[fitsbits] CDELTn

[Steve Allen]

On Tue 2007-09-18T12:26:21 +0100, David Berry hath writ:
> Section 2.1.2 of FITS-WCS paper 1 says "Furthermore, all CDELTi must
> be non-zero." Even if NAXISn is 1, that one pixel should still have
> some width. If any CDELT is zero then the transformation from pixel to
> WCS cannot be inverted.

In the lingo of RFC, I think this is a SHOULD and not a MUST.

More explicitly, CDELTi SHOULD be non-zero, and FITS intepreters
SHOULD NOT fail if CDELTi is zero.

The subject of non-invertible transformations does arise in
FITS images.  The simplest and most historic example would be
a photographic plate exposed with an objective prism.
The draft of the WCS Distortions paper already says more on this
subject.

--
Steve Allen                 <sla at ucolick.org>                WGS-84 (GPS)
UCO/Lick Observatory        Natural Sciences II, Room 165    Lat  +36.99855
University of California    Voice: +1 831 459 3046           Lng -122.06015
Santa Cruz, CA 95064        http://www.ucolick.org/~sla/     Hgt +250 m