Hi Marco,
your trick seems to me potentially 'dangerous', though: if you assign an altloc identifier to a whole residue, its atoms won't 'see' any other atom that does have a different altloc identifier because it is like they don't exist at the same time - is this right?
that's true. Although, I don't know what is actually correct... If the occupancy of the whole ion is less than 1.0 that means there are situations (unit cells) where this ion is simply absent at that particular location and therefore there is nothing to "see" for the surrounding atoms.
If so, in principle, a sulfate labelled with an altloc identifier 'A' should not see neighbouring atoms which are labelled as altloc 'B', which may well be the whole side chain of a residue in close contact, modelled in double conformation. That way, no restraints will be applied on close non bonded atoms, etc...
Correct. Anyway, I was just pointing out an alternative option for group occupancy refinement, and thanks for emphasizing the implications for geometry restraints. Pavel.