[fitswcs] I've got this FITS WCS matrix

[Malcolm J. Currie]

Thanks for the responses Doug and Mark.

> Are you saying that any 2x2 matrix can be factorized as a product of a
> scale and a rotation?  Your code does not seem to admit otherwise.  What
> factorization does it produce for Malcolm's matrix?

I just copied ds9 from the SAO site.  It shows N down and E to the right
for these data.

I think I follow Doug's code without reference to any documentation.
It derives the Cartesian offsets corresponding to absolute sky
increments from the reference position along each axis.  From those
offsets it generates east and north position angles, using the latter
only to define the rotation.  It compares the two angles to determine
the handedness of the axes.  The rub is what exactly wcs->xinc and
wcs->yinc derive from a CD matrix.  In essence how do they obtain what
appears to be CDELTi?

Malcolm Currie
Starlink Project