Luecx wrote: ↑Fri Jul 23, 2021 10:11 am
It feels like many people here dont understand the difference between whats legal and what is allowed. The legal part is decided by official courts where as what is allowed is decided by the chess community. Especially the unique engine authors, tcec etc.
According to common chess community standards, it is not allowed to:
- steal code
- not correctly credit other people if parts of it are used
- using other peoples data when training NN's
- ...
I do not care if FF is actually legal or not, all i know is what they did is not allowed by common chess community standards and they need to face the consquences for that.
one more point:
- restricting engine to run on specific rating-lists/tournaments only
Hmm, can this work out? Enforcing moral standards on an GPL project which are not covered by the GPL? Maybe people should come up with an "unique engine authors licsense"? I mean it serious. If you define some rules for playing the game, why not enforce it via your own kind of licsense. Crafty had its own kind of license, its point was that taking ideas is OK but taking code is not, as an example.
--
Srdja
PS: in context of all these recent issues I concluded myself to use MIT for my next releases.
smatovic wrote: ↑Sat Jul 24, 2021 7:54 am
From the ChessBase thread:
Luecx wrote: ↑Fri Jul 23, 2021 10:11 am
It feels like many people here dont understand the difference between whats legal and what is allowed. The legal part is decided by official courts where as what is allowed is decided by the chess community. Especially the unique engine authors, tcec etc.
According to common chess community standards, it is not allowed to:
- steal code
- not correctly credit other people if parts of it are used
- using other peoples data when training NN's
- ...
I do not care if FF is actually legal or not, all i know is what they did is not allowed by common chess community standards and they need to face the consquences for that.
one more point:
- restricting engine to run on specific rating-lists/tournaments only
Hmm, can this work out? Enforcing moral standards on an GPL project which are not covered by the GPL? Maybe people should come up with an "unique engine authors licsense"? I mean it serious. If you define some rules for playing the game, why not enforce it via your own kind of licsense. Crafty had its own kind of license, its point was that taking ideas is OK but taking code is not, as an example.
--
Srdja
PS: in context of all these recent issues I concluded myself to use MIT for my next releases.
I am not sure if your post is supposed to be offensive to us (the Koivisto Devs). Considering that we refused at some point that specific rating-lists test us. We believe that the existence of specific engines in some rating lists simply inflate the ratings and make the rating list lose its credibility. Furthermore rating FF2 as #1 in some rating lists just supports their behaviour which we do not want to support.
It was not supposed to be offensive. You are free to do with your engine whatever you like. But the points given, I see that there are moral standards on top of what the GPL covers. These moral standards are not unique to the Kovisto team. If different kind of teams intend to set up their own kind of moral standards on top of GPL, my suggestion is to ponder on to use an own license to cover these and to be able to enforce these.
In fact, what is legal is allowed. If credit is given where credit is due, there should be no problem. Derivatives can be interesting projects and have merit, and sometimes even little changes are welcome. I personally find the projects of Honey, Shashchess, Sugar, useful.
I see that some young authors who are not aware of the complicated history of chess programming try to implement ethical rules while themselves being in a grey area. A lot of ideas they routinely code were never intended to be shared, and, on an ethical plane, there are no differences between stealing code and stealing ideas. The fact that these ideas were stolen by others does not make things suddenly ethical.
Then, there is a fundamental misunderstanding of the GPL: this license exists to allow reuse of code, not to forbid it, to promote sharing.
Luckily moral standards are, as the name suggests, standards which are widely accepted in the chess community. And I think one needs to distinguish between moral standards and legal standards. In case of Koivisto, what could there possibly be for us to add to our GPL?
matejst wrote: ↑Sat Jul 24, 2021 12:25 pm
In fact, what is legal is allowed. If credit is given where credit is due, there should be no problem. Derivatives can be interesting projects and have merit, and sometimes even little changes are welcome. I personally find the projects of Honey, Shashchess, Sugar, useful.
I see that some young authors who are not aware of the complicated history of chess programming try to implement ethical rules while themselves being in a grey area. A lot of ideas they routinely code were never intended to be shared, and, on an ethical plane, there are no differences between stealing code and stealing ideas. The fact that these ideas were stolen by others does not make things suddenly ethical.
Then, there is a fundamental misunderstanding of the GPL: this license exists to allow reuse of code, not to forbid it, to promote sharing.
So, a bit of tolerance would be welcome.
I think you are very wrong here. Something not beeing forbidden doesnt make it allowed by common understanding in this community.
Luecx wrote: ↑Sat Jul 24, 2021 12:26 pm
Luckily moral standards are, as the name suggests, standards which are widely accepted in the chess community. And I think one needs to distinguish between moral standards and legal standards. In case of Koivisto, what could there possibly be for us to add to our GPL?
@Milos: I am not a lawyer.
1. steal code
2. not correctly credit other people if parts of it are used
3. using other peoples data when training NN's
4. restricting engine to run on specific rating-lists/tournaments only
1. not clear defined
2. 'correctly' is matter of interpretation, not defined
3. correct me if I am wrong, but the output of an GPL program is not restricted in any use
4. restriction of the usage of an GPL binary?
As mentioned, Kovisto team is not the only one with such standards.
Luecx wrote: ↑Sat Jul 24, 2021 12:26 pm
Luckily moral standards are, as the name suggests, standards which are widely accepted in the chess community. And I think one needs to distinguish between moral standards and legal standards. In case of Koivisto, what could there possibly be for us to add to our GPL?
@Milos: I am not a lawyer.
1. steal code
2. not correctly credit other people if parts of it are used
3. using other peoples data when training NN's
4. restricting engine to run on specific rating-lists/tournaments only
1. not clear defined
2. 'correctly' is matter of interpretation, not defined
3. correct me if I am wrong, but the output of an GPL program is not restricted in any use
4. restriction of the usage of an GPL binary?
As mentioned, Kovisto team is not the only one with such standards.
--
Srdja
1. Apart from stealing line for line I agree this is not easy to define. My experience is that everyone wants to share and even see their code/ideas be adapted in a wide variety of engines, so I think it's mostly fine as long as proper credit is given -> 2.
2. This is very tricky. In Koivisto many "well known"/"widespread"/"whatever term u wana use" ideas are implemented, but often we don't know who to actually give credit.
On another note (although related), I'm troubled by some commercial engines. The commercial engines get pretty much all their strength from other peoples work, I don't know if that is ethical? Idk, thoughts?
Koivisto wrote: ↑Sat Jul 24, 2021 10:29 pm
The commercial engines get pretty much all their strength from other peoples work, I don't know if that is ethical? Idk, thoughts?
GPL says it is OK, the engine authors who licensed their engines in GPL (or open source license really), implicitly accepted this. That is the sole purpose of GPL, to share code, and make everybody benefit from it, even to sell it.
There is no open source license which prevents commercial engines from taking ideas, or formulas, or numbers. Copyright covers only the implementation, the code, not the technical solution itself, what in some cases can be protected by patents. If authors do not want to share their ideas and concrete implementation, they go commercial (or private/free) with their engines. And then the commercial engines are a subject of reverse engineering and disassembling. That is what was prev mentioned as complicated history of computer chess.
