Hi Steve, I'm glad that it worked as well as it did!

But that's actually a good example of the kind of thing that you should not really expect to convert well.

The main problem there is that your file there has a tremendous number of little facets in it that are essentially trying to replicate curved surface pieces.

This type of conversion is meant for objects that are actually intended to be faceted, stuff like the polyhedron mentioned earlier, or things like diamonds which are supposed to have all flat faces in them.

For areas that are supposed to be smooth surfaces, in a NURBS type object those areas are intended to be created out of large smooth spline patches with one big surface for a broad smooth area, not a whole ton of little flat facets.

When you have such a large number of little facets you're basically replicating a polygon mesh type object structure in NURBS, but there is more overhead for the trim boundaries in a NURBS object structure than there is in a polygon mesh structure - for example the trim boundaries of every NURBS surface are constructed of curve objects, not just a single implicit line between vertices and so they're much more generalized in structure (with additional overhead), and on top of that there's both a UV parameter space curve in the 2D space of the surface as well as a 3D edge curve.

The NURBS structure works well when it's used as intended with larger areas of the model being made up of just a single broad curved spline surface.

You probably know all this already, but I thought I'd better explain this in some more detail since someone looking at your post may think that's the kind of conversion that is supposed to work - I just want to make it clear that it's not - I don't expect for a high poly count object like that to convert well into a NURBS object using a separate little trimmed planar surface for every single polygon.

It does make for a great torture test though! :)

- Michael