I've been wondering for a while what determines "quality" in GP-software.

I've previously used the proprietary product Discipulus, but didn't like it due to it's very closed nature.

I'm wondering if there are differences in how different software produce their results, and if the difference will matter.
Could there be some solutions that either ECJ or Discipulus (or other product) will find, while the other never will?

I've been using ecj for quite a while btw, and think it's great stuff.
Big thanks to SeanĀ  (And everyone else involved) for all his efforts :-)