GPL Hindering Two-Way Code Sharing?

An anonymous reader writes "KernelTrap has some fascinating coverage of the recent rift between the OpenBSD developers and the Linux kernel developers. Proponents of the GPL defend their license for enforcing that their code can always be shared. However in the current debate the GPL is being added to BSD-licensed code, thereby preventing it from being shared back with the original authors of the code. Thus, a share-and-share-alike license is effectively preventing two-way sharing." We discussed an instance of this one-way effect a few days back.

You are in a maze

You are in a maze with twisty little passages.
Do you tag this article:
* noshitsherlock
* duh
* wateriswet
* slownewsday
* cowboynealsayalloftheabove

Sigh.

For fucks sake, it's forking...

This is nothing new. Provide a permissive license and expecting everything to be returned to you is contradictory to the very license you've chose. Forking happens all the time, usually around licensing or management issues. So aside from the little dust storm we've seen recently regarding the wifi driver and the copyright clause I don't see how this is news.
The GPL and BSD type licenses coexist perfectly, so long as both parties take the time to understand each other. Which is mostly the way it's happened. Kind of making this a none story.

Re:

Maybe the BSD license should be altered to say code can be closed-sourced but not open-sourced without retaining the original BSD license(adding an additional license to the code would probably be fine). Seems like BSD's intent is to allow code to be used

This is not the point

The point is that we have the GPL camp and we have the BSD camp. The GPL camp takes code from the BSD camp and the BSD camp is not able to merge those changes back into BSD code. The issue here is not that this is a license violation; it is not. BSD people, like me, want other people to use our code. The complaint here is about the hypocrisy of the GPL camp, who claim that they don't want anyone to use their code without giving back the changes, but then turn around and do just that to the BSD people's code. Again, I emphasize that this action is not a problem to us; we want it and we expect it. The problem is with the GPL camp saying how they are somehow "more free".

Re:

The GPL camp takes code from the BSD camp and the BSD camp is not able to merge those changes back into BSD code.

And that's the inherant problem with the BSD license, people can mod your code and not give it back to you.

The complaint here is about th

Ball of confusion

There's no hypocrisy in that. Anyone can use the changes that where GPL'd, but you just have to adhere to the GPL license for those changes. The hyprocisy is the BSD camp saying "be free to use our code any way you want" and when people take them up on the

Re:For fucks sake, it's forking...

The way I see it, it takes nearly no effort at all to contribute the changes back to the BSD camp that provided you with the base for the code. All else being equal, that seems to be the ethical thing to do.

It takes more effort to change the licensing in such a way that the BSD camp can't use the code. So it's kind of a slap in the face. I think that's where the animosity comes from, especially since the GPL camp proclaims to be all about freedom and sharing.

Re:

It's not rude.

Is it nice to give back under the BSD? Sure! But not doing so is not rude. They could have used the GPL if wanted it to be given back under the same license. Instead by using BSD they explicitly give you the legal right and moral OK to no

Simple solution

Why do they feel obliged to remove the BSD license from the Linux port of the driver? If they just keep it dual-licensed, there isn't a problem. Or did someone issue an edict that Linux kernel code can't be dual-licensed, at some point when I wasn't payi

Re:

It seems to me that some Linux developers want to deny derivative works to the original authors. I guess they think that their not a part of a community, but a members-only club, and damn anyone not using the home team kernel.

Re:Simple solution

Or did someone issue an edict that Linux kernel code can't be dual-licensed, at some point when I wasn't paying attention?

I think the point of the story is the following:
1. Developer A writes some code for OpenBSD (or whatever)
2. Developer B says "that's cool, I wish Linux had that"
3. Developer B ports developer A's code to Linux
4. Developer B then starts improving on A's code

However, developer B doesn't want to release his changes under the BSD license, so the improved version goes out GPL-only. Developer A says "hey, wait, that sucks", because now he can't incorporate those changes back into OpenBSD, which does (I assume) have a policy that all code must be BSD-licensed.

One one hand, it's unfortunate, because OpenBSD loses out. On the other hand, the original author wrote the code knowing that someone could take it and not release changes (for instance, incorporate it into Windows or Mac OS X or SunOS or something like that), and this really isn't all that much different.

