I'm keeping my STL files in a git repository and noticed that a few of these models appear to be unusually large. As an example TST CLM DGN2r.stl is 118M and TST WL-PT3 DNG1.stl is 96M. II wonder if there are more data in the STL files distributed than is needed to describe the surface. I don't mind the size but when using hosted git services, such as GitHub, 100M files in the repo gets you into having to do things a little differently. I will try to use Fusion to see if I can reduce the complexity without reducing the resolution but I expect that to be a fairly large learning hurdle.