Choosing an OSS license
choosealicense.com
choosealicense.com
This is a site for beginners to OSS licensing. If this is your first time, just choose the MIT. If it's not the first time, you get to worry about whether you want to restrict reuse with the GPL, whether you prefer the BSD (2-clause or 3?), whether you agree with Stallman's philosophy enough to put up with the downsides of imposing GPL on your users, whether patents affect you, whether you need copyright assignment, whether to set up an entity to own your code, and all sort of other license bullshit.
First timers should use MIT, every time.
For an example of embrace and extend look at some of the libraries usurped by Microsoft such as Kerberos. I bet they wish they'd have used the GPL.
Open Source is free labor, but only if you MIT.
I spent a little time and made a decision. I think everyone should do the same. Don't let someone embrace and extend your work unless you want them to.
But, yes, if they're hell bent on not learning about the differences between the licenses, they probably don't want the GPL.
My personal view is that I should optimize for people getting involved. If it gets so successful that a company wants to use it without contributing back, its a nice problem to have (and it doesn't particularly bother me).
I know you have a lot more in OSS than I. Has this affected you much? Would love to hear any experiences - it seems you're sour on it and there might be a story there?
From another point of view (also one of mine), your license says something about the kind of community you see yourself being a part of. And, when you make this choice, you are excluding certain kinds of participants and inviting others. But community is not just a license, it's also how you treat others, how you see them, and how they see you, and the kinds of activities they find interesting (and/or sustainable - sometimes but not often the same thing).
A license is also a cost. For some people this will be a moral or ethical cost. The more desperate the startup the bigger an issue cost starts to become. Related is whatever small bit of return the startup can get - this is the babyfood that it will need to survive and grow.
Ultimately, I think our systems of economics (and if you pretend we only have one you are fooling yourself) demand diversity. There cannot be only one good choice when talking about all startups. There can be one good choice when talking about a specific startup (but for a few, multiple licenses might be a good thing).
Not that I have ever spent any time thinking about these kinds of issues...
Finally, note that while what I am saying here is so overly general as to be entirely useless, there are other [almost conflicting] ways of thinking about these issues which can also have value for a startup.
Carry on...
Unless Microsoft tracked down and managed to wipe out all places that were distributing the code that Microsoft started from, they did not "usurp" the code. Anyone is still free to start with the same code Microsoft started with.
Since this erupted into a fireball, here is a link explaining how code licensed under a permissive license can be included in a project that accepts GPL contributions:
https://www.softwarefreedom.org/resources/2007/gpl-non-gpl-c...
Second, changing the license is a permission granted only to the copyright owner. Otherwise I could take BSD-licensed code, relicense it to WTFPL, and then include it in my code without any credit to the original author.
Most licenses do not permit any sort of modification or version upgrade. The standard GPL boilerplate is extremely unusual for giving third parties a range of permitted versions.
> almost all non-GPL open source code can be relicensed as GPL.
For what it is worth, this is where this argument often ends up.
You can reason that developers are users if you can reason that software is a tool.
You can reason that developers are not users if you can reason that software is a product.
You can reason that developers are both users and not users if you can reason that software is an ingredient.
And that is where it gets tangled.
There is absolutely no incentive to build GPL software into products because of the covenants it requires.
There is little incentive to not use GPL tools because the covenants result in a lower cost for the tools if they are covered by the GPL.
The effect of this is that the majority of users of products which contain GPL'd software, are developers. [1]
[1] Or family members of developers, my Dad has a router with OpenWRT running on it but he could not have installed it, nor would the company that sold the router ever use OpenWRT as the supported firmware for the product.
Basically, you can convert the same code from BSD to LGPL to GPL to AGPL, but you cannot go in the other direction.
> Okay, so changing a license to GPL does not violate the
> conditions of the BSD.
> You just have to maintain the little notice and the
> copyright.
> If you want you can call it GPL+BSD but it's
> obviously identical to GPL.
When two works are combined, it is very important to know that their licenses are combined.
> Basically, you can convert the same code from BSD to
> LGPL to GPL to AGPL, but you cannot go in the other
> direction.
There is no such wording in the BSD license.
https://en.wikipedia.org/wiki/GPL#Compatibility_and_multi-li...
