With thanks to Dan Hoey for getting me on the right track, I
have finally got most of the squares group evaluted. The big
breakthru was developing a checksum for a squares position.
I know it's been done before, but I wanted to prove to myself
I could do it on a mere 386 with 4 megs of memory. My
latest program (rubik5.exe) took 24 hours to number the squares
group up to 8 moves deep. The point of all this was to create
a squares group database to aid in developing an optimal
solver for the cube. Ultimately the database will have an entry
for every squares group position, along with it's optimal
solution. I would be interested in hearing from any others who
have created such a database, and what type of compression or
checksum was used for the arrangement.
Also I've received a call from Richard Schneider. He is publishing
a comprehensive book on square 1, plus a follow-up book on
pretty patterns and shapes. This will be available in the States
I haven't been seeing anything from cube-lovers in a while,
I hope it's still up and running.
To: Mike Reid --- Hope you see this! Any further progress been
made on God's Algorithm? I'm still trying to
catch up. I'm still interested in that code of yours to find
improvements on some pretty patterns I've discovered.
Anyways here is what my program has found so far:
Squares group (u2, d2, l2, r2, f2, b2) Moves Deep Number of patterns ---------- ------------------ 0 1 1 6 2 27 3 120 4 519 5 1932 6 6484 7 20310 8 53000 (and counting) :->
Got to improve it's speed....
-> Mark <- -- Canada Remote Systems - Toronto, Ontario World's Largest PCBOARD System - 416-629-7000/629-7044