Marcus has Highlighted an interesting paper:http://www.arxiv.org/abs/gr-qc/0509049 But I think it should be here in the Quantum Physics thread, and it definetly should have some discussion? Quote here from the paper: The first one is what observables (besides the volume) should be specified to actually describe a macroscopic state in a given region. Since the contribution to the volume from a given 4-valent vertex grows like j1 ·j2·j3 it is important to know whether these observables provide suppression of the contribution from very large spins rendering final expectation value of the volume finite. The second problem is connected with the first one and concerns the bulk entropy: there are many sets of graphs with many different assignments of spins on legs that give (macroscopically) the same volume. If we specify both the volume and the other observables we are left with a certain number of graphs. The question is, whether the logarithm of this number is connected with the bulk entropy much as sequences of spins are connected with the black hole surface entropy . There is/must be a 'fix' to the authors concerns, any idea's?