Wisdom is knowing how

   little we know -- Socrates

About Company

Started in 1980, retired in 2004 REBEL was baptized into ProDeo, latin for gratis according to Dutch tradition.

Other Information

 

    Rybka-ICGA fiasco

 

 

    External links

 

 

 

 

 

Marcel van Kervinck, Netherlands

 

in reply to David Levy

 

I would like to draw your attention to two points. First: in part one of his recent interview with ChessBase David Levy states the following: "Your question implies that Rajlich might only have copied a small and insignificant amount of Fruit, while the conclusion of the ICGA investigation was that a lot of code was copied." This does not represent my opinion as an active and voting panel member. Not now and not at the time of the investigation.

 

Code (=literal elements) was proven to be copied in a Rybka version that didn't compete in ICGA tournaments. But that didn't violate rule 2 because it didn't play in ICGA tournaments.

 

Formulas (=non-literal elements) was proven to be copied in Rybka 2.3.2a, which did compete in an ICGA tournament. Proven there was that more than just "ideas" were taken, but many Fruit-specific design choices that are rather arbitrary by themselves and unnecessary. But not proven was code copying in this version.

 

As such VR [ Vasik Rajlich] still violated the letter of the first sentence of rule 2. The rule's first sentence doesn't talk about "code", but about "program" and "work". An important distinction to make in representing what happened. (The logic of the later sentences of rule 2 makes it clear that "code" must not be taken as a synonym for "program" or "work", as is sometimes done in other contexts)

 

I therefore request you to rectify the above statement at ChessBase.

 

You don't need it for your case and in your second part of the interview you are much more precise in your wording: there you say that Vasik Rajlich "used code derived from Fruit". I can accept that phrasing if it is intended to mean "Rajlich competed with code that he wrote, but which was derived from Fruit's design".

 

And honestly, if that is what it means, that is how it should be phrased because it is too easy to misread otherwise. The most likely reason for the evaluation overlap in my opinion is that VR designed Rybka's evaluation first-most as an as faithful as possible replica of Fruit's without copying any lines of code in that process. From that point on, he worked to improve this evaluation.

 

But still many traces of Fruit's formulas are clearly visible in version Rybka 2.3.2a. Second, I would also like to inform you that with the knowledge I have today, I would have voted differently in the investigation process. I will explain why: I'm not sure anymore that information provided for making a conclusion about the breaking of the 'spirit' of rule 2 was sufficient. In particular, the entry forms of prior entrants were not considered and should have been. I feel the investigators were insufficiently instructed about how prior entrants have declared their origins.


My specific interest is in the application form of the original entry of the Deep Thought team.

It was well-known at the time that DT was based on the design of Belle, and that Ken Thompson had given permission to the DT team to use that design as a template. It is not known wether this was written on the entry form. Why does this matter, you would say, because Fabien Letouzey should have given permission anyway. I will explain:


It was at the time well-known, by his own statements on his website, which was referenced on his entry form BTW, that Rajlich had studied Fruit extensively and had taken many things. And in my current opinion, Vas Rajlich did have Fabien Letouzey's permission to use Fruit's formulas in a WCCCh program.

 

The GPL license does allow that, unlike your statement in the ChessBase interview. The GPL can protect reproduction only. Entering a derived work in a tournament is not reproduction by the EU's copyright directive for software, which was applicable in Amsterdam. The fact that commercializing Rybka 2.3.2a might break the GPL is a civil matter between FL/FSF and Vasik Rajlich, and the ICGA should be no party in that.

 

I regret I didn't persist more at the time the comparison of Rybka's entry form with that of DT's. Because my vote should have been affected as follows:

 

  1. If DT's team did declare, on the entry form, that Belle's design was used, I would have voted Vas Rajlich guilty.
      
  2. If DT's didn't declare such on the form, I would have voted "not guilty, because there is no difference between DT's entry and VR's. Rule 2 should be applied the same."
      
  3. If DT's entry form was not to be found anymore, I would have voted "not guilty, because the information was insufficient to make an informed decision."

 

Copyright « 2012 Ed Schr÷der