Tweet
I've been reading this four part series on chessbase.com's website called
"A Gross Miscarriage of Justice in Computer Chess". You can find it at the following address:
http://www.chessbase.com/newsdetail.asp?newsid=7813
As many of you know Rybka stands accused of being a copy of Fruity. Now Rajlich, Rybka's author, has said from the very beginning he studied Fruity very closely and took ideas from it (Fruity started off as an open source program)..but the accusation here seems to be that he literally cut and paste Fruity's code.
I should note the chessbase article's author, Soren Riis, says he is a Rybka forum moderator, and I am not sure if to consider that a personal connection to Rybka and IM Vasik Rajlich and so, the reader is left with "some" distrust of his objectivity. However, at least one of the arguments he presents will leave any open minded reader with questions about how the ICGA conducted its investigation.
During the investigation there was a committee of (I think) 34 people who were presented research from 3 of its members. After the presentations were given the members voted to strip Rybka of its world championship titles under their banner. A media storm then followed basically labeling Rybka a product of cheating/stealing.
Now I'd like to reiterate a single key point made in the chessbase article just to get people thinking about this issue:
During one of the presentations a paper was handed out to the committee with lines of code under two columns. The left hand column was labelled "Fruity" and the right hand column was labelled "Rybka". As you look down the two columns in comparison you see many of the paragraphs on both side are exactly the same or very similar.
One thing that struck me is that to the many lay people on the committee, heavily dependent on the more experienced researchers, this was the critical moment in swinging their vote to "guilty". Here they had something they could understand...two things side by side that looked similar.
But it turns out that the code on the right under the heading "Rybka" was completely fabricated for the presentation. There were many possible lines of code that would output the results of a decompiled version of Rybka they had, so they purposefully wrote in code that would be visually similar to what was on the Fruity side...Even though there are many possible permutations which look very different to Fruity which would have given the same result.
In Soren Riis' words, "There are many different ways to write functionally equivalent code that looks nothing like the Fruit code. Dr. Miguel Ballicora posted source code to the Rybka forum that generates Rybka and Fruit PSTs but looks completely different from Fruit code."
Now it may turn out that Rybka is a pirated version of Fruity. There are many other arguments back and fourth on the internet of course and I have not read all of them. There is an entire website dedicated to showing Rybka's guilt now (http://rybka-is-fruit.wikispaces.com/) and I have not gone through all of its claims to weigh them against chessbase's four part series...which was in itself a fairly long read. (And Im sure Ive missed or misunderstood at least some of its details).
What I can say is that one little part of the ICGA presentation does very little to demonstrate Rajlich's guilt. Quite the contrary, it makes the IGCA investigators seem dishonest.
"A Gross Miscarriage of Justice in Computer Chess". You can find it at the following address:
http://www.chessbase.com/newsdetail.asp?newsid=7813
As many of you know Rybka stands accused of being a copy of Fruity. Now Rajlich, Rybka's author, has said from the very beginning he studied Fruity very closely and took ideas from it (Fruity started off as an open source program)..but the accusation here seems to be that he literally cut and paste Fruity's code.
I should note the chessbase article's author, Soren Riis, says he is a Rybka forum moderator, and I am not sure if to consider that a personal connection to Rybka and IM Vasik Rajlich and so, the reader is left with "some" distrust of his objectivity. However, at least one of the arguments he presents will leave any open minded reader with questions about how the ICGA conducted its investigation.
During the investigation there was a committee of (I think) 34 people who were presented research from 3 of its members. After the presentations were given the members voted to strip Rybka of its world championship titles under their banner. A media storm then followed basically labeling Rybka a product of cheating/stealing.
Now I'd like to reiterate a single key point made in the chessbase article just to get people thinking about this issue:
During one of the presentations a paper was handed out to the committee with lines of code under two columns. The left hand column was labelled "Fruity" and the right hand column was labelled "Rybka". As you look down the two columns in comparison you see many of the paragraphs on both side are exactly the same or very similar.
One thing that struck me is that to the many lay people on the committee, heavily dependent on the more experienced researchers, this was the critical moment in swinging their vote to "guilty". Here they had something they could understand...two things side by side that looked similar.
But it turns out that the code on the right under the heading "Rybka" was completely fabricated for the presentation. There were many possible lines of code that would output the results of a decompiled version of Rybka they had, so they purposefully wrote in code that would be visually similar to what was on the Fruity side...Even though there are many possible permutations which look very different to Fruity which would have given the same result.
In Soren Riis' words, "There are many different ways to write functionally equivalent code that looks nothing like the Fruit code. Dr. Miguel Ballicora posted source code to the Rybka forum that generates Rybka and Fruit PSTs but looks completely different from Fruit code."
Now it may turn out that Rybka is a pirated version of Fruity. There are many other arguments back and fourth on the internet of course and I have not read all of them. There is an entire website dedicated to showing Rybka's guilt now (http://rybka-is-fruit.wikispaces.com/) and I have not gone through all of its claims to weigh them against chessbase's four part series...which was in itself a fairly long read. (And Im sure Ive missed or misunderstood at least some of its details).
What I can say is that one little part of the ICGA presentation does very little to demonstrate Rajlich's guilt. Quite the contrary, it makes the IGCA investigators seem dishonest.
Comment