Re:Simple solution

Everyone seems to be completely missing the point here. As someone else pointed out, GPL supporters love to claim the moral high ground when it comes to comparing the GPL to anything proprietary and they love to say how the GPL promotes sharing and openness. So how do you claim the moral high ground when you just took someone else's project and forked it so that they can't use it the way they originally intended?

So what if that's what if that's what the BSD license allows people to do! It's about moral hypocrisy, pure and simple. How can you claim to be free and open when you just basically told the original author that he/she needs to follow your rules in order to benefit from anything you add to it. It wasn't your project to begin with, but you're arrogant enough to fork the project and slap your own license on it, for what? Just because you don't like the BSD license?

Re:Simple solution

The BSD developers got what they wanted. Their code is in use. The BSD license intentionally trades away protection from inclusion in differently licensed projects in return for the increased likelihood that the code can be used.

The GPL developers got what they wanted. Their code is protected from proprietization (And ONLY their code. Anyone can take the original BSD licensed code and do what they want with it).

There is no story here. The GPL and BSD licenses try to achieve different goals and both work as advertised. If you want an analogy: BSD is like the girl who sleeps with everybody. She gets a lot of sex and is invited to every party, but nobody respects her. GPL is like the girl who is selective about her partners. She doesn't have quite as much "fun" and has earned herself a little bit of a hard-to-get reputation, but the people who know her treat her well. Proprietary licenses usually require payment.

Do the BSD proponents understand "Alternatively"

Reyk Floeter (et al) [madwifi.org] put the following license on their code:

* Alternatively, this software may be distributed under the terms of the
* GNU General Public License ("GPL") version 2 as published by the Free
* Software Foundation.

If you think adding this to Linux would do anything the code's original authors did not want to happen, you don't understand what "alternatively" means.

Clue: it doesn't mean "as well as".

Isn't closing them out the point

Having talked to a few BSD licence fans most of them like the licence because it allows another group to take their code and close it off.

This is exactly what the Kernel and other guys are doing, they are taking the code and putting a GPL header in there, closing it off from the BSD developers.

The only difference here seems to be that because the BSD developers can see the changes and improvements being made they want to include them. Whilst putting the GPL on may be against the spirit of cooperation it seems to me to be exactly the kind of closing off of the code that the BSD developers want to allow.

Options

People release code under the license they like. If you can't abide by the license then you should not use the code. It's not a complex issue, you either play by the rules or you can't use the equipment supplied by the other team. If you don't like their r

What the original author of the code has to say:

Quote Sam Leffler, the original author of the code, in http://uwsg.indiana.edu/hypermail/linux/kernel/070 9.0/0159.html [indiana.edu] (emphasis added by me):

I dual-licensed the code so folks could adopt and use it however they saw fit. As I've said before I don't care what people do with the work I give away so long as they don't claim it's their own.

[...]

I am speaking up as the author of the code that set the dual license in place. I have the definitive say and I have said that any of my code that is dual-licensed can be made gpl only.

Sam

So Theo and the rest of his OpenBSD-Trolls better shut up.

I'm already seeing "except for GPL" licenses

Netrek 2006, for example, has a BSD/MIT style license that says "Do what thou wilt except re-license under a (L)GPL or similarly viral license". The author of that license specifically identifies GPL as reducing the freedoms of the developer, which to be fair I'm inclined to agree with.

Re:I'm already seeing "except for GPL" licenses

Netrek 2006, for example, has a BSD/MIT style license that says "Do what thou wilt except re-license under a (L)GPL or similarly viral license". The author of that license specifically identifies GPL as reducing the freedoms of the developer, which to be fair I'm inclined to agree with. So, what the author of the license is basically saying is, it's even okay if you re-license the BSD code under an anally-restrictive proprietary license which allows that restrict every kind of freedom for everyone (users and developers included), just as long a those dirty, dirty GNU/hippies don't share the code their way. Because... it's... restrictive sharing in a way that we snobbishly disapprove of! Yeah. Because no sharing at all is better than GPL-style sharing.

