chooyu.astro.utoronto.ca!/d/palung/hall/cnoc/howto_decr

**CNOC de-cosmicraying notes

When done with decosmicraying (or any particular stage):
--modify summary file (compress & uncompress to change permissions if needed)
--before copying files to archive:  newgrp cnoc & umask 002

--the cnoc97b package WILL NOT run properly if the ctio package is loaded.
  (our version of the ctio package is old, and contains a "printf"
       task which conflicts with the newer IRAF task of the same name)

--the pppdir.dat file MUST be generated before running the dodecr.cl task
  (but not before the pppdecr.cl task, though it's a good idea to do so)

--the dodecr.cl task CANNOT run on chooyu if ximtool/saoimage are running
  with any images in their buffers, but ximtool should be running if display
  is desired.

--the pppdecr.cl task & others were edited to run ppp from unix, not from IRAF.
  (for some reason the task thought ppp was complete before it really was)

--crview.cl (and maskedit.cl and contreplace.cl) were edited to have
  "radius=0." as one of their command lines arguments, as required by
  images.imutil.imreplace (new for IRAF V2.11 I think)


**for the 2148a0L mask, the auto-decr choice was IRAF, but it was very poor.
  I chpixtype'd image #2 to real, subtracted the bias, multiplied by 0.9331637,
  added the bias back in, chpixtype'd back to short, and proceeded normally.
  When decr'ing was done, I chpixtype'd the final image #2 to real, subtracted 
  the bias, divided by 0.9331637, added the bias back in, and chpixtype'd back 
  to short.  The resultant image shows 0.6% average deviation from the original
  in the spectral regions (excluding cr's), presumably due to roundoff error.

**using the above procedure (ipppdecr.cl batch file),
  typically one image has sky lines erroneously marked.
  use "imexam" and "b" to mark corners of boxes to replace

2148c5B: 1040 1467 ok -- true cr atop emline