Hello to All,
To put in perspective your lost thoughts on this matter, I think my three posts somewhere else might be useful in this thread. So, here they are.
---------------------------------------------------------------------------------------------------
---From "OpenChess • View topic - Attack of the Clones (ChessVibes)" at
http://www.open-chess.org/viewtopic.php?f=3&t=1138
ANY CHESS ENGINE PROGRAM RELIES ON THE ADVANCES MADE BEFORE
Post by MichaelIsGreat » Sun Feb 20, 2011 1:02 am
Hello to All,
This article by David Levy entitled "Cloning Chess Engines" at
http://www.chessvibes.com/reports/attack-of-the-clones/ is interesting in the sense that it shows that a program like a chess engine program always has to rely on the advances made previously before it by other chess engine programs. It is true of Rybka but of all the other programs.
The same could be said of most programs written today. For instance, word processors, spreadsheets, you name it, all these types of programs have what I could call a history. Many ideas in these programs have been made available by programs whose source codes are readily available on the Internet. Therefore, in order to write a "new" program of a particular type (word processor, spreadsheet, etc.), the programmer will surely read the source code of programs of that particular type and he will surely be influenced a lot by many good ideas from these previous programs whose source code he read.
There is this false idea in many people's minds that, in order for a chess engine program to be "legitimate" or not a clone, it would have to be written from scratch and nearly without adopting the advances made by others! It simply does not make sense at all. Why would a programmer want to recreate the "wheel" when the way to create the wheel has already been clearly explained all over the Internet I could say! This same idea applies to chess engine programs!
Even if a programmer would made only extremely very few changes to the full source code of a previous chess engine program that he would have found on the Internet but he would somehow manage to increase by say 100 ELO the strength of this chess engine program, this programmer should have the right to claim full credit for what he has accomplished with his very few changes. Why that?
Many (if not most!!) discoveries in the past have been made by inventors who have changed only a simple detail to the work of many others before them but this tiny change made a huge difference at creating a discovery whereas before the discovery did not exist!!!
Therefore, I could say the same for a programmer who would manage to increase a chess engine program that was at 2700 ELO to say 2800 ELO. This programmer could be considered the true author or the true inventor of this new chess engine program even if he would have changed only very few ideas or very few parts of the source code of this chess engine program! This programmer has added the indispensable "ingredient" that had eluded many other programmers before him and that transformed completely the strength of the chess engine program that he modified, even if it would have modified this chess engine program only very slightly.
CONCLUSION:
Those who want to accuse a chess engine program to be a clone of a previous chess engine program are blind and they certainly have no idea at all how discoveries are made. In any invention or in any discovery or in any new chess engine program whose strength increases dramatically, it is not the amount or the number of changes made to what was previously known that matters, it is the result that is obtained by any amount or by any number of changes made to what was previously known that matters!
In other words, even if a programmer would change only a few lines of code to an existing chess engine program whose source code is available on the Internet, if the new program increases dramatically in its strength by increasing dramatically its ELO, then this programmer should be entitled to claim full credit for his accomplishment, NO MATTER THE FACT THAT HE CHANGED ONLY A FEW LINES OF CODE TO AN EXISTING CHESS ENGINE PROGRAM!!! That is the difference between making a discovery and being stuck with a not satisfying solution!!
Best Regards To All
MichaelIsGreat
---------------------------------------------------------------------------------------------------
---From "OpenChess • View topic - Attack of the Clones (ChessVibes)" at
http://www.open-chess.org/viewtopic.php ... 8&start=10
NO NEED AT ALL TO SCRUTINIZE ANY CHESS PROGRAM AT THE WCCC!!
Post by MichaelIsGreat » Fri Feb 25, 2011 1:01 am
Hello to All,
David Levy in his article "Attack of the clones : ChessVibes" at http www chessvibes com reports attack-of-the-clones is quite pathetic. He wants to transform himself into a kind of inquisitor and he would like to scrutinize the few other programs that would like to participate to the World Computer Chess Championships (WCCC) but that would be FALSELY or not accused of being clones or of being illegitimate.
And the best, he said it at the end of his article by asking people to waste their time going over the code of the incriminated (FALSELY or not) chess programs! I am sure he could find many volunteers who would like to read the code of competing chess programs and find out, the easy way, how on earth did they manage to become so strong!
