Hi Paul,

On Thu, 2011-12-01 at 14:24 +0000, Paul Emsley wrote:

> On 30/11/11 15:34, Judit Debreczeni wrote:
> >
> >>>  Coot, however, ignores such records entirely, so the chirality remains
> >>>  unrestrained, cannot be edited in the restraints editor or flipped by
> >>>  a keystroke. Using 0.7-pre-1 (revision 3792)  [with guile 1.8.7
> >>>  embedded] [with python 2.7.0 embedded].
> >>>
> >>>  Bug? Oversight? Feature?
> >>>
> >>>
> >>  Oversight.  I didn't think that anyone would be so contrary as to format
> >>  their chirality in such a way (it seems to me that they must have gone out
> >>  of their way to do so - I wonder why...)
> >>
> >
> >  Contrary? -- We are talking about the genuine and vanilla RCSB cif
> >  parser which I thought should be the gold standard?
> 
> It took me a little while to understand what you meant.  Yes, I suppose
> that this could well be the "RCSB blessed" way of handling such
> situations - and grade is merely following the rules.  I take it back.

The current STAR and CIF-family standards are in International Tables
volume G. I have not found any distinction made between:

# Form 1: Loop with exactly one packet
loop_
_name1
_name2
value1 value2

and:

# Form 2: <name> <value> pairs
_name1 value1
_name2 value2

As far as I can tell, these two forms are exactly equivalent, and
applications are free to choose whichever way they want to write the
data out. An mmCIF could be written using only form 1 for all data. This
might be a perverse choice from some points of view, but would not
violate the standard.

A conforming parser should handle both forms on input with identical
outcomes. Any parser that only handles one of them has implicitly
specified a dialect of (mm)CIF, and will be incompatible with software
whose developer has chosen to only handle the other form. An application
(like grade) that writes out mmCIF data then cannot satisfy both
dialects simultaneously. The best it could do is have an option to
control which dialect it writes out. It is far better that applications
that read mmCIF's do so in a standards-conformant way.

I don't think that this question has anything to do with the RCSB. I
can't see any reason why they should prefer either form 1 or form 2,
although someone else with more specific information may correct me
here :-)

> >>  Anyway, it's something that I should fix.  I'll add it for 0.7.

Yes, that is clearly the best thing to do. I hope that this makes things
clearer.

Regards,
Peter.

-- 
Peter Keller                                     Tel.: +44 (0)1223 353033
Global Phasing Ltd.,                             Fax.: +44 (0)1223 366889
Sheraton House,
Castle Park,
Cambridge CB3 0AX
United Kingdom

Reply via email to