Re: Fits Headers Again

[Cary Chleborad <bennu@ns.net>]

aah@nofs.navy.mil wrote:
> 
> Chris and Tom suggested programs that add keywords
> and information to the fits header.  I come from
> the old school that says to leave the raw images
> alone.  They should be archived as they are read
> from the CCD.  Processed images are another story.
> I have no problem with dark subtracting, flatfielding,
> and WCS generation for processed frames, which can
> then be archived separately.
>   The last thing you want to do is make a mistake
> with the raw image and lose the information forever.
> Arne


This is a valid Concern, but this project has quite a staggering Data
flow output: as I understand: 8Meg per field * approx. 200 images per 10
hr night per camera = 1.6+ Gigabytes * total amount of cameras. I think
that the "safety" factor vs. data transportation & storage capacity
needs to be considered.

Any well written routine that manipulates a data set can and should use
a procedure to "prevent" data mishaps.  i.e. copy the data file to a new
file, add the required headers etc., do a comparison of the data field
with the old, then delete the original.


- Cary