Matt Thompson

what is an expression is written in a different order? i.e. r12 vs r6 term being first, would that break this?

Please add any other good starting points @mosdef-hub/mosdef-maintainers and consider chipping away at these tasks @mosdef-hub/mosdef-contributors

Yes, it's a major motivation of this package and something we necessarily intend to support. Not quite sure _how_ to do it, as most of the `sympy` stuff should be...

I mean these are typically going to be 1-D arrays of length few hundred to maybe a few thousand. This would be 3 (`r`, `U(r)`, `F(r)`) of these arrays for...

I'm pretty sure `ParmEd` stores the cross-interactions only if it is told to, and can generate the cross-interactions at the point of the writer, if necessary. Or at least this...

Are there any engines that require only `r` and `U(r)` but not `F(r)`? Because a starting point would be easier to put together if we could get away with that.

Could you elaborate on the last point? Is it too risky to use a finite difference to estimate forces? If this is problematic, we could just require that any tabulated...

Maybe we should separate out these ideas into discrete modules/functions. So the tabulated potential class would need to be passed in some minimum amount of information to be reasonable (at...

PySIMM has a much smaller scope than this package but at some point conversions to/from it could be supported https://github.com/polysimtools/pysimm

We discussed this briefly in person, but just to put some general ideas to paper - * Supporting new potentials is a task for both the XML schema and some...