This is another thread that has disappeared. I'll recap the gist of it. Paul suggested that GF 4 could enhance sharpening properties. Jonathan suggested Focus Magic was a better deal.
I've tried both. I bought GF a few years ago, partly because Michael R. recommended it and partly because I'd read of this fascinating theory of 'fractals' which describes a similarity of structure in the small and the large. (Ie. a tree trunk and branches has a similar structure to a branch and smaller offshoots, which in turn has a similar structure to the veins in a leaf. Find the recurring structure and you can do marvels.)
It sounds great in theory and the name'Genuine Fractals' leads one to believe this is an implementation of the theory. Maybe it is, to some degree. How would I know? All I can do is look at the results, and when I do, I get a sense that I'm looking at very, very marginal improvements. Differences between GF and PS bicubic are of the highest order of pixel peeping. They are there, of course, at 400% magnification on the monitor. Or, if one takes a low resolution image or a small crop, interpolates it with GF, sharpens it in PS, repeats the process a couple of times, one can end up with a fair representation of an impressionist art work, which might be quite interesting. One could add this process to the 'artistic filters' section, along with poster edges and pallette knife.
Focus magic does a similarly botched job, but at small print sizes (8x12" and perhaps larger) the images really do look sharper. Do a bit of pixel peeping, however, and the enlarged crops look just awful; black dots, black lines and misshapen forms. However, one thing in favour of Focus Magic, it appears to be able to create an 'impression' of greater sharpness without the increase in grain and color noise that the normal 'unsharp mask' produces.
For this reason I tend to favour it in preference to GF as regards sharpening methods.
As a general comment, we seem to be scraping the bottom of the barrel with regard to sharpening programs. As far as I can see, there are no breakthroughs yet.