I've been reading some material by R. Kiehn, developing a topological approach to non-equilibrium thermodynamics through Cartan forms, where the fundamental claim is that irreversible processes are best usefully modeled by changes in topology of a state space and that many requirements classifications can be described topologically. What I've read so far seems reasonable (at least, the math looks correct). However, Kiehn seems to take his approach to be extremely universal, and has later books apparently detailing applications to cosmology and nuclear physics. While thermodynamics is general enough that this is semi-believable, it does sound like a 'crackpot' sort of claim to make, and almost any search on such a topological/Cartan form approach to thermodynamics comes up with Kiehn and essentially no one else. Does anyone know if the approach is legitimate and useful?