we keep two different branches of this code,
which share common subroutines.
We try to minimize the code files in the
different branches, and keep as much overlap as possbile in
the common subroutines.  The reason we have two different
branches is that the data structures get messy when
computing correlation functions such as S(q,t)

1) Sq_Average_FFT:
   computes average structure factor, either charge density or number density
   structure factor.  For charge density, can be split into cation, anion, and
   total contribution.  For number density, can be split into partial atomic
   structure factors.  This code returns 3D structure factors, S(qvec), which
   may be used for analysis of anisotropic systems, or subsequently averaged
   as a function of |qvec| using perl scripts.
   

2) Sq_Ct_FFT:
   In addition to computing averaged structure factors, this branch computes correlation
   functions, e.g. S(q,t) for structure factors.  In this code we assume isotropic system
   so that we don't have correlation functions for all qvecs, and thus we internally average
   over qvecs to output S(|qvec|,t)  

Both of these branches share code in
"source_shared" directory