README
==========================
This program computes max likelihood estimates for SV models.
It can also smooth and filter. See paper for details of algorithm.

The code is structured as follows:
   sv.f90          main program
   sv_like.f90     computational module
   sv_models.f90   model definitions
Set compiler flag to use double precision for default reals.


The following libraries are included:
   gbd    : various utilities. Set compiler flag to use double precision for
            default reals.
   Lapack : used for linear algebra. a subset is included with this code. The
            full library is available at www.netlib.org.
   BLAS   : required by LAPACK.
   DCDF   : computes cumulative distribution functions and their inverses
            for many distributions.
   SLATEC : Various numerical routines. A subset is included with this code.
            The full library is available at www.netlib.org.

Notes:
 * At each iteration, the optimizer outputs time in seconds, followed by the
   parameter values, likelihood, and gradient.
 * The makefile is set up to work for digital fortran. Type 'nmake' at the command line.
