FSF: New Apache License not GPL-Compatible

__past__ writes "It seems that the XFree86 issue is not the only licensing battle currently fought in the FLOSS world: An update to the FSF's list of Free Software licenses lists the new Apache License, Version 2.0 (which has been discussed on Slashdot before) as not being GPL-compatible, due to a clause related to software patents." (Read on for more.)

__past__ continues "The new version of the Apache license will apply to all Apache projects, including the popular web server and many Java libraries like Xerces and Log4J, and making it easier to integrate Apache- and GNU-licensed code was one of the primary goals for its development. With the new license being GPL-incompatible (just like the older Apache licenses were), it is not possible to distribute programs that use libraries covered by under it and others covered by the GPL.

Apparently, the FSF does not actually consider the patent-related clauses a bad idea, let alone non-free - it is just that they impose a restriction that the GPL does not, and that makes the license automatically incompatible. It might even be that GPL Version 3 will include similar statements or at least allow them, as a message from FSF legal counsel Eben Moglen indicates. Additionally, prominent Apache hacker Roy Fielding claims that it doesn't really matter what the FSF thinks about the matter, because according to the Apache Software Foundation, derived works can just be distributed under the GPL."

From the FSF site

[FuryG3]

This is a free software license but it is incompatible with the GPL. The Apache Software License is incompatible with the GPL because it has a specific requirement that is not in the GPL: it has certain patent termination cases that the GPL does not require. (We don't think those patent termination cases are inherently a bad idea, but nonetheless they are incompatible with the GNU GPL.)

The old license was incompatible too

[Kelson]

From the FSF page linked in the article:

The Apache License, Version 1.1.

This is a permissive non-copyleft free software license with a few requirements that render it incompatible with the GNU GPL.

We urge you not to use the Apache licenses for software you write. However, there is no reason to avoid running programs that have been released under this license, such as Apache.

No falling sky here. Move along.

Re:The OLD license wasn't either...

[__past__]

Because GPL-compatibility was one of the reasons to develop ASL V2 in the first place (as the article states, by the way). That it didn't work out is kinda disappointing.

Re:Is anyone else getting worried here?

[Directrix1]

Apache license has never been compatible with any GPL besides LGPL.

Insightfull, I say.

[Anonymous Coward]

Apache's license isn't compatible with it? tough shit. find another server that is

Exactly. Linux does not hinge upon the inclusion Apache. People can download and install Apache on their linux machines if they like to run a server.

I quite like how Debian makes this destinction between free and non-free.

not exactly a new problem

[tverbeek]

Incompatiblity between the Apache licence ver 2.0 and the GPL is nothing new. The FSF cites other problems with the 1.x Apache licences. They also say that "there is no reason to avoid running programs that have been released under [the 1.1 Apache] license".

The question of GPL compatibility becomes a problem only when a package contains links directly to GPL code, as seems to be the case with XFree86. If the packages are distinct enough, any "free" licence (which is the term the FSF uses for Apache's) is OK for the two to coexist in a distro.

Re:GPL

[__past__]

No, it is about the GPL. Whether a license is GPL-compatible or not can be mechanically deduced by looking at its terms, and neither the FSF nor anybody else has the power to declare some license GPL-compatible because they think it is morally or politically good. The FSF doesn't seem to have much of a political problem with the new Apache license, but it is still not GPL-compatible.

This is not the first case when the FSF had to declare a license they actually liked GPL-incompatible, the Affero GPL is another.

Re:Motivation?

[johnnyb]

Apache was never GPL-compatible. The only shift _away_ from GPL was XFree86. Apache is actually shifting _towards_ GPL-compatibility, but the news story is that it still has terms that prevent that compatibility.

Re:oh no, rms doesn't approve!

[__past__]

This is not about an individual, or even the FSF as an organization, enforcing their feelings; it doesn't matter if they like it or not. The GPL's wording is clear in that derived works may not distributed under a license that imposes any restriction not found in the GPL itself, and the Apache license supposedly imposes such restrictions (even if these restrictions only affect people who abuse software patents, and the FSF definitely does not like people abusing software patents).

Still, it is certainly the case the it is the GPL that prevents combination of ASL- and GPL-licensed code in a derived work, the ASL is fine with that.

Re:The old license was incompatible too

[sys49152]

Exactly! Why all the hullabaloo all of a sudden? The Apache license and many others have long been incompatibile with the GPL, that does not make them any less free. I was under the impression that we "elected" the OSI to be the final arbiter of what is and is not an open source license.

