In marketing it's a standing joke that anything can be proven with a Powerpoint graph. In politics anything can be proven by a poll. And in equipment forums, anything can be proven with some web search. The whole idea is
to be so nonspecific with your parameters that everything is plastic, and can be moulded any way one wishes.
And over and over again in questions just like the one at hand, "sharpness" can mean any number of things. A good sample of an ancient 90 Angulon can be deadly sharp in the middle but pretty mushy near the margins if
movements are involved. Referring to an Apo Sironar S is meaningless in this conversation because it wouldn't
even cover 4x5 if it was made in a 90. Same concept per Nikon: the 4.5 has a larger image circle than the f/8, which might or might not determine which is sharper for a given application. In architectural work it might make a big difference. All kinds of variables here. And nobody has even brought up analogous lenses adjacent to 90,
like the 80 or 110 Super-Symmar with an aspheric elements. It's not like working with a traditional 35mm lens
where you just point it a certain direction. View camera lenses have different performance standards with various kinds of swings, tilts, rise, etc, plus the regular f-stop related issues. Determining what is appropriate for your personal application is often best learned from people using such lenses in analogous applications, not by
crunching relatively meaningless numbers like the pixel-counters are addicted to.