I don't see it as it's been discussed so far. There is no such thing as a lens that's been designed for digital or not. There's some lens element coating technology that has been used for use with digital sensors because sensors have a different amount and kind of reflectivity but nothing more than that.
A lens (any lens) is either well corrected or it's not. On film a poor lens in that regard causes just as many problems as it does on digital. The main difference is that some of the colored fringes don't show up as much when printed as they do on screen but that's true with digital images as well. Same same.
The bits about in-camera correction are only a way of selling poorly designed lenses for more. It might save the user some time in editing if they care about such things but the user doesn't seem to benefit otherwise. I mean the lens in question gets sold for more than it would otherwise - meaning the user pays for it in the end.
So IMO a good lens design is a good lens design - period. It has nothing to do with digital or not. And there are plenty of WA/UWA lenses designed prior to the advent of digital which are superb and therefor support this thinking. And in-camera correction is only a hidden tax - whether or not you want to pay that tax or spend your time at the computer is for each individual to decide. Im my case I'd rather have the good design rather than paying higher prices for the poor design because the in-camera correction is not always the best quality and robs me of the ability both to get the most out of it and to see exactly what a lens can and can't do.
A lens is well designed if it produces the desired quality in the system it is designed for.
The 7-14 Panasonic zoom and 20/1.7 are not good film camera lenses, but that is irrelevant because they are not designed for nor are ever likely to be used on a film camera. On the digital cameras they were designed for they produce exceptional results while maintaining a small size and reasonable costs for their performance. The parameters that can't be corrected in software right now, such as astigmatism and coma have been corrected for extremely well, but distortion and CA have been left to 'float', because those things can be corrected.
The number of wide and super wide lenses designed before digital that have decent performance is extremely small, and cost a lot of money. The 38 Biogon in the Hassy SWC, and the other pre-digital Biogons for that matter are eclipsed by the performance of the 7-14. I have some of those Biogons, as well as the 7-14, and the value in the 7-14 is outstanding.
I also have 4 Leica 21mm lenses, including the 21 Summilux, and on the m43 sensor these lenses are useless. They are good lenses, but they were not designed to be part of this system, and that is telling.
The user doesn't 'pay' for this correction in processing time as it's done by the camera automatically. The user benefits in all sorts of ways.
Remember, there never were any absolutes in lens design. A 'good' lens was always part of a system. Apo repro lenses were often useless for general photography, as they were commonly designed for 1:1 reproductions, and only behaved wonderfully at their three specified wavelengths. A more generally useful lens might not have or need the superb resolution or distortion correction that those lenses had at 1:1 at the three primary colours, but would run rings around them performance wise for general photography. And what about 'floating elements'? The first creeping intrusion of providing lenses with corrections that they weren't capable of on their own in fixed configurations?
Lenses are never to be regarded 'in a vacuum', as objects in themselves unless they are to live out their lives on a shelf. Lenses are designed to serve a purpose, and that includes the whole system. If a lens doesn't serve that system properly and to your satisfaction, get another lens. But don't blame designers or manufacturers for giving us the best effective lens for this system. Right now that includes the possibility of software correction.
Mr. Kobayashi is not interested in this right now, so he is not doing it and the lenses he has designed to date do not have the contacts and firmware to interact with the camera in this way. He has however designed the lenses for the m43 sensor, which requires a degree of telecentricity which Leica and other rangefinder lenses lack for the most part. Therefore I fully expect higher performance in the corners from the 17/0.95 than from my 21 Summilux on the m43 cameras.
His lenses are excellent value, partly due to not constraining the size a lot. Leica will tell you that part of the cost of producing lenses for the (full frame) M9 and film Leicas comes from making sure the lenses are not too large. Remember, scaling the 25/0.95 up to full frame would result in a lens ~6" long, weighing 6 to 8 times as much and with a filter size of 105mm. Then we would have a 50mm/0.95 Nokton.
He has more in store for us, so let your imaginations run wild.
Henning