Dividing the score by number of data points #71
Description
Recently, Q2MM was changed so that the score value for each data point is divided by the total number of that type of data points. It has some advantages, but also leads to some problems. First of all, it's much harder to get an overview when the numbers are small; with the old printout, any single data value score above 1 was a candidate for inspection, now it veries depending on data type, and is buried deep in the decimals. Second, the 1st and 2nd derivatives printed out after the numerical differentiation step are now too low to allow for analysis, the values can easily be too low to show in the 4 decimals shown. A critical part of the post-parameterization analysis is to verify that for each parameter, the 2nd derivative is positive and higher than the absolute value of the 1st derivative. I would suggest reverting this feature, take away the division, or, if you think it's motivated, divide by it only before the total score is printed, after the summation for each type is done (possibly printing both, calling it "raw score" and "normalized score").