Judging

Just wanted to let everyone know that I am sorting through all the submissions. It is going to take a little longer than expected to judge all of them (there are alot of good ones). So, expect a winner in about 2 days...

Till then..

It's drawing near! I have selected my list of finalists and am just going through them very carefully. Expect a winner very soon!

Oh boy, oh boy!!

The winner is posted! Congrats to Scott!!

Hi Matt,
could you please post a short explanation, in what this maze-solver was better than the others. That will also explain on eval criteria you use.

OUTRAGE
Hey give the winning script this maze:



The solution to this maze has 9 steps as you will be able to see. The winning script will not display the solution and also say that the shortest is 12. What is going on?

This is how I judged the submissions:

1. Could it process all 4 mazes? With the correct result?

2. Basic speed test. If a maze took more than 10 seconds or so, it was obvious that it wasn't going to win.

2. I turn on all error and warning reporting. If any errors or warnings are present (that aren't explained by system differences), the number of warnings are recorded. I got a lot of undefined indexes on this one.

3. Detailed speed test. Each remaining submission had each of the 4 mazes run through it 10 times. The average time per a maze was taken. To do the detailed speed test, I used the same code to test each submission. The code was added to the submission and only took into account the actual maze processing time, not the display time.

After all these factors, Scott's was the winner. He had no errors or warnings and his detailed speed tests were consistently below all the other submissions....

Is not a valid maze. The challenge specifically states that the outside of the maze will only be a wall or an exit. The starting point cannot exist on the outside edge.

fair enough.

It does work if the start is inside the outside wall.

Why that regulation? What difference does it make to coding?

Really the only difference it makes is that it was stated as part of the objective. Just as it was stated that walls were represented as W's, they could have been just as easily represented by #'s.....

Heh... wouldn't say I'm "outraged" like someone else posted earlier... but a little confused. Scott's program is very well done. Short and damned fast. Well done, Scott! But it doesn't seem to find multiple paths to the exit properly? When I posted a question early in the contest about the same number of steps leading to the exit, you said both would be the shortest path. But a maze such as:



No idea if that posted right... If not, it's available here: http://mcsherry.org/maze/maze-sm.txt

Anyway, I have no delusions that I would have won after a) Seeing Scott's code and b) seeing the examples that others have posted (mine didn't even allow uploaded files! ), but on the maze above there are 2 ways to reach the exit in 52 steps. Did the number of turns become important and I missed it in the forums?

Haven't had time yet to look over his code thoroughly to see how or if he deals with this issue.

Thanks, can't wait to see what's next.

would be cool if you listed who the finalists were...

Actually I said that either would be the shortest path ;)

Did your error checking include E_NOTICE?

gg

It's me again. The 'gg' in the end of the message seams to idicate that I'm somehow mad at the judge if he inlcuded e_notice. This is not true. I'm sorry. I was going to say that it were a good contest, and I liked it very much (to be short worded doesn't always work)