[chimera-dev] case-dependent chain IDs

[Eric Pettersen]

Hi all,

David Konerding has pointed out to me that treating chain IDs in a 
case-independent manner is becoming a problem with the very big 
structures that are now being deposited.  Both 1fnt and 1ml5 have chain 
IDs that differ only in case.

Right now PDBio uppercases chain IDs, so there is no possibility for 
the rest of Chimera treating them correctly.  I guess the question is 
what do we want to do?  We could just stop PDBio uppercasing them.  
This mostly works, but has the drawback that chain IDs in atom specs 
now have to be typed in the correct case, which is inconvenient 
compared to the way it is now where you can just type in all lowercase 
(which is perfectly fine for 99.99% of structures where there are < 26 
chains and the lowercase chain ID is unambiguous).

My thinking is that there could be a model-level function that declares 
whether chain IDs are case-sensitive for that model, and the atom-spec 
code could query that function do decide how to test chain IDs on a 
per-model basis.  Other than this, I am stumped for a good solution 
other than making people type the correct case for chain IDs.

Opinions?

--Eric