Don't know if this will help: https://github.com/Starlink/perl-Astro-FITS-HdrTrans/blob/master/lib/Astro/FITS/HdrTrans/WFCAM.pm for the header definitions. I can't find the file ICD just now.

I'm assuming you have FITS files rather than the original HDS containers.

I have FITS files from CASU. Are you familiar with these? I have a few questions:

• What do the numbers in the *_conf.fit files represent? They are of order 100. Are they percentages? (If so, why don't they write a BSCALE so that they would actually be read in as a fraction?!)
• Very bright stars have holes in the middle, and this isn't clearly reflected in the *_conf.fit files. How should we attempt to mask these bad pixels?
• What are the gains of the individual detectors after the CASU coadd?

Sorry. The telescope generated NDF and CASU then did whatever they wanted to them to archive them as FITS. There were two distinct WFCAM pipelines. One was built using Starlink by JAC to work on the native NDF format and the other was written by CASU using the FITS files. WFCAM has large pixels so the observing mode involved stepping "x and a 1/3" pixels (I think a third) and then interleaving the coordinate grids to oversample. The FITS headers reflect that stepping. Sounds like you've got coadds though so this is irrelevant and I don't know how CASU merge headers from individual exposures or handle provenance.

John Swinbank if this is still of use for HSC processing should we set it to team External?

Yes.

I've had zero contact with the people working on this, so I suggest we close it for now.