my understanding is that af could be different among several copies of the same lens? so the af accuracy of one lens wouldn't necessarily represent how accurately they all autofocused... the af of each lens has to be checked and possibly trimmed on the body that it's going to be used on... that happens with mirrors in cameras.
granted, lenses are like snowflakes, the sharpness of one fa28 certainly isn't going to represent the sharpness of all fa28s, but sharpness problems generally show up as decentering.
here is the takinami rez tests:
http://www.takinami.com/yoshihiko/photo/lens_test/pentax_28-30.html
his copy of the fa28 didn't show radical differences at f/2.8, if anything it's stronger on the sides than most of the other 28s... things change as the lenses stop down, tho.