Series of minor questions about OpenPGP 5

Robert J. Hansen rjh at
Wed Jan 28 18:36:10 CET 2009

Peter Thomas wrote:
> I've read about special hardware devices that (claim to) give true
> random numbers, some based on thermodynamics some even on quantum
> mechanics.

True randomness exists in nature, but so far we're unable to detect it.

Imagine you have a Geiger counter and a radioactive sample.  Over each
time frame, the Geiger counter reports how many decays it measures.
That number becomes your random value.  So far, so random, right?

But a Geiger counter has a reset time.  Once it clicks, there's a small
time window in which it's unable to detect new decay events.  This has
the effect of introducing a bias into your random number generator: some
decay events will be transformed into non-events.

There are some tricks of physics and mathematics you can use to get very
high quality values out of this kind of radioisotope setup, but the
basic problem remains: even when measuring a totally nondeterministic
event, determinism in the detection mechanism will undercut you.  You
can get really, really close to truly random numbers, but you can't get

> Are there any supported in Linux and would
> it make sense or is it even possible to use them with gnupg?

Linux has support for some hardware RNGs, yes.  I don't know offhand
which ones.  OpenBSD apparently has support for a lot of them.

More information about the Gnupg-users mailing list