Here's a to-do list of updates/additions for the next major version of the package...

• Implement continuous-part kriging into all of kriging classes. (This approach doesn't force the kriging solution to converge to the measured value at a measurement point, so this is a way of dealing with error.)

• I've noticed some problems in the statistics calculations (specifically statistics coming out as NaNs); need to fix those problems.

• I discovered a problem with the variogram estimation scheme while talking to one user via email. Currently, the code doesn't enforce bounds on variogram model parameter estimation. I certain cases, with certain datasets, the automatic variogram model estimation can yield parameters that give nonsensical variograms. So, realistic bounds on the variogram model parameter estimates should be put in place.

• Also, there's room for improvement and optimization in the variogram model parameter estimation scheme.

• I have some ideas for optimizing the solution of the kriging system even more (specifically, trying out LU decomposition of the matrix system). Need to explore this idea.

• The interpolation scheme for the scalar grid drift currently utilizes a python-level loop. Definitely room for optimization here. When I tried using scipy interpolation schemes before, I wasn't pleased with the results; however, I think there's more experimentation with those tools that should be done.

• Use sphinx to make nice web-based docs for the package.

• Look into potential numerical datatype problems.