July 16, 2004

This is the final version, which uses calloc and
floats, instead of arrays and doubles. The range of
application is greater with this code. The source is
the same as 
  XnumCalloc/FloatCalloc/*

Since it is tricky, and in case future bugs arise,
I have saved 
  FINAL_0
which is the last stable code using arrays and doubles.

Finally, just in case other problems arise, I've saved all earlier
versions.
