--- Marcel Martin wrote:

>

> In a previous post, I wrote:

> >>2^64695-2^15-1 19476 x43 2002 Unmentionable!

>

> >It could be archived as "3-bit prime".

>

> That's anywhat. I didn't take in account the minus signs :-))

>

Yes the signs are important for these trinomials. The most genral

case I have is x^n+-x^k-1 (x,n,k all positive integers x>1,n>2,n>k>0

and discounting Mersenne and "Fermat-type" (2^n+1) numbers. Call the

trinomial f. I have not found a composite f:f|x^f-x. For instance

this for f=x^3-x-1 passes 5 rounds of Miller Rabin at least for all

x<19700000001 (1007728207 PrPs).

I forgot to say in my previous post that Primo was used for proving

the primailty of some of the factors of N+1. I used bc for my

arithmetic ;)

Paul