Should splines be part of smoothed FD?

[pavelkomarov]

Splines are piecewise functions and have derivative values at the sample points, which we can get with spline.derivative()(points), so should we be finite differencing on top of a spline? Does this method belong in the smoothed finite difference module? Is there a benefit to doing FD instead of using the inbuilt derivative functionality?

[pavelkomarov]

In the paper I've put the first five methods of smoothed FD as a section and put splines with the sliding window polynomial fit and Savitzky-Golay as polynomial-based methods.