> I admit this is beyond my knowledge, but maybe the following is rather
> intuitive and not too incorrect.
Another way of looking at it: RAM is normally implemented as a flipflop.
(The EEs insist on calling them "bi-stable multivibrators," [1] but I
think that's just too kinky for a family-friendly mailing list.) The
way a flipflop works, the contents are refreshed and/or changed with
each clock cycle. Each and every clock, the former contents are
replaced with whatever the current state should be. If a bit held 1
before and it holds 1 now, that still counts as a bit erasure for
thermodynamic purposes.
[1] No, I'm not kidding. See, e.g.,
http://en.wikipedia.org/wiki/Bistable_multivibrator
> Physics and computation at this level are pretty unintuitive, I think.
*Very* unintuitive, yeah. You flat-out can't trust your intuition: you
have to take refuge in math and physics.
To really understand computation at the limits of physics requires
general relativity (Riemann geometries, tensors, really high-end
calculus), quantum mechanics (matrices, Dirac brakets, eigenvalues,
probabilities), computational theory (discrete math, state transforms,
etc), statistical entropy, thermodynamic entropy, Shannon entropy, and
more. It's hard. I wasn't kidding about this field making me feel like
a dog sitting at a table with Ed Witten and David Deutsch. Woof woof.
Niels Bohr is supposed to have said anyone who is not shocked by quantum
mechanics clearly has not understood it. The same can be said about
computational limits.
_______________________________________________
Gnupg-users mailing list
Gnupg-users@gnupg.org
http://lists.gnupg.org/mailman/listinfo/gnupg-users
