|
|
Post by edtrice on Jun 2, 2016 3:48:43 GMT
I thought this would be a good place for members to introduce themselves. Feel free to tell people about yourself. Well, whatever you are comfortable telling others about anyway  My name is Ed Trice. I like to design high performance computers, and write high performance software programs that align with my interests. I've written programs that play chess and checkers, and co-authored a program that plays a chess variant I invented back in the year 2000 known as Gothic Chess (10x8 board with a Chancellor that moves like a Knight or Rook, and an Archbishop that moves like a Knight or a Bishop). I also enjoy creating databases of perfect knowledge for the games I have programmed. For example, my Gothic Chess program could announce a forced checkmate from a distance of 268 moves, playing perfectly along the way, forecasting the distance to mate on every move until the inevitable conclusion.  Shown above: A mate in 202 moves in the endgame of Archbishop vs. Bishop and Knight. White played Kb2 after being in check by the black knight. The white piece on square j7 is the deadly Archbishop. It can move like a Bishop or a Knight on any turn, but the enemy King must avoid both checking auras if caught in them.I also wrote endgame databases for the game of checkers, where the winning side could announce a win in 253 plies (moves counted for both sides, so half as many full moves). Papers I have published also appear in textbooks and at least one other book I know of. That book is "One Jump Ahead: Computer Perfection At Checkers" by Dr. Jonathan Schaeffer. I also wrote software that can solve the 4x4x4 and 5x5x5 versions of the Rubik's Cube. I created this discussion board primarily for those interested in discussing programming techniques for cube solving software, and also to encourage cube aficionados that are subject matter experts to weigh in and offer insight that could be helpful to programmers seeking to made their creations better.  I learned a great deal from other programmers and cube enthusiasts at a discussion board called SpeedSolving.com and my programs would be much worse than they are now if it were not for their input. But that discussion board is aimed more for speed cubers looking to find faster finger tricks, or algorithms that are easier for humans to perform, rather than an outlet that would cater specifically to programmers. There is also another reason why I started this forum when I did, but I was thinking about creating this long before the date issue became sooner rather than later. So, if you are a programmer, a want-to-be-programmer, speed solver, or a published artificial intelligence researcher with an interest in Single Agent Search, this place was created specifically to welcome you. |
|
abunickabhi
Junior Member
I am a 5x5 enthusiast and want to make efficient human methods in it
Posts: 64 
|
Post by abunickabhi on Jul 28, 2017 16:17:03 GMT
Hello , I am Abhijeet Ghodgaonkar
I have been interested in 5x5 solving programs for a while, and I have always loved the 5x5 cube as I like to do it blindfolded.
Since the 5x5 blindfold record is now out of reach, somewhat for me, I have turned to using 5x5 solvers to find very good algorithms in doing center clusters and wings in onder to shorten my blindfold solving process.
I hope I may be able to progress in this matter as much as I can
Cheers!!
|
|
|
|
Post by edtrice on Jul 28, 2017 17:00:19 GMT
Hello , I am Abhijeet Ghodgaonkar I have been interested in 5x5 solving programs for a while, and I have always loved the 5x5 cube as I like to do it blindfolded. Since the 5x5 blindfold record is now out of reach, somewhat for me, I have turned to using 5x5 solvers to find very good algorithms in doing center clusters and wings in onder to shorten my blindfold solving process. I hope I may be able to progress in this matter as much as I can Cheers!! Hello Abhijeet, Welcome to the board. I am Ed. We've been chatting through the Facebook messaging interface. The brute force version of my program was able to solve the centers-only problem you showed me:  I changed the notation of my solver (again) because I grew tired of the ambiguous notation. Also "M", "E", and "S" never made any sense to me! Some of those moves turn in the opposite direction you think they would. In my notation, + means clockwise, - means counterclockwise, 2 is two turns of course. There are only 2 "face" designations: Top, Front, and Right. The prefix to T, F, and R just means the slice number counted from the top, front, or right. So: 1T+ = U 1T- = U' 1T2 = U2 2T+ = either 2U in SiGN notation or u if that means a single slice turn or uU' if u means turn the outer layer and the first inner layer both as a block turn, then just turn the outer layer back. So much confusion in the notation is why I came up with this one! The thing to remember is 3T+ spins in the same direction as 1T+, as does 4T+ and 5T+. Usually, speed cubers reverse the direction of rotation once crossing over the "halfway" mark on a cube face, so that 4T+ = 2D' in SiGN, and 5T+ = D' in SiGN. I found myself having to "pause" to "think from the left" (or the bottom, or the back) when reading WCA notation, but using my notation, it's all "right handed" and seems simpler. |
|
Ben
New Member
Posts: 1
|
Post by Ben on Mar 7, 2018 20:27:42 GMT
Hi!
