@Ray,
Thanks Ray, but I am not that knowledgeable.
@Peter
Hi Peter,
Nice try for the nice try, but the paper does contain the proof of the uniqueness in section 5, independent of the mixing of classical and quantum systems.
The story goes like this. In 1974 Grgin and Petersen wrote a seminal paper: "Duality of observables and generators in classical and quantum mechanics" J.Math.Phys. 15, 764.
The following year they wrote another major paper about composability (E. Grgin and A. Petersen, Commun. Math. Phys. 50, 177 (1976)), clarifying the earlier ideas, but this one is not well known. Sahoo was colleague at that time with Grgin and Petersen at Yeshiva University, and he wrote 2 papers in this area, one being this http://arxiv.org/PS_cache/quant-ph/pdf/0301/0301044v3.pdf.
What Grgin and Petersen observed is that QM and classical mechanics have 2 products: a symmetric product \sigma: (anticommutator for CM, and regular function multiplication for CM) and an antisymmetric product \alpha: (commutator for QM, Poisson bracket for CM).
The 2 products obey 3 identities: Jordan identity for \alpha (because it is a Lie algebra), Leibniz (or derivation identity) for \alpha and \sigma (making the 2 products compatible), and an associator identity where LHS is in \sigma, and RHS is in \alpha and the proportionality constant in the RHS could be -1, 0, or 1.
The origin of those identities is the so-called composability principle: take any 2 physical system described by CM (or QM), put them in contact, and the total composed system should be described by the same formalism (CM or QM). -1, 0, and 1 correspond to 3 independent composability classes: (1 = QM, 0 = CM, -1 = hyperbolic QM). What Sahoo is doing is working the composability paper: E. Grgin and A. Petersen, Commun. Math. Phys. 50, 177 (1976). in reverse, but because there are 3 independent composability classes, his main result of proving the impossibility of combining QM with CM is rather trivial. (Enrico Prati proved the same thing in the essay contest from the C* algebra point of view.)
The argument for the uniqueness of PC is rather trivial as well: the +1 dimensionless parameter is proportional with (1/4) \hbar^2, and stability under composability demands (1/4)\hbar^2 to be the same always (this is exactly how Sahoo is doing it).
I told you earlier that today Grgin's 1974 paper is well known. This paper started the study of the so-called Jordan-Lie algebras. Augmented with the norm property, they gave rise to the study of the modern JB (Jordan Banach) operator algebras for QM.
About the experimental evidence, I do not recall it now exactly, but I think there were comparisons between different elementary particles and comparing experimental measurements between them. (See the references in Sahoo and also: http://www.springerlink.com/content/67238242437h73g4/)