Originally posted by Blue Thanks Blue. Here is a
Flickr set with the remaining pages.
Some of the text is as informative as it is whimsical. Consider:
"[W]ithout giving priority to resolving power, MTF values and other numerical evaluations, they attain a level of correction in actual photo capture that remains in your mind."
This article makes plain the optical design goals, which I find to be admirable, and which certainly resulted in something special. I have in the past used this information to define "pixie dust" as the following:
Fully correcting the astigmatic difference in both meridional and sagittal subject planes, while allowing small amounts of field curvature to remain.
This means that sharpness is not as absolutely perfect as is possible. It also means that field curvature is a given, hence the decline in sharpness at wide apertures towards the periphery of the image. However what this gains is "a gentle transition from the solid subject to the out of focus portion" of the image. This reflects the priority given to rendering real-world dimensional objects and not the idealised focus plane.
This is what can be achieved by a designer who looks beyond mere numerical measurements into areas which are important despite being qualitative rather than quantitative. Certainly Jun Hirakawa did not achieve this with every lens he designed, but he might well have been constrained by different design priorities and budgets.