Are you disagreeing with the FSF's lawyers here?
> Are you disagreeing with the FSF's lawyers here?
If you say that it's legal to change the license of BSD-licensed code to GPL, that means you think it would be legal to:
1. Download the FreeBSD kernel.
2. Replace the LICENSE file with the GPL.
3. Replace the authorship and copyright comments with your own name.
4. Publish the resulting code online.
Do you really believe that doing that would be legal? If so, that is such an extremely unusual position that I do not know what evidence would be required to correct you.
https://www.softwarefreedom.org/resources/2007/gpl-non-gpl-c...
I'm sure that one could contrive a GPL wrapper project to contain the BSD kernel. Nobody ever said anything about removing copyright notices or breaking license terms.
> Nobody ever said anything about removing copyright
> notices or breaking license terms.
Your earlier posts stated:
> almost all non-GPL open source code can be relicensed as GPL.
> changing a license to GPL does not violate the conditions of the BSD
When I say relicense a BSD project under the GPL, I mean take the existing codebase and go through whatever legal steps are necessary (without additional permission from the copyright owners) such that the resulting distribution can accept future contributions under the GPL.
So do you now understand what I mean, and see that it's possible with the right legal wrangling, per the link from the Software Freedom Law Center?
Not requiring this downstream violates the BSDL, but requiring it downstream, violates the GPL, at least in GPLv2, which does not itself require it, and prohibits additional terms to its own.
I think the GPLv3's allowance of additional terms is broad enough to actually make some of the licenses that the FSF promoted as GPL compatible under GPLv2 but which were clearly not on a reasonable reading of the terms of the licenses compatible with the GPLv3.
http://www.quora.com/Can-you-convert-code-under-a-simplified...
While the FSF has actively promoted this view, for many actual open source licenses for which they promote the view, it requires a rather creative (at best) reading of one or both of the two licenses involved to support this claim.
> While the FSF has actively promoted this view
Nope, only those that want to distribute their changes of GPL licensed projects. You're free to do whatever you want and keep the changes to yourself as long as you never distribute outside your machines or your organization (this allows closed in-house development).
Which non-GPL open source licenses allow relicensing of the code? I've heard this claimed about the BSD code in a GPL project, but I was unaware the license permits that code to be relicensed as GPL.
Think about it: if you can release BSD code as closed source software, how is that okay but relicensing as GPL is not?
You are confusing source with binary on the close source side. The source is BSD and cannot be relicensed without the copyright holders permission. The closed source project has BSD-licensed source code that has no restrictions on distribution of the binary including adding a EULA.
The GPL project that has included BSD code cannot relicense the code to GPL without permission of the copyright holders. The BSD gives you no permission to relicense.
https://www.softwarefreedom.org/resources/2007/gpl-non-gpl-c...
Relicense has a very specific meaning.
No, it would not be legal to add a clause unless you were the copyright holder of the BSD code.
> You may not impose any further restrictions on the exercise of the rights granted or affirmed under this License. For example, you may not impose a license fee, royalty, or other charge for exercise of rights granted under this License, and you may not initiate litigation (including a cross-claim or counterclaim in a lawsuit) alleging that any patent claim is infringed by making, using, selling, offering for sale, or importing the Program or any portion of it.
The BSD does not say anything like that, so you are allowed to add restrictions. This is why you can place the GPL alongside the BSD, because you are adding restrictions. The BSD has nothing that isn't already present in the GPL (it is "GPL-compatible"), whereas the GPL has lots of additional restrictions, but since the BSD does not prohibit this, it is fine.
That is not how copyright or the BSD license works. You don't get to change the terms, since BSD doesn't give you permission to change the terms. Only the copyright holder can change the terms.
http://www.groklaw.net/article.php?story=20070114093427179
The analysis is clear that it is contrary to the consensus opinion, and there is further discussion refuting it in the comments. It has never been tested in court.
Since the commenters over at Groklaw are much more informed than either of us, and there is much more written than we could ever hope to figure out going back and forth like this, I propose we call a truce.
