[daip] Mosaic observations

[Eric Greisen]

Dan Marrone writes:
 > 
 > Okay, ignore any previous terms and draw definitions from the following:
 > The data set consists of 100h of observations spread over 6 files
 > downloaded from the VLA archive. I am imaging four fields, each 1.5deg^2
 > in area. The four fields are separated by roughly 6h in RA, centered at 3,
 > 9, 15, and 21h. So the observing strategy is to observe each field for 6h
 > and then move to the next. This means that each file contains data from
 > 2-4 fields, depending on the length of the observation. 
 > 
 > Each field is covered by a mosaic of 9rows of 20pointings each. Each
 > pointing is visited 9 times over the course of the 100h.
 > 
 > At present I can calibrate all the data in any one of the files. I can
 > SPLIT the data and end up with single-source catalog entries for every
 > pointing visited during the observation represented by the file. However,
 > since every field appears in almost every file, someday I'll need to SPLIT
 > from all the multi-source data files at once, or else I'll have to
 > concatenate all the single-source catalog entries for a given pointing
 > derived from the 6 multi-source data files. Remember that I have
 > 4fields*9rows*20pointings = 720 pointings, so I cannot manually select the
 > single-source catalog entries for each pointing for concatenation.
 > 
 > Clearer?

Yes - that is clearer.  I suspect that you will need to be conversant
with writing scripts.  That will require that the "source" names for
each pointing be in some way unique so that the script can select the
split files and run the DBCONs.  With 180 pointings per field, it is
unfortunate that AIPS does not have the ability to do joint
deconvolutions as some packages claim to do.  The problem in my
perspective is that I would have the AIPS version be more elaborate
than existing programs in other packages and so be rather hard to
code.  IMAGR cannot even read toseparate UV data sets for 1 pointing
as yet.

Eric Greisen