In as few words as possible

[Michael Sherwin]

We are going in circles here. If I never saw another chess program other than Fruit and I wanted to write a chess program, so I used Fruit as a guide then my program is derived from Fruit. If anyone asked me, 'where did you derive the knowledge from to write your own program', then I would have to answear that I derived it from Fruit. Therefore Rybka is a derivitive of Fruit. However, the GPL does not apply, because, there is nothing original in Fruit and I can quote the author himself as saying that there is nothing original.

So, in the case of Fruit at least the GPL may not be valid under the law.

If this were true, then anyone who had read the source of Fruit (or any other GPL program) and writes a program that uses something he has learned is writing a GPL program. And if that is true, then GPL is viral in a horrible way beyond any imagining.

It is like saying, "If you have read my book and later write your book on the same subject, then your book also belongs to me."

Yes, that is what it implies! But, only pertaining to original never before known content.

alpha-beta, null move, history huristic, hash tables, common move ordering, PVS, IID, futility, etc. None of these things are licenseable as they are not original, but rather derived from someone elses work that has been freely submitted to the public.

However, this does not imply that you do not own copyright to your program derived from known ideas. In the same way, a poem or song is written using known words, patterns, etc. However it belongs to the author in its implementation.

But having written a song about a bluebird's beautiful color, I do not own the rights to all songs about the color of a bluebird even though the content will be approximately the same

That is why I made the claim that Fruit is protected under copyright law, but not under any contract aspect of the GPL.

But GPL *IS* a form of copyright.

It is also an agreed to contract!

No it isn't:
http://lwn.net/Articles/61292/

This article is wrong. There is a reciprical obligation involved. I can redistribute but, I must hold up my end of the bargin and pass on the license as well as honor it. It is a contract.

[hristo]

We are going in circles here. If I never saw another chess program other than Fruit and I wanted to write a chess program, so I used Fruit as a guide then my program is derived from Fruit.

Nope.
This is not necessarily true. It very strongly depends on what is in your code.

If anyone asked me, 'where did you derive the knowledge from to write your own program', then I would have to answear that I derived it from Fruit.

Yes.
This doesn't prohibit you from using the acquired knowledge in a novel way such that your program is not a derivative.

Therefore Rybka is a derivitive of Fruit.

Unsubstantiated conclusion.

However, the GPL does not apply, because, there is nothing original in Fruit and I can quote the author himself as saying that there is nothing original.

This doesn't mean that the GPL copyright doesn't apply, in fact it does. You simply cannot take verbatim code from Fruit and claim it isn't under the GPL. Fruit, as an expression, is copyrighted and if you want to use this expression you must follow its license.

So, in the case of Fruit at least the GPL may not be valid under the law.

I don't believe that you have shown any evidence that the GPL would be invalid.

The only question is about "derivative works" and this is independent issue from the GPL.

Regards.

[jdart]

This issue is getting thrashed out pretty thoroughly in the litigation SCO is involved in with IBM and Novell.

SCO is claiming IBM copied their IP out of System V Unix into Linux, but since they have had great trouble finding examples of literal copying they are resorting to all kinds of devious arguments about similarity of ideas.

The problem is, copyright law is all about similarity of expression, not ideas. If the code doesn't look like the original then it is pretty hard to argue copyright violation. Or, for that matter, that it's a derived work in a sense that matters to the GPL.

Maybe SCO will prevail in their liberal view of what similarity or derivation means, but right now that doesn't look very likely.

(They also have other issues that may moot the whole point, like it is not clear they even own any copyrights).

[Michael Sherwin]

We are going in circles here. If I never saw another chess program other than Fruit and I wanted to write a chess program, so I used Fruit as a guide then my program is derived from Fruit.

Nope.
This is not necessarily true. It very strongly depends on what is in your code.

If anyone asked me, 'where did you derive the knowledge from to write your own program', then I would have to answear that I derived it from Fruit.

Yes.
This doesn't prohibit you from using the acquired knowledge in a novel way such that your program is not a derivative.

Therefore Rybka is a derivitive of Fruit.

Unsubstantiated conclusion.

However, the GPL does not apply, because, there is nothing original in Fruit and I can quote the author himself as saying that there is nothing original.

This doesn't mean that the GPL copyright doesn't apply, in fact it does. You simply cannot take verbatim code from Fruit and claim it isn't under the GPL. Fruit, as an expression, is copyrighted and if you want to use this expression you must follow its license.

So, in the case of Fruit at least the GPL may not be valid under the law.

I don't believe that you have shown any evidence that the GPL would be invalid.

The only question is about "derivative works" and this is independent issue from the GPL.

Regards.

Vasik admitted that he took many things from Fruit. So, Vas himself substantiates what I said.

