Maze Maker — First Test Results

Test #1

This is the first test of the MazeObject interface.

The goal is to test setting the walls of a given cell. The small "maze" was created by hand-written test code.

Also tested, setting the cell contents (the blue and green cells).

Test #2

This is the very first test of the random path maker cutting its very first path.

In this first test, the "path cutter" halts as soon as it hits an obstacle — in this case, the north wall.

Test #3

The path cutter now has a direction-change routine it can take.

Whether it changes direction or not (and which direction it takes) is random.

Test #4

This test of the full operation of the path cutter created a funny-looking maze.

That's due to the parameter that controls the probability of the cutter randomly changing direction. Clearly, it needs to change more often!

path dump

Test #5

This is looking much better, but there are too many "cubicals" in the maze.

The parameters aren't tweaked quite right just yet, but it's getting close.

log file

Test #6

I think we found the sweet spot!!

The thing that really surprised me was how a very simple algorithm worked so well generating a decent maze!

I was (pardon the pun) amazed. I expected it to be much harder to find a decent process.