Crystallography Online: the website of the International Union of Crystallography

Here are some questions and ideas about CIF/STAR. I have seen a few
comments with older list messages about whether CIF should remain STAR
compliant, or just become it's own format. The addition of MIME headers
breaks STAR compliance, because multi-line string marks can be masked by
the MIME boundaries, and binary CIF breaks STAR rules even more. SO, it
may n

I looked at the definition of GLOBAL_, and it seems easy to implement,
although I'm not sure how sensible the scoping rules are. It seems like
it might be useful at the beginning of a file for a few identifiers,
such as which dictionary is in use. The current data_global hack that
has been mentioned might be formalized a bit by having a specially named
GLOBAL_ data item that gives the name of one or more data blocks that
are intended to have global scope. That way, 'data_global' is not a
hard-wired name, and multiple globals could be listed in order of priority.

Does anybody have any interest in using the $<name> save frame
references? If not, maybe the reserved leading $ rule should go away?
Or, perhaps a more generalized definition of using a leading $ to
indicate a data-block name reference.

Has anyone considered nested save frames? Hierarchical data can be
useful, but CIF avoids even one level of save frame. This should not be
hard to implement, in the absence of the $ references.

Also, I was thinking that a set of data blocks written as save frames
could be an alternative to the 'data_global' hack. All of the data
blocks in the file could be contained within one parent data block as
save frames, and the parent block's data items would correspond to the
ones currently defined inside 'data_global'.

Of course, the usefulness and difficulty of implementation all depend on
the current set of CIF software. Perhaps CIF users would rather just
formalize the current data_global hack.

Joe Krahn