Bug with normf and renorm()

[Vito Spadevecchio]

I've noticed that the way in which data is normalized, and subsequently unnormalized, is flawed in neurolab (but it might just be a np problem?) 

For example, I have the following column as my target column: 

-7.933
-9.633
-7.853
-9.113
-8.263
-9.023
-10.333
-7.853
-9.343
-9.153

After the data is initialized (using nl.tool.Norm()), the following is the output 

0.97899649941656963
0.89148191365227547
0.87281213535589275
1
0.51108518086347732
0.33255542590431747
0
0.93932230102442871
1
0

As it should be first apparent, this normalization isn't correct (e.g. both -9.343 and -9.113 are normalizing to '1', when neither is the maximum value"). 

If you then renormalize the same column, you get the following output: 

-7.890386231038506
-8.0461621936989491
-8.0793943990665102
-7.8529999999999998
-8.7232683780630094
-9.0410513418903147
-9.6329999999999991
-7.9610063041765162
-7.8529999999999998
-9.6329999999999991

While close, these values *do not* match the values before being normalized and then de-normalized (they should be exactly the same). These small errors are causing me a major headache, as errors are within the range of accuracy that I'm looking for.