Add "Fresnel spin-splitting" to plots

[bmaranville]

It has been noted that the traditional spin-asymmetry $\frac{(p - m)}{(p + m)}$ plot is misleading about the uncertainty in the information contained in the spin-splitting. In particular, by dividing by the sum the error bars are amplified for (p,m) combinations that have greater uncertainty.

Alex G. has suggested that we offer another plot (instead?) of $\frac{(p - m)} {R_{\textrm{Fresnel}}}$, so that the error bars are less dependent on random fluctuations in the measured value (p + m) in the denominator.

[pkienzle]

Is there any physics in the spin asymmetry curve? Or is it just a visualization technique to magnify the difference on the log scale? If the latter, than $R_F$ or $Q^4$ norm of the difference should be fine.

It might be interesting to divide the cross sections by $R_N$, the nuclear profile without any magnetism. I suspect it will be disturbing since the data itself will appear to wander as you move through parameter space.