Thanks for your explanation, Mike.

However, I still do not understand—and the lack of understanding is my own problem—why, if a program can "see", parse, understand a curve, why it doesn't see the tension and bias values of bounding control points and just use them when a gylph is "decompiled into a simpler vector form.

In any event, it's a chore to eliminate unnecessary control points!

My Best,

Gary