PRNG design document?

bear bear at
Fri Aug 22 19:22:50 EDT 2003

On Fri, 22 Aug 2003, John S. Denker wrote:

>The mentioned technology is what I classify as a
>_stretched_ random symbol generator, because it
>outputs an entropy density greater than zero but
>less than 100%.

The accepted (?) terminology among the relatively
few writers who are distinguishing between all three
classes of RNG's is that

- "True random number generators" require entropy
density of 100%,

- "Semirandom number generators require entropy
densities between 99% and 1%,

- "Pseudorandom number generators" do not consume
entropy at all and eventually (disregarding such
intervening events as the heat death of the universe)

In practice however, many writers aren't distinguishing
between semirandom and pseudorandom generators at this
time, despite the fact that their properties are drastically
different (semirandom generators for example are unsuitable
for applications such as stream ciphers). When they're not
distinguished, folk tend to use "pseudorandom" as nomenclature
for both.


The Cryptography Mailing List
Unsubscribe by sending "unsubscribe cryptography" to majordomo at

More information about the cryptography mailing list