I agree that using this fix to remove rotation regularly is unphysical and can cause issues. Still, when applied to some reduced group of atoms (i.e. not the whole system but, for example, a molecule) this fix does not remove only momentum, it also removes vibrations. However, it is clear to me now that this was never the intention of this fix, thank you for that clarification.
I’m beginning to think that maybe presenting this as modification to fix momentum was not a good idea. I still think there is some value to adding a compute for the equilibrium coordinates of the molecule in the reference frame that rotates with the molecule, or maybe the omega of this rotating frame.
I don’t know if there is even demand for this kind of result. I do, at least. Like I said most of the code is already done, it’s just a question of how to package it. What do you think, Axel?