[EM] A good way to fit B/(q+A) to state density function

[Michael Ossipoff]

Fit the definite integral, from 0 to Q (where Q is a particular state's q) 
of B/(q+A) to the data points consisting of the states' q and their 
cumulative numbers. The smallest state's cumulative number is 1. The 2nd 
smallest state's cumutive number is 2.

That is, fit B*ln(Q/A+1) to the data points consisting of the states' q and 
their cumulative numbers.

Mike Ossipoff