Pam,
Actually what was the issue with the cif files with multiple phases/data sets? 
gsas2cif writes it out fine, I presume. Is it that the cif checkers can't 
handle the complexity? I guess my comment is that they'd better as there is 
going to be a lot more of this kind of thing in the future. Even mmCIF will 
need to adapt to this (that's a one data set only system). For the cifer's - 
are there fundamental design issues in cif that makes this a problem? Can cif 
adequately deal with the multiple histogram (x-ray, neutron, single crystal, 
powder & restraints all mixed together) capabilities of GSAS?
Pam, as you probably know, gsas2cif writes out template cif files which you can 
"edit once" and reuse. Would this help your problem?
Bob

________________________________

From: Whitfield, Pamela [mailto:[EMAIL PROTECTED]
Sent: Wed 7/12/2006 6:23 AM
To: [email protected]


Vincent and co
 
Unfortunately I don't yet have the software to do a real VCT data-collection 
(not sure many people do, hence 'VCT-type').  Until I can do it properly, the 
best I can do is chop the pattern up into pieces with different count-times 
(and sometimes step-size at high angles), and treat them as multiple datasets 
:-(  
Easy enough to deal with in the refinement but more than a couple gets clumsy 
to deal with in the CIF....
 
This is the first one of these files I've had to make up, and because I intend 
submitting to a IUCr journal I can't skip too many of the fields (which 
includes the data, calculated, reflections, etc, which do work BTW).  To cover 
all my bases I've put this through every CIF-editing/checking piece of software 
I can get my hands on.  Some of them give no errors whereas some light up like 
a Christmas tree, e.g. Platon.
 
Next week I have to make another file with resonant diffraction and neutron 
data thrown in with anisotropic broadening and some very complex occupational 
constraints - I have to say that after this experience I'm not looking forward 
to it, although that one will be going to Elsevier so maybe they won't miss a 
few terms!
 
Pam

________________________________

From: Favre-Nicolin Vincent [mailto:[EMAIL PROTECTED]
Sent: Wed 12/07/2006 5:30 AM
To: [email protected]



        Hi,

On Tuesday 11 July 2006 19:13, Whitfield, Pamela wrote:
> After spending over 2 days making up a single file, I'd like to hear some
> other opinions on the practical aspects of CIF files for structures from
> powder data.  This is partly a moan from trying to get a 11000 line file to
> pass the CheckCIF when all of the items it complains about are actually
> there from what I can tell.  Although optimizing data collection using
> VCT-type approaches is nice from a statistics point of view, it's absolute
> hell when it comes to creating the CIF file, and multiple phases just piles
> on the grief.  I almost wish I hadn't bothered with the internal standard.

  As for VCT, is there really a specific need to write everything ? The only
useful information is, for each point "2theta (or d or t), Iobs and
sigma(Iobs)".
  After all, the 'VCT' information is entirely included in
sqrt(Iobs)/sigma(Iobs), which can be constant or not, so why bother writing
the exact counting time for each point, even if the powderCIF dictionnary
allows it ?

        Vincent
--
Vincent Favre-Nicolin

CEA/Grenoble                 http://www-drfmc.ceng.cea.fr/
DRFMC/SP2M/Nano-structures et Rayonnement Synchrotron
17, rue des Martyrs
38054 Grenoble Cedex 9 - France

tél: (+33) 4 38 78 95 40                fax: (+33) 4 38 78 51 97

--
Vincent Favre-Nicolin
Université Joseph Fourier
http://v.favrenicolin.free.fr <http://v.favrenicolin.free.fr/> 
ObjCryst & Fox : http://objcryst.sourceforge.net 
<http://objcryst.sourceforge.net/> 




Reply via email to