[improbable22]

Statistical mechanics is basically the study of how nature stores information. Or at least, what this means for certain kinds of physics.

This article is about relatively new things, but if you rewind 100 years, you find people thinking hard about this in simpler contexts. Gibbs is the big name, who essentially re-wrote thermodynamics (the study of heat, which until then had been thought of as some kind of invisible fluid) in the statistics of microscopic particles. This often involves counting the number of different possible states, for instance of all the atoms in a gas. Gasses whose molecules have two atoms (like N_2, compared to He) have larger heat capacity precisely because there are more different ways that these bi-atoms can be oriented, i.e. more information is needed to write down their states completely.

The older, thermodynamic, description is in a sense an optimally compressed representation of this microscopic picture. It keeps only what little information is visible to giants like us, who cannot see the individual atoms.