It may (or may not) be helpful to this discussion for people to
remember that the entropies tabulated in the backs of chemistry
books do often include what we're referring to as "quenched disorder".
Many crystals at T=0 have what chemists call a configurational
entropy; ice is a common example. If the configuration entropy
were not included in the tabulated S value for H_2O, then the
tabulated Delta-G and Delta-H values would appear to be inconsistent
with each other. Yet each of these can be experimentally measured.
Therefore, if I'm not mistaken, configurational entropy can be
measured experimentally, at least in some cases.
(I don't know whether this comment is really relevant to the
discussion or even whether someone might have already made it.
Sorry--I haven't had time to read all of the very long posts.)