START  
   _________________________________________
|      	|f7		  	 	   |
|_____ A|________________ G	|_______   |
|    B   		|  _____|	   |
|    |__|____________|C |	|  ________|
|    |			|____   |	   |
| _  |			|	|_______   |
|  | |  |____________|  |  _____| H	   |
|f1| |  |	     |	|	| _   _____|
|__| |  |  _______   |  |____	| |   |    |
|f2| |  |  |   f6|   |          | |        |
|    |  |  |	 |   |______| F | |_____   |
|____|  |  |  ___|   |	  f5|	| | f8 |   |
|       |  |         | |____|   | |    |   |
|    |__|  |__ E_____| |        | |        |	
|    |	      |        |  ______| |________|
|    |_____ D |________|	| |        |
|		     f3|      f4|	   |
|______________________|________|_|____    | 
					GOAL

	We can traverse this maze in two different ways.  Every time that we
have a choice to make, we can either do one of two things: 
(1) always go the left (or right) but consistently. I.e., always go left when a
choice.
(2) stay at the same level.  If a choice (at say A), then first go left until 
another choice (say B), then backtrack (back to A in our example) and do the 
other choice (at A) until one finds a choice.  Then go back to the choice from
the first branch at A (B in our example) and continue there until a choice or a
dead-end.

	The choice of paths can be diagrammed as follows: 

					 A
				       /   \
				      B      C
				     / \   /   \
				    f1 f2 D     F
					 / \   / \
					E  f3 G  f4
				       / \   / \
				      f6 f5 f7  H
					       /  \
					     GOAL f8

	In (1) we would visit the nodes as :
A,B,f1,f2,C,D,E,f6,f5,f3,F,G,f7,H,GOAL, 
while in (2) we would visit the the nodes as : 
A,B,C,f1,f2,D,F,E,f3,G,f4,f5,f6,f7,H,GOAL.

	In (1) we use a stack to store the elements I have not yet visited in
the depth first search. (In fact, this is a NLR (preorder) search.)  In (2)
we use a queue to store the elements we have not yet visited, which is a
breath first search.