This is my perspective as someone who was once inside of the Bukkit community as an outside developer (not one of the core team), but Wolverness has been a very toxic maintainer of the project, especially as it regards to responding to pull requests. Either your code complies 100% with the contributing guidelines (including confusingly written and sometimes unwritten rules about spacing, formatting, naming and pull request formatting) or it gets rejected out of hand. Any mention of forks of the Bukkit project (such as Spigot, Glowstone or SportBukkit) could get you banned from IRC, the forums and your comments deleted on Github. This upon many, many instances of just caustic behaviors and treatment of interested developers and contributors.

I'm still subscribed to the projects on Github, and I was never burned by these rules personally (so these feelings are not out of spite--they're out of sorrow for the developers who were turned away) but I don't think I'd ever have a chance of getting any non-trivial PR integrated into the project. The standards were just too insane. I'm no fan of how Mojang has been handling the minecraft community recently, but anything that would drive this guy out is just fine by me.

(This takedown is BS btw. CraftBukkit has ALWAYS linked against the minecraft official server library--a release of CB made by a Mojang employee doesn't change ANYTHING about the GPL status of Minecraft. Note that the proprietary minecraft server DOESN'T link against in any way shape or form)

But unless Bukkit/CraftBukkit had a copyright assignment policy, which it doesn't appear to have had, the developers Mojang hired could not transfer the entirety of Bukkit to Mojang's ownership, and thus the "obtained rights [to]" language by the COO seems either inaccurate or meaningless.

Doesn't Mojang have lawyers? Their behavior seems baffling and their public statements even more so.

To me this reads as "obtained [some] rights in" which seems correct.

The context of that quote is:

> As Jeb announced on twitter, Mojang obtained rights in the project over 2 years ago

The tweet from Jeb says that Mojang "bought" Bukkit, and a subsequent tweet says that they "dug up the receipt to be sure".

The previously hired Bukkit devs have also said Mojang "acquired" and "owns" Bukkit.

Mojang is to blame for not letting @EvilSeph shut down the project? Despite his reason for shutting it down was that he was worried Mojang was going to issue a takedown against them—something Mojang showed no interest in doing, and then was done by one of their own developers?

I'm curious how you think they should have handled this. EvilSeph wanted to take down Bukkit, something that would have freaked out a whole lot of server admins (and did freak out a whole lot of server admins, now that Wolverness managed to actually do it). Clearly they didn't want it taken down. They owned it (although what that means is unknown). What else should they have done?

--EDIT-- fixed spelling.

* Wolvereness's contributions to CraftBukkit are licensed under the GPL, and therefore may only be redistributed on that license

* CraftBukkit itself is released under the GPL

* ...but Wolvereness claims that CraftBukkit includes Mojang's Minecraft server, which is proprietary (apparently this is disputed by Mojang themselves, though)

* Therefore CraftBukkit's GPL license is illegitimate, and the CraftBukkit developers' right to use Wolvereness's contributions is terminated.

Even if the factual claims in the DMCA takedown are correct (which is hard to verify now that the repository is offline), it strikes me as an abuse of the process. When you file a lawsuit, you have to demonstrate that you have standing because of some adverse effect that specifically affects you. By that logic, Wolvereness shouldn't be able to use a third party's copyright to enforce a takedown.

[1]: https://github.com/github/dmca/blob/master/2014-09-05-CraftB...

Violating any clause of the GPL revokes your copy of the license. So if I'm a company that distributes GPL'd binaries without providing corresponding source (a legal term in the GPL that means the preferred version of the source that a developer would edit, so decompiled code does NOT satisfy the license), I lose that license and am violating the copyright of the original author.

In this case, Wolvereness' code has been incorporated into a derivative work, with other code (the Mojang decompilation) that isn't licensed compatibly with the GPL.

Note that the form of the copyrighted work (the Mojang Minecraft server) is irrelevant. Mojang's copyrighted work is there, whether it's decompiled or in binary form or in any other form. The work is the same, much in the same way a vinyl record, a digital file, and a live performance all can contain the same copyrighted work.

I don't think this is weird at all; this is a very straightforward case of GPL incompatibility. If Wolvereness wanted to include an exception for linking to the Mojang server, he could have done so using the infrastructure in the GPLv3, or with a special exception in the GPLv2. Mojang has definitely violated the GPL, and as such they have no claim to Wolvereness's license. Wolvereness is the party with standing here.

If I'm understanding this correctly, CraftBukkit purports to be GPL, but includes disassembled, deobfuscated code that they did not own and was not available under a GPL compatible license.

You're right that the original CraftBukkit project equally violated the GPL, but now that Mojang is distributing it, they're the ones liable for the violation.

