Finite-Difference Notes

[feathern]

These are some potential issues that I'm making note of while working on the initial finite-difference documentation.

1. It looks as though when a nonuniform grid is specified through dr_input, there is no consistency check to ensure that radius(1) is equivalent to rmax. It's possible this could cause issues in other parts of the code where rmax is used explicitly.
2. We should also set dr_input to -1 by default, rather than 0 so that we're not doing an exact comparison of two real numbers. I worry about the logical check of IF (dr_input(n_r) .eq. 0.0) possibly failing in some instances otherwise.