First timer or not, either they care or they don't about said legal details. If they don't care, then the choice doesn't matter. Pick what ever license is used by libraries, or pick randomly by dart. If they do care, then they surely care enough to read up on the subject and pick based on their (or their community's) liking.
http://www.gnu.org/licenses/gpl-faq.html#NoMilitary
What it ensures is that the source code will be released along with any uses. If a products uses your code and someone requests that code from them, they should be able to point to your repository, or theirs, and provide a copy.
Now let's say one day you change course and want to close the source for future development. That is completely okay. Only the code you have already released will be out there under the GPL and the future work will be covered under your new license. Any open source libs you use, you of course have to provide the source for. What this does for all of us is it ensures the source code is available to the developer/consumer and they can modify it as needed.
Just some background, I released an OSS SMTP library in 2000 and I've never had a problem using the same library in any commercial product and in fact I went on to build a small component business on it by offering new closed-source features, such as S/MIME, and support contracts to go along with the purchase.
Here it is: http://sourceforge.net/projects/opensmtp-net/
On the subject of:
Now let's say one day you change course and want to close the source for future development. That is completely okay.
It may not be okay though. If your code incorporates other people's GPL code, you are unable to close that. And due to the "viral" nature of the GPL, the fact that your code may work together with someone else's GPL code could mean that you can't close your own code either. So you have to be very careful.
It is true that the GPL does not explicitly prohibit commercial use, but this is not the same thing as being broadly compatible with commercial use. In general, the GPL is hostile to commercial use by a whole variety of restrictions, including the requirement to share source code. This may not be immediately obvious to OSS newcomers who just want to share their code and have people use it freely.
I liked that site and for the life of me cannot find it again. Things like "can I copy Apache licensed code into this LGPL code" were answered really simply, even when the answer included comments around making changes available, impact on copyright notices etc.
Does anyone know of anything similar these days? Don't get me wrong, the site linked here is great, but for me, at least, the fuzziness is in the legalities of combining code...
I was looking up OSS licenses the other day and happened upon it.
Edit: It looks like sandyarmstrong also posted it below.
Thanks!
As a guy releasing code, I don't want you to feel obligated to me in any way. I don't want you to have to attribute back to me, or even acknowledge in any form that you've used my thing. I certainly don't want to do anything silly like force you to open source the rest of your code.
I just want to give stuff away. It doesn't seem like that's very possible these days.
"Wait," you're thinking, "I didn't say sell. I said give away!"
Ah, yes, but were you planning on accepting pull requests? Because exchange of money is not the only way to sell something. There is also the concept of "In Kind" payment (http://en.wikipedia.org/wiki/In_kind). So, you make a library with something like a WTFPL license attached. I download your library and use it for my product. Then, to show my gratitude, I submit a pull request with a new feature to one of your other projects (an "In Kind" payment). Later, I find that your first library has a bug that's causing problems for my customers...
...so I sue you for violating the Implied Warranty that you didn't disclaim in your WTFPL license.
This is why all of those open source licenses have the ANNOYING ALL CAPS clauses explicitly disclaiming any implied warranty.
The More You Know!™
So, as far as in kind is viewed by the state, software has been left alone.
For my own code, I prefer the BSD license.
Anyone who wants to take it and use it in their own project (whether open source or not) is free to do so.
For me, this provides more "freedom" than the GPL does. I don't care if others provide their changes back to me or if they take it and incorporate it into a successful commercial project and make a few dollars from it.
I'm a big fan of OS X (as are many others on here). I don't think it would exist today if not for the BSD license.
I write free software, and share it with other people who write free software. I don't share it with people who write closed source software. They are free to write their own versions of my software. After all, that's exactly what they expect me to do, write my own versions of their software.
I'm a big fan of Linux (as are many others on here). I don't think it would exist today if not for the GPL license.
> After all, that's exactly what they expect me to do, write my own versions of their software.
This part is particularly well explained, thank you.
Of course you do. GCC was shared to great effect with Apple.
Furthermore, when somebody writes MIT/BSD code (for example Apple), they share it with you because you can put it under the GPL. But when you write GPL code, you don't share it with them because they can't relicense it as MIT/BSD code.
"Furthermore, when somebody writes MIT/BSD code (for example Apple), they share it with you because you can put it under the GPL. But when you write GPL code, you don't share it with them because they can't relicense it as MIT/BSD code."
