> The quality of the average thermodynamics textbook doesn’t help. The very beginning of Huang’s statistical mechanics text is a really good, really short summ

> The quality of the average thermodynamics textbook doesn’t help. The very beginning of Huang’s statistical mechanics text is a really good, really short summary of axiomatic thermodynamics. The bulk of the book is 60s-era stat mech models, but the first couple dozen pages contain a surprisingly lucid exposition of the classic story of Clausius, Thompson, and Carnot. Rumer and Ryvkin is also a good book that explores things purely from a perspective of phenomenological axiomatics before plunging into statistics, and includes for example a discussion of how the potentials determine the equilibrium state under various conditions. Unfortunately, neither of these subsumes the other, and I wouldn’t really recommend either as a first-time introduction, so that mental spot is still empty for me. But as a way to demystify thermodynamics (which I find myself reaching for every few years with disturbing regularity) they serve well. (I also have a stack of various people’s notes that finally made the Legendre transform click for me lying around here somewhere, if anybody wants them.)