Re:I'm already seeing "except for GPL" licenses

The problem I thnk a lot of people have with the FSF and GPL is that it's moving away from a "this will protect your code whilst allowing others to use it" licence, to a "we have a philosophy about how the software should be used" license. This is why I have removed the "and any later version" clause from my GPL v2 code, and why these restrictions are being placed in other licenses.

It's not that people like me don't want to share the code, just that we don't want to join the Cult of the Gnu either. For it is, almost, turning into a religious issue of whether you swallow the FSF dogma, rather than a practical one of whether you just want other people to benefit from the code.

Bob

Guys, i dont get the whole discussion.

I summrize my view -developer develops code and is kind enough to say that people may use this code in GPL way or in BSD way. -Linux developer derives work fron this code, take away BSD licensing terms, and by that the rights of the people who wanted to use the code and derived works under BSD license. If I, as a developer, for whatever reason, license code under my copyright to somebody, I demand that he agrees with the terms of the licence which I put, because after all I am still the copyright holder. Since GPL and BSD mainly collide in the handling of derivative works in respect to dristributing final products, it would seem to me that only the distributor in the end may chose not to distribute the source code of the device. And since the linux developer cut this right when he removed the license from the file, he is definetly violating the spirit of the dual licensed approach. The dual licensed apporach in nothing else but a "keeping both doors open" policy. While I wont comment on the legal terms i find this behaviour rude. When developing cond in our lab i several times encountered a similar spirit. People who do not honour the idea under which I gave them code which they modified (sometime actually causing work for me). If i give code to anybody it is not an invitation to missionate me into any license

a brief FAQ on this controversy

Q: What happened?
A: A contributor of a patch to the linux kernel didn't notice that it contained both dual-licensed and BSD-only code, and posted a diff that GPL'ed the whole thing.

Q: What happened then?
A: Several things. 1) The mistaken (and clearly incorrect) change of license on BSD-only code was rectified. 2) Theo de Raadt leaped upon this golden opportunity to accuse the linux kernel developers of stealing code and eating babies 3) Separate issues of the legal and ethical obligations related to license changes, dual-licensing, proprietary software, and the price of peanuts in Perth were immediately injected in the discussion and a classic internet blizzard of bullshit blanketed the land of free software.

Q: Latest news?
A: Several developers involved have attempted to help the situation by saying they want collaboration and harmony and dual-licensing their code, but these positive efforts have gone mostly unnoticed as everyone on all sides proceeds to get angry and confused. Apparently high intensity behind the scenes consultations with Eben Moglen have resulted in a daring mission to dual license an OS/2 + Novell Netware application stack under GPL 3 as translated into Babylonian Cuneiform, thus simplifying the situation for everyone.

Q: What's the moral of the story?
A: Sometimes, cooperation is harder to achieve than competition, or "the greedy fox gets stuck debugging the rotten oysters".

Yes, but!

In one sense, the GPL does hinder two-way code sharing. You can't distribute, modify, etc. a project as a whole under the terms of the BSD license if some code in the project is under the GPL. So adding GPL-ed code to a BSD-licensed project does hinder two-way sharing.

However, the fact to the matter is that it is the _BSD_ license that allows you to do this. The BSD license simply does not require you to share your changes.

So, if you are asking yourself why changes aren't being shared back, the answer really is that the original authors (who put their code under the BSD license) said it was OK to use their code without sharing back.

Of course, you can still call into question the behavior of people who take something willingly shared with them and then put up obstacles for sharing back with the original authors.

Re:Yes, but!

Here's an analogy for you: Say that I have a magic jellybean, and that magic jellybean can make as many red jellybeans as you like, but only five black ones each day. So I take my magic jellybean to the market, where I see Theo DeRaadt, and try to exchange my magic jellybean for a cow. It has a bit of a limp, but it makes chocolate milk on Thursdays. That's pretty nifty, so I offer him fifty black jellybeans. Then he says he'd also like a date with my sister, and I say, "I have two, and you'd better not mean the married one," and he fires back with, "Hey, you promised this analogy would be relevant to this discussion."

