On Rybka/Crafty and licensing/copyright

[BB+]

well, remind you, yet again, for the umpteenth time, what you now call Rybka 1.6.1 was an entirely legal version of Crafty, doing all the things Hyatt wants, eg learning about chess programming and playing around and testing Crafty source code which is freely downloadable, changing it, recompiling it, testing the chessic results.

he never distributed the program to the public for money nor gave it away to the public. his mistake was to enter it in a tournament which is counter to the Hyatt license. This is not "copying" or "plagiarism", this is entering the program into a tournament in breach of the licence, some might say trivial or technical breach, but nevertheless a breach. In the tournament, btw, it played weakly and without effect.

This appears to contain a misunderstanding of licensing and copyright law. HG Muller explained it better than most awhile back in TalkChess. Rather: "copying" Crafty, even under the licensing agreements, is still copying per se -- the license simply gives the end-user additional rights, in particular precluding the copyright holder from undertaking legal action for said copying (presuming any license conditions are followed). Seen a different way, the Crafty license was the only thing stopping Rajlich from being subject to copyright law for his copying/modifying of Crafty -- and this disappeared when he entered Rybka into various author-based tournaments (such as ChessWar).

Once the license conditions are transgressed, the situation devolves to copyright law, with the "license violator" unable to make an appeal to the license as a warrant for any actions, but still possessing traditional defences such as "fair use", etc. I would agree that "Rybka 1.6.1 was an entirely legal version of Crafty" right up to the moment when the license was violated, at which point copyright law becomes effective. But I would say that it was "derivative" of Crafty all along. Nor would I denote such a license breach as "trivial" or "technical", but constitutive of the purpose of the license.

My position remains clear. The "copying" and "use" of Crafty in Rybka 1.6.1 was entirely lawful and encouraged by Hyatt. There was a relatively trivial license breach by Vas when he entered Rybka 1.6.1 into the basement CCT tournament. It is perfectly valid to use the term "breach of license" in connection with this, but it is misleading and inappropriate to use the term "plagiarise".

First, I can note that "Rybka 1.6.1"/CCT6 is wrong here, as it seems to mix-and-match various facts again. R161 was entered into Olivier Deville's tournaments (among other places and/or "basements"). The closest approximation to the CCT6 entry the Panel obtained was "Rybka 1.4" or something, I think. Rajlich did not address whether his CCT6 entry was also derivative of Crafty, and the Panel/Secretariat never pushed the issue -- the only mention of CCT6 in the Report occurred in the opinion of Peter Skinner.

On the note of the post in general, it makes the same error as above. Hyatt allows copying under conditions XYZ. Rajlich did not follow these conditions. Ergo, Hyatt doesn't allow such copying/distribution, and Rajlich is subject to copyright law. Hyatt could presumably take legal action against Rajlich for (partial) misappropriation, which is typically not lawful [particularly in locales such as Europe that are big on the author's "moral rights"], and not something that I would call "relatively trivial" on the whole. Given that "plagiarise" includes (gasp) "wrongful appropriation" as its primary definition, it thus seems to be an adequate term IMO.

Since my principal point seems to be a sticking point for many, I will reiterate it yet again: the Crafty license is merely one part of the copyright/licensing issue for pre-Beta Rybka versions. Once the license was violated, the entirety of copyright law then applies. By the very nature of the situation vis-a-vis copyright [and I might agree that "license" is perhaps not the best word for the concept, given its variegated construal in other contexts] it is hard for me to say that any license violation is "technical" or "trivial", particularly when the licensor makes the conditions rather precise, and the licensee inarguably violates them. Saying that the license is silly, misguided, or whatnot, entirely misses the point -- rather, the license ameliorates copyright by providing specific conditions under which the licensee can copy/modify (and perhaps distribute) the work, and (thus) the licensor is free to name these conditions even if they are onerous, imbecilic, or just plain crazy.

[wgarvin]

A license is a grant of permissions. Copyright laws reserve for the copyright holder the rights to distribute the work, make derivative works, etc. The copyright holder of an original work can do anything with it, but everybody else who wants to do these things, needs to get the author's permission to do them, or else they must rely on exemptions in copyright laws (such as Fair Use, or the equivalent in your country if it has one).

Permissions are usually given by a license. Anybody who receives a copy of the work under the copyright holder's license, gets permission to do the things spelled out in the license. But only if they comply with the "terms" of the license. If they start out complying, but at some point they do something that violates the terms of the license, from that moment on they can no longer rely on the license, and any distribution, etc. is now a copyright violation.

Licenses like the GPL are enforced in courts all the time... But not directly. The FSF usually approaches violators and offers them a chance to come back into compliance (by releasing the source code, for example). If the violator refuses to cooperate, the FSF (or whoever is the copyright holder) would file a lawsuit for copyright violation. If the defendant claims that he was relying on the promise of the license, they just have to show that he violated the license, which is usually easy to do if the case got to this point. At that point, defendant is toast.

Needless to say, most license violations don't make it into court and to trial. Few defendants are dumb enough to try and fight that battle. Nearly all of them voluntarily repair the breach of the license terms, or at least settle out of court.

[wgarvin]

By the way, only the copyright holder has standing to enforce their copyrights in court. One reason why open source authors sometimes assign their copyrights to the FSF, is so that the FSF will be the owner of those copyrights and be able to enforce them against license violators. Another reason is so that lots of people can contribute to a copyrighted program and all of the copyrights in it can be assigned to the FSF, instead of leaving them with the individual authors and trusting them to not become unreasonable later. E.g. Linux is under a GPLv2 license, but the copyrights are held by hundreds of individuals and companies, so changing the license would require approval by ALL of them, so it will probably never happen. But FSF-owned copyrighted programs can be unilaterally re-licensed under e.g. GPLv3 by the FSF. This is either a strength or a weakness, depending on your point of view.

Even if they don't own the copyrights, I think the FSF is willing to advise authors of GPL-licensed software about how to enforce their rights against license violators. I assume Fabien still owns the copyrights to Fruit, so he would have to sue Rajlich himself. Or I suppose he could donate his copyrights to the FSF and then they would be the party that could sue over them.

[BB+]

Even if they don't own the copyrights, I think the FSF is willing to advise authors of GPL-licensed software about how to enforce their rights against license violators. I assume Fabien still owns the copyrights to Fruit, so he would have to sue Rajlich himself. Or I suppose he could donate his copyrights to the FSF and then they would be the party that could sue over them.

My information is that Fabien assigned the Fruit copyright to the FSF in Apr 2007 (as "changes and/or enhancements [...] to GNU CHESS"). It is unclear whether he can sue in his own name for direct compensation for the period Dec 2005 to Apr 2007, which was a question that was raised when I was in Lille 2 weeks ago. He could also sue for indirect compensation (Rybka gained "status" while being Fruit-derived, Fabien had his contract talks with Convekta suddenly evaporate, and then Fruit commercialisation was curtailed, etc.), which would be closer to "unfair business practices" than copyright per se, but I don't claim any legal expertise regarding the issue(s).

In any event, the FSF seems to be taking the case seriously:

I reviewed the analysis of the similarities between Fruit and Rybka, and found it pretty compelling. I think this is something that we could absolutely pursue as a compliance case. I'm guessing you've met with the ICGA researchers by now [...]

[kingliveson]

.
.
.

In any event, the FSF seems to be taking the case seriously:

I reviewed the analysis of the similarities between Fruit and Rybka, and found it pretty compelling. I think this is something that we could absolutely pursue as a compliance case. I'm guessing you've met with the ICGA researchers by now [...]

This is an interesting development.