Matt Thompson

Related: * https://github.com/ParmEd/ParmEd/pull/1082 * https://open-forcefield-toolkit.readthedocs.io/en/0.6.0/smirnoff.html

You may wish to consider residue templates instead of force fields for cases like this. For small, custom force fields that are only valid for one molecule - and likely...

It seems that in [general](https://stackoverflow.com/a/4005412/4248961) `__hash__` should return the same value for objects that are equal, so that's why I was thinking about moving the logic in between functions. I...

It seems imperfect but I think the best approach is above: override both `__eq__` and `__hash__` but at some point when accessing the high-level tracking, calling `IndexedSet(foo)` will remove any...

> have doubts about how "modular/generalizable" this sort of thing could be. It's probably a fool's errand to try to handle this generally, but I think it may be useful...

One way this could take place is with some arguments that specify parts we do and don't want to be updated, i.e. ```python3 # if you wanted to use positions...

Can't get it to work with `unyt` at the moment but here's the basics: ```python3 >> expr = sympy.sympify('sigma*epsilon') >>> params = {sympy.symbols('sigma'): 1, sympy.symbols('epsilon'): 2} >>> expr.evalf(subs=x) #@{**x, **y})...

I wonder if we can get `sympy` to intelligently compare functions that are similar but not quite exact, like the case of harmonic bond potentials sometimes having a 1/2 coefficient...

> Bonds and Angles can be rigid, so maybe this property be in Connection? But Dihedrals are generally not rigid. I think we should just start with this in `Bond`...

I think abstracting the and readers into classes is a good idea. It's becoming increasingly clear that the value of this package will be how well it converts to different...