Today I got most of the way to expressing an equation for a Gaussian Newton optimizer to minimize error and produce a result. I had to re-learn Jacobian matrices and partial derivatives. It's been a looooong time since Diff Eq. Fun stuff though.

Interestingly, it looks like an Android phone can minimize error with two variables from 10 samples in about 1/4 second... in Java. Pretty sweet, as I only need to do this about twice/minute, and no problem running it in another thread. I thought I was going to have to bust out the NDK and implement the tough stuff in C++.