Not computer aided, as far as I know.
I use a 17mm Tamron SP (my second) although I'd buy a 21mm Leitz lens for my M3 if I needed a lens like this for commercial use (with film). I much prefer the 21mm Leitz lenses but would buy a Cosina eqivalent if the budget was too tight however I can't justift it these days as I rarely use 35mm.
David, in Germany the first computers to calculate lens designs were applied in the late 50's.
However these first applications were rather calculations than designing. As the early computers were more related to what we now call calculators than those selfsustained computers aiding in lens design today.
Olympus introduced its own optical computational metrics by proprietary computer in mid-1973 and began the task of creating the retro-focus inverted 6E/5G optic of the XA, released in July of 1975. The design of that camera was evidently unorthodox, and the compromises that had to be made to fit a very sharp lens into a small space were formidable for that time (the question asked is, who else did the same thing, eliminating just about all of the sever forms of distortion and aberration, AND which wasn't a fold-out design?), though commonplace today (and often better) in many RFs.
I think at the time the designers designed a zoom lens based on intuitions and knowledge/witchcraft, then had to make calculations to verify/understand lens behaviour. With zoom lenses (zoom proper, lenses that perform at all focal lenghts) these calculations take a huge amount of time, considering the great number of lenses and the very great number of "positions" (zoom, focus, aperture), and computer helped make this calculation allowing the designers to figure out much faster at which focal length a problem might have appeared, modifying the design, calculating again etc. This "trial and error" procedure was certainly much faster with computers.
I used to think of the Pentax Super Takumar 35mm f/2.0 (with 67mm filter ring) as a rectilinear wide angle lens. Was I mistaken? Or is this really an attribute of that lens?
After it was stolen, I "replaced" it with the Olympus Zuiko 35mm f/2.0 lens. I wasn't thrilled that I was introduced to barrel distortion. So I never really took a liking to that lens, and I kept thinking I should get the shift lens... not at all for the shift feature... but because I had the impression it was bound to be rectilinear.
Later computing power improved to the point that they could support the programming necessary to become design tools.
Large cost reduction has also come from a reduced need for prototypes.
Very understandable about the advancement of zoom optics and what I really thought: that this computerization of design was largely gradual, emanating (seminally) from 'helping' to calculate and verify the manual efforts, then gradually segueing into actually BEING the 'person' who formulates the data.
Might I ask another question: I heard long ago (two decades) that varifocal lenses were actually sharper than zooms because they had to be refocused with each change in focal length. Has the quality of today's zooms negated that 'advantage'? - David Lyga
As a side notation, the best lenses as far as correction of distortion and of field curvature are usually lenses for "reproduction stand" normally to be found in 50mm and 100mm, often called "bellows" lenses because they don't have a focusing helicoid and must be used in conjunction with a bellows (they would normally focus at infinite, with their bellows).
In case is true that Parfocal lenses have less abberations, the ability with Varfocal lenses to focus with a long FL and to zoom then back to a short FL, and by this gaining higher focussing acuracy, might compensate in some cases for any better lab-performance of Parfocal lenses.