Dear Lachlan,
I really don't believe that there is any design incompatibility but I am sure there is a lot of quality control incompatibility. After all, what incentive had they to change away from the Leica register?
The lens flange on any screw-compatible camera is commonly shimmed, and equally commonly, those shims have been lost during 'repairs'. The nominal flange-to-film distance for Leica screw is 28,8mm (27.8 for bayonet mount) and those who have measured many examples have apparently that early Feds were barely standardized at all while late Feds and Zorkiis pretty much were.
'Those who have measured many cameras' includes some who are personally known to me, and some who are not. Among the former is Dr. A. Neill Wright (of Vade Mecum fame) who does not think there is any design incompatibility with either 39mm x 26 tpi or for that matter Contax/Kiev.
A difference of around 0.05 mm -- call it two thousandths of an inch -- can worse than halve image image resolution: see Lipinsky, Precision and Miniature Cameras, London, 1955, page 28. As he points out, the thickness of a single page of his book is around 0.004 inches...
As for quality, yes, both the Jupiter 8 and the Jupiter 9 were great lenses in the 1930s when they were designed but it would be more than a little surprising if there were not qiote a few more modern lenses with higher resolution or more contrast or both. Both were designed before lens coating, for maximum contrast at the expense of resolution (in the same era, Leica chose resolution over contrast).
The Jupiter-8 (50/2) is 6 glasses in 3 groups, 1-3-2 while the the Jupiter-3 (50/1.5) and Jupiter-9 (85/2) are both 7 glasses again in the Sonnar layout, 1-3-3 (source: Foto Lyubiteli, Minsk 1964). At its best, the f/1.5 Sonnar is 'stretched' and with Soviet quality control you'd need to be really lucky to get one that was at its best.
Twenty or thirty years ago, Jupiter-8 and Jupiter-9 lenses were readily available at very modest prices. Over the years I must have had at least six or eight of the former and three or four of the latter. They aren't bad lenses. But they certainly can't stand comparison with most of the later lenses I have owned or tried.
Obviously, the better the manufacturer, the sooner the good, new designs appeared, but I'd back the Vigtlander Nokton 50/1,5 (designed about 1950) against the f/1.5 Sonnar, and over the next decade or two more and more good new designs appeared. Among the lenses I have owned or used for more or less extended periods are Zeiss, Leica, Schneider, Voigtlander (original and Cosina), Nikon, Canon, Taylor Taylor Hobson, Minolta, Vivitar Series 1, Sigma, Hexar, Kiron, Zuiko and Tamron SP. I'm talking about primes here. Zooms are another matter: until the last 5-10 years, only the best were worth using anyway, though modern cheap zooms are often astonishingly good for the price.
Of course there are inferior and later primes than the Jupiters, such as most of Meyer's output, a good deal from British Optical and many more; but even by 1960, Jupiters had been overtaken by many manufacturers and by 1970s I'd say that had been overtaken by most.
Actually, you can use a 50/1.2 on a Leica IIIa -- there's a picture of exactly that set-up on page 57 of Rangefinder, Hicks/Schultz, GMC 2003 -- and you can even just abiut focus with the rangefinder, but you'd better have the flange set up well before you do; you'd better hope the RF is accurate; and you'd better not knock the camera, as the shutter crate ain't all that strong.
Cheers,
Roger