Document electrostatics 1-4 scaling factor changing from 0.833333 to 0.8333333333 #120
Description
Is your feature request related to a problem? Please describe.
Sometime between releases 1.2.0 and 1.3.0, the 1-4 electrostatics scaling factor was updated with a few extra digits from 0.833333 to 0.8333333333. This value is roughly inherited from Amber family force fields which use 1/1.2 without a particular intent to round. IIUC this change was never retroactively applied to force fields but simply applied during iterations and between releases/refits. (In other words force fields fit to 0.833333 have 0.833333 encoded, force fields fit with 0.8333333333 have 0.8333333333 encoded, never shall they cross.) I'm not sure the error from rounding at six digits affects any results, but I'm pretty sure rounding at 10 digits produces an error that is a complete wash relative to other errors in force fields.
Describe the solution you'd like
Publicly document when and why this change happened, optionally with reassurances that it's probably a marginal change.
Describe alternatives you've considered
Just say it myself elsewhere with a "trust me" attitude.
Additional context
I couldn't find this documented publicly (searched this). There is some chatter on Slack from July 2020 (search "It uses 1/1.2." to get in the middle of it) but I can't link to that in public docs.