[fitsbits] CDELTn

[Mark Calabretta]

On Thu 2007/09/20 10:04:16 +0200, "LC's NoSpam Newsreading account" wrote
in a message to: fitsbits at donar.cv.nrao.edu

>will be no doubt they are applicable. But maybe it is worth adding 
>ranges of legal values to table 8.2 of thereabout (Bill are you keeping 

Good idea.  It would also be worth emphasising that the transformation
matrix must not be singular.

>The only other point is that CDELTn were around before WCS, and we have 
>in 3.0 the rule that a change of a standard cannot make invalid what 
>done before ...

...maybe, but the behaviour of a WCS with CDELTn == 0 wasn't defined
before either.  In any case, if you use any of the new WCS constructs
then you have to follow all of the new WCS rules.  This also means
disallowing some other things (e.g. CROTAn) that were previously
allowed.

Cheers, Mark