I know that using doxygen one can create call graphs of blocks in fortran code. These call graphs show the relationships between caller and called blocks. Doxygen in my experience is organized in a modular fashion. In Fortran this means each subroutine, function or module can have documentation including what variables are in it and how it relates (via call graph) to other functions, modules and subroutines.
My question is, can Doxygen handle code at the variable level? Meaning I would like to generate call graphs or flow charts of actual variables within a subroutine, such that the interrelationships between the variables can be drawn?
If Doxygen cannot do this, what can? I would rather have a systematic and efficient way to map out software rather than debugging things and having to always commit them to human memory...