An Interesting Q TT.
Car mags initially used 'carbon wipe' dizzy heads, and they're 'OK'. But quite soon - well pre-WW2 - Lucas, BTH and other largely British manufacturers pioneered the spark gap rotor arm set-up as per coil ignition. Much better for cleanliness in the dizzy head, air gaps came into their own when nickel electrodes became a thing, and they also, it is said, extended spark plug life (or at least the intervals between having to clean them). Which probably made it less critical than it had been to test magnetos using various 'utility factors' (big unwelcome resistances) to try to replicate real-life situations in tired dirty engines, although test equipment always retained that facility.
As to 'Why Not?' in non-distributor types, dunno, but I absolutely can't imagine that Lucas, or particularly BTH and Bosch, others too maybe like ML and CAV, hadn't thought about it. I'd guess it was an idea discarded as opposed to one never thought of, but WTHDIK? Some of the folk who wrote Learned Papers on ignition systems for the various learned societies were extraordinarily clever.
Don't think 'speed of rotation' comes into it because two cam-lobe ringcam 4 cyl car mags run at engine speed, 6cyl at 2/3rds etc, and 'camshaft mags' (a straight swap for a distributor and running at 1/2 engine like ours) used however many lobes there were cylinders, obviously. So sparks galore. And slipring diameters on most rotating coil magnetos are not far off the internal dimensions of many distributors of the day.
Could it be that the risk of spark erosion on sliprings that aren't that easy to replace (the way rotor arms are) has anything to do with it? None of the above manufacturers used an air gap on rotating coil magnetos to get the spark from sliprings to the distributor HT feeds. Or maybe 2 gaps to jump would be a step too far at low rpm when the mag is at its feeblest. Not a non-dizzy mag issue though.
I think I'll amuse myself with an experiment when I get A Round Tuit. I no longer have a good means of measuring HT voltages at my disposal, but at least I'd be able to see what size gap any sparks would jump which is a pretty good proxy for voltage - and compare to the same mag running with the standard bits. Worth a try just for curiosity's sake. (We all know what happened to the cat though.)