My name is Ben. I've just gotten back into twisty cubes after a long hiatus. I'm a web/mobile developer with a background in 'hard' languages.
As I learned the various cube types, I became more and more intrigued with designing a solver for various cube sizes. I'm really enjoying the 5x5x5 right now...
I'm having a little trouble following some (ok, most!) of the terminology and was wondering if you could point me in the direction of some help or research papers. I'd also love to learn more about how mathematicians use the cube to demonstrate various group theory concepts, if that's something of interest here.
Part of the challenge when diving into this topic is that each twisty cube group (speed cubers, researchers, programmers) seem to have varying terminology for some of the same concepts. Or that's my excuse anyway for feeling so lost!
Hope to learn a lot and thanks for setting up this forum!
Cheers,
Ben
|
|
|
|
Post by edtrice on Mar 12, 2018 2:01:58 GMT
Hi! My name is Ben. I've just gotten back into twisty cubes after a long hiatus. I'm a web/mobile developer with a background in 'hard' languages. As I learned the various cube types, I became more and more intrigued with designing a solver for various cube sizes. I'm really enjoying the 5x5x5 right now... I'm having a little trouble following some (ok, most!) of the terminology and was wondering if you could point me in the direction of some help or research papers. I'd also love to learn more about how mathematicians use the cube to demonstrate various group theory concepts, if that's something of interest here. Part of the challenge when diving into this topic is that each twisty cube group (speed cubers, researchers, programmers) seem to have varying terminology for some of the same concepts. Or that's my excuse anyway for feeling so lost! Hope to learn a lot and thanks for setting up this forum! Cheers, Ben Hello Ben, There was a flurry of activity "programming-wise" and "published papers-wise" in the 1980s. Some names you might want to google for should bring up links to their papers. They are a bit heavy on the "math side" and some knowledge of number theory will go a long way. 1. Herbert Kociemba 2. Tomas Rokicki 3. Morwen B. Thistlethwaite I've written a 4x4x4 and 5x5x5 solver myself, and I still don't use the correct nomenclature when describing stuff, so don't let that be a barrier to asking questions and trying to find out more information. And welcome aboard! |
|
abunickabhi
Junior Member
I am a 5x5 enthusiast and want to make efficient human methods in it
Posts: 64
|
Post by abunickabhi on Jun 30, 2018 10:58:29 GMT
Hello Abhijeet, Welcome to the board. I am Ed. We've been chatting through the Facebook messaging interface. The brute force version of my program was able to solve the centers-only problem you showed me: I changed the notation of my solver (again) because I grew tired of the ambiguous notation. Also "M", "E", and "S" never made any sense to me! Some of those moves turn in the opposite direction you think they would. In my notation, + means clockwise, - means counterclockwise, 2 is two turns of course. There are only 2 "face" designations: Top, Front, and Right. The prefix to T, F, and R just means the slice number counted from the top, front, or right. So: 1T+ = U 1T- = U' 1T2 = U2 2T+ = either 2U in SiGN notation or u if that means a single slice turn or uU' if u means turn the outer layer and the first inner layer both as a block turn, then just turn the outer layer back. So much confusion in the notation is why I came up with this one! The thing to remember is 3T+ spins in the same direction as 1T+, as does 4T+ and 5T+. Usually, speed cubers reverse the direction of rotation once crossing over the "halfway" mark on a cube face, so that 4T+ = 2D' in SiGN, and 5T+ = D' in SiGN. I found myself having to "pause" to "think from the left" (or the bottom, or the back) when reading WCA notation, but using my notation, it's all "right handed" and seems simpler. Yeah thanks for helping me out getting introduced to your program. I agree the speedcubing notations just cannot be coded directly in a solver, and it better to adopt a new notation system. Your notation system seems good and intuitive. |
|
|
|
Post by edtrice on Jun 30, 2018 18:56:23 GMT
Yeah thanks for helping me out getting introduced to your program. I agree the speedcubing notations just cannot be coded directly in a solver, and it better to adopt a new notation system. Your notation system seems good and intuitive. You might be the first speedcuber to actually agree with me on the notation issue |
|
|
|
