preinflation and calculating normals tip

[David Dufresne]

Just in case anyone else has a similar issue, I've found that I get better normal values along the control curve if I preinflate, and then calculate the normals.  I was doing it in the other order originally, and it made a big difference when I fixed it.