[Nexus] Re: Geometric information in NeXus

[Bull, MJ (Martyn)]

Hi,
 
I too have been thinking about the content of the NXSample group and how to
convey more information about the sample. 
 
For example, the definition of the unit cell is currently defined as a 6
element array, but there is no indication of the order in which elements
should be stored. Is it a, b,c, alpha, beta, gamma? Or alpha, beta, gamma,
a, b, c? Given that everything can be stored sensibly using XML and the HDF
format, it would be better to define each of these quantities separately:
 
<NXsample name="whatanicesample">
<unit_cell> 
    <length_a units="Angstrom" type="NX_FLOAT32"> {Unit cell a-axis length}?
</length_a> 
    <length_b units="Angstrom" type="NX_FLOAT32"> {Unit cell b-axis
length}?</length_b> 
    <length_c units="Angstrom" type="NX_FLOAT32"> {Unit cell c-axis
length}?</length_c> 
    <length_alpha units="degrees" type="NX_FLOAT32"> {Unit cell alpha
angle}?</length_alpha> 
    <length_beta units="degrees" type="NX_FLOAT32"> {Unit cell beta
angle}?</length_beta> 
    <length_gamma units="degrees" type="NX_FLOAT32"> {Unit cell gamma
angle}? </length_gamma> 
</unit_cell>
</NXsample>
 
Crystallographers have already gone to great trouble to define a huge number
of sample specific properties, including the geometric location of sample
faces, in quite some detail. 
 
The current CIF dictionaries ( http://www.iucr.org/cif/
<http://www.iucr.org/cif/> ) are exceedingly comprehensive, even dealing
with storing raw and processed data, and I believe there is a great deal of
knowledge and wisdom that can be ported over directly into NeXus. The CIF
structures look as if they will port to XML relatively painlessly, and
additionally would make conversion of NeXuS files to and from CIF a breeze.
 
My suggestion then is that NeXuS should aim to incorporate the CIF
definitions, or certainly a large subset of them, into the NXSample class
definitions.
 
Any thoughts?
 
Regards, 
 
Martyn Bull
 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.nexusformat.org/pipermail/nexus/attachments/20021127/2fca2f09/attachment.html