A lawyer on the GPL and Fruit/Rybka

[JM_UK]

Seems that Alan Sassler brought up this article back when Rybka/Strelka was in vogue: http://www.btlj.org/data/articles/21_04_04.pdf
Or in a nicer format (to me): http://www.cs.berkeley.edu/~tlavian/pub ... 060403.pdf

Dangerous Liaisons—Software Combinations as Derivative Works?
Distribution, Installation and Execution of Linked Programs under Copyright Law, Commdercial Licenses and the GPL
By Lothar Determann

I wrote the original post on talkchess.com. This is an interesting article - thanks for posting. Unfortunately, it is rather US-centric and I don't think (although it's difficult to be entirely sure) that any claim here would be heard in the US.

I wrote the talkchess post after reading a number of posts on talkchess querying whether there had been a "GPL breach." However, in legal terms, Vas could be sued for copyright infringement, not for a "GPL breach." Legally, I think the GPL may be of limited relevance here. The main point of reviewing the terms of the GPL was to determine whether there are grounds for reporting the matter to the FSF. I think there clearly are such grounds, even though the drafting of the GPL is unclear in places.

The more fundamental issue is whether there has been copyright infringement. This will determine whether Fabien and/or the FSF could successfully sue Vas. If I were Fabien and I thought someone had sold a derivative of my engine for profit, I would want to know whether I could recover part of that profit from him. I avoided looking at the possibility of suing for copyright infringement in my earlier posts (principally because it was easier just to read the GPL, which is a single document!) However, I've set out some further thoughts below.

- The first step should be to determine what the FSF's attitude is. Are they prepared to take the matter to court and, if so, what remedy would they seek: damages, the withdrawal of Rybka from sale or merely the release of source code? If they are prepared to seek damages in court, are they prepared to allow the damages to be paid to Fabien, as the author of Fruit? I am assuming that, as a question of law, any damages would belong to the FSF as copyright owner, rather than to Fabien.

- On the strength of what I have read about the Fruit/Rybka similarities, I think there is a strong argument that copyright infringement has occurred. In my opinion, it would definitely be worth paying for legal advice on whether a claim for copyright infringement would have merit and, if so, what remedies would be available; in particular, what (if any) payment by way of damages could be recovered from Vas. At a reasonably educated guess, I expect that if a court does award damages, it would consider the (no doubt tortuous) question of how much of Rybka's playing strength derives from Fruit, to quantify how much of Vas' profit derives from Fabien's work - and would be influenced by experts' views on this. There is a difficulty for Fabien/ the FSF here. As I understand it, the principal issue with R1 beta was its use of Fruit eval and the plagiarised eval was replaced by Larry K’s eval in R3. If Vas’ profits from R3 and R4 are only minimally attributable to Fruit, any damages payable to the Fruit owner for copyright infringement may not include a material share of the R3/ R4 profits.

- The next question is: which jurisdiction would the claim be made in (and where to seek legal advice)? The rules for EU member states are governed by the 2001 European "Jurisdiction Regulations." The general rule is that a defendant domiciled in an EU member state should be sued in the jurisdiction of his domicile. I cannot say with any certainty where Vas is domiciled. His wikipedia page states that he is currently living in Poland but this is not enough to be sure that he is "domiciled" there. However, it may be possible to pursue the claim in a different (and possibly more favourable) jurisdiction. The Jurisdiction Regulations will determine whether or not this is possible - there are various rules, which I will not go into here.

- The best approach _may_ be simply to determine where Vas' domicile is and go to a law firm in that jurisdiction. I expect the FSF will have their own legal advisers, who would tell them what to do. If there is a desire to take legal action but uncertainty on how to proceed, I can speak to a specialist on an informal basis to find out which jurisdiction(s) to take legal advice in and how much an initial view on the chances of success/ available remedies would cost.

[benstoker]

Seems that Alan Sassler brought up this article back when Rybka/Strelka was in vogue: http://www.btlj.org/data/articles/21_04_04.pdf
Or in a nicer format (to me): http://www.cs.berkeley.edu/~tlavian/pub ... 060403.pdf

Dangerous Liaisons—Software Combinations as Derivative Works?
Distribution, Installation and Execution of Linked Programs under Copyright Law, Commdercial Licenses and the GPL
By Lothar Determann

I wrote the original post on talkchess.com. This is an interesting article - thanks for posting. Unfortunately, it is rather US-centric and I don't think (although it's difficult to be entirely sure) that any claim here would be heard in the US.

I wrote the talkchess post after reading a number of posts on talkchess querying whether there had been a "GPL breach." However, in legal terms, Vas could be sued for copyright infringement, not for a "GPL breach." Legally, I think the GPL may be of limited relevance here. The main point of reviewing the terms of the GPL was to determine whether there are grounds for reporting the matter to the FSF. I think there clearly are such grounds, even though the drafting of the GPL is unclear in places.

