home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!europa.eng.gtefsd.com!paladin.american.edu!news.univie.ac.at!hp4at!mcsun!sun4nl!cwi.nl!dik
- From: dik@cwi.nl (Dik T. Winter)
- Newsgroups: comp.programming
- Subject: Re: how about PRIME numbers ?
- Message-ID: <8752@charon.cwi.nl>
- Date: 28 Jan 93 02:21:48 GMT
- References: <peterd.727578786@tscc2> <1jvfnq$rvn@agate.berkeley.edu> <4671@sicsun.epfl.ch>
- Sender: news@cwi.nl
- Organization: CWI, Amsterdam
- Lines: 21
-
- In article <4671@sicsun.epfl.ch> guglielmetti@elia.epfl.ch (Philippe Guglielmetti) writes:
- > about primes, I've read somewhere that it is possible to check if a
- > really large "candidate" number like 2^245-1 is prime or not in a
- > reasonable time
- Yup. A few minutes for a number in that range. The actual technique takes
- a bit more space to explain.
- >
- > I remember that this what makes modern "revealed key encryption" feasible
- > since the public key is the the product of 2 large prime numbers (which
- > form the private key used in decryption).
- Actually it is in general not really necessary that both components are
- primes. Being strong pseudo-primes will also help. But you are less
- secure, and communication might fail.
-
- > If I understood correctly, it
- > was formally prooved that no polynomial time algorithm exists to
- > decompose a number in its prim e factor...
- There is no formal proof. But it is widely believed.
- --
- dik t. winter, cwi, kruislaan 413, 1098 sj amsterdam, nederland
- home: bovenover 215, 1025 jn amsterdam, nederland; e-mail: dik@cwi.nl
-