You don't think a court would use some equitable doctrine such as estoppel to find that if I explicitly contribute my code to a project with the intent that they include it and distribute it, I am giving them whatever licenses is necessary to allow that?

Unless you as a copyright owner explicitly allow use under other licenses/conditions, then that's it.

This is basic copyright law, it doesn't matter if it's GPL or BSD etc.

A non-exclusive copyright license in the US does not need to be explicit or be written. It can be implied.

One court noted, "An implied license can be found where the copyright holder engages in conduct from which the other party may properly infer that the owner consents to his use" [1].

I'd say that a copyright holder submitting his code to a project for inclusion in the project allows the project to properly infer that the copyright owner consents to the project using his code, and so supports an implied non-exclusive license to use the code.

That implied license would be revocable, so if the author changes his mind he can tell the project to stop using the code, but would almost surely not be able to sue for damages over the past use. (Major exception: if the copyright holder was paid for the contribution, that might make the license irrevocable).

Implied non-exclusive licenses are limited in scope, but as far as I have been able to tell there is no good objective way to tell what the scope is. If I post a comment on a forum (let's assume nothing in the TOS that I purportedly agreed to said anything about copyright), it is generally believed that I am granting them in implied non-exclusive license to display my content on the forum, and giving an implied non-exclusive license to other forum posters to quote my post in theirs without limit. But what if the forum wants to publish a "best of" book and include my post? Does the implied non-exclusive license cover that?

In "Goldstein on Copyright" (one of the top two references on US Copyright law--the other being Nimmer), he says of the various cases on implied licenses and their scope, "The fact patterns of these cases suggest a close connection between implied copyright licenses and the doctrines of laches and estoppel", so I think I was on the right track suspecting that estoppel could play a role. Unfortunately, for more info that cite a section in the book that was not covered in the Google Books preview I was reading, and much as I would like owning a copy of Goldstein, it is way over my law book budget (it's almost $1000...about 1/3 the price of Nimmer).

[1] Field v. Google Inc., 412 F. Supp. 2d 1106, 1115-16 (D. Nev. 2006)

Further, the actual license that the original author used was the GPL. I think it'd be very difficult to say that there was an "implied license" that was somehow not the GPL.

It's true that he can terminate his right to use the contributions (particularly if no CLA was signed granting rights), contrary to their assertion that he can't because he knew what the licensing was.

He can also sue for copyright infringement in previous versions. But that's it.

At least in the US, the only one who could sue for copyright infringement is a copyright holder. That means, in this case, the folks who own copyright to CraftBukkit. Right now that includes him. They could remove his contributions, but that would not put them in compliance for the versions already distributed, and they could be sued for those.

(For the curious, Breach of contract wise, third party beneficiary arguments around the GPL have repeatedly failed, AFAIK, so others trying to get source as beneficiaries have not succeeded).

Mojang is not in a very business friendly situation right now. The only 100% not-lawsuit-libel approach is to either open source minecraft or rewrite the entire thing. They are still open to lawsuits related to the existing distributions but the traditional settlement is a combination of stopping to infringe and medium cash payment.

This is why, when folks say "CLA's are not necessary, you'd just remove the code" (like bradley kuhn recently did), i think they are very very wrong :)

A few relevant facts about the political landscape involved:

Bukkit is used by a large number of Minecraft fans to run servers, due to the availability of plugins and other improvements in the modded server.

The developer in question contributed significantly to Bukkit, and has stated that without the full Bukkit product being open sourced (including the base Minecraft binary it's based around), it's a copyright violation to be using his contribution.

This is legal, because of the nature of his code licensing, however, it is not a demand that the Bukkit team can comply with, as they don't hold the original Minecraft code. As such, their only option is to comply with the take down (and not use that code) until such time as Mojang releases the source to the Minecraft code that Bukkit uses.

Mojang is under no obligation to do so, because they're not using any of the dev in question's code, as he only contributed to Bukkit.

However, Mojang risk upsetting a large portion of their tech savvy and more active user base (including YouTubers and streams), because relying on Bukkit servers is common in those circles.

So Wolverness isn't just using the popularity of Bukkit as leverage. Rather, Mojang is redistributing his code without having sought his permission.

Of course, Bukkit itself was a derivative work of the Minecraft Server (the decompiled java bytecode) in violation of that piece of software's license, so the whole matter seems to be a tangle of license violations.

In any case, Mojang is (or WAS, rather) distributing Wolverness's copyrighted code without his permission, and so are apparently in violation of his copy rights. I think that's where the leverage comes in. Wolverness seems to be hoping that it will be easier for Mojang to relicense and release Minecraft code to make it compatible with his bukkit code than to work around his code.

