Re: New UCD1+ for SIA protocol

From: Jonathan McDowell <jcm-at-head.cfa.harvard.edu>
Date: Thu, 25 Mar 2004 11:01:31 -0500 (EST)


Apologies if this is already covered, I missed some emails.. My biggest problem is "pos.wcs" which generalizes poorly. Seems to me that naxis and naxes are
properties of the pixels, and indeed of the image and not of pos itself. What if you have an x,y,time cube? I would argue it would be

 obs.image.naxes
 obs.image.naxis

or perhaps
 obs.image.wcs.naxes
 obs.image.wcs.naxis

(or array.naxes or whatever the top level thing, but not pos)

but not
 pos.naxes = 2
 time.naxes = 1
 pos.naxis = 512 512
 time.naxis = 30

Similarly, the other wcs.* should probably be image.wcs.* not pos.wcs.* ?

I agree with others comments about instr.scale; Should be
 image.wcs.scale

I am generally confused about what is in "wcs" and what is not: "pos.frame" not part of wcs, but "pos.wcs.naxes" is?

> > meta.bullshit
> meta.unknown?

I think the first is a more accurate description; how about

   meta.badly_defined
or

   meta.obscure
:-)

Received on 2004-03-25Z16:02:32