No, Theo, I promised no such thing. Just like nobody promises to share their changes with the BSD team when they take advantage of BSDL'ed code. The BSD'ers say people ought to be able to do what they like with their code. Well, what the GPL'ers would like to do is protect their modifications from being appropriated by people who won't share the code. If they automatically hand their changes back to the BSD folks to distribute as BSD code, then they lose the protections they wanted from the GPL in the first place.

Theo is basically saying, "The Linux people are hypocrites because they say they believe in software freedom but they don't believe in my definition of software freedom." Which is pretty lame.

Re:

I'd disagree with that assessment.

If dual-BSD/GPL-licensed code is used, then changes can be distributed under the GPL or BSD by the author of derivative works. However that applies to the derivative work; the COPYING file in any source code tarball dictat

Everybody seems to have missed the key part of TFA

The Linux code is being patched to fix the license problem, says TFA. Here's the content of the patch [marc.info].

Note what the patch is doing, very carefully. The patch is changing the copyright notices on top of the modified files to say that these files are licensed under the GPL, but are also based upon an earlier work licensed under the BSD, and then reproduce the copyright and license statements as required by the original BSD licenses. This makes completely transparent the following things:

1. The new work is released under the GPL license only. Anybody who uses, modifies or distributes this new work must abide by that license. They don't have any other license to that work.
2. The new work is based on older work whose authors released under the BSD license, and the authors of the new work received the original under that license. In order for the authors of the new work to comply with the license that allows them to release a derivative of the original work, they reproduce the copyright and license notices of the original. These license notices only apply to the portions of the new work that are taken from the original one.

BSD is hindering two way sharing.

In related news a debate that the BSD is being added to public domain code, thereby preventing it from being shared back with the original authors of the code.

-

Thats exactly whats BSD made for

Thats exactly whats BSD made for. Get everything you like and then throw back into the developers face everything you hate. No need to say thanks. Apple did it, Linux did it, dozends of others do it all day.

Seriously, bragging about this is a sign of total ignorance about the BSD philosophy: Giving away everything without asking for anything. They should feel honored that they are getting ripped like they wanted always to be.

Think of BSD license like citing sources

If you think about the BSD license in terms of an academic citation, it makes more sense.

In the original patch, it appeared that some Linux folk took some code, stripped the BSD copyright notice and put it under a GPL license. Viewed through an academic mindset, it sounds less like "building on existing research" and more like plagiarism. Were they legally entitled to do what they did? I suspect probably so. Still, it seems like bad form not to cite your sources.

-Peter

BSD Alternative

Is there an alternative to an MIT or BSD license that does the same thing, but doesn't allow GPL people to use it? I release just about everything I do under the MIT license, but I'd consider a license that prohibits GPL people from using it after see some of the BSD hate in these comments and over at kernel trap.

Yes, it is and does!!

The GPL has hindered code sharing. Do you remember the Broadcom wireless driver debacle wherein Theo recieved a nasty gram for porting the code to BSD? No, well here goes. Theo recieved a warning from the Linux project maintainers of a Broadcom wireless driver. This was really in poor code-sharing taste and ultimately caused the cancellation of the porting efforts. In the end, both Linux and BSD users lost out. Was OpenBSD really going to profit from it or commercialize it, no! It is simply in the spirit of hardware support. Now, let us look at OpenSSH, a fine product from OpenBSD. Linux and GPL people use it all of the time. I am hard pressed to find a more significant contribution to Linux. After all, OpenSSH is the foundation of secure remote administration, logins, tunnelling, and more. Now, someone tell me what Linux has contributed to BSD of similar significance as I cannot think of any.

Re:

Exactly. If you don't like people taking your code and not giving back, don't use BSD.

Re:BSD

I've read Theo's rant, and I found the section about not sharing code back to be pretty humorous, considering that's the way the BSD license is written. If you wanted to ensure that code be shared back into your projects, you'd use a copyleft-style license instead of a BSD/MIT-style license, wouldn't you?

I personally prefer the GPL, but I've been around Slashdot for a few years and understand the "more freedom" argument from BSD fans. That "more freedom" is the freedom to relicense or even completely close up the code, returning nothing to the original project.

