[NeXus-committee] obsolescence and evolution: NeXus needs a historic dimension

[Pete Jemian]

Let's test the assertion, that validation is sufficient for 
interoperability, against a data file with only the required NeXus 
structure.  One data field is presented with the requisite groups:

/:NeXus HDF5 file
   entry:NXentry
     data:NXdata
       data:NX_NUMBER
         @signal=1

Data files that validate against this structure satisfy the contract for 
interoperability of all NeXus files.  That contract is that the data 
file provides the data for default visualization.  In a NeXus data file, 
the data for default visualization is indicated by the unique attribute 
of signal=1 on a single data field within a NXdata group within a 
NXentry group at the root level of the HDF5 file.  Interoperability 
towards any other goal is not asserted by the required minimum structure.

Pete

On 7/16/2014 8:05 AM, Pete Jemian wrote:
>
>
> On 7/16/2014 7:13 AM, Joachim Wuttke wrote:
>> Anyway, formal validity is perhaps a necessary, but certainly not a
>> sufficient condition for interoperability.
>
> This is actually a goal for any NeXus application definition.  A NeXus
> data file that is valid against the relevant application definition
> meets the structural criteria declared sufficient for interoperability.
>
> Perhaps another measure of interoperability is the data content.  The
> data might be arranged in a NeXus data file with valid structure but
> useless content for the processing step (such as no valid range of data
> satisfying the Guinier approximation in small-angle scattering or no
> intensity peaks for powder diffraction analysis).  Standardization of
> data content or quality is not addressed by NeXus.
>
> Pete
>