The more fundamental issue is whether there has been copyright infringement. This will determine whether Fabien and/or the FSF could successfully sue Vas. If I were Fabien and I thought someone had sold a derivative of my engine for profit, I would want to know whether I could recover part of that profit from him. I avoided looking at the possibility of suing for copyright infringement in my earlier posts (principally because it was easier just to read the GPL, which is a single document!) However, I've set out some further thoughts below.

- The first step should be to determine what the FSF's attitude is. Are they prepared to take the matter to court and, if so, what remedy would they seek: damages, the withdrawal of Rybka from sale or merely the release of source code? If they are prepared to seek damages in court, are they prepared to allow the damages to be paid to Fabien, as the author of Fruit? I am assuming that, as a question of law, any damages would belong to the FSF as copyright owner, rather than to Fabien.

- On the strength of what I have read about the Fruit/Rybka similarities, I think there is a strong argument that copyright infringement has occurred. In my opinion, it would definitely be worth paying for legal advice on whether a claim for copyright infringement would have merit and, if so, what remedies would be available; in particular, what (if any) payment by way of damages could be recovered from Vas. At a reasonably educated guess, I expect that if a court does award damages, it would consider the (no doubt tortuous) question of how much of Rybka's playing strength derives from Fruit, to quantify how much of Vas' profit derives from Fabien's work - and would be influenced by experts' views on this. There is a difficulty for Fabien/ the FSF here. As I understand it, the principal issue with R1 beta was its use of Fruit eval and the plagiarised eval was replaced by Larry K’s eval in R3. If Vas’ profits from R3 and R4 are only minimally attributable to Fruit, any damages payable to the Fruit owner for copyright infringement may not include a material share of the R3/ R4 profits.

- The next question is: which jurisdiction would the claim be made in (and where to seek legal advice)? The rules for EU member states are governed by the 2001 European "Jurisdiction Regulations." The general rule is that a defendant domiciled in an EU member state should be sued in the jurisdiction of his domicile. I cannot say with any certainty where Vas is domiciled. His wikipedia page states that he is currently living in Poland but this is not enough to be sure that he is "domiciled" there. However, it may be possible to pursue the claim in a different (and possibly more favourable) jurisdiction. The Jurisdiction Regulations will determine whether or not this is possible - there are various rules, which I will not go into here.

- The best approach _may_ be simply to determine where Vas' domicile is and go to a law firm in that jurisdiction. I expect the FSF will have their own legal advisers, who would tell them what to do. If there is a desire to take legal action but uncertainty on how to proceed, I can speak to a specialist on an informal basis to find out which jurisdiction(s) to take legal advice in and how much an initial view on the chances of success/ available remedies would cost.

Re: Jurisdiction.
If I live in Alaska, steal source code from somebody domiciled in UK, and then made numerous and continuous contacts with the UK from sales of my ill-gotten gains in the UK via software distributors, and even if I never set foot in the UK, you can be damn sure I can be hailed to court in the UK for inringement.

[hyatt]

It is not so clear. It is often a matter of "where did you commit the crime." And since you committed the crime, the location where you did this is the point of primary jurisdiction. Many people do this stuff in locations that do not have extradition agreements/policies in place. National sovereignty makes it hard to get at someone in that circumstance and transfer them to your location, so you are stuck with going to where they live and employing the legal system there. If there is one, and if the alleged act is criminal in that country.

[G.Paredes]

Either off-topic or not, a decent punishment in the event of rybka (note is not "she" just it, as retarded as it IS among its supporters) should be one of the same magnitude as its success as a clone.

Thankfully, Vas reputation is already placed in doubt, which in some way applies some justice to the situation.

[JM_UK]

If I live in Alaska, steal source code from somebody domiciled in UK, and then made numerous and continuous contacts with the UK from sales of my ill-gotten gains in the UK via software distributors, and even if I never set foot in the UK, you can be damn sure I can be hailed to court in the UK for inringement.

A recent high-profile example of a cross-border copyright infringement claim is Lucasfilm Ltd v Ainsworth. See http://news.bbc.co.uk/1/hi/entertainment/8418333.stm.

A British engineer (Andrew Ainsworth) produced the stormtrooper helmets and armour for the original 1977 Star Wars film. In 2004, he started selling replica helmets in the US, using design casts from 1977. Lucasfilm's lawyers argued that he was not entitled to sell the replicas because he did not own the IP. A US court upheld the claim and awarded Lucasfilm $20m in damages, although Lucasfilm's real aim was to stop Ainsworth selling the replicas.

However, the judgment could only be enforced against Ainsworth's US assets, and he didn't have any. Lucasfilm sued again, this time in the UK courts, claiming (inter alia) for copyright infringement under UK law and that the US law judgment should be enforced in the UK. So far, they have lost on both counts, although the case will be heard in the UK's Supreme Court in March.

