When sentence world is initialized, this number reflects the maximum possible penalty
on word identifiation time (which is useful for knowing the bounds of the state space).
Allows a command-line based reinforcement learning simulation based on the contents of a maze.prop
properties file (see Details below for more information on maze.prop
Simple example to navigate through a grid world for for tutorial/intro
purposes; does not load anything from a file but is
instead fully self-contained.