# probable prime tests for gmp

**Jason Moxham
**
J.L.Moxham@maths.soton.ac.uk

*Sat, 2 Nov 2002 02:06:56 +0000*

On Thursday 31 Oct 2002 12:14 am, Kevin Ryde wrote:
>* Jason Moxham <J.L.Moxham@maths.soton.ac.uk> writes:
*>* > Again if a user
*>* > wants a pure strong psuedoprime test then he will have to write one f=
*or
>* > himself , or gmp will have to provide both ?
*>*
*>* Don't know.
*>*
*>* I guess the potential strengthening of tests in tricky ways is a good
*>* argument for not exposing particular components like mpz_millerrabin.
*>* An interface that's highly specific to a particular algorithm is
*>* instantly obsoleted by an advance in the theory.
*
But if an advance of theory takes say 10 years , then we will be stuck wi=
th a=20
generic interface.=20
Something like "mpz_probable_prime_p" should be independant of the parti=
cular=20
test , so that we can take advantage of any new theory , and could also h=
ave=20
"mpz_sprp" and "mpz_rqft" for people or algorithms that need a particular=
=20
test for whatever reasons. mpz_sprp allready has good reason for being=20
exposed (crypto)
=20