> A short question. Can NCL do a probability scale? I have a cumulative
> distribution function (CDF) of something which I expect it to be a normal
> distribution. Is there anyway I can plot the function on a probability
> scale and make the graph looks linear? I am searching for "error function"
> and its inverse on the ncl website, but it seems like ncl doesn't have it.
> _______________________________________________
The error function codes are contained in the underlying code.
The following accesses these functions.
We will try to make the erf* functions available
as built-in functions.
---- %> WRAPIT erf_sub_ncl.f This creates a shared object ===> erf_sub_ncl.so Then run the test code: %> ncl tst_erf_so.ncl good luck
This archive was generated by hypermail 2.2.0 : Mon Jul 10 2006 - 09:18:19 MDT