Maze generation

[Gordon E. Greene]

Perhaps a clarification is required about my posting.  I received several
messages about it.  My contention that a non-tree maze is not solvable by
trailing hand refers to mazes that don't have both the entry and exit on
the outside.  
 
My starting point for this was an old book of mazes I once had.  Many
of the mazes had both entry and exit inside the maze.  Some of them were not
planar: that is they had corridors that passed over and under each other. It
wouldn't be so tough to encode these mazes by labelling all the intersections,
putting the maze into some graph representation, and then doing a path
search through it from the start node to the end node.  I think you'll agree
that this would take a lot of the challenge out of doing mazes though.
 
Breadth-first and depth-first searches would both find a path through, and
variations to find the best (whatever best means) path could be implemented
too. 

Sorry for any confusion that may have arisen from my incomplete description.

--
--------- You can never have too many ferrets. -----------
gordon at rpi.edu    USERF023 at RPITSMTS   USERF023 at mts.rpi.edu