Dragon Age Premium Content

The battle of the clones

a couple of weeks ago the International Chess Master David Levy, wrote an interesting article about what is happening in the world of game-playing programs science. Notably, Levy became famous when starting a bet on where he said science could not create a computer that will win in a formal meeting during the next ten years. The bet was known by other scientists and chess and increased the amount wagered. The end of the story is that at the time, Levy won the bet, to my knowledge, never been able to collect.

But the interesting thing about this is that the Scottish Levy has been involved throughout his life with the world of computers and programs that play chess. Now has come out with an article called "The Cloning of Chess Engines, a phenomenon that has begun to rise with concern for two reasons. The first is that chess programs can be a good income if they play well. Amateurs and professionals are now looking for this type of software because you can analyze the games with much more precision, and second, because many programs that play with great force, have been made available to the community computer, complete with source code available, ie that the author or authors of some other very good at playing chess, have donated their work to the community and anyone can then study how these little programs come to play so well . A International Association for Computer Games (ICGA, for its acronym in English), who sanctioned tournaments that take place between computers. Each year the best program, in a fierce competition between computer processors, becomes world chess champion cyberspace. This title is suitable in terms of trade since the winner has a spectacular sales pitch: buy his chess program champion between machines world! ... By itself, the cloning of other programs or code be incorporated into the third for a "new" program can sometimes fall into the most basic plagiarism.

For example, in 1996, a German programmer, Herr Langer, enrolled her Quickstep program, which had a sensational start in the competition between computers. However, Richard Lang, then programmer Mephisto chess computers revealed that Quickstep interface was identical to its own program. The case was investigated and interrogated Langer, who rejected all the accusations. However, it was the Richard Lang, who showed an error in your code, which is reproduced exactly the Quickstep program. Langer then confessed and was expelled from the tournament. All in all, the German was a honeymoon and his wife saw it and disqualified unmasking. Sad story.

Another more recent case was the SquarknII program, FIDE Master Jonadry Gonzalez Espin, of Havana, Cuba, who participated in the Computer World Championship 2010 in Kanazawa, Japan. After making many efforts by the organization to which Gonzalez could attend the fair, due to visa problems, which were finally resolved, it was found that their software was a clone of Robbolito 0.75g3, with only three values \u200b\u200bchanged at all code. Espin was disqualified and punished by their inclusion in any event ICGA for life.

But the worst problem is the case of Rybka, it is argued, is a copy of an open source program called Fruit. Rybka, programmed by International Master Vasik Rajlich, has won the world championship machines in 2007, 2008, 2009 and 2010. For three years there have been rumors in the world of computerized chess Rybka is based on the Fruit code programmed by Fabien Letouzey, who finished second in the 2005 world championship in Reykjavik.

To consider whether Fruit has greatly influenced the Rybka code, we begin by examining the history of both programs. The results of these world championships in chess programs, in February 2004, Rybka ranked No. 53 of 54 possible. In his appearance in 2004, Fruit had basic evaluation functions and search, but each new version was improved about 100 points of rating, so in version 2.1, even open source, it was a very strong program. The latest version of Fruit is no longer open presumably because the author does not want to generate clones with their code.

According to Wikipedia, Vasik Rajlich Rybka started work in early 203. The first beta version of Rybka made its debut in December 2005. At that time the program did in the tournament Paderborn Computer Chess, 5.5 of 7 possible points, up from Gandalf, Zappa, Spike, Shredder and Fruit.

So, Rybka won his first major result in December 2005, six months after the start of open source Fruit 2.1. And while there are similarities, many experts believe that computer chess Rybka's success is due to the source code of Fruit.

But all the above were only rumors. However, more recently the rumors have become stronger allegations, made by experts and supported by evidence that at least superficially, seem to prove them right, because of the nature and volume of evidence. Levy, however this is book your own opinion.

First, Zach Wegner, who is developing a new version of Zappa, with permission from Anthony Cozzie, the original developer of this system. Wegner attended the 2010 world championship with its program Rondo. Their arguments are as follows:




Rybka assessment has been subject to much speculation since its occurrence. To put it in simple terms, the evaluation of Rybka is virtually identical to the Fruit. There are a couple of important changes, however, that should be considered in this argument:

The most obvious is the translation the data structure of bitboards Rybka (boards of bits). In some instances, the evaluation of bitboards, in regard to the pawn may be slightly different from the original software (Fruit). But the high-level functionality is equivalent in these cases. The changes seem to be with respect to a natural representation of bit boards.

logic ends and draws in Fruit has been replaced by the final tables Rybka. This has already used almost all programs.

All weights on the criteria should be established specifically to play by hand. Due to unnatural values \u200b\u200bof Rybka, apparently they are organized automatically. However, there are a couple of places where the original values \u200b\u200bof Fruit are the same in Rybka.

Vasik Rajlich, author of Rybka has rejected, since it was said that his program was Fruit a clone of these allegations and hehco wrote in 2007: "The speculation Osipov (creator of other chess program) is incorrect. Rybka is and has always been original code, with the exception of several fragments of low-level are in the public domain. "

Shortly thereafter, Vasik wrote to Levy, saying: "The source code is original Rybka. I used many ideas of Fruit, as I have said many times. Rybka Fruit and used all sorts of ideas common in computer chess. " But when asked if his answer meant that the source code of Rybka 1 was all original, Vasik replied:" All versions of Rybka are original the sense that I always wrote the source code (with the exception of several fragments standard low-level) "

So, the doubts persist, but it is clear that it has cast a shadow over Rajlich Rybka and his creator. How probra that Rybka is a copy and not an original work? This is something that other developers are looking to find to discredit and perhaps eliminate one computerized powers of most successful programs in this field.