List of potential things to be refactored

[Pbellive]

Hi All,
jInv is in pretty good shape at the moment but there are still many potential areas that could be improved. Thinking ahead to a possible hackathon next week, here are some things to look at:

• Generalizing the globalToLocal interpolation to better handle inversions where a single forward problem has more than one parameter type, especially where a single inversion model parameter might effect multiple physical parameters of a forward problem.
• Better document the forward problem API so we're clear exactly what new forward problems need to be fully compatible with jInv. This goes hand in hand with cleaning up getDataParallel.jl
• Rethink implementation of supported regularization methods, with a view to clarifying input and adding flexibility
• Clarify when functions modify their input arguments.
  • Stop explicitly returning arguments that are modified in place.
  • Add ! to function names where inputs are modified? This might be a pain to add but I think it would
    improve the clarity of the code significantly.
• Update prepareMesh2Mesh so calling signature is consistent with prepareMesh2MeshOT from JOcTree. Consider using optional keyword arguments to replace overloading.
• Think more generally about where optional keyword arguments might be better than overloading?