BTW, surf over to gnu.org and take a look at the long list of GPL incompatible licenses [gnu.org]. These include Apache, Mozilla, PHP, Zope, Apple, and more. Again I ask, why all the excitement?

Not too serious a problem...

[Goonie]

If I'm reading this correctly, GPL incompatibility isn't a huge issue with regards to Apache.

What this means is it can't be linked (like a library is linked) with GPL'd code. But that's not an issue anyway, as Apache doesn't need to link to any GPL'd code. Pretty much all the libraries on a Linux system are LGPL'd (or under even less restrictive licenses like the BSD license), which can be dynamically linked to anything, including proprietary code - yep, that's right, Microsoft Word could be legally linked to an LGPL'd library.

Where it does matter is if somebody wants to add a piece of code from a GPL'd project in Apache, or a piece of Apache to a GPL project. So, this would be nice to sort out, but it doesn't have the urgency of the XFree86 issue, where all the end-user apps on the system link to the X libraries.

Uh, dude...

[Svartalf]

You just invalidated your entire rant.

If you use the BSD, I can relicense the whole damn thing under the GPL unless you're using the oldest version of the license that requires the advertising.

Better yet, you're confused about the licensing...

BSD licensing gives you absolute freedom- including proprietarizing the code. Nothing wrong with that. I just don't like the thought of someone taking my hard work and making money off of it- I want to reserve that privilege for myself, thank you.

GPL licensing gives you the freedom to do whatever you want with the code, so long as you give the people recieving the binaries the same rights you got, including on any enhancements you did to the code.

That's a dramatic difference from what you're claiming of things. Is it all that hard for BSD licensing proponents (the moment you labeled me a GPL fanatic, you resorted to ad-hominem, and I will not stoop to that level, thank you...) to get the fact that BSD isn't incompatible for the most part? Is it all that hard for a proponent of the BSD or similar license to figure out that there's going to be people that do not want to use your preferred license for varying reasons?

(Here's a clue for you, me bucko- I've got stuff licensed under GPL, LGPL, BSD, AND MIT/X out in the world right at the moment. I know all about all the popular Open/Free licenses and I tend to pick the one that works or makes the most sense for each piece of code I license to the rest of the world.)

Re:So Let me Get This Straight

[hexene]

If he finds his patented work incorporated into the original project without his consent and sues over it, that person suddenly loses the patent license to the original work. It's like putting 2 trains aimed at each other on the same track, except that one of them (the patent holders) is made out of paper.

In the above example he should be sueing for copyright infringement, not patent infringement. As a result, the patent clause doesn't take effect.

Borked Section References in Moglen's Statement

[ortcutt]

I'm like many of you was having a little difficulty understanding what exactly isn't compatible with the GPL. I think part of the problem is that Moglen gets all of the section numbers wrong. He says that the only incompatibility is with Section 5. He writes:

Second, with respect to section 5:

FSF, having long warned against the dangers of software patents, entirely shares the goal of providing protection against inappropriate and destructive uses of patent claims to interfere with software freedom. The Foundation expects version 3 of the GPL to incorporate one or more license provisions designed to increase the collective defense against patent abuse. The two provisions contained in section 5, however, embody two quite different theories of response.

But Section 5 isn't about software patents at all. Section 5 of the APL v.2.0 reads:

5. Submission of Contributions. Unless You explicitly state otherwise, any Contribution intentionally submitted for inclusion in the Work by You to the Licensor shall be under the terms and conditions of this License, without any additional terms or conditions. Notwithstanding the above, nothing herein shall supersede or modify the terms of any separate license agreement you may have executed with Licensor regarding such Contributions.

Furthermore, he complains earlier about problems with Section 3.c. There is no section 3.c There is a section 4.c which has roughly the provisions which he seems to be talking about.

I'm slightly disappointed. I usually expect better work from Eben Moglen. N.B., I have checked his references against the section numbers in the GPL and they don't refer to sections in the GPL.

Re:Not necessarily...

[be-fan]

The GNU C libraries are LGPL'ed, which include a specific clause that says that linking is different from copying code. This clause is necessary, because in reality, (static) linking isn't any different from just copying code.

Re:No.

[ross.w]

Actually, there's nothing stopping anyone from incorporating BSD licenced code into a GPL project any more than it stops Microsoft or Adobe.

It can't be done the other way though. Neither proprietary nor GPL code can be put under a BSD licence.

Re:No.

