Problems with system variables

["Brad L. Fisher" <fisher(at)trmm.gsfc.nasa.gov>]

This problem requires a little bit explanation.  I am extracting data
from an image window with the cursor command, which I am using to
generate two different windows (the cursor coordinates are used to
select the central pixel in one of two windows).  One window is
displayed with TV; the other with PLOT (draws a simple graph). The idea
is for user to be able to select different points in the main data
window and carry out analysis on the data selected within the other two
windows. Consequently, the program was designed to interact with the
main data image multiple times using the cursor command.

Before adding the PLOT window everything was fine, but after generating
a PLOT, which was drawn in a unique window opened by the program, it
seems that the window system info from the main image is getting lost or
replaced.  For instance, when I select a point with the cursor the first
time through everything is fine, for at first, there is only one window
drawn, but the second time through the cursor return bogus data
coordinates.  It turns out that the data selected from the main window
was actually data in my PLOT window.  I did notice that the system
values in !p were getting changed.  I then tried just saving the
original !p  and reassigning it after generating the PLOT, but that
solution did not rectify the problem I described above.  Can anyone help
with this?  If so, please send responses to fisher@trmm.gsfc.nasa.gov.

Brad