jdart wrote:ICGA can set up whatever rules they want. But I'm not sure a tribunal is going to help things. I think a program that would get judged in this way would probably just not participate.
Also, I agree there are some difficulties with programs that are closed source. Technically these could have improperly used some parts of another program and that may not be detectable. So you are really punishing inexpert "cloners" it seems to me.
And what if - hypothetically - someone were to modify an existing engine that had a permissive license (not GPL)? Would this be acceptable to ICGA, or would it just be considered a non-original work? What if the author made it 200 ELO stronger?
You hit the nail on the head.
Assuming you define the clone criteria well enough, and then require closed source tournament participants to reveal source code for inspection, and settle on the conditions for such disclosure and examination, you still
must convince closed source particiants to submit to this regime.
The ICGA is a self-policing type of tribunal. Its authority is as wide and strong as the number of people who voluntarily submit to its regime. Nobody can be forced to enter, of course! If it is to have any teeth, you would need to have the big commercial chess software companies back it and support it, along with other prestigious, recognized computer chess leaders and contributors.
If the ICGA tribunal does not have a well-defined set of published criteria for determining impermissible derivations, then it's back to the Law of the Jungle.
If the ICGA is unwilling to enforce its rules by requiring suspected closed source authors to submit their code for examination, then it's back to the Law of the Jungle.
If prestigious closed source engine authors have no incentive to voluntarily submit to the ICGA regime in order to win its prize, then the prestige and authority of the ICGA is itself put into question, and its back to the Law of the Jungle.
If a prestigious closed source engine author believes he can make as much money without an ICGA title as with one, because he believes the Internet will be all the advertising he needs, then the regime is not much of a regime, and it's back to the Law of the Jungle.
If prestigious top closed source engine authors drop out of the ICGA, then the tournament becomes one for open source authors and second tier closed source authors, and the top chess engine community becomes all the more balkanized. And, back to the Law of the Jungle.
Thus, the ICGA has to measure whether losing the participation and defection of the top engines, like Rybka, is worth it. Does the ICGA really want to hold a tournament that does not have Rybka participating? Probably not. Will Chessbase sponsor ICGA if it demands that Rybka submit its source code for examination in order to participate?
Back to the Law of the Jungle.