The Court of Appeal judgment can be found here: www.bailii.org/ew/cases/EWCA/Civ/2009/1328.rtf. The Court of Appeal is the UK's second highest court, behind the Supreme Court.

[BB+]

I was actually planning to get back to legal issues, now that the ICGA process seems to be moving.
In particular, Letouzey, as the copyright owner, has the exclusive adaptation right of his work:

The adaptation right refers to the exclusive right a copyright owner holds to prepare derivative works (adaptations) based on his own protected work. For example, the author of a book holds the exclusive right to make a movie based on that book. As such, if you make a derivative work without having the permission of the owner of the copyright for the underlying work, you are committing copyright infringement because you have violated the author’s adaptation right.

The same site also lists remedies for copyright infringement:

In a lawsuit where a copyright owner proves that there was copyright infringement, he can get an injunction from the court ordering the other part to cease any further infringing activities. The copyright owner is also entitled to financial damages. This monetary relief can be one of two types. First, it can be any actual damages the copyright owner has suffered, along with any profits the infringing party obtained. [...]

Distributors could conceivably be liable for vicarious infringement (see also this).

This is an interesting article - thanks for posting. Unfortunately, it is rather US-centric and I don't think (although it's difficult to be entirely sure) that any claim here would be heard in the US.

Indeed, though I do find (at least from a layman's perspective) that the author does do a nice job of stressing some philosophical principles from which the law derives. For instance, the question of "why does copyright exist?" is something that needs to be addressed before one can hope to understand why some legislature enacted some statute, or why "sweat of the brow" is meaningless.

I expect that if a court does award damages, it would consider the (no doubt tortuous) question of how much of Rybka's playing strength derives from Fruit, to quantify how much of Vas' profit derives from Fabien's work - and would be influenced by experts' views on this.

Firstly, Letouzey rejects "strength" as the over-riding selling point, so it is not only "playing strength" that matters in this quantification.
Perhaps the best bet would be something like: both sides agree there is infringement (maybe already hopeless), and a knowledgable binding arbitrator (e.g. Dann Corbit ?!) could oversee an analysis of: this sales impact of Rybka version X derived Y% from Fruit [repeated for all X], with the sides able to argue for their best guess about Y. I think Jeremy asked in a different thread whether there looked to be any hope of resolution, so this is my attempt at an "innovative solution" as it were [of course, I assume "good faith" in such arbitration].

[benstoker]

If I live in Alaska, steal source code from somebody domiciled in UK, and then made numerous and continuous contacts with the UK from sales of my ill-gotten gains in the UK via software distributors, and even if I never set foot in the UK, you can be damn sure I can be hailed to court in the UK for inringement.

A recent high-profile example of a cross-border copyright infringement claim is Lucasfilm Ltd v Ainsworth. See http://news.bbc.co.uk/1/hi/entertainment/8418333.stm.

A British engineer (Andrew Ainsworth) produced the stormtrooper helmets and armour for the original 1977 Star Wars film. In 2004, he started selling replica helmets in the US, using design casts from 1977. Lucasfilm's lawyers argued that he was not entitled to sell the replicas because he did not own the IP. A US court upheld the claim and awarded Lucasfilm $20m in damages, although Lucasfilm's real aim was to stop Ainsworth selling the replicas.

However, the judgment could only be enforced against Ainsworth's US assets, and he didn't have any. Lucasfilm sued again, this time in the UK courts, claiming (inter alia) for copyright infringement under UK law and that the US law judgment should be enforced in the UK. So far, they have lost on both counts, although the case will be heard in the UK's Supreme Court in March.

The Court of Appeal judgment can be found here: http://www.bailii.org/ew/cases/EWCA/Civ/2009/1328.rtf. The Court of Appeal is the UK's second highest court, behind the Supreme Court.

choice of law != jurisdiction

[BB+]

For the question of copyright infringement, here is a relevant opinion (link from Jeremy):
http://lawdigitalcommons.bc.edu/cgi/vie ... ntext=bclr

For a court to find infringement, the defendants must have copied not just the plaintiff's ideas, but the plaintiff's expression of the ideas. Accordingly, the substantial similarity that constitutes copyright infringement must be expression-level similarity, not merely idea-level similarity. [...]
Professor Chafee, writing in 1945, offered further guidance about line-drawing, stating that "the protection covers the 'pattern' of the work.' Professor Nimmer combined Hand's observation that any work can be broken into patterns at different levels of abstraction with Chafee's suggestion that substantial similarity should be determined by comparing common elements at a level that is somewhat abstract but still concrete enough to constitute an expression. Levels of abstraction provide answers to the question "What is the work about?" The most abstract answer might be, to use Romeo and Juliet as an example, "boy meets girl." A slightly less abstract answer is "boy meets girl, and the two come from hostile groups." An even less abstract answer is "boy meets girl at a dance, the two come from hostile groups, and they eventually take the marriage vows." The least abstract answer is the exact language chosen by the author Shakespeare to express that theme of "boy meets girl." Nimmer's treatise uses a comparison of Romeo and Juliet and West Side Story to illustrate this approach to the idea/expression distinction, identifying thirteen common elements and concluding that the common elements form a pattern sufficiently concrete to provide a basis for a finding that the two works are substantially similar. The pattern of the works are similar, involving a meeting and romance between a boy and girl from hostile groups, with resulting killings and misunderstandings.

