2048 or 4096 for new keys? aka defaults vs. Debian

[Robert J. Hansen]

On 10/27/2013 8:21 AM, Johan Wevers wrote:
> Well, both are not broken after substantial research. Further, a break
> of ElGamal would also break RSA but not the other way around.

If you can compute discrete logs in a finite field, then you can factor,
yes, and the reverse is not guaranteed to be true.

However, Elgamal is not necessarily the same as the discrete log
problem, much in the same way that RSA is probably not the same as the
integer factorization problem.