[seqfan] Re: Software for searching for Wolstenholme primes?

[israel at math.ubc.ca]

It is wrong. If n is composite and not equal to 4, then (n-1)! == 0 (mod 
n). So computing (n-1)! (mod n) would indeed be a deterministic primality 
test.

Cheers,
Robert Israel

On Sep 29 2015, Georgi Guninski wrote:


>I don't this gives deterministic primality test,
>since if $p$ is composite the algorithm will return
>"random" results AFAICT.
>
>Would be glad if this claim is wrong :)