Aside from MacroParse, all the code is at the package root, and it is not amazingly well-organized. The chief difficulty to be overcome is that there are at least two interesting axes along which code might reasonably be split:
the area of theory to which it applies, or
when the code is needed (run-time, preparation-time, or both).
Aside from MacroParse, all the code is at the package root, and it is not amazingly well-organized. The chief difficulty to be overcome is that there are at least two interesting axes along which code might reasonably be split:
Probably some naming convention could help here.