Hopefully TA will comment on the nuances of mast / fabric interaction. You might be right about the effects of RDM on the sail - I was mainly theorizing as follows:
As one increases the diameter of the mast from a theoretical 0 (infinitely thin shaft), to a diameter with an upper limit = the diameter of the luff sleeve (i.e., so there is no excess fabric not touching the mast), I THINK (but don't know) that the effect will be to make the seam where the sail joins the luff sleeve to be retracted towards the mast. I don't know if that's increased tension, or a change in conformation at the same tension. Whatever it is, it seems to me that pulling this seam forward can only flatten the sail.
I've always assumed that RDMs are more "rangey" because at any given downhaul tension, there will be more shape, and thus better bottom-end for an RDM. If this also adds to the ability of the sail to be floppy / flat at the top end, then an RDM would help at both ends of it's range. Hard to beat that!
As to what happens to the CoE with an RDM, I would welcome comments from TA or Jeff. In particular, one question in my mind is if the CoE moves up and down, not just fore and aft on the sail.
And...as to the effect of cams...I think there are a lot, a LOT of camless sails that are wicked fast. But the sails on the race podium almost always have cams in them.
I don't know how designers think about cams, but I think of them in three modes: 1) underpowered, mechanically deflected, 2) fully- to over -powered, fully deflected by the wind, and 3) way over-powered, mechanically restrained by the cam mechanism.
In contrast, RAF sails in the analogous modes are: 1) underpowered, under-deflected, less drag but little lift either, 2) fully- to over-powered, fully deflected by the wind, and 3) way overpowered and distorted without no mechanism to control batten position (aside from the fabric).
RAF sails are thus superior in waves because phase #1 is useful when surfing and maneuvering. RAF sails are inferior in underpowered slalom / B&F / course conditions, because energy is lost during pumping as the fabric rotates and stretches and doesn't have a stable foil shape. So the main factor in determining RAF vs cambers, at least for me, is what I want the sail to do in phase #1 (luff or be stable for pumping and gliding through lulls).
In the fully-powered condition, #2, there is very little, precious little difference between otherwise similar sails. I think you could probably determine a winner on the speed course, but it would be a damned good race and I would bet more on the specific sail and how well it rigs than on the cams.
In the over-powered condition, cams are advantageous because they help control the foil shape, which is exceeding design parameters and beginning to distort. In this situation, shape control can substantially prevent degradation in performance because any shape diverging from the designed foil shape is bad (more drag and harder to handle the shifting CoE).
That's not to say that sailmakers haven't learned a lot about controlling foil conformation with just the fabric, spar, and battens. And for sure, the smaller the sail, the more the fabric material and battens are likely to do just fine. Long ago, I had a 4.7 NP Mark IV with a bunch of battens and cams. It was very flat, and I never thought the cams added much to the sail, but to really test the theory I probably should have sailed it in 40 kts (but back then I wasn't good enough to do that). Today, I would probably think about cams in a speed sail in that size, but not for anything else. My sense is that the smallest sail that clearly and demonstrably benefits from cams is about 6.0. I could see, for instance, use for a single or twin-cam 6.0, but would probably want 4 or maybe more cams in a 10.0.
Again, the choice of cams or no-cams is mainly dictated by the intended use - I can't see having cams in waves (been there, tried that, never again!). Conversely, I prefer the feel of cams when I have to pump hard to a plane and when I'm trying to keep planing in a lull.
That's just my anecdotal opinion on it; don't know of any data (I suspect HSM and others have some, but it is proprietary). Would gladly hear other thoughts on the matter!