This document describes all the configuration stuff we want to put in xml format so we 
have a single place/language.

How are we going to store this, make it easy to understand and should we make some of these
configuration dynamic.

Heuristics
==========


Initial state
=============

How is that going to be different from having everything coming via the goal-loader

Reactor settings
===============

- association between token/predicate and Reactor 

Planner settings
================

- Horizon settings 
- Skipping off timelines
- Revocable decisions

Goal planner settings
---------------------

Reactive planner settings
--------------------------

- blowing away the future
- forgetting the past
- latency
- token stretching

Controllability of tokens
-------------------------

Communication settings
======================


AgentRelay settings
===================


PlanRunner settings
===================

Command line settings
=====================

Simulator stuff
===============


Output filtering
===============

What do we put in the log when we print a plan

Modules to load
===============

Shared libraries we load

Time settings
=============

- unit of time
- 

