FIDE Ethics Commission ruling on ICGA/Rybka complaint

[hgm]

If Marco et al. entered it, yes of course it would be theirs. If the ICGA enters it just to make the competition more "complete", then I'd say no. In my view you cannot force a title upon someone that is not interested in the title in the first place. It would anyway be meaningless.

The actually proposed case, however, was what would happen when you would enter your syzygybases-enabled version. You would obviously be co-author of that, although perhaps you would consider your contribution only minor. But the whole case would then boil down to a disagreement between authors on whether to enter, and whether the ICGA would in such a case side with the author that wanted to participate, or the one that would not want it. It seems natural that the ICGA in that case would make the decision that serves their own interests best.

The ICGA could for instance adopt a policy that would give priority to the wish to enter, and allow authors that explicitly inform ICGA that they want to have no part in this, in which case they would not share the title. So it would be that (say) "Ronald de Man and Marco Costalba won the WCCC with Stockfish, which was co-authored by Tord Romstad and Joona Kinski". (Fat chance! )

[michiguel]

Note how Vas had to bear the responsibility for his decision to enter those tournaments. Such responsibility cannot be forced upon Marco/Joona/Tord without their explicit agreement. Arguing that the GPL implies such agreement is nonsensical.

Note that the GPL is also very explicit in denying any liability for the usage of the program. So Tord et al. could never be hold accountable for any rule violations, in the same way that Vas was held accountable. This liability would fall solely on the person(s) entering Stockfish.

So the ICGA would have to say: we organise a tournament for programmers in the usual way, BUT we'll add some engines to the mix in our own name. If the tournament has already been announced, the ICGA should in my view first seek the agreement of those that already entered. Participants may rightfully expect to play against entities that fully comply with the rules. (How else could the ICGA have felt obliged to investigate Rybka after receiving a complaint. At that time it seems even non-participants could complain that ICGA rules were violated. So those rules were not considered to be at the discretion of the ICGA.)

Or the ICGA could say: "We have this entry of Stockfish, by a guy who made a significant contribution to the code, and we judged this contribution large enough to accept him as someone who has the right to do this. After all, he complies with all the rules we published in advance, concerning mentioning the other authors and having their legal consent to use their code!"

I'm curious. Suppose Rybka had still not been surpassed by other engines. Would you or anyone else have wished that it was entered, by whatever means, into ICGA tournaments?

I would have no objection whatsoever against participation of a 'clean' Rybka. Even a Rybka that is so heavily based on Fruit that it is to be considered a derivative can participate, as far as I am concerned, provided that it lists Fabien as one of the authors. There would be a problem, however, for a Rybka version that violated the GPL. In that case the use of the Fruit code it included would not be 'legally used', and it could not be made legal without publishing the entire Rybka source (which Vas would never do). So I don't think there would be any way to enter that.

You insist on missing the point. Very well.

Or perhaps you had no point, or one that I just didn't agree with?

There is not such a thing as a Rybka program violating GPL. R was not penalized for that. Also, a program that uses a GPL code in the competition, does not have to be open source as long as it does not distribute it, and as far as ICGA is concerned, the competition is not about distribution. The wording is legal to use, and it means of course legal to use in the tournament since ICGA has no jurisdiction outside of it. GPL is always legal to be used, but it may not be to be distribute it.

In other words, under current rules, someone can enter a modified version of SF w/o even publishing the source code.

Miguel

[hgm]

That two chess programs from the same author cannot be entered in the chess competition. It is not explicit.

OK. But a registration form that is used for both the WCCC and the Olympiad cannot really be used as evidence that this is possible. One would really have to look up the WCCC rules. I know the ICGA board thinks it is not allowed, from private e-mail correspondence. Presumably they have some basis for this.

[syzygy]

The actually proposed case (...)

Something must really be bugging you.

[Roger Brown]

However, despite my keen interest in it happening, I suspect that the title of Greatest Lover in the World will never be forced on me.

So, it is equally meaningless to attribute - or even desire - titles where it will never happen as well.

The nice thing of titles is that everyone can make them up.

The official ICGA World Champion Computer Chess
The offical <insert> Greatest Lover in the World

Just need to find one organisation or person willing to award you the title.

If Roger wishes to be innovative I am sure this could be arranged by the ICGA in Leiden. Dr. Levy has written a book called "Sex with Robots."

Harvey Williamson!!



Later.

[hgm]

There is not such a thing as a Rybka program violating GPL. R was not penalized for that.

According to Bob it was also not investigated for that. I am just exhaustively listing all possibilities, hypothetical or not.

Also, a program that uses a GPL code in the competition, does not have to be open source as long as it does not distribute it, and as far as ICGA is concerned, the competition is not about distribution. The wording is legal to use, and it means of course legal to use in the tournament since ICGA has no jurisdiction outside of it. GPL is always legal to be used, but it may not be to be distribute it.

This is a good point. It would pose a major ethical problem for ICGA, though. It would look quite hypocritical to me to allow participation of someone who openly admits he is stealing from another member, by selling the same binary as he participates with.

In other words, under current rules, someone can enter a modified version of SF w/o even publishing the source code.

Sure. I even considered doing that, (e.g. equipped with on-the-fly EGT generation), just to outrage people. I would then not be distributing/selling that same binary, though. (That would be part of the fun.)

[hgm]

Something must really be bugging you.

How so? I just saw a great opportunity to make a lot of people happy, and pointed it out. It was just unfortunate (for them) that it depended on you.

[Rebel]

Note how Vas had to bear the responsibility for his decision to enter those tournaments. Such responsibility cannot be forced upon Marco/Joona/Tord without their explicit agreement. Arguing that the GPL implies such agreement is nonsensical.

Note that the GPL is also very explicit in denying any liability for the usage of the program. So Tord et al. could never be hold accountable for any rule violations, in the same way that Vas was held accountable. This liability would fall solely on the person(s) entering Stockfish.

Sure, clear for us insiders.

Claims that are proven to the satisfaction of the ICGA will result in sanctions being imposed by the ICGA on the offending persons, who will be named and shamed on the Internet. -- © David Levy

Surely it will backfire on Marco/Joona/Tord, the Stockfish name will be smeared in the perception of the vast majority of non-insiders.

[hgm]

Do I understand correctly that you accuse Marco/Joona/Tord of having stolen code in Stockfish?

[syzygy]

Note how Vas had to bear the responsibility for his decision to enter those tournaments. Such responsibility cannot be forced upon Marco/Joona/Tord without their explicit agreement. Arguing that the GPL implies such agreement is nonsensical.

Note that the GPL is also very explicit in denying any liability for the usage of the program. So Tord et al. could never be hold accountable for any rule violations, in the same way that Vas was held accountable. This liability would fall solely on the person(s) entering Stockfish.

Sure, clear for us insiders.

Claims that are proven to the satisfaction of the ICGA will result in sanctions being imposed by the ICGA on the offending persons, who will be named and shamed on the Internet. -- © David Levy

Surely it will backfire on Marco/Joona/Tord, the Stockfish name will be smeared in the perception of the vast majority of non-insiders.

Even worse, their own names would be smeared.

That it would be unjust is beyond dispute, whatever the license. Pretending toward the public that Marco/Joona/Tord participate in an ICGA tournament, which would be the result of letting Stockfish compete unless it is made absolutely clear that SF's participation is unauthorised, would violate their personality rights. They get to choose in which events they participate, not the ICGA and not anyone else. They don't get to choose what other people do with SF, but this is not about SF but about their own rights.