I wrote:
Yep. You don't need unpredictability or other attributes that are usually necessary for crypto PRNGs. Basically all you need is a good uniform distribution. A quality PRNG like the ones used for simulations and games will work fine. No need to use Blum-Blum-Shub or other fancy (and slow) generators. A linear-congruential generator or an LFSR would probably do the trick.
This really doesn`t seem a good idea to me seeing as not only are we going to be wasting time on an LFSR runtime wise but we will be covering a lot of duplicate keyspace, if you look at the mersenne prime finding exercise which has its homepages somewhere on compuserve (i`ll give the address if you need it but I can`t remember it at the moment) they have a set of numerical blocks of exponent space that are allocated to people then removed from the list, not an option of course if we are going to be "stealing" runtime, unless we do it on networked workstations using some sort of UDP packet or similar. Datacomms Technologies web authoring and data security Paul Bradley, Paul@fatmans.demon.co.uk Paul@crypto.uk.eu.org, Paul@cryptography.uk.eu.org Http://www.cryptography.home.ml.org/ Email for PGP public key, ID: 5BBFAEB1 "Don`t forget to mount a scratch monkey"