[caseih]

You're missing the point with regards to the GPL. Having a license that's not compatible with the GPL is fine. It's just that there are ramifications that need to be considered. If a library is licensed in a way that's not GPL compatible, all that means is that GPL'd programs cannot link against it. If that's the design of the author, then that's fine (a la Microsoft's licensing schemes for their WinCE SDK). In the case of XFree86 this is a big deal for distro makers because the majority of their software is GPL'd and under the new licensing terms there's no way to legally link the gpl'd code against the XFree86 libraries. In the case of Apache, there's not really a problem here; Apache has never been GPL compatible. You just have to be aware of the terms when you develop.

As for your falacies regarding freedom and the BSD and GPL licenses, those have been dealt with thoroughly before, except to say that your attitude works well for a small project, but for a large project (especially a large, self-contained project like a whole OS), the GPL offers you the developer more freedom and protection from greedy folks than the BSD would.

No, the GPL isn't the problem here. And it's hardly a virus. Just because Apache is not GPL compatible doesn't mean the Apache foundation is going to be pressured (or mysteriously forced by way of some magical IP beast) to go GPL. Let's end that FUD right here and now.

Re:Is anyone else getting worried here?

[pla]

So, you are suggesting that you be able to place restrictions on code that the original author did not place on the code?

I said nothing of the sort.

If you contribute to a "pure" GPL project, you agree to release your code under the terms of the GPL.

Similarly, if you contribute to Apache, you agree to release your code under the Apache (2.0?) license.

If, however, you release code to a GPL project which later adopts a different license, than, by the GPL, you have every right to demand either the removal of your code, or that the project remain GPL.

This doesn't seem all that complicated, IMO; I wonder what I said that gave you the wrong idea...

Even assuming, for whatever reason, you cannot withdraw your code, by someone releasing it against your will under a more restrictive license, you haven't lost anything at all - You can still release it on your own under the raw GPL, with no such restrictions. But offhand, I can't think of any scenarios where that would even apply, since my fourth paragraph pretty much covers the precursor to such a situation.

Re:Is anyone else getting worried here?

[pla]

Sure, we can still get Apache for free... but what happens when we want to make a few modifications and such? We get canned.

No. If you want to make modifications in violation of the non-GPL portion of Apache's license, you have every right (since they do base 99% of the license on the GPL) to release your changes under the pure GPL. You just can't contribute it back to Apache unless you agree to their additional terms. Nothing more, nothing less. They have even publically stated as much.

Re:oh no, rms doesn't approve!

[Ice_Balrog]

RTFA. RMS does approve of the Apache 2.0 license. But that doesn't make it any more GPL compliant.

Re:Is anyone else getting worried here?

[Losat]

Of course, a future version of the GPL may or may not apply to your favorite software. Quite a bit of GPL software omits the "or any later version" provision. This includes the Linux kernel itself, if I remember right.
Of course, once people see the new version, they may be willing to add "or Version 3" (provided that all relevant copyright holders can be tracked down).

Re:Is anyone else getting worried here?

[Losat]

It already has such a provision.

Finally, any free program is threatened constantly by software patents. We wish to avoid the danger that redistributors of a free program will individually obtain patent licenses, in effect making the program proprietary. To prevent this, we have made it clear that any patent must be licensed for everyone's free use or not licensed at all.

...

7. If, as a consequence of a court judgment or allegation of patent infringement or for any other reason (not limited to patent issues), conditions are imposed on you (whether by court order, agreement or otherwise) that contradict the conditions of this License, they do not excuse you from the conditions of this License. If you cannot distribute so as to satisfy simultaneously your obligations under this License and any other pertinent obligations, then as a consequence you may not distribute the Program at all. For example, if a patent license would not permit royalty-free redistribution of the Program by all those who receive copies directly or indirectly through you, then the only way you could satisfy both it and this License would be to refrain entirely from distribution of the Program.

However, it has this available geographic restriction.

8. If the distribution and/or use of the Program is restricted in certain countries either by patents or by copyrighted interfaces, the original copyright holder who places the Program under this License may add an explicit geographical distribution limitation excluding those countries, so that distribution is permitted only in or among countries not thus excluded. In such case, this License incorporates the limitation as if written in the body of this License.

Re:I disagree

[foandd]

If I write a program utilizing a GPL'd binary lib, the above lines state that I've just written a GPL'd program.

Wrong, wrong, wrong, wrong, WRONG!!! How the fuck do people manage to screw this up so completely and so consistently?

