I'm interested in understanding how (mathematically) to do the transformation required to rectify an observed longslit spectrum image to a wavelength-spatial scale.
Basically, I'm asking how IRAF's "transform" task works.
I don't expect anyone to explain it to me (particularly not here, unless you're very keen!) but rather wonder if someone could point me to a good reference which explains how the task is done. Less desireable would be some code (in something other than IRAF's spf, or whatever it is) that does this which I could reverse engineer.
Basically, I'm trying to decide if IRAF really does the optimal job of this, and would also like more control to play with exactly how the associated "fitcoords" part of the task is done.