[daip] trouble with mapping

[Prof. Sara Beck]

Dear Aips-people,

I need advice on some problems with producing maps from calibrated data.
I am working from a data-set which has produced very nice maps when
reduced and analysed with 31DEC00 from beginning to end. (This was done at
UCLA on a Sparcstation) However, when I
take UV data that has been calibrated (in 31DEC00, at UCLA) but not split,
sorted or imaged, and split-sort-image with my 31DEC03 version (running
under Redhat), the sources disappear.  the noise looks reasonable (this is
at U,Q, and X bands) but the fields appear blank. including the
calibrators, by the way.  peak flux in the fields is at least a
factor 10 lower than the peak flux in the maps already generated.

my first thought was some dumb error in running imagr, so i opened the
history file of one of the sucessful maps and reproduced the steps
exactly; same value for gainuse and flagver and everything else.
>original) and produced the maps again, and again saw nothing.  here are
the relevant lines from the prthi files:
 IMAGR TFLUX =  8.38551E-05
(when done on 31DEC03)
IMAGR TFLUX = 4.84090E-03 (the original maps that were 31DEC00 all the
way) .

there are some differences in the prthi for the 31DEC00 and 31DEC03 runs,
in the way things are reported in prthi: the 31DEC00 prthi says

 SPLIT GAINUSE =  2 / CL table
>  SPLIT / Weights not calibrated
>  SPLIT /            Previously flagged   flagged by gain         kept
>  SPLIT / Partially               0                   0            0
>  SPLIT / Fully                   0                   0         195769
>

>and the 31DEC03 gives

>        SPLIT SUBARRAY =   0
>   SPLIT FLAGVER=   1  / Edited using FG table version #
>   SPLIT / Output data in compressed format
>   SPLIT / Weights not calibrated
>   SPLIT /    195769 Visibilities written

(although CL was set to 2 in this version as well).

I am at a complete loss.  I can't believe that the 2 aips versions are
incompatible so that data which is calibrated in one cannot be dealt with
in the other. I also can't believe that copying the data over to a CD,
which we did in LA, managed to corrupt the data files but not the maps.
I could easily believe it is human error and that I have made some mistake
so epochally dumb that it will go down in history and provide endless
amusement for the analysts. Which is fine with me, laugh all you like,
just tell me how to get decent maps out of this run!

many thanks,

sara beck