Why's everyone got their panties in a bunch over something which the license allows? (I also understand the origin of this anger being the removal of the attribution and BSD text from the wireless kernel patch proposed, but it was just proposed, not accepted, and the situation was immediately resolved.)

Re:

Personally I don't think either one is more free than the other. I think it comes down to the GPL keeping code free and the BSD license keeping people free.

Re:

When I read the original OpenBSD thread, the author of the driver (originally dual-licensed BSD/GPL) was the one who submitted the GPLed driver to the Linux kernel, so he's not denying himself anything. Additionally, the original BSDed code is still availa

No, you are wrong...

The situation is this:

The code in question had the BSD license included, plus the author (copyright holder) added a condition saying that "alternatively," the GPL could be used.

One is therefore free to select the alternate license and ignore the BSD lice

Why is it BSD proponents can't read?

You and Theo are disingenuous idiots. First, you make an unsubstantiated claim - '"Alternatively" means that the GPL could be used, that is, it can conform to the GPL.', then you redefine a term you just pulled from your ass to mean what you want it to mea

Re:BSD

I've read Theo's rant, and I found the section about not sharing code back to be pretty humorous, considering that's the way the BSD license is written. If you wanted to ensure that code be shared back into your projects, you'd use a copyleft-style license instead of a BSD/MIT-style license, wouldn't you?

The reason I like using the BSD license is that it's supposed to be progress. When you write good BSD code there's no reason anyone will need to rewrite that code, and it's the kind of code that you hope everyone will use because it makes applications better, or an application developed on top of it will be high quality itself.

Contrast that with the GPL, where code written with the GPL has to be rewritten if you want to use it for proprietary purposes. If you want to end all proprietary code then that's obviously a good thing, but it's not so good if you want the best code to be used, and for no-one to have to needlessly rewrite the code you're writing.

Take SQLite for example. It's in the public domain which is only slightly less restrictive as BSD. Anyone can use SQLite for any purpose. If I'm developing proprietary software and I need a lightweight database engine I know SQLite is available. It saves me time and money, and the software is going to be better as a result. That's why drh chose to make it available so freely: It's the best there is, and if someone can improve on it good luck to them.
Both licenses have their purposes of course, but I hope you can at least see why someone would prefer the BSD license.

So if it's supposed to be extra free then what's wrong with relicensing it as open source? Well there are two possible outcomes of releasing GPL changes to a BSD/MIT/public domain project:

• The developers which wrote the software take the changed code and start using the GPL; the code can no longer be used in proprietary software. Whether you think that's a good thing or not it's a decision made against the original developer's will.
• The developers which wrote the software continue developing their code and ignore the changes; the software is now either forked, causing compatibility issues, or changes are made in parallel, causing duplicated efforts. BSD is chosen so that no-one has to duplicate efforts, because the code is free.

Both of these outcomes just don't help. They don't help the software improve, they don't help anything. Rather than ask "Why choose BSD?" (which I hope I explained well enough earlier) why not ask "Why relicense BSD code under a different open source license?" Does he think that the developers of the BSD licensed code will start using his modified code along with the GPL license?

Take SQLite for example. Imagine if someone took SQLite and made some nice updates to it, but released the changes under the GPL. Now there are either duplicate efforts and compatibility worries, or the developer of SQLite is forced to use GPL and his software can't be used everywhere like he intended. A stable piece of code that used to be universal no longer is universal.

Using the GPL to close the original developers out is using a share-alike license to avoid sharing, and using an open collaboration license to prevent collaboration.

Re:

The developers which wrote the software take the changed code and start using the GPL; the code can no longer be used in proprietary software. Whether you think that's a good thing or not it's a decision made against the original developer's will.

And if so

Re:

Proprietary modifications are the problem of the person doing the modification. Because they're proprietary, we (the open source community) never see them again. You can mark this down as a loss, in a sense, but because it's effectively invisible to us

Re:BSD

Telling me I can do something and then rebuking me for doing it is kind of a shitty practice, isn't it? If you want me to share code with you, put it in the license. Microsoft and Real won't contribute back, either.

