Dave, old boy, not having a go at you - this is a good piece of analysis. I'm just applying critique honed from years of looking at this sort of problem. Part of my job is to determine whether my own analysis is in error.
What is mean by "flat aerofoil" is that the Cornu design was fabric stretched over structure. I do not know whether 8-H-12 profile captures the flat, but cambered, section accurately - lets assume that it does. What the Ray Prouty model will not catch is the local turbulence caused by structure exposed to the airflow.
Also there is the Reynolds number factor. Most pioneers of this era (including the Wright brothers) did wind tunnel testing on small models, but found that it did not scale up to a large structure. This is because flow transitions from laminar to turbulence the same distance along a surface, so does not scale with the model to actual.
What i am basically saying is that the Cd vs AOA will likely be higher than the data suggests...