`filewriter`: `isis_vms_compat`

[Tom-Willemsen]

Mantid currently loads some information from the isis_vms_compat group.

Primarily this is UDET and SPEC, information which doesn't appear elsewhere in the .nxs file, but if the isis_vms_compat group exists at all then mantid apparently needs all of:

['IRPB', 'NSP1', 'RRPB', 'RSPB', 'SPB', 'SPEC', 'UDET']

Documentation for what some of the old VMS names mean is available at:

\\isis\shares\ISIS_Experiment_Controls\data for tickets\ds_76

Potential differences from existing files

• Extent to which we emulate the isis_vms_compat group to be discussed with Mantid team and other relevant consumers.

[martyngigg]

A summary of recent findings regarding isis_vms_compat:

SPB, RSPB

Relates to sample geometry.

The code already checks for alternate locations in the standard NeXus locations - the if just needs switching around.

IRPB, RRPB

Relates to run details: proton charge, source frequency.

The code again checks in alternate, more standard, locations - the if just needs switching around.

NSP1, UDET, SPEC

Relates to the spectrum-detector mapping.

The code is a bit misleading as there is a fallback but it only applies to Muon Nexus V2 files and does not work for neutron NeXus files.

This is an area where changes in Mantid and in the structure of the file would be required. Currently detector_1/spectrum_index lists the spectrum numbers but the UDET field has no corresponding entry elsewhere. It looks as if the NeXus component NXdetector can contain a detector_number field that looks suitable for this purpose.

[Tom-Willemsen]

Need to check with other analysis programs e.g. GudRun for SANDALS-2, to what extent they need isis_vms_compat. My understanding is they currently use .raw files but also have ability to use .nxs.

[martyngigg]

I remember checking with Tristan during the MNeudata requirements gathering and he said:

"Gudrun actually reads both. On SANDALS they use NeXuS files as standard now - on NIMROD we are still in the RAW file era, but there isn't really a reason why we couldn't easily shift to using NeXuS.

From the perspective of MNeuData I would agree and say that RAW files shouldn't be an option, for what it's worth. I have my reservations with HDF5, sure, but I appreciate that RAW is not a sustainable format moving forward."

[Tom-Willemsen]

Tris has confirmed that GudRun does not use information from isis_vms_compat.