Via ChessViibes: The International Computer Games Association claims that Vasik Rajlich lifted the code of Crafty and Fruit.
The claim is not a new one (it’s been discussed on rybkaforum for some time), but ICGA bans Rajlich for life and strips Rybka of the 2007-2010 titles. (I hope Rybka takes the news well.)
The Rybka-Fruit comparison is damning. They show examples where the code is very close, and the variable and structure names are identical.
Caught!
Hmm – when you said an entrant had been disqualified from a computer tournament, it conjured up images of a computer cheating by consulting a human. 
Bill Smythe
I haven’t read Rybka forum in a couple of months, but for a while, I was following the debate on the Rybka forum.
It doesn’t suprise me. In fact, many of the forum regulars came to the same conclusion around February.
The reason is that since the inception of the Rybka forum, it was heavily censored to talk about any program that the forum moderators felt was a clone of Rybka (regardless of which version). Then in Feburary, suddenly there was almost nothing banned on the forum. (That is named supposed “clones” of Rybka).
I can say without a doubt that in the 3rd week of Jan, it was banned to speak the name the Houdini chess engine. I’d heard of the chess engine at the local chess club, and mentioned it on the Rybka Forum, and it was promptly banned/removed. When I complained that there were other threads that talked about Houdini… one of the moderators retroactively went back through the forum and deleted any and all posts regarding Houdini chess engine.
Anyway, it wasn’t long after that, when discussions about Rybka being plagurized from Fruit became a hot topic again. Although its never died down completely.
I wanted to mention. Someone on another forum asked why it took so long (5+ years) to come to the conclusion that Rybka was derived from other programs.
Not really sure why, but clearly, this spring, legal action was taken in Europe to force VAS to hand over the actual source code of Rybka.
If you look at the “evidence” that you can download, it clearly shows source code: something VAS was very reluctant to hand over. Thats why I’m assuming that he’d only do that upon a court order or some similar reason.
I’ve used Rybka for several years, and I know that honorable folks worked on the project: Larry Kaufman on the valuation function, for example. It’s sad.
I’ve switched to Houdini, which works nicely under ChessBase and Aquarium.
Can’t beat the price of Houdini, either.
The investigation involved looking at disassembled code from the rybka binaries not looking at source code for rybka, and that is a very time consuming process.
Edit: Apparently it won’t end with the ICGA as the complaint has been forwarded to the FSF for investigation of GPL violations.
That was done for the Rybka - Crafty comparison, it was from a really old version of Rybka, and from a quick glance it wasn’t obvious to me that they were from the same origin. I didn’t see the point they were making there.
But the Rybka - Fruit comparison looked like C code to C code, with similar formatting of the code and identical data structure names.
The point with regards to crafty was that he copied over code into Rybka that couldn’t possibly need as it was a special case dealing with a bug cocerning en passant in an obsolete form of tablebases that Rybka does not use.
Look at section 2.1 of chessvibes.com/plaatjes/rybk … gation.pdf . It is all based on executable file analysis, and they used some techniques to take the disassembled code and interpret it into c code.
Just finished scanning the Rybka forum. First impression was the number of “hero worshippers” attacking the messengers. Hmmmm. Deja Vu.
That doesn’t sit well with those that have a lot of forum posts over there. The house of cards that Vas built is quickly becoming a house of cards built on a shaky foundation.
Indeed, I can’t read further after this
To state the obvious … Seems to me that Bob and Fabien should have been credited as co-developers of Rybka, and then according to their comments, the judges would not have been nearly so harsh.
It would appear that Chessbase is becoming selective in it’s news reporting when it could effect their bottom line. Over 24 hours and still no story there, which is funny considering Frederic Friedel and their person in charge of the news Albert Silver both applied and were accepted to take part in the ICGA panel that did the investigation.
They need to talk to counsel, as they (innocent AFAIK) may now have significant economic exposure.
It wouldn’t be the first time that Chessbase was accused of being biased in its news reporting.
True, but that exposure would be due to continued distribution of a product that is potentially in violation of the GPL, not exposure for reporting the story.
Don’t assume that laws in this or any particular area of commerce work exactly the same in Germany or other countries as they do throughout the US. (And even in the US, laws vary from state to state, although there is something called the Uniform Commercial Code to help keep this from varying a lot, if I understand this area correctly.)
Until put on legal notice, it would seem to me that resellers would not be liable, and promoters would not be liable, either. In any case, if commercial disputes are going to follow this beyond ICGA and other event organizers, there may indeed be a costly and lengthy amount of chaos in the field of chess engine sales.
Ownership issues are commercial matters; authorship is about integrity and prestige, too. Throughout all of this, it appears that scientific progress has been occurring, too, and none of this detracts from that separate component, does it? From the perspective of perhaps beleaguered chessic John Henrys, these are all variants on a basic machine that drives spikes into the ground to lay railroad track.