[daip] Re: new vs old imagr

[Eric Greisen]

I am baffled.  I can easily point at a very long list of serious bugs
that were fixed in the period from April 98 to the present.  Some were
so serious that we basically told people that 15APR98 had to be
replaced.  From what you have provided I see a few differences - the
input data were not the same - one a UVSRT and the other the output of
UBAVG.  The new one does not show a flagver in the HI but that should
then be the largest version if there is one at all.  The meaning of
RASHIFT and DECSHIFT changed subtly but this is a fairly small area
near the equator so I would not think that that would matter.  I note
that you do not use the OVERLAP parameter - that is generally a
mistake since your fields clearly do overlap.  The messages from the
two task - or at least the first 1000 or so (until the Cleans are
quite different) would still be some help.  Maybe even the UBAVG data
set although I have no way to go back to 15APR98.

Eric