Uly wrote:It hasn't been proven that Rybka 3 or Rybka 4 broke ICGA rules. They just went "it's Vas's responsibility to prove those versions are innocent". [not a direct quote, AFAIK]
For that matter, it wasn't even proven that R3/R4 even competed in an ICGA event [Rajlich never addressed that question].

In essence, Rajlich was declared to be in default of Rule #2 for 2008-10 by (willfully) failing to respond to the inquiry regarding the originality of his entries. If you want a more "legalistic" argument: the history of prior incidents for Rajlich combined with the typical direction of engine development sufficed as
prima facie evidence, and he chose not to rebut it.
In short, if investigating R3/R4 was a low-cost ordeal, it would have been more likely for the Panel to attempt this. As others have noted, at some point it becomes the responsibility of the author to address the situation, and the Board determined that this point had occurred in the given case. Obviously others disagree concerning this.
Jeremy Bernstein wrote:The fact that he [Rajlich] was/is so contemptuous of the entire proceeding doesn't really help with the sympathy/mercy vote, either.
Let me just say that the R161/R1/R232a analyses were a lot of work for all involved [including those who read/debated the documents that were produced]. R3 might have been easier for me personally, due to my previous investigation with IPPOLIT, but I can't think that the Panel would/should just take my word that it was "clean" with respect to Fruit [if for no other reason, due process toward the nominal "plaintiff" Fabien Letouzey]. So another document would need to be produced, and read/verified by various Panel members. I'm not sure all of them have purchased R3/R4 [current commercial products], and particularly for the latter I would be unwilling to discuss it too closely in any event. As RV indicated, it's also much easier to show plagiarism than to be sure that none occurred via an investigatory process (it would likely be somewhat easier via a dialogue process). Finally, R4 would firstly be a large amount of work to do a proper RE, and then it seems [from recent RV comments] that one might have to try to determine its relation to the IPPOLIT history -- and then it might turn out that R4 wasn't really that close to the version that competed in Kanazawa.
An alternative way to proceed (chosen by the Panel) would be to address only R232a and previous. If Rajlich responded to the Board, some of the questions he was asked to address concerned R3/R4, and the Board would then choose how to proceed from there. If he didn't respond to the Board, there would be little incentive to continue the investigation in any event [essentially, if he doesn't care, why should anyone else? -- particularly if it takes hundreds of man-hours
in toto?].
When I was in Lille on Friday, the gathered programmers agreed that Rajlich would likely have received much less punishment [perhaps very little indeed] had he not followed the path of silence, denoted by DD in the Report as "another manifestation of disrespect" to the community of games programmers as a whole.