However, my feeling is that a jury might well convict if it was presented with:
1. Evidence that indicates that Vas has a history of copying.
2. Evidence that an enormous increase in strength of the very weak Rybka 1.x series occurred shortly after Fruit was released.
3. The histories and credentials of the signatories of the letter to the ICGA.
4. Well prepared and properly annotated documents that clearly show areas of overlap between Fruit and Rybka and Crafty and Rybka.
First, note that Polish civil courts are almost exclusively "bench trials", so there is no jury (one can also note that there are no "standards of proof" in Polish civil law, and the judge uses whatever he thinks is most reasonable for the case at hand). As noted
above, Polish law is perhaps the most stringent in the world regarding copyright, allowing triple damages in some cases, and additionally allowing the plaintiff to require a defendant found guilty of culpable infringement to "donate" not less than twice any damages to a "Creative Fund".
I think #1 and #4 will be the main exhibits for the plaintiff(s). #1 follows quite irrefragably from the Rybka 1.6.1 evidence with Crafty. The fact that Rajlich was "found out" concerning this (rather than making an open admission) is also a factor here. For #4, the "substantial similarity" between the evaluation features
Rybka 1.0 Beta and
Fruit 2.1 is quite evident. It does not seem to be precisely at the "code" level (even in a transliteration, from what I can tell so far), but it is an almost exact match at the next level of abstraction (see the opinion in the
PDF the Jeremy found concerning such abstractions, Section II-D in particular). There is also the fact that evaluation functions even written by the same person (such as Kaufman with Rybka/Komodo and Romstad with Gothmog/Glaurung) tend to have many more differences in features than R1/Fruit. This should all be quantified more for the civil copyright infringement case -- with the ICGA Panel, most programmers (chess and otherwise) already understand these things don't "just happen" accidentally.
I'm not sure which "Rybka 1.x" series is meant in #2.
![Very Happy :D](./images/smilies/icon_e_biggrin.gif)
Even if the comparison is to the post-Fruit Rybka 1 series, Letouzey has pointed out that he was busy with other (non-strength) features in the second half of 2005, and that Fruit 2.1 was almost a development snapshot, with some room for rather easy Elo gains. For the purposes of damages, it might also become relevant whether most "Rybka 1" subscriptions were sold soon after the
Rybka 1.0 Beta release (when the Elo advantage over
Fruit 2.1 was not so large), or later in the Rybka 1 schedule (but as I say, Letouzey rejects "strength" as the dominant point, so it would be more of whether later Rybka 1 versions had less copyright infringement).
#3 would probably have its most impact with respect to "credibility" of Rajlich (I think I pointed out previously that a good lawyer could easily make hay of some his more loopy statements over the years, such as not having any back-ups, and his inanity in "node-counting" [particularly claiming that Shredder/Hiarcs might be doing something similar]), though it must be pointed out that many of the 16 signatories probably have not studied the evidence sufficiently closely to sign an affidavit.