Request for discussion: dual-version CIF conformance

Bollinger, John C John.Bollinger at STJUDE.ORG
Tue Jul 18 21:15:31 BST 2017


Dear Colleagues,

I have lately been scrutinizing the CIF 1.1 specifications in ITvG and carefully considering a variety of unusual cases.  One of the less obscure questions I have examined in the course of this process is whether a CIF can simultaneously comply with both CIF 1.1 and CIF 2.0.  Up to now I have supposed that CIF 2.0's requirement for a leading version-identifying comment prevents a file conforming to CIF 2.0 from also conforming to CIF 1.1, but I am having trouble supporting that with the letter of the CIF 1.1 specifications.  In particular,

- momentarily disregarding the question of version-identification comments, there is, in fact, a common subset of CIF 1.1 and CIF 2.0 syntaxes on those systems where local text conventions line up with CIF 2.0's encoding and line-termination requirements.
- CIF 1.1 does not require a version-identification comment to appear at all.
- Although CIF 1.1 defines a format for a version comment that claims conformance with CIF 1.1, it does not specify that files claiming conformance to a different version thereby fail to conform to CIF 1.1.
- In fact, the text seems to suppose that a file may conform to multiple CIF versions, for it remarks that "A CIF may be considered compliant against the most recent version for which it in practice it satisfies all syntactic and content rules [...]," and it characterizes the version comment as embodying only a creation-time conformance assertion.
- If it were necessary, a version comment specifying CIF 2.0 could be interpreted as a plain comment in CIF 1.1.

Does anyone want to mount a counterargument?  Or am I maybe late to this party?

As a concrete example, consider a machine whose text conventions rely on an ASCII-compatible encoding and use a line feed character as a line terminator (maybe some flavor of Linux).  If the following text is recorded on that machine in a file complying with its text conventions, then does the file simultaneously conform with both CIF 1.1 and CIF 2.0?

#\#CIF_2.0
data_example
_item value
### end of CIF ###

Note that this is a different question from whether a processor that supports only CIF 1.1 should assume that it can handle such a file.


Regards,

John

--
John C. Bollinger, Ph.D.
Computing and X-Ray Scientist
Department of Structural Biology
St. Jude Children's Research Hospital
John.Bollinger at StJude.org<mailto:John.Bollinger at StJude.org>
(901) 595-3166 [office]
www.stjude.org<http://www.stjude.org/>


________________________________

Email Disclaimer: www.stjude.org/emaildisclaimer
Consultation Disclaimer: www.stjude.org/consultationdisclaimer
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.iucr.org/pipermail/comcifs/attachments/20170718/53674de6/attachment.html>


More information about the comcifs mailing list