Re: FITS tables

[Jure Skvarc <jure.skvarc@ijs.si>]

I promised to be busy with my work, but let me have some further questions.

In the star list fields, there are quantities FWHM[XY], ROUNDNESS and SHARPNESS.
FWHM is a parameter which can be naturally derived from the PSF function, like
2.35*sigma for a (1D) gaussian etc.  I wonder what are the PSF functions used
for images with noticeable coma like TASS images and how do you define FWHMX and
FWHMY. Are X and Y meant along x, y axis of the images or are they along the
minor and major axis of the blob (star)?  

Now I will risk to expose all my ignorance, but I am interested in this
subject.  How is the roundness defined?  Probably everybody else knows this
because no one complains by saying "my software does not compute this!" and this
field is supposed to be required.   Next, how exactly is the sharpness defined? 
The closest thing to this that I remember is expressed by <x^4> and <x^2>, where
<> means expectation values, but I don't know the exact formula right now.  In
any case, unless these things are well defined, I see no sense to make them the
requested quantities.

Also, do all software packages aspiring to process Mark IV images compute both
aperture magnitude and PSF fitted magnitude?  If not, they are out of the game
until they are updated to do this.

Now to the Tom's comment:


> I have one comment.  Should not the header include the processing steps
> applied to the data?  For example, we might have a star list, then a star
> list corrected for air mass.  I would show this by simply listing in the
> etc... 

Yes, it is quite common to indicate any data manipulation using HISTORY keywords
in the FITS header.  It would be a good idea to continue with this praxis, maybe
like:
HISTORY ACTION:  Calculate sky coordinates
HISTORY Performed by  XYZ program version 5.3
HISTORY Date of operation 2000-11-29
 
> If the below is acceptable to all, then I would like to make it a TN, and
> get on with processing data.  Please make objections as soon as
> possible.  Note this is just the format for exchanging star lists.  There
> will be other formats for images.
> 
I think we should wait with this definition until major players in this
processing business accept (or reject) this.


Also, I will repeat my question from one of the previous emails.  Is it a
realistic scenario to shuffle half-processed star lists with millions of stars
all around the planet until they are well analyzed or is it better to do
complete processing at the site which produced the data?  I understand Arne's
concern that the whole and correct processing is a complicated task which must
be performed carefully.  But isn't this one of the things that operators will
just have to learn or accept other people's pipelines?


				Jure