Honestly, I can't say if this is a practical method in this day and age, but I'm glad you guys are sharing your thoughts.
I do wonder how "low you can go", in terms of minimum quantities of emulsion. You could use up the remainder with one of the tried and true coating methods though.
Imagine though if you had a continuous web of film that you could run over something like this, and then afterwards you could introduce a twist into the film to bring it right side up.
Or, you could just have the leading end go up... then back.
" Any suitable material", So reads the patent. I think that the success or failure of this kinde of gizzmo would depend on the surface characteristics of the roller. Having worked with coatings rolled on via multi roller machines, the interaction between the coating and the roller is crucial to how the coating is deposited on the substrate. Varrious synthetic "rubber" materials are used. I think that George probably knew, from experience, what kind of rubber worked best for the emulsions of the day, but considered that proprietary info.
One thing I have seen described in a few places is a counter rotating roller followed by a meyer rod to precisely control the thickness of the coating. The meyer rod is suspended over the trough so that any runoff is caught and reused.