A delayed reply to the CNC thoughts.
What sort of CNC machine would work for finishing cylinders? I would guess a cylindrical grinder, but I suspect few toy engine cylinders are built on 1/2 million dollar (new cost) CNC grinders. I would guess most are still finished on manual (automated) grinders. Some discussion a while back suggested OS burnishes their sleeves. At OS the production volume is ~2 orders of magnitude larger than diesel engines. Sometimes the finishing method is evident to the eye, other times it's not. Using a CNC grinder for cylinder sleeves is certainly possible. Some equipment can resolve .0001° angles in the grinding head. This is more than adequate for a ringless engine. I doubt that many model engines are made on equipment this advanced, except maybe the high volume or high quality mfrs, OS, Saito, YS.
The piston to cylinder "feel" could be quantified with a specific taper angle (on the bore and piston) and materials, etc., but piston and sleeve pairs must still be hand selected for the proper mating fit. The entire idea behind successful high volume manufacture is to get the operators feelings out of the design equation. This is done by putting tolerances on the parts. This isn't to say a modern CNC machinist doesn't need "feel". In the case of CNC, operator feel is knowing how much the dimensions change as the machine warms and how much tooling wears and being able to account for that with tool offsets to keep parts within spec to avoid making scrap.
I understand CNC and the tolerances they work in. There is a difference between a diamond turning machine with air bearing spindles and ways with 0.01 nanometer resolution and machines which make engine parts. But CNC is just as much program, tooling, operator, as a manual machine. My father programs, sets up, and runs old clapped out CNC lathes at his place of work and can obtain ±.0002" when called for. This, while programming the taper due to worn ways out
of the part. When running ±.0002" parts he can't let the machine stand idle more than a few minutes before dimensions begin to change. The machine cuts air for an hour or two in the morning if possible as well, otherwise the changes are hard to keep up with as the machine warms from cold overnight. Making scrap on a CNC is all too easy.
For those who haven't seen, Peter Burford goes into great detail explaining how his little engine is made. http://www.peterburford.com.au/manufacture.php
There is also this bit on Tom Ridley. http://modelenginenews.org/cornell/oliver_reprod.html
For reproduction engines I don't have much to say. To me, anyone who makes a repro is taking advantage of the heritage of the name. The vintage competition rules make it a necessity. The rules do limit engine designs to technology (materials, porting) of the era, why also limit the external appearance? I don't compete and I'm certainly not nostalgic for these types of engines, but it seems such competition rules created a market. And, like everything else there is good and bad.