Solvent padding description about openfe

[mikemhenry]

Thanks for this!

If I understand this correctly, the first thing you are asking is for us to update the documentation. Do you have a suggested doc string? At first I was going to summarize the openmm doc and then include a link to it, but it sounds like even the openmm doc isn't great, so do you have a suggestion? Or, what would you like to see mentioned?

[mattwthompson]

Yeah sorry this is all over the place; it'd be much easier if it was an OpenFE or OpenMM documentation issue and not a mixture of both

The best suggestion I've thought of would be a dressed-up version of

"""
Minimum distance from any solute atoms to the solvent box edge.

This argument is passed directly to OpenMM's addSolvent (link), which
typically produces boxes of size twice the padding distance plus the approximate
solute size, but for some small solutes may treat the solute as having zero size.
"""

This would communicate the quirk I'm observing in a way that squares the responsibility with OpenMM but doesn't assign blame. I think paraphrasing the OpenMM docs risks this docstrings falling out of sync each other, whereas just a link wouldn't be as useful to the OpenFE user. (If the behavior changes in OpenMM, this is a different issue altogether). That covers everything I have been thinking of, the most important thing being that it describes what happens

[IAlibay]

Thanks @mattwthompson. We could/should add it to this PR: #673, unfortunately it didn't make it into 1.0, but the ability to use the wider set of inputs from addSolvent alongside better docstring would be great.