Look at STARPAC ftp://ftp.ucar.edu/starpac/ and Statlib
http://lib.stat.cmu.edu/ for more ideas

0. Clean up the documentation to reduce the proliferation of
prototypes for all the different types (float double etc). e.g. Just
document the double versions and explain how to call the others by
analogy.

1. Try using the Kahan summation formula to improve accuracy for the
NIST tests (see Brian for details, below is a sketch of the algorithm).

      sum = x(1)
      c = 0
      
      DO i = 2, 1000000, 1
         y = x(i) - c
         t = sum + y
         c = (t - sum) - y
         sum = t
      ENDDO

2. Allow weighted values, mean(x) = \sum_i w_i x_i / \sum w_i,
mean_weight(x) = \sum w_i, etc

3. Prevent incorrect use of unsorted data for quartile calculations
using a typedef for sorted data.

4. Rejection of outliers

5. Time series. Auto correlation, cross-correlation, smoothing (moving
average), detrending, various econometric things. Integrated
quantities (area under the curve). Interpolation of noisy data/fitting
-- maybe add that to the existing interpolation stuff.What about
missing data and gaps?

6. Statistical tests (equal means, equal variance, etc).