I don't find the semantics question particularly compelling, or whatever aside the ed staff had to birddog, but I do think the topic has a LOT of merit.
I'm glad Eric posted what he did. For years the dogma has been "set happens" as if it is something inevitable, something inherent and not introduced. Fact is that set is CAUSED. If you "overbuild" a bow enough, it will not change is profile during construction or use. While I haven't had a selfbow project turn out this way, I'm convinced it can be done. Badger pointed out a very useful technique years ago. Check your draw weight at various intervals of tiller, say 20", 22", 24", 26". Generally you'll find you get the same draw weight at say 20" until you get to about 26". At this point you've exceeded the elastic limits of the material and you'll not ever get the same reading at 20". This obviously assuming no wood removal, no change in moisture content, anything. What you'll find is that you once you've broken past the elastic threshold, you'll never again achieve the efficiency potential once in the project. Bow building for me has become a dance with the stave, to find and not cross this threshold, or at least not by too much. It's a whole new way of looking at bow building, although not new. Jim Fetrow preached the same thing 10 years ago, and was soundly ridiculed for it. A bow can be made with no set, but not by the methods usually advocated, and taken for granted. And, like Eric said, when it comes to bow wood whether species matters remains a question, but specimen definately matters.
I've actually had r/d bows increase reflex as belly wood is removed, where I had a lot of "spring back" out of the glue up. The stronger belly was actually holding the relfex down. I assume a similar phenomenon is taking place on "zero" set stave bows. As belly wood is removed tension wood on the back is offseting the crushed belly wood.