RE: Fits Headers Again

[<aah@nofs.navy.mil>]

Tom,
  There are three levels of archiving in my thoughts, and I may not have
made things clear.
  First, I don't expect your qbasic program to do processing,
even at the star matching level.  It should be a basic acquisition
program with effort put into scheduling and fast archiving rather
than processing; otherwise, things just get too complex.
As such, it should save the image with rudimentary header additions.
As you state, each image should be able to stand on its own,
without a log file.
  Second, what I was lobbying against, is to take that raw image
*after it has been stored*, and modify the header to reflect
things like improved WCS.  My Norwegian heritage sez you run the
risk of overwriting important parameters or corrupting the image
itself, since you now have no original copy.  It happens all of
the time, and you really kick yourself, but you have lost the
data.  As long as you do not do this step, I will stay quiet.
However, it appears you are advocating this step from your comment:
>I think a first project for me will be to ask if someone will write a 
>program to read an image and add WCS to it.  I assume that this is a simple 
>script that I could do on my Linus machine which has both a read and write CD.

  Third is a processed image; perhaps flatfielded, perhaps with
astrometry.  Andrew or Chris or Michael or Arne may have acquisition programs
that can do this before storing, but it is not easy and you
should think about three times before trying it yourself (and spending
many hours that could be better utilized in improving the hardware).
My plans have been to save this processed image in compressed form,
along with extracted starlists, and not to save the raw image.  That
is a goal, but is nowhere near reality at this time.  I end up saving
both the raw and the processed images right now, on the assumption
I may want to reprocess when I have improved algorithms.
For my other CCD data, I find it *extremely* rare to go back to the
raw images, and relatively rare to even go back to processed images
except for inspection,  but then the processing pipeline for
my standard CCD images is well in hand.
Arne