[ccp4bb] Proper way to handle multiple conformations of ions/waters

[Paul Paukstelis]

We've been working on a high resolution (1.0 A) DNA structure that has 
several coordinated magnesium ions that have complete octahedral 
geometry via water or phosphate oxygens but are clearly in multiple 
conformations (which is related to the multiple conformations of the 
coordinating phosphates). Is there a _proper_ way to handle this 
scenario? Should the ions/waters be modeled as independent atoms with 
fractional occupancy, or should they be identified as atoms with 
alternate conformations? The latter case can get complicated for the 
waters as the overlap can make it quite ambiguous in some cases as to 
which two water pairs should be associated as the alternate conformers. 
Perhaps it is a minor point, but I thought there might be some kind of 
convention in the small molecule community where this is sure to crop up 
more often.

Thanks in advance,
--paul