[Nexus] problems with current stable of nxbrowse and nxdir

[Mark Koennecke]

Eugen,


On 09/10/2013 10:54 AM, Wintersberger, Eugen wrote:
> Hi folks
>    I got some reports that nxbrowse and nxdir do not work with the Nexus
> files we produce here at DESY. I can confirm this on my Debian system
> where libnexus0 is installed in version 4.2.1.
>
> I suspect that this still has a problem that we are using variable
> length strings and thus cannot read attribute data correctly. I have
> attached a file for testing. By the way can someone may have a look at
> the file if we got it at least a bit right ;)
>
> @Benjamin - this file does not collect non-standard fields into
> NXcollection as we discussed in Warwick. However, our new setups do. I
> am still working on the draft for a kind of 'best-practice' paper and I
> hope I can submit it to this mailing list this week.
I had a look at your file.

You got many things right, more then most synchrotron people as fas as I 
can see.
Some comments:

1) In NXmonochromator you appear to be defining a double crystal 
monochromator. NeXus does not
     yet know how to do this. Perhaps this is worth a proposal?
2) In NXsample I found many funny named fields. Units of mm indicate 
that they might be translations.
     I am feeling lost: either use standard NeXus names or at least use 
the vector, type, depends_on etc stuff
     to annotate what this is doing.
3) There seem to be many more unknown names.
4) I also wonder if there should not be more then one NXdata; one for 
each detector. of which there appear to be
    10: channel1-10 and detector proper.

Have a nice weekend,

             Mark


> regards
>    Eugen
>
>
>
> _______________________________________________
> NeXus mailing list
> NeXus at nexusformat.org
> http://lists.nexusformat.org/mailman/listinfo/nexus