in principle you can use any polynomial to describe that. The most common ones for X-rays and CW neutrons are polynomials in tan(theta). The choice of the variable is related to the functional form of the instrumental aberrations. Be warned, however, that a parameterisation is sufficiently good if your machine is well behaving. If (axial) divergence plays a major role, then you need to consider asymmetry in a proper way.
If you do a modelling using the fundamental parameters approach, then you don't need any parameterisation, as the aberration functions already provide for the correct breadth and shape (and their variation with the angle)
Yes it is the Caglioti, Paelotti and Ricci, Nucl. Instrum. Methods, 35,223 (1958). But that expression can be modified for including addtional effects.
Yes you're rigth in the spelling, I will need to correct my glasses for close-reading. Also the second point it's true but the formula is so-called as the CPR., isnít?