Warren responds to me:
quote:
quote:
Take a look at the latest Boeing 777. It was completely designed by computer.
Yeah, ID had nothing to do with it. Give me a break!
I am.
Are you saying the computer was intelligent?
Or perhaps you're saying that the design for the Boeing 777 was "stacked" inside the program that created it and the computer programmers that put together the design program were just dicking around for all that time and sucking money out of Boeing in a conspiracy to keep themselves employed? "Pssst! Joe! I've got this great design for a new airplane. Yeah...down to the last wire. The bosses want a new plane but if we give them the design right now, we'll be out of a job in two weeks! Can you come up with a fancy program that'll spit this design out after taking a long time looking like it's doing calculations? Great."
I recall an assignment in my computer science class from my undergrad days. It was to develop a program that found a knight's tour of a chessboard and then indicated whether or not it was cyclic. That is, given an input of a starting square and using only moves made by a knight, attempt to find a path that lands on every square exactly once and then see if you could land on the starting square from the ending square.
So the first thing I do is model the chessboard. And I notice that not all squares on a chessboard are created equal. For example, the corner squares only have two possible moves for a knight. The further you get into the middle of the board, the more freedom of movement you have. Thus, to save time, I'll restrict movements based upon physical possibility so that the program doesn't try to move the knight off the board.
But in doing so, it occurs to me that certain moves might be more likely to result in a knight's tour than others. Through some practice on my own, I happen to find a cyclic knight's tour of the chessboard and thus, I could constrain the results of all the squares to just two possible moves...one of which is not allowed since it's moving to the square you just came from.
And thus, the program I turned in truly was a "stacked" program. No matter where you started on the chessboard, my program found a cyclic knight's tour.
The same cyclic knight's tour.
Every time.
I got an A because the assignment didn't say that you had to develop an algorithm that would find the tour on its own, though many people wrote just such a program. You'd place the knight on a space and the program would systematically start moving the piece around, keeping track of the resultant path and if it found that it got stuck with no more moves but still not a complete path, it would backtrack to a spot where it had another choice and continue on the new path until it found a tour.
Now, is this program "stacked"? I'd say no, though I'd point out that there are two possibilities for such a program.
One is that it has a rule for how to go about finding paths. That is, when it lands on a square, it might always do an "up two, right one" move first and if that move is not allowed, do a "right two, up one" move, and so on around the square, trying to find a legitimate move.
Such a program will always find you the same tour given a specific starting square. Why? Because it has a methodical approach that is always used in exactly the same way. There is no randomness involved. However, this program is not "stacked" because the tour you find is dependent upon where you start. My program always gave you the same tour no matter where you started. The choice of square, since the tour returned was cyclic, simply noted where on the cycle you started, not which cycle you found.
But there is a way to have the program find unique tours even for identical start positions: Have it randomly choose squares. That is, instead of having it always start by trying the "up two, right one" move, have it randomly choose one of the eight possible moves until it finds one that it can use.
Since the knight is moving randomly (though constrained by the rules that no square be landed on more than once and not moving the knight off the board), you can find multiple tours of the board given a single starting position.
In neither case does the program know what the tour is like it did in my program, though. The program is figuring it out for itself. The only thing the program knows how to do is move the knight and how to keep track of which squares have already been landed on.
So tell me, Warren: How do those programs count as having "stacked the deck"? I thoroughly admit that my program was stacked. If the boss had come to me and asked me to find a cyclic knight's tour of a chessboard, I should have simply written down the path that I found and turned it in. For me to spend another three days writing a program to spit out what I had already found would be a huge waste of time.
------------------
Rrhain
WWJD? JWRTFM!
™ Version 4.2