He says:
"The first thing we need is someone willing to set up and operate a bulletin board where members of the forum can “meet” and exchange views. Will someone volunteer to do this to help the ICGA on its way to stamping out these insidious practices?"
And he wants to do this kind of inquisition when even the author of Rybka acknowledges very honestly that he benefited from a lot of ideas that were present in the chess engine program Fruit whose source code was made available on the Internet by its French author Fabien Letouzey.
Once again, and it is worth repeating what I have already clearly said in a previous posting:
-----------------------------------
Many (if not most!!) discoveries in the past have been made by inventors who have changed only a simple detail to the work of many others before them but this tiny change made a huge difference at creating a discovery whereas before the discovery did not exist!!!
Therefore, I could say the same for a programmer who would manage to increase a chess engine program that was at 2700 ELO to say 2800 ELO. This programmer could be considered the true author or the true inventor of this new chess engine program even if he would have changed only very few ideas or very few parts of the source code of this chess engine program! This programmer has added the indispensable "ingredient" that had eluded many other programmers before him and that transformed completely the strength of the chess engine program that he modified, even if it would have modified this chess engine program only very slightly.
CONCLUSION:
Those who want to accuse a chess engine program to be a clone of a previous chess engine program are blind and they certainly have no idea at all how discoveries are made. In any invention or in any discovery or in any new chess engine program whose strength increases dramatically, it is not the amount or the number of changes made to what was previously known that matters, it is the result that is obtained by any amount or by any number of changes made to what was previously known that matters!
In other words, even if a programmer would change only a few lines of code to an existing chess engine program whose source code is available on the Internet, if the new program increases dramatically in its strength by increasing dramatically its ELO, then this programmer should be entitled to claim full credit for his accomplishment, NO MATTER THE FACT THAT HE CHANGED ONLY A FEW LINES OF CODE TO AN EXISTING CHESS ENGINE PROGRAM!!! That is the difference between making a discovery and being stuck with a not satisfying solution!!
-----------------------------------
If David Levy reads this posting, I strongly advise him to change completely the way the World Computer Chess Championship (WCCC) is done where the hardware is the same for each chess program competing.
1) Each chess program must play a match of 24 games against each opponent. This requirement means that the way the WCCC is done must change completely. As these 24 games against each opponent would take time, the way the WCCC is done should be changed.
One way to make this change is, for instance, by doing most of the games between each chess engine program BEFORE the official WCCC. This way, there would still be an opportunity for the WCCC to be held and it would not last longer but, at least, a reasonable number of games between each chess program would have been played so that the WCCC could be really a true test of the strength of the chess programs that are competing.
The number of games played during the WCCC is simply too low to determine the true strength of the chess programs that compete in this WCCC.
2) The chess programs that would like to compete in the WCCC where the hardware is not the same for each opponent should be entitled to participate only if they would have entered the WCCC with the same hardware for each opponent.
In 2010, Rybka only participated in the WCCC where the opponent did not have to stick to the same hardware for each opponent but he avoided participating to the WCCC where opponent should use the same hardware. He had, by a wide margin, the best hardware platform and therefore he did not have any problem winning the WCCC with unlimited hardware.
To conclude, there is NO need at all to prevent any chess program from participating to the WCCC with the same hardware for each opponent. Why? Even if a chess program entering the competition would have only one line of code changed compared to the original chess program, if this tiny change would show a dramatic increase in the ELO compared to the original chess program, this simple change should give the author the right to claim full credit for the dramatic ELO improvement he allowed the original chess program to accomplish, no matter how few changes he made to the original chess program. By simple honesty, the author could be completely honest by simply crediting the source code that he benefited from and/or the ideas from other chess programs that he benefited from. Same as the programmer of Rybka did!
Best Regards to All
MichaelIsGreat
---------------------------------------------------------------------------------------------------
---From "OpenChess • View topic - Attack of the Clones (ChessVibes)" at
http://www.open-chess.org/viewtopic.php ... 8&start=20
TCEC=TRUE WORLD COMPUTER CHESS CHAMPIONSHIP.
