Yeah, so, at DigiPen, we had a class in Advanced Animation Techniques, one of which was "hierarchial models and animations", which involved treating a model as composed of "joints", each with its own local translation/rotation matrix. Our first project involved implementing this, which I attempted to do with DirectX. Apparently, they had a very similar but differently named idea called "skinned meshes", where a model is composed of "bones", each with its own matrix. I spent a lot of time and code effort translating between my two interpretations of the concept. For some reason, it didn't hit me that the two were the same until just recently. How I make it out of bed each day, I'll never know...