SMIRNOFF: Cut-off electrostatics not allowed

[mattwthompson]

Using cut-off electrostatics at all, and especially a hard cut-off, is scientifically dubious, but can be useful for testing purposes. PME, reaction field, and plausibly other methods each have their own quirks and special parameters. I don't think there's any method that's as simple and straightforward, which is useful for testing that systems of many sizes (single-molecule to condensed phase, solvated protein-ligand, etc.) evaluate their charges consistently. I think this was used in the InterMol paper.

Note that OpenMM does not support using a hard cut-off, but other engines (at least GROMACS and LAMMPS) do.