State variables provide a convenient mechanism for managing additional
model parameterizations such as mesh density, simulation convergence
tolerances, and time step controls. These are typically fixed
parameters for a given %Dakota run.

By default, only parameter studies and design of experiments methods
will vary state variables.  This can be overridden for other methods
by specifying \c active \c state or \c active \c all.

When a state variable is held fixed, the specified \c initial_state is
used as its sole value.  If the state variable is defined only by its
bounds, then the \c initial_state will be inferred from the variable
bounds or valid set values.

If a method iterates on a state variable, the variable is treated
as a design variable with the given bounds, or as a uniform uncertain 
variable with the given bounds.
