Periodic splines?

[tobydriscoll]

Do we want to use periodic splines? Why not specify them using trig interpolants? I'm concerned about the limitation to fourth order accuracy for smooth boundaries.

Perhaps (at the risk of complexity) there needs to be a periodic-closed-curve class with subclassed implementations.

[ehkropf]

The periodic splines are there because I was using them for various experiments. I've found them useful for being able to specify arbitrary smooth-ish boundaries. They probably should be done using trig interpolates; the current code was based on some older stuff I used trying to save time.

Any parameterised closed curve should probably be considered periodic. It makes sense to me to create a periodic/parameterised subclass of closed curve.