On the choice of “Geometric” thermodynamic models