"Although the timing is different, this is exactly the situation with GPL libraries like Readline. You can't use them in a proprietary product.
"
Actually, this is disputed by a number of good open source lawyers. In particular, it's not clear what the line for derivative works is, despite what the FSF/etc claim.
"On the other hand, I suspect if this went to court, Mojang could argue that since that's all bukkit ever did, and this dev contributed to it while that was the case, the license should be interpreted as including something like the classpath exemption, because it's absurd to assume that people wrote code and distributed it but did not legally intend for it to be used in exactly that way."
I'm not sure what you are trying to say here.
Let me give you an official open source lawyer view: There is a roughly 0% chance this guy could force Mojang to open source anything.
He can sue for CraftBukkit for copyright infringement over previous versions. He can stop them from distributing new versions with his code in it.
That's it.
This is because his rights are in his code, not Mojang's. At least in the US, he cannot sue for copyright infringement of someone else's source code, no matter what license it's released under.
So once they removed his contributions, they would be fine, except for whatever damages already exist. If CraftBukkit included decompiled source without Mojang's authorization, the most likely outcome is that they would pay some small amount of money damages, and maybe be enjoined from distributing older versions.
>There is a roughly 0% chance this guy could force Mojang to open source anything.
>He can stop them from distributing new versions with his code in it. That's it.
He seems to be effectively saying "license the whole thing as GPL" or otherwise you can't use my contributions. That seems reasonable and could very well lead Mojang to at least license as GPL the decompiled source of their server, which they seem to have been happy to allow to continue to exist in the repository.
He may instead be trying to get them to release the original code as GPL, which is probably not needed to comply with the GPL in this case.
"He seems to be effectively saying "license the whole thing as GPL" or otherwise you can't use my contributions."
I think this is what he is saying, and agree that this is reasonable to do :)
"He may instead be trying to get them to release the original code as GPL, which is probably not needed to comply with the GPL in this case."
It's a bit of a grey area, depending on how it's linked, used, etc.
There's one thing I'm curious of from the perspective of a lawyer, which is the part I didn't describe clearly. If someone were to post some piece of software online with a description saying, "go ahead and use it for anything you want" but then also include a LICENSE.txt that says it's only licensed for non-commercial use, would that original statement of "anything you want" make it harder to sue a commercial user?
In this case, Bukkit and CraftBukkit are effectively the same project. Bukkit is the collection of plugins, CraftBukkit is the essentially the abstraction layer to allow those mods to run in the memory space of the Minecraft server.
Regardless of what FSF thinks is a derivative work, and really regardless of what really is a derivative work under the law, everyone contributing to Bukkit seemed to be acting as if they thought plugins in the same memory space communicating over APIs was a perfectly legitimate use of their code, at least until this squabble.
Actually, this is disputed by a number of good open source lawyers. In particular, it's not clear what the line for derivative works is, despite what the FSF/etc claim.
Could it be argued that the definition of a derivative work used in the GPL is a condition for distributing the GPL code? In other words, if you distribute readline you can't link against it for sure, but if you don't distribute it then whatever legal definition of "derivative" applies?
"Could it be argued that the definition of a derivative work used in the GPL is a condition for distributing the GPL code? "
GPL does not redefine derivative work, or else, yes, you could argue that ;)
When a program is linked with a library, whether
statically or using a shared library, the combination
of the two is legally speaking a combined work, a
derivative of the original library. The ordinary
General Public License therefore permits such linking
only if the entire combination fits its criteria of
freedom. The Lesser General Public License permits more
lax criteria for linking other code with the library.
The key concept to take away from legal theory is that in the end, judges decide.
The uncertainty and doubt that DannyBee comments seem to instill is a bit regrettable, since its not that complicated. A derivative work is a legal concept built on a few vague concept, neither which has been clearly defined, and the precedential case law for software derivatives is almost non-existent. What does exist is an industry standard, and a defined intent written in the license text.
When a judge look at a copyright case, intent and industry standards goes a long way. Judges does not seem amused when people trying to find loop holes or game the system. Add that with the commonly commercial nature of cases that get brought to court, and I would not bet on the infringer getting a pass.
"The key concept to take away from legal theory is that in the end, judges decide.
