This was by far the shortest program in Tomas Rokicki's Cube Solver Contest in 2004 (see the results), the whole source code easily fitting on a single screen (it's 528 bytes the way Tom counted). Yes, this is a complete solver, you enter a scrambled cube state and it solves it in a fraction of a second. But it takes a few hundred turns, which is why its overall score was bad. Here's some documentation. This program btw is a predecessor of my blindsolving method.
My C++ implementation of the Thistlethwaite algorithm (except a little modification to make it easier for me) which won the second place and judge's prize in Tomas Rokicki's Cube Solver Contest in 2004. The way Tom counted, its source code is 1311 bytes and for the cubes in Tom's test set it took 16.72 turns on average to solve them (those weren't all random, though, for random cubes expect 30-35 turns on average). Here's some documentation.
A program to analyze the quality of the standard 25 moves (and longer) scrambles we're using to scramble the 3x3. See its results. In contrast to some other analyses by other people which generated and analyzed a small number of randomly generated scrambles, my program analyzes *all* scrambles. It is thus not a random experiment but an exact computation (except negligibly tiny rounding errors) of the expected outcomes of scrambling with this kind of scrambles. It was inspired by a post of Lars and I posted my first results post shortly afterwards.
I like to print out computer generated scrambles so I can use them whenever I want and don't have to turn on the computer and have it running next to me. This program generates a PDF document with two pages, each showing 50 computer-generated scrambles of 25 moves each. The font is scaled for each scramble so they all have the same width. There's some space on the left where I like to write down the result after the solve. Every time you reload it, it should give you different scrambles (if not, your browser cache or a proxy keep showing you the same version). If you're interested in how it's done, have a look at the PHP source code.
Just like my above 3x3 scrambler, but for scrambling/solving in <r,u,f>. For scramble length I chose 40 moves because that results in about the same number of possible different scrambles as 25 moves for the regular 3x3, and because the number of possible states for <r,u,f> is about the same as for regular 3x3 (half, to be precise). If you're interested in how it's done, have a look at the PHP source code.
My new megaminx scrambler/notation.
Last modified: March 14 2009, 13:48:38