Post by MichaelIsGreat » Fri Feb 25, 2011 11:30 am
Hello to All,
Let us be clear on one particular point. The ICGA World Computer Chess Championships with same hardware for each opponent and the version for unlimited hardware are completely useless and totally irrelevant because the number of games played between each opponent is simply too small. And David Levy does not want to make the necessary changes to make these two tournaments relevant!
The ICGA could keep calling their two tournaments with a title similar to the World Computer Chess Championships but the reality is that these tournaments are a very bad joke and an insult to the title of "World Computer Chess Championship" and to the meaning of the word tournament. The current titles for the two tournaments that they organise have been modified a little.
The recent tournaments on "TCEC (Thoresen Chess Engines Competition)" at
http://www.tcec-chess.org/archive.php are much more meaningful and relevant than what the ICGA does with the World Computer Chess Championships! That is the current reality.
We could consider that THE TCEC TOURNAMENTS ARE THE TRUE WORLD COMPUTER CHESS CHAMPIONSHIPS. The Elite match between Rybka 4 and Houdini 1.5a was the true World Computer Chess Championship and, with 40 games played against each other, we had the opportunity to see very clearly which chess program was the best (Houdini 1.5a).
The various chess programs rating lists are useful but, as I have shown, a few of them have a dishonest hidden agenda of avoiding testing particular programs that could compromise their preferred pet programs (especially Rybka 4). That is what happened to the CCRL and CEGT chess programs rating lists where they purposefully avoided taking into account Houdini for quite a long time when Houdini was clearly the top chess program because Houdini was too strong and endangered their pet programs. When Houdini was taken into account in most of the other chess programs rating lists, CCRL and CEGT recently decided they were forced to take into account Houdini or face the ridicule of not listing the strongest chess program in their biased chess programs rating lists!
To conclude, let the ICGA waste other people's time playing the inquisitors. I am sure they will find many willing would be inquisitors, willing to read the code of other strong chess engine programs only to steal their ideas! The ICGA World Computer Chess Championships are completely irrelevant and useless and basically dead!!
The idea that a programmer would have to provide his source code to play at these bogus ICGA World Computer Chess Championships is very scary to me. I hope no serious programmers is going to allow such a ridiculous request.
Best Regards to All
MichaelIsGreat
---------------------------------------------------------------------------------------------------
And let me repeat it again to make it crystal-clear that your issues on this matter are a pure waste of time.
CONCLUSION:
Those who want to accuse a chess engine program to be a clone of a previous chess engine program are blind and they certainly have no idea at all how discoveries are made. In any invention or in any discovery or in any new chess engine program whose strength increases dramatically, it is not the amount or the number of changes made to what was previously known that matters, it is the result that is obtained by any amount or by any number of changes made to what was previously known that matters!
In other words, even if a programmer would change only a few lines of code to an existing chess engine program whose source code is available on the Internet, if the new program increases dramatically in its strength by increasing dramatically its ELO, then this programmer should be entitled to claim full credit for his accomplishment, NO MATTER THE FACT THAT HE CHANGED ONLY A FEW LINES OF CODE TO AN EXISTING CHESS ENGINE PROGRAM!!! That is the difference between making a discovery and being stuck with a not satisfying solution!!
Reading of few of your comments, there are people who seem to have no clue whatsoever on what an algorithm is in Computing and why it is necessary to copy it exactly as it was in the source code where it was used and why it is necessary to keep using it if there is no better algorithms existing!
I also remind you that the programmer of Rybka, Vasik Rajlich, honestly fully acknowledged that he used many ideas from the open source program Fruit whose source code is widely available on the Internet. At that time, basically no one accused Vasik Rajlich of having created a clone of Fruit!
And yet, with Houdini, you would like to accuse Robert Houdart of being a liar or of being dishonest when he states what he honestly and openly stated on his web site for all to see.
Are you, like children, jealous of Robert Houdart's amazing achievement with Houdini 1.5a? Well, it would seem so!!
Truly, you should really tone down your language because even the best cloners would be incapable of doing what Robert Houdart achieved with Houdini 1.5a!
Having said that, you have the right to disagree with me. But let us keep being civilized and friends, no matter our different points of view on this matter!!
Again,
congratulations to Robert Houdart, the programmer of Houdini 1.5a, as he went where no one went before by reaching the pinnacle among chess engine programs and by a wide margin.
Best Regards to All
MichaelIsGreat