Check my ftp directory:    ~ftp/pub/Users/mw
There are 4 files there -- copy them

mdp.f            A Fortran program to fit "Dirichlet mixture models" (usually 
                 wrongly called "Mixtures of Dirichlets" or MDPs) to univariate 
                 data : a flexible class of normal mixture models

mdp-rand.f       Fortran code for some random variate generation and other 
                 functions used in mdp.f

mdp.dat          An input file for mdp.f containg initialisations and parameter
                 values, and an example data set -- from neurophysiology (the 
                 data in the paper by West and Turner in the course-pack) 


mdp.s            Some crude S-Plus code to read in the file mdp.d that is 
                 created when you use the Fortran program. This plots estimated
                 density functions -- Bayesian predictive densities and also
                 cdfs. 

At the start of mdp.f there are some details of how to compile and run using
the Fortran 90 compiler (or Fortran 77, substituting f77 for f90) 

In you spare time, try it. Run it a few times with varying values of "alpha"
-- the first parameter in the mdp.dat file. e.g., try alpha=0.1, 1, 5 for 
this data. You may want to change the MC sample size too (Burn-in set at 500
should be OK here).