renamed PDEs

[bmcdanie]

sync up fokkerplanck PDEs in C++ with renames to the MATLAB versions.

fokkerplanck1_4p1a -> fokkerplanck1_pitch_E(1) ,
fokkerplanck1_4p1b -> fokkerplanck1_pitch_E(2) ,
fokkerplanck1_4p2-> fokkerplanck1_pitch_C

[hbrunie]

@bmcdanie @dlg0 @lwonnell There is a difference between the description in the comment and the actual code I think.
The description on top shows:
df/dt == d/dz ( (1-z^2) df/dz ) code
But the description in the Matlab code shows df/dt == -d/dz ( (1-z^2)f. Notice the sign is different and f is not derived.
But in code the description is different. Same file but different comment. They should tell the same story I think. Which one is correct?

[hbrunie]

Here is the branch corresponding to the changes I am making to solve this issue: branch

[dlg0]

fokkerplanck1_4p1a -> fokkerplanck1_pitch_E (case=1) 
fokkerplanck1_4p1b -> fokkerplanck1_pitch_E (case=2) ** this is a good adaptivity test case
fokkerplanck1_4p2-> fokkerplanck1_pitch_C
fokkerplanck1_4p3-> fokkerplanck1_pitch_R

It looks like the C++ does not have _4p1b so we will need to add it. It is a very minor variant of _4p1a and I'd like to add the case ability as we have in the matlab as a command line option in the C++, so this seems a good opportunity to do that?

@bmcdanie @hbrunie

[hbrunie]

Yes I'm on it. The pr I'm building will contain a new option to choose the pde case (1 by default). So on long term for each pde we will be able to add new cases.
But @dlg0 you did not answer about the difference between the PDE in matlab and the one in C++.
The description on top shows:
df/dt == d/dz ( (1-z^2) df/dz ) code, but the description in the Matlab code shows df/dt == -d/dz ( (1-z^2)f.
Which one is the right one?

[dlg0]

It looks to me like in src/pde/pde_fokkerplanck1_4p1a.hpp the actual PDE coding matches the matlab, but the comment is a copy and paste error - so the matlab comment is correct, the C++ comment is incorrect. You can confirm this by noting only one grad operator, whereas anything of the form d/dz df/dz would have two.