnomer though? If randomness is reinjected to
prevent the system from falling into a period, then it won't
be possible to generate the same sequence of bits twice -- so
you can't use such a system for a PSEUDO-random generator, in
applications like a stream cipher or whatever. Programs rely
on the same sequence coming out of the same initial state with
a PRNG -- otherwise things like stream ciphers can't be decrypted.
What you describe above, I'd have termed an RNG - not a PRNG.
Bear
