FIDE Ethics Commission ruling on ICGA/Rybka complaint

[APassionForCriminalJustic]

Well having Komodo and Stockfish both in the up and coming ICGA computer-chess tournament would certainly make it feel more like a real world championship. Could not anyone enter Stockfish into the tournament given that it is open source, and publicly available?

Hello Adam,

Should you mean anyone not an author of the engine, say me for instance, the answer is no.

The Stockfish authors have been clear about not entering the engine.

Open source does not mean that the engine has no owner, or that there are no rights over its use...

Later.

Ah, that is what I thought. Thank you...

[syzygy]

I can't imagine why the SF guys don't enter that event. They don't even have to attend. Then they could win, and actually have another title in hand.

You would let them enter?

[hgm]

Of course they could. Why do you doubt it?

In fact I suppose you could enter it, based on the fact that you made a non-trivial contribution to the code through the tablebase probing.

I did explicitly pose this question to the ICGA a few month ago, how they would deal with collective projects. The preliminary consensus of the replies was that for a collective open-source effort like Stockfish, the GPL would be a valid substitute for written permissions of all contributors. The person entering it should have made a significant contribution to the code, though. So they would not accept me to enter it after editing (say) the name, even though I could claim that the GPL implies permission of all other authors to do so.

There was some doubt on whether participation of Stockfish would exclude anyone who has submitted a patch from participating with an engine of his own. The sentiment was that for a minor patch this should not be the case, but of course strict interpretation of the current rules would forbid it. But in your case this would not be a problem.

So why don't you enter it? I would even be willing to pay the entrance fee for you, if that was a problem. And Hans would probably be happy to operate it. Wouldn't that be fun?

[Henk]

That event ceased after the 1994 Edition, sadly, as the ACM, like everyone else not directly involved in computer chess, considered that "computer chess had been solved" after the DB project had so much success.

I'll personally always remember those events as the BEST years of computer chess. Quite unlike todays clone wars...

So if GO is not solved then there should be much interest in GO engines.

[hgm]

There certainly is. There will be 9x9, 13x13 and 19x19 Go tournaments at the Olympiad.

[syzygy]

Of course they could. Why do you doubt it?

The ICGA would no doubt be more than happy with SF as entrant and, as you explain, might go out of their way looking for some surprising interpretation of the rules that would allow it to happen.

So my question was to Bob: would he allow it?
http://talkchess.com/forum/viewtopic.ph ... 612#469612

No idea. I have not looked at their code and compared it to anything. Some have reported similarities between Stockfish and Fruit. That doesn't concern me one bit at the moment since both are GPL and both are completely legal from a license point of view. But could stockfish enter the WCCC? No idea until it becomes an issue, because it will take time to answer.

[hgm]

Well, I don't think Bob has any jurisdiction here. Of course he cannot answer your question if it complies with rule #2 without having studied the code. I don't think there is any actual doubt that Stockfish is OK.

Why do you think this a 'surprising interpretation' of ICGA rules? I would say it would be very surprising if an obviously clean engine like Stockfish would be prevented from entering by the rules. Such rules would be very sick indeed!

And the GPL is very explicit in that there cannot be any additional conditions attached to using GPL'ed code. In particular the copyright holder(s) cannot require that participating in a tournament is subject to explicit written permission from them. As soon as you stamp 'GPL' on something, you implicitly permit any conceivable use that complies with the GPL. So I think it is the only natural interpretation. It would be highly impractical and illogical when a thousand people have contributed to a project that might not even keep explicit track of who contributed what to require that all contributors sign an explicit permission, when they already have agreed to relinguish any control over their contribution.

But why don't you just put it to the test? Just to see where everyone stands, rather than allowing them to dodge questions that are merely hypothetical. I am sure that whatever way it goes, we would have so much fun it would be more than worth it!

[Roger Brown]

[SNIP]

But why don't you just put it to the test? Just to see where everyone stands, rather than allowing them to dodge questions that are merely hypothetical. I am sure that whatever way it goes, we would have so much fun it would be more than worth it!

Hello H.G.,

What an absolutely, wickedly delicious idea!

I second it whole heartedly.

Regards.

[michiguel]

Well, I don't think Bob has any jurisdiction here. Of course he cannot answer your question if it complies with rule #2 without having studied the code. I don't think there is any actual doubt that Stockfish is OK.

If the same criteria ICGA used for Rybka is applied, SF would be certainly ok if Fabien name is included in the list of authors. Why? For instance, Glaurung PST's (and many versions of SF afterwards, probably gone by now) were much more similar to Fruit than R1 was using the criteria ICGA used. I would disagree with applying that criteria, but it was what ICGA used against R1 (that did not even participate in ICGA tournaments).

That is why, I kept saying, the ICGA case set a really bad precedent.
Well, that is the mess of open source for this type of things. You need to list the name of authors, but how are you sure you are not missing any?

Miguel

Why do you think this a 'surprising interpretation' of ICGA rules? I would say it would be very surprising if an obviously clean engine like Stockfish would be prevented from entering by the rules. Such rules would be very sick indeed!

And the GPL is very explicit in that there cannot be any additional conditions attached to using GPL'ed code. In particular the copyright holder(s) cannot require that participating in a tournament is subject to explicit written permission from them. As soon as you stamp 'GPL' on something, you implicitly permit any conceivable use that complies with the GPL. So I think it is the only natural interpretation. It would be highly impractical and illogical when a thousand people have contributed to a project that might not even keep explicit track of who contributed what to require that all contributors sign an explicit permission, when they already have agreed to relinguish any control over their contribution.

But why don't you just put it to the test? Just to see where everyone stands, rather than allowing them to dodge questions that are merely hypothetical. I am sure that whatever way it goes, we would have so much fun it would be more than worth it!

[hgm]

Well, it is clear that the purpose of the rules is not to exclude successful open-source projects. If they were made without the foresight of this possibility, so that they form an unintended obstacle to this, they should be changed immediately. But I don't think the ICGA interpretation of the rules would go beyond what is written in the copyright notice / acknowledgements of the project itself.