[1] Reportedly. I've seen various, partially contradictory characterizations of how this happened, but there seems to have been at least some sort business deal between Mojang, and a subset of the Bukkit team not including Wolverness

[edit: added footnote]

The fact that Mojang now wants to distribute GPL code without providing full source would, I suspect, stand badly against that since a claim of license violation has been made now. Generally, you can't ignore copyright by yelling "two wrongs make a right".

Incidentally, the link has been changed since I began commenting. I find this assertion from the now-linked post to be odd: "Additionally, we believe that Wesley has no right to prevent the continued use of his contributions to the project – which he gave freely, knowing them to be subject to applicable open source licenses. We believe these licenses continue to bind all those who contributed to the Bukkit Project." This doesn't really address the primary charge from Wolverness, to paraphrase, that essentially the bukkit and related projects are distributing code incompatible with its license and therefore are not permitted to distribute the copyrighted code contained in the bukkit project without express permission. If the license for the code is not adhered to, then you don't have permission to distribute the copyrighted material... So one party is saying that the license was broken, and the other party responds by saying that there was a license. That's kind of a non sequitur.

Does Mojang own the issue tracker, the website, etc.? In that case, sure, they can keep those going, although without any developers I don't see the point.

Plus Mojang keep contradicting themselves about the relationship between Bukkit and them.

Yes this is an interesting issue, I find it hard to believe that decompiling Java code would possibly count as a clean-room implementation, even if you later rewrote everything.

Back when I did clean-room reverse engineering (which was admittedly long time ago), it was one guy (whom I never even spoke to) who disassembled code and wrote down what how it functioned, this was then sent to me and another programmer and we made an implementation based upon his descriptions (which of course contained no code whatsoever).

Granted it was like 7 years ago I did clean-room reverse engineering last, but AFAIK there's been nothing new in how this is applied.

The changes link doesn't point to the DMCA request response which is here http://www.spigotmc.org/threads/our-dmca-response.28772/ which I think is equally important to the story.

1. HN already consists of pointers to stories, and pointers to pointers don't work as well.

1. Bukkit is GPL, developed without help from Mojang, but with a questionable use of the decompiled source.

2. Mojang likes Bukkit, hires some of the devs and buys name. And I'll assume the these core devs would be willing to assign copyright to Mojang.

3. (Eventually) One of these core devs, now a Mojang employee, makes a new release of CraftBukkit, including Bukkit and linked against the Mojang official server blob.

Although the timing is different, this is exactly the situation with GPL libraries like Readline. You can't use them in a proprietary product.

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.

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.

>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 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.

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.

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?

https://www.gnu.org/licenses/old-licenses/gpl-2.0-faq.html#T... - "Subclassing is creating a derivative work"

https://www.gnu.org/licenses/old-licenses/gpl-2.0-faq.html#T...

Also the preamble of LGPL2.1 (https://www.gnu.org/licenses/old-licenses/lgpl-2.1.html#SEC2):

    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 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.

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.

> 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.

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.

1. Many (most?) of the servers running with Bukkit also had some form of payment for items, skills, ranks, etc. that can affect game play, also known as "pay-to-win." Mojang recently announced that they were going to be cracking down on this since it's a violation of the license, though I don't know what if any enforcement has actually happened. Enforcement was due to start August 1.

2. Mojang has a couple of multiplayer options without Bukkit - the Minecraft Server piece that is part of this discussion and Minecraft Realms. If Bukkit is dead (which may be the case no matter what happens because W. is not the only outside developer who could pull this kind of stunt), there are still multiplayer options and Mojang could likely implement some of the features currently in Bukkit but under their control - perhaps under a licensing model that lets them sell the server software as well.

Basically this guy has managed to shut down something being used by folks with business practices that Mojang was already having some issues with, while leaving available some less feature-rich options fully approved (and controlled by) Mojang. The odds of that resulting in their giving in and open-sourcing something that's included in Minecraft as a whole are somewhere around nil - worst case for Mojang is that the Minecraft server community that they're not a direct part of takes a feature and popularity hit for some time while they get to point at this guy as the reason why.

Second, a quick explanation of the issue, as I have it: Bukkit is a server made by the community to make up for the simplicity of the official server. It uses deobfuscated, reverse engineered java code from the official server, and has never been given the actual source from Mojang. While Bukkit was sort-of folded into Mojang (They own the name, and a few of the developers went to work for Mojang), no one at Mojang is paid to work on Bukkit and they still don't get source access. Now, a disgruntled Bukkit Dev, Wolverness, has thrown out this DCMA request in some sort of retaliation / blackmail attempt.

Mojang is definitely claiming they "acquired" some kind of rights to Bukkit, but I can't find any description of what those rights are -- and, for a community project, I don't understand what they could possibly own that isn't public knowledge. You can't own secret trademarks.