Re:BSD

Telling me I can do something and then rebuking me for doing it is kind of a shitty practice, isn't it?

No, not really. It's a way to reconcile moral views and freedom: give people the right to do something you disagree with, while retaining the right to disagree with them if they take advantage of that right.

For example, I believe people should have the right to troll on Slashdot, because I believe that's the only way to ensure that people who genuinely hold controversial views feel able to argue their case. Does that mean I shouldn't have the right to mod trolls down? If so, why?

If you want me to share code with you, put it in the license.

No, that's for if you want to demand that code be shared. If you would like code to be shared, and intend to complain if it isn't when it reasonably could be, but absolutely do not want to take away the licensee's freedom to decide for themselves whether to share or not, then you should not add anything to the license. It would probably be appropriate to mention your attitude in the README, though.

(One very good reason not to put this kind of stuff in the license is that the GP apparently was specifically interested in allowing large companies to use his code. Large companies hate custom licenses. Stick with something standard if you want the commercial world to even bother looking at your code.)

Re:

But until that world comes, BSD just hinders freedom

So the license that allows the most use of the code hinders freedom?

Re:

So the license that allows the most use of the code hinders freedom?

Yes, in a roundabout and counterintuitive, yet valid, way. GPL code stays "free" no matter what happens to it, whereas BSD code can always be rendered un-free.

Personally, when I want to

Freedom for WHOM? To do WHAT?

Honestly, the way some people talk about "Freedom", you'd think it was something you could buy by the wheelbarrow load. Freedom isn't something that exists in and of itself. It only exists in relation to people and activities.

To say that the BSD licence hinders freedom is just insane: it grants close to maximum freedom TO users, developers, and distributors TO DO pretty much whatever they like. v The GPL on the other hand deliberately restricts the freedom of one of those stakeholder groups - the distributors - in order to preserve the freedom of the users and developers in the longer term.

If you say that the GPL licence is more or less free than the BSD licence, all you are really doing is criticising a group of developers for their failure to share your own priorities. That always strikes me as an ugly, intolerant, narrow minded way of thinking.

This whole mess has the stink of FUD about it. There are a lot of people who would like nothing better than to get the GPL devs and the BSD guys together and say "hey, why don't you and them fight?"

I have a suggestion to make: let's disappoint them.

Re:

This is something that the Linux community doesn't need, because if another open source project starts accusing Linux of using stolen code, us Linux promoters will lose our "moral" highground.

I have to disagree with you on this one, we don't lose our high

Re:

You're parable doesn't fit. It neglects the changing nature of software and hardware. Your parable would work if everyone had stock-standard hardware that never, ever changed, and if software weren't allowed to be modified. It neglects to factor in how com

Re:

That's nice, but it doesn't really have anything to do with this situation. Code doesn't grow on trees, and especially not magical, self-replenishing trees.

Actually, it has everything to do with the situation. Software is like that apple tree. No matter ho

Re:The Apple Tree

Actually, it has everything to do with the situation. Software is like that apple tree. No matter how many times you copy software, it remains undamaged. So there is no reason to "protect" it with restrictive licenses.

I think the problem here is that you haven't told the parable in its entirety:

Then one day, a powerful lord heard about the magical apple tree and sent his soldiers to build a wall around the orchard. "These are MY apples now", he told the villagers. "Anyone who wants one of these apples must buy it from me". And the villagers thought this to be most unfair, but they could do nothing against all the lord's soldiers.

However, in the village there was a young boy who saved some seeds from the magic tree, and in later years he moved to a new village and planted a tree of his own. And around this tree, the villages set up a fence and a notice that said "Anyone is free to take these apples. They may eat them, sell them, give them away, or plant magic apple trees of their own. The only thing you may not do is build a wall around the tree to prevent others from picking their own apples".

There you go: fixed that for you :)

Re:

Does the BSD license allow people to make extensions and GPL the base code plus those extensions? Absolutely. Do BSD-style developers, then, have a right to be miffed if this is what happens? It's a hard question.

I would say that on the basis of what I've