Even when cheating's not cheating, there's cheating!

IYSWIM

According to this article (about Rybka possibly being plagiarized), cheating has NOT been proven, the article also suggests the possibility of a "witch hunt". The article informs that in order for Vas Rajlich to prove his claim that the work is all original, he needs to reveal his source code, which is a trade secret and therefore a lose-lose situation for him.

Vasik Rajlich, the designer of Rybka, was a regular customer of mine when I owned Chess Express in Ann Arbor(1997-2000). He used to use my store as a meeting place for his lessons. He took one 4-hour lesson every week from Ben Finegold (now a GM). I was lucky to be able to witness most of these lessons. Finegold's teaching method was very demeaning to the student, but Rajlich put up with the put-downs and soaked in whatever he could. Vas always had a friendly demeanor and a very strong work ethic.

I hope that he is acquitted of these accusations!

I hope that the outcome is a fair one. *If*, and it's a huge if, he's innocent, I hope that can be demonstrated - all he needs to do is undergo a clean room examination of the code by mutually agreed upon third party, that's not "revealing" his code. There's no reason for any source to leave the clean room. If he plagiarised, then I hope he returns his trophies and prizes, and then compensates those who feel defrauded by his insistance that what he was selling was 100% original.

As someone who spends a lot of his time (a large chunk of today, for example) poring over compiler-generated machine code, the arguments in the papers linked to are very compelling. As a comparative linguist once said to me - one single slice of cucumber cannot be made to balance on its side, but you pile enough of them beside each other, and they happily will.

And the 'revealing his code' argument is bullshit anyway. He's revealed it. Everyone who runs is has a copy of all of it. OK, in compiled form, but it's *exactly* the same algorithms as the more human readable source code form that he's so worried about "revealing". Did he not notice how easy it was for people to reverse engineer his algorithms? There is nothing secret to anyone who knows a bit of x86 assembly language (and it only takes a bit, as x86 compilers tend to generate a fairly restricted stream of machine code) or who has access to a C decompiler.

http://www.multiupload.com/VF2VBS01QT

An interview with Vas.

476 Megabytes? He must have a lot to say.

I thought I'd bump this as chessbase, chessvibes have updated coverage.

Sorens Riis (moderator at the Rybka forum) wrote a series of articles which chessbase (sellers of rybka) publish calling it a miscarriage of justice.

This has sparked discussion in various forums and the publishing of rebuttals from the accusers.

Pointers to the ChessBase articles and the rebuttals can be found here: http://icga.uvt.nl/