"
This depends. If whether it is a derivative work turns on a question of fact, the jury will decide it, not the judge.
"The uncertainty and doubt that DannyBee comments seem to instill is a bit regrettable, since its not that complicated. "
If you honestly believe this is "not that complicated", you have apparently found a way to solve a problem that has plagued software copyright law for 20+ years!
You then go on to describe that it is not clearly defined, and there is no case law, which seems complicated to me!
". A derivative work is a legal concept built on a few vague concept, neither which has been clearly defined, and the precedential case law for software derivatives is almost non-existent"
With you so far.
However, you then claim intent and industry standards go a long way:
"When a judge look at a copyright case, intent and industry standards goes a long way."
Which is it?
Is there no precedential case law, or is the precedential software case law saying "intent about derivative works is what drives what a derivative work is". Because it can't be both!
In any case, there is literally no industry standard here, and no intent in the written license text. The GPL says nothing about it (the word "derivative work" appears twice in the GPL 2.0, and neither is a definition). Large numbers of authors using the license disagree on what it means (happy to give specific examples if you like, i deal with this literally every day).
"Judges does not seem amused when people trying to find loop holes or game the system."
I agree with this, but it seems, at least to me, irrelevant to this discussion.
A judge, a jury, it depend on legal system. In the legal system that Mojan resides in, Sweden, the lowest court is made up by 3 politicians and one judge which then vote on the issues at hand. To simplify, lets call it "a judge decide", but in truth, it depend on the justice system in the nation that the case is put forth in.
> plagued software copyright law for 20+
For something to plague, it has to inflict it. The question about derivative work is not brought up in cases around copyright law. Lawyers are not arguing about it in court, judges are not deciding about it, and juries are not asked to vote on it. At most, it has plagued software legal scholars, but then that claim would have to be supported by statistics from academia.
>Is there no precedential case law...
Industry standards is not precedential case law, its just the defined status quo. If one for example read the US supreme court decision around Aereo, one find that industry standard and status quo has a considered impact on judgments. The legal system in Sweden commonly uses industry standard when defining legal theory in unprecedential areas (See the Pirate Bay trial for example).
> In any case, there is literally no industry standard here
I strongly beg the differ. Larger commercial companies both uses GPL and LGPL, and the praxis is to differentiate the licenses on linking. A company like HP and Apple will distribute GPL software, and will combine LGPL and proprietary software. Someone could argue in court that the perceived praxis do not exist, but I suspect a number Amicus curiae would then be sent in arguing the opposite.
> no intent in the written license text
"The GNU General Public License does not permit incorporating your program into proprietary programs." - GPLv2 and GPLv3
If a client of yours take GPL licensed software and incorporate it into a proprietary program, I would strong doubt arguing the definition of derivative work in court will yield successful result. Its the primary reason in my view why there is no precedential case for derivative work in GPL licensed software.
The FSF, while the license writer, is not the copyright owner, or the contracting party. Their view of what should/should not be a derivative work, listed somewhere in an FAQ nobody here has probably read, would mean precisely nothing in court :)
In fact, in the only US cases to have a say on the GPL and definition of derivative works, the FSF's view was never asked for or cared about.
Actually, this is disputed by a number of good open source lawyers. In particular, it's not clear what the line for derivative works is, despite what the FSF/etc claim.
"On the other hand, I suspect if this went to court, Mojang could argue that since that's all bukkit ever did, and this dev contributed to it while that was the case, the license should be interpreted as including something like the classpath exemption, because it's absurd to assume that people wrote code and distributed it but did not legally intend for it to be used in exactly that way."
I'm not sure what you are trying to say here. Let me give you an official open source lawyer view: There is a roughly 0% chance this guy could force Mojang to open source anything.
He can sue for CraftBukkit for copyright infringement over previous versions. He can stop them from distributing new versions with his code in it. That's it.
This is because his rights are in his code, not Mojang's. At least in the US, he cannot sue for copyright infringement of someone else's source code, no matter what license it's released under.
So once they removed his contributions, they would be fine, except for whatever damages already exist. If CraftBukkit included decompiled source without Mojang's authorization, the most likely outcome is that they would pay some small amount of money damages, and maybe be enjoined from distributing older versions.