Constraints: we have already agreed on a filenaming system for the raw
data with Alan Pickup.

UKIRT has a policy of resetting the run number each night back to 1.....

At telescope files will be called wayyyymmdd_12345.sdf,
wbyyyymmdd_12345.sdf ......
where the a,b,c,d correspond to detector, w stands for wfcam and the 5
digit run number is just that

We will create 2D raw MEF files with names like

w_yyyymmdd_12345.fit

We are proposing a simple prescription for stacked, superduper frames
or whatever....

The run number comes from the first run in the sequence and has a unique
few letter code appended to denote if its a stack, interleave ........

so standard superframe

w_yyyymmdd_12345_sf.fit

stacked dither

w_yyyymmdd_12345_st.fit

stacked superframes

w_yyyymmdd_12345_sf_st.fit

tile from superframes

w_yyyymmdd_12345_sf_tl.fit

and so on.......

These are all MEFs so we drop the a,b,c,d.  If its been processed dark
correction and flatfielding are implicit and the processed and raw data
are stored on different systems.  This way we have a simple trace back
to the original filenames at the telescope.

Catalogues generated from frames will be rootname_cat.fits and confidence
maps for frames rootname_conf.fit.  For generic confidence maps ie. one
for each passband for each night for non-stacked data we could just use
soft links rather than duplicate data or not bother since info regarding
correct confidence map is in the header.

Last modified: Wed Mar 5 12:26:53 2003