>But doesn't this test tell you only about relative sharpness/corner
>performance of a lens? It tells you nothing about geometry, color purity,
>ultimate sharpness, freedom from coma, flare resistance, and other factors
>that play important roles in determining image quality.

Yes - but it is a reasonable assumption with all but the original Nikkor
set that flare is not a problem (and can be checked in the finder with a
bare light source with a dark background); and all single FL Nikkors 60mm
and up have no linear distortion (checkable in an F3 finder - no need for
film) and most others have moderate distortion; Nikkor color rendering is
consistent (and can be checked on the few non-Nikkors that passed the
sharpness checks, using slide film); contrast variation (rare for Nikkors)
is often detectable in the sharpness check, or in use later; ultimate
sharpness is usually hit around f5.6 in the center for practical purposes,
and it is reasonably consistent across the Nikon lens line (the exceptions
occasionally cabolix the comparisons, but rarely); coma shows in the
corners in the sharpness checks. The sharpness checks establish the basic
needs: that the lens is sharp enough for use at the apertures I want to
use the lens at, and that the lens has no obvious misalignments. After
that, the subtleties can be explored (and probably will be, through use
and gathered experience). If a lens is not sharp enough, stellar
performance in other areas is useless (the 28mm f2.8 AI [not-AIS] is
unusually flare resistent, but the corner performance knocks it out as a
lens I would want to use.).