Post by IAssemble on Oct 1, 2018 17:01:03 GMT
Good afternoon all. dwalton76 suggested I might like to join and take part in the fewest moves challenges this year :-) You may be aware of me and my activities from other forums or internet activity. Probably my most well-known activity is co-designer of the LEGO "CubeStormer" robots: www.youtube.com/watch?v=X0pFZG7j5cE More recently I've been working on faster versions of my 4x4x4 robot solvers. |
|
|
|
Post by edtrice on Oct 1, 2018 17:33:14 GMT
Good afternoon all. dwalton76 suggested I might like to join and take part in the fewest moves challenges this year :-) You may be aware of me and my activities from other forums or internet activity. Probably my most well-known activity is co-designer of the LEGO "CubeStormer" robots: www.youtube.com/watch?v=X0pFZG7j5cE More recently I've been working on faster versions of my 4x4x4 robot solvers. Hello, and welcome. Yes, I remember you, and Dan was nice enough to extend the invitation that I was unable to do. I am wondering if your 5x5x5 solver was completed before mine? One of us is probably the first to have a 5x5x5 solver unless you know of an earlier effort. This year, I'd like all participants to consider having their results published in a paper. I have already started describing how my program operates, and when I'm done with it, I'll let those interested in doing the same copy my format. Think about this, and get back to me. In the meantime, the scrambles have been posted. Solutions are due October 31st. |
|
|
|
Post by IAssemble on Oct 1, 2018 18:08:15 GMT
I am wondering if your 5x5x5 solver was completed before mine? One of us is probably the first to have a 5x5x5 solver unless you know of an earlier effort. This year, I'd like all participants to consider having their results published in a paper. I have already started describing how my program operates, and when I'm done with it, I'll let those interested in doing the same copy my format. Think about this, and get back to me. In the meantime, the scrambles have been posted. Solutions are due October 31st. Thanks. Do you mean solver program or robot solver? I published a video of my first 5x5x5 robot in Feb 2010: I think this initially used my first 5x5x5 solving algorithm but was the point at which I began to create a generic NxNxN algorithm. I will think about the idea of publishing my algorithms in a paper but do not make any promises ;-) EDIT: looking at my source code control history, it looks like I didn't start my latest "optimised" 5x5x5 solving algorithm until Apr 2014. |
|
|
|
Post by edtrice on Oct 1, 2018 21:31:53 GMT
Thanks. Do you mean solver program or robot solver? I published a video of my first 5x5x5 robot in Feb 2010...looking at my source code control history, it looks like I didn't start my latest "optimised" 5x5x5 solving algorithm until Apr 2014. Ok your program was years before mine. I started my brute force solver right around the same time as your optimized solver. I saw your posts about that effort and mistakenly concluded that was your initial project. Nice graphics in the video by the way. OpenGL? |
|
|
|
Post by edtrice on Oct 2, 2018 9:57:22 GMT
I will think about the idea of publishing my algorithms in a paper but do not make any promises ;-) I am publishing the paper on this event, the programs involved, and the names of the program authors. I am not looking for an "algorithm-level" delineation of how each program works; just an overview. Example: Omnia Obtorquebantur 5x5x5 is primarily a brute force solving program for the Professor's Cube. At present, it can always find the fewest moves to solve any scramble up to 14 turns away from the solved state. It has been modified to perform as a stage-solving program, capable of solving any scrambled cube. It uses a Single Agent Search coupled with large hash tables to apply trunk-level pruning to the search tree in Phases 1 through 3 of its 4 phases. The 4th phase uses over 5.5 billion pre-computed algorithms to finish solving the cube. |
|
|
|
Post by IAssemble on Oct 2, 2018 18:56:20 GMT
Ok your program was years before mine. I started my brute force solver right around the same time as your optimized solver. I saw your posts about that effort and mistakenly concluded that was your initial project. Nice graphics in the video by the way. OpenGL? Thanks. Yes, OpenGL |
|
|
|
Post by IAssemble on Oct 2, 2018 18:59:15 GMT
I am publishing the paper on this event, the programs involved, and the names of the program authors. I am not looking for an "algorithm-level" delineation of how each program works; just an overview. That sounds like a possibility :-) |
|
