Well, if you could make a perfect point source, you would be able to image it with a Standard (conic asphere) surface with just R and k. BUT...

All real sources are extended, and indeed imaging systems are always a fairly wide field, which makes ‘perfect’ imaging of a single point irrelevant. You need the performance of the system averaged over all fields and wavelengths. That’s why we evolve ever more complex designs.

All this ‘perfect optics’ stuff is great on paper but it’s generally of little practical value. Plus, of course, diffraction limits the performance anyway so you could never resolve the ‘perfect’ image point. The image is always smeared over a Point Spread Function.

• Mark