In the Fruit-Rybka situation, the evaluation features would seem to meet this "substantially similar" criterion.

[BB+]

'copying isn't a big deal as long as you copy the stuff that doesn't make any difference'

Yes and no. Copyright infringement is usually a "balance of probabilities" argument (as with most civil actions).
So suppose there is indisputable evidence that "irrelevant" code was copied -- does this make it more likely that "large similarities" seen in other parts of the code also infringed copyright? OTOH, for the purpose of actual damages, the relevancy (or "value") of the code does need to be considered -- for punitive damages (in locales where such exist for copyright issues) I'm not sure I can say one way or the other.

In the Fruit/Rybka case, what would you call the single most irrefutable piece of evidence that shows code copying?

In my opinion (and I don't claim to have looked at every Fruit/Rybka part, so there might be something even more clear), the evidence outlined in Appendix A of my PDF is quite compelling, particularly the second part of the code segment given there (starting with the "EasyThreshold" usage on page 17). At the "ideas" level, the use of bad_2, bad_1, etc. is already a framework that to the best of my knowledge is rather unique to Fruit [see also my comment in footnote 17 (page 20) regarding the implementation/idea question here] -- in my mind, this outweighs the distinction that (say) Rybka doesn't have explicit "depth_is_limited" and "time_is_limited" conditions. See also the bit about the allocation of variables at the bottom of page 20. I would also be wary of anyone terming iterative deepening code as "irrelevant". The end of the UCI parsing code (bottom page 15 and top page 16) is also quite strong evidence of "code copying" in my opinion.

A (much) more difficult question would concern evaluation features, and whether the evidence is sufficiently strong to support a finding of "transliteration" from mailbox to bitboards. This is particularly conflated by the fact that the idioms of the latter might lead to a different implementation being preferred in a few cases (I suspect the clearest would be things like backward/candidate pawns and pawn storm/shelter, which can also involve complicated constructs to boot) -- I suspect that these should simply be considered "distinct", leaving ~20 cases still to be considered.

[gaard]

'copying isn't a big deal as long as you copy the stuff that doesn't make any difference'

Yes and no. Copyright infringement is usually a "balance of probabilities" argument (as with most civil actions).
So suppose there is indisputable evidence that "irrelevant" code was copied -- does this make it more likely that "large similarities" seen in other parts of the code also infringed copyright? OTOH, for the purpose of actual damages, the relevancy (or "value") of the code does need to be considered -- for punitive damages (in locales where such exist for copyright issues) I'm not sure I can say one way or the other.

In the Fruit/Rybka case, what would you call the single most irrefutable piece of evidence that shows code copying?

In my opinion (and I don't claim to have looked at every Fruit/Rybka part, so there might be something even more clear), the evidence outlined in Appendix A of my PDF is quite compelling, particularly the second part of the code segment given there (starting with the "EasyThreshold" usage on page 17). At the "ideas" level, the use of bad_2, bad_1, etc. is already a framework that to the best of my knowledge is rather unique to Fruit [see also my comment in footnote 17 (page 20) regarding the implementation/idea question here] -- in my mind, this outweighs the distinction that (say) Rybka doesn't have explicit "depth_is_limited" and "time_is_limited" conditions. See also the bit about the allocation of variables at the bottom of page 20. I would also be wary of anyone terming iterative deepening code as "irrelevant". The end of the UCI parsing code (bottom page 15 and top page 16) is also quite strong evidence of "code copying" in my opinion.

A (much) more difficult question would concern evaluation features, and whether the evidence is sufficiently strong to support a finding of "transliteration" from mailbox to bitboards. This is particularly conflated by the fact that the idioms of the latter might lead to a different implementation being preferred in a few cases (I suspect the clearest would be things like backward/candidate pawns and pawn storm/shelter, which can also involve complicated constructs to boot) -- I suspect that these should simply be considered "distinct", leaving ~20 cases still to be considered.

Thanks for answering my question posted on the other forum. Bob has put a huge emphasis on semantic equivalence as being the gold standard for determining code copying. However, in the codes you reference here, there does not appear to be an equivalence. What have I misunderstood?