Not so mysterious: time can be divided into discrete units, and therefore measured quite easily.
Voltage can, too: using quantum Josephson junctions, voltage can be quantized into nV steps; or more accurately, converted in a direct fashion into GHz frequencies.
Current, at least in a sufficiently small (quantum scale) system, is quantized, and easily measured as such; in a small superconducting loop, the current flow / magnetization (same thing, since they're directly proportional) is quantized because only discrete pairs of electrons can be flowing around the loop (or put another way, the ensemble wave functions have to make
Aww hell... that might be an awesome analogy...
So, physicists love wave mechanics and transmission lines. If a system can be expressed in that way, it's very powerful, because a huge toolkit becomes available. It's not always a very clean system ("deep" gravity waves are dispersive and nonlinear, for instance), but it's something.
So, normally in a metallic crystal, you have electrons zooming through the lattice. Occasionally they bump off atoms or impurities, and this is more likely to occur at higher temperature (the atoms are bumping around more). If you imagine an electron not as a particle but a sort of presence, a pinching of the local wave function, then you get the feeling that, as an electron propagates, it has a sort of wake, which pulls on atoms (electron-phonon interaction), and occasionally gets scattered by their thermal motion, or other abberations (like impurities). Which is all consistent with metallic conduction. But normally, the velocities of those waves are all over the place, thermally scattered.
Suppose the lattice were shaped in such a way that it had one resonant velocity, plus or minus: well then, you'd have a transmission line, which permits waves at only one velocity, and you'd get propagation and reflection (from mismatch) and quantization of waves (a finite sized crystal is a resonator of some sort). You might still have losses, but perhaps that could be reduced arbitrarily or through some other mechanism. As it turns out, Cooper pairs are just the phenomenon needed: it seems the spins of electrons traveling in coherent pairs cancel out in suitable materials, at suitable temperatures, that they don't disturb the lattice but skate right on through instead.
Aaaanyway, I digress... Suffice it to say, there are a number of phenomena that can be measured with discrete accuracy (i.e., as exact integer numbers rather than real numbers with inevitable error), but that are very difficult to do in most situations due to uncontrolled circumstances (most notably temperature, which scrambles readings so that even relatively moderate (~10^9), let alone huge (~10^23) integers apparently get scrambled into "real number plus noise" figures.
Tim