fValue = Logp ( fNumber AS Float ) AS Float
Computes Logp(x) = Log(1 + x).
This function is supposed to keep the precision when fNumber is near zero.