I have always said that copyright law applies. I was only talking about the contractual part of the license. The GPL is not law. The GPL is a license that incorporates a copyright notice. It is copyright law that protects Fruit from being copied. The question then is the GPL a contract or not. I say that it is because by accepting it I am obligated to pass it on. I believe that the contractional part of the GPL is very weak. That is why the FSF lawyers are now saying that the GPL is not a contract.

[hristo]

We are going in circles here. If I never saw another chess program other than Fruit and I wanted to write a chess program, so I used Fruit as a guide then my program is derived from Fruit.

Nope.
This is not necessarily true. It very strongly depends on what is in your code.

If anyone asked me, 'where did you derive the knowledge from to write your own program', then I would have to answear that I derived it from Fruit.

Yes.
This doesn't prohibit you from using the acquired knowledge in a novel way such that your program is not a derivative.

Therefore Rybka is a derivitive of Fruit.

Unsubstantiated conclusion.

However, the GPL does not apply, because, there is nothing original in Fruit and I can quote the author himself as saying that there is nothing original.

This doesn't mean that the GPL copyright doesn't apply, in fact it does. You simply cannot take verbatim code from Fruit and claim it isn't under the GPL. Fruit, as an expression, is copyrighted and if you want to use this expression you must follow its license.

So, in the case of Fruit at least the GPL may not be valid under the law.

I don't believe that you have shown any evidence that the GPL would be invalid.

The only question is about "derivative works" and this is independent issue from the GPL.

Regards.

Vasik admitted that he took many things from Fruit. So, Vas himself substantiates what I said.

I meant this (unsubstantiated) only with respect to the hypothetical argumentation that preceded it.

I have always said that copyright law applies. I was only talking about the contractual part of the license. The GPL is not law. The GPL is a license that incorporates a copyright notice. It is copyright law that protects Fruit from being copied. The question then is the GPL a contract or not. I say that it is because by accepting it I am obligated to pass it on.

To be more accurate, by accepting the GPL you aren't obliged to do anything so long as you are the user of the software. In this case, from a user perspective, the GPL is not a contract.

If you decide to become more than just a user -- a distributor -- then you are obliged to provide to others the same freedom that you had when obtaining the original software.
Yes, this is a contract, or at least it has all of the underpinnings to be considered a contract. The arguments against this are semantical, IMO, and aren't useful -- but again, this only applies to those who chose to be distributors and no one else.

The catch is that what you chose to distribute can be modified so it doesn't fall under the GPL and hence you wouldn't have to obey the contract. The contract is limited by the definition of "derivative works" and if you have strong enough desire to avoid disclosing your own software you can do so, it will only take you time and money.

In all cases, your code remains yours and you can't be forced to release it if you don't want to do so, the worst that can happen is for you to be prohibited from distributing the software in its current form.

I believe that the contractional part of the GPL is very weak.

This could be true. Since there are options available to all those who distribute software to 1) release the source code and 2) invest the time and money to make their work original, I don't believe that people are interested in challenging it.

Of course, people also have the option to not use GPLd code and instead go with any other license that works better for them.

That is why the FSF lawyers are now saying that the GPL is not a contract.

Perhaps, they are disambiguating some of the aspects of the GPL -- where you, as a user, are not obliged to do anything at all if you have accepted the GPL. 99% of all people are users and it is probably a good thing for them to know that they can do as they please with the software, without having to give their modifications to anyone, ever.

Regards.

[Michael Sherwin]

Hi Hristo,

Sounds like we are mostly back in agreement again. I for one am glad of that and so I wish to leave it at where you left it. Now I have work to do.

Thanks,
Mike

[Dann Corbit]

If using GPL software is entering into a contract, that means that we can enter into a contract without a signature or even a handshake.

If true, then I think it is the most dangerous legal precedent in history.

[hristo]

If using GPL software is entering into a contract, that means that we can enter into a contract without a signature or even a handshake.

If true, then I think it is the most dangerous legal precedent in history.

Dann,
I don't believe it is that bad.

The GPL is absolutely clear with respect to what your obligations are if you decide to become a distributor of GPL derived code. Whether you prefer to call these obligations a 'contract' or something else is irrelevant, IMO.

If you look at it from a pure business standpoint then you can determine that 1) the monetary value you get from GPL is higher than the amount you spent developing your modifications to the software and so you release under the GPL (therefore abiding by the contract) or 2) your code is more valuable than the benefit you got from GPL software and therefore you spend more money to abide by the contract and become a none derivative work.

The similarity with contract comes from the fact that GPL requires you to relinquish the rights to your own work (money), which becomes GPL if you want to continue and use the original software -- it is a clear tit-for-tat situation, in the context of a distributor.

This is not a 'bad' situation, though, and is actually required in some cases. So, if anything, it isn't all green pastures on the GPL side, but at least you know where the crevices are. (if you ever have to deal with contracts, sub-contracts, EULA, NDAs and other such you should be happy about the GPL)

As a mere user there are no 'contractual' (like) obligations whatsoever when using GPL software.

Regards.