[Sane-analysis] Update to CVS for g_analyze_scaler_bank.f

[O. A. Rondon]

Hi,

Here is the updated table of non-ammonia parallel runs, with exact
errors (they don't change at the 5th. or 6th. percent decimal place).

I have also applied the charge normalization and livetime corrections
with the assumption that the helicity labels for the corresponding
scalers should be swapped (there is an ambiguity, since only one
helicity was recorded for LT). With the swap, the false asymmetry goes
away, so swapping may cure some of our issues. To get zero false
asymmetry the swap must be:

- for the charge asymmetry, swap hel. plus with minus
- for the livetime, assign gscaler(538) to hel. plus.

The other three swap combinations don't kill the false asymmetry.

http://www.jlab.org/~rondon/sane/analysis/asym/a_para_unpol_beta4.pdf

In any case, we need to analyze ALL non-NH3 runs to make sure this
really works as advertised. In my view, whether this procedure works or
not for the perp runs, it needs to be applied to the parallel ones, to
get the correct asymmetry (apply at the level of counts asymmetry).

Cheers,

Oscar