If you write a program utilizing GPLed code, and you distribute said program, you must release it under the GPL in order for that distribution to be legal. Should you fail to do this and get caught, you basically have three options:

1. Release your program under the GPL
2. Quit distributing the program until the GPLed code is removed
3. Contact the original author and get him/her to let you use the code under a different license

Under no circumstances does your program automatically become GPL; copyright law doesn't allow for it, and the GPL doesn't try to do it. It simply says that if you wish to use GPLed code, you can do it under certain conditions. If those conditions aren't acceptable, then you may not use the code. Should you do so anyway, all they can do is make you stop using that code! They cannot make you change the license on yours. Is that clear enough?

Re:Is anyone else getting worried here?

[JohnFluxx]

I was going to mod you down, but decided to reply instead.

This is why when you add a GPL license to your code, you either say "Placed under the GPL version X" or you say "Placed under the GPL version X or later".

Since the license is versioned, you can change the GPL, and not run into problems with changing the license on code people didn't want the license changed on..

Re:Is anyone else getting worried here?

[TKinias]

scripsit tiger99:

Seems to me that as it is feasible to make non-Linux based distros (BSD and maybe others) of similar functionality to a typical Linux distro, with and without GNU bits and/or X or Apache, there maybe should be a new overall brand name, emphasising the word "open" or "free" or similar. Under that general heading it would say on the box, in quite big letters: "Contains Linux, GNU/FSF, X, Apache..." so that the emphaiss was not entirely on the kernel, or indeed on any one other major component. Maybe the FSF need some marketing men to work out the branding.....

Such a thing has already been done: It's called Debian [debian.org]. While the only mature kernel at the moment is Linux, you can also run Debian with HURD [debian.org], FreeBSD [debian.org], and NetBSD [debian.org] kernels. It's all Debian, though, regardless of what kernel you run or whether a particular system has X, Apache, or whatever.

Re:Is anyone else getting worried here?

[Phillup]

Well, I read the quoted part to say that you thought "standing sphexishly by" section six to be a problem... and that you thought additional restrictions should be allowed.

Which sounded a lot to me like: after the code is released we should be able to place additional restrictions on it.

Then you say...

Even assuming, for whatever reason, you cannot withdraw your code, by someone releasing it against your will under a more restrictive license, you haven't lost anything at all...

To which I say, yes... you have lost something. You have lost the ability to insure that the end user has access to the source code that you wrote.

Someone may have specifically choosen the GPL because of the inability to restrict access to the code. (Which is my point.)

Note that I'm not talking about changes by the original author. If they release the software under a specific license, then that is done. It is their code and they can always release it under another license if they want. (Keeping in mind that it is still out there under the other license, also.) What I am talking about is how the software is released by someone else based upon the original license.

So, for example, taking a chunk of BSD code and converting it to GPL would be OK... because the BSD license let's you do just about anything with the code. But, going from GPL to BSD would be a no-no because you could have a binary only release and thus restrict who and how the original GPL source is furhter used... against the wishes of the GPL code author.

And my point (again) is that the original GPL code writer may have choosen the GPL because they wanted to make sure that the end user always had access to the source code... that the access was not restricted.

(Perhaps our different POVs are due to the term restricting?)

Re:Is anyone else getting worried here?

[GigsVT]

Are you sure you are both talking about the same kind of ASP here? (Active server pages vs Application service provider)

The way I understand the problem, a GPL server-side app might send out pieces of itself, either in the form of static HTML which is GPLed, or snippets of what could be considered executable code (Javascript, etc), which are also GPLed.

If this counts as distribution of GPLed code, then many people who make modified works of GPLed server-side web apps might be in violation of the GPL by not distributing the rest of the source. It seems to be a gray area right now.

OpenBSD rejects new license.

[JonMartin]

OpenBSD ships with Apache. Not two hours ago Theo de Raadt sent this message to the OpenBSD mailing list:

And another license issue..

The new apache license is not acceptable. Code written under that new license will never go into our tree.

Look, I am quite frankly getting sick and tired of this. It is time for the user community to tell these software developers who have gotten themselves involved with lawyers to stop it. They are NOT making their software better, they are NOT protecting anyone, and they we NOT making their software any more free when they add new terms.

As of this moment in time, therefore, it looks like the httpd in OpenBSD has now become a fork. It will continue to be managed under the existing license.

XFree86 on Monday and now this. What project will screw their license up tomorrow?

Re:the problem is, that we have to be this specifi

[zsau]

But you fail to grasp the difference between commercial and propriatary. Ximian Gnome is a commercial program. It is free. Many freeware Windows apps a propriatary programs. They are non-free.