That's fine. They're happy for me to not share any changes I make with them. They explicitly told me this when they released it under MIT/BSD. That's their choice, and they are free to make it.
I wouldn't be happy with that arrangement if it were my code, so I choose a different license. As is my choice. If I spend time working on their code, my changes are going to be released under the GPL, because I don't want my code to be used in closed source software. Where's the problem?
So yeah, it sucks that they can't re-use my code. But that's their choice. And I'm assuming they're happy with their choice because of the license they chose. if they weren't happy with this arrangement, they'd have chosen the GPL.
[edit] Also, even if I do this, I am still sharing my code with them, as long as they are happy to keep it out of closed source products. If they think it's more important that people can write closed source software with their code, than integrate my code into theirs, that's fine. Again, their choice. Everyone is free to do what they want. Even people who write closed source software. They are free to open up their code to me if they want to use mine.
Thank you for putting it so well. I'm amazed by the number of comments on this page that disparage GPL developers for not sharing reciprocally, while promoting licenses that intentionally don't require reciprocal sharing. It's nuts.
My biggest concern in this project is "abandonment." There are lots of fantastic books that have since been abandoned by their authors and are no longer suitable to assign for a class---sometimes the author died, sometimes they moved onto other projects; a few key chapters might still be relevant, but I can't expect students to buy 9 different textbooks and read one or two chapters from each one. There are even completely free textbooks (Bruce Hansen, an extremely well-respected econometrician, has one[3]), lots of teachers have online lecture notes that are quite good, some of the stuff on MIT's OCW are excellent, etc. But I don't know what anyone's plans are for maintaining those resources, what direction they plan to go, whether they will keep them free in the future, and so on.
What I'd really like to do is be able to maintain my own copy of this material so that I can continue to add to it if the original author loses interest, correct typos, etc. For most of these books that's impossible because they aren't licensed for that (even if the book is out of print). But even for Creative Commons books it's not feasible, because they only distribute the pdf. If you look at Hansen's book (link below; it's not CC, but it's representative of the material), you'll see that there's a lot of math in these books, so backing out a human editable version of the text from the binary would be a lot of work.
So, I think that there should be free textbooks and I know that textbook projects are often abandoned (or eventually made non-free) by their authors. So it's pretty clear to me that the best way to have there be free textbooks is to write one (unfortunately, since it's a lot of work, but I do have lecture notes from when I've taught these classes as a starting point) and to use a license that entitles anyone who gets a copy of the book to also have a version they can edit and maintain. The GNU Free Documentation License[4] is the only license I know of that does it (it's the GPL of documentation), so I chose it instead of any of the CC licenses.
The FDL is more burdensome than I need it to be; if someone prints out a chapter of the book and hands it out to his or her class, I don't really need every student to get a printed version of the FDL (I'm fine with the professor just telling them about the FDL), but on balance I think the license makes the project stronger. No one needs to worry that I'll abandon or derail these books, because the license guarantees that they can fork them (no matter how they originally got a copy of the books).
I don't think that people who use CC licenses for educational material are unprincipled. Even though I think that the FDL is a better for projects like mine, it's great that they're making the material widely available (even if I worry that it will eventually be abandoned, the fact that I can redistribute their binaries is very nice); and I definitely don't see why they'd think that I'm greedy for choosing the FDL.
[1] http://www.econometricslibrary.org [2] https://github.com/EconometricsLibrary [3] http://www.ssc.wisc.edu/~bhansen/econometrics/ [4] http://www.gnu.org/copyleft/fdl.html
The FDL also allows commercial use and textbooks can be expensive; I think it's likely that a publisher would try to bundle or repackage a free book if they could make money off it, and I don't think they're too likely to make it obvious that the book could be downloaded for free unless they're required to. CC NonCommercial would be less free than the FDL; I don't mind someone else making money off the books per se (especially if they are providing real value, but even if they're not), but a lot of the strategies that many textbook publishers use to generate revenue don't serve students very well and would weaken the project.
Like I said, I appreciate the effort put out by people making educational material CC, and I appreciate it even more if they make the source available. But for this type of project I think the source being available is the most important thing so I've chosen a license that makes it mandatory. For other projects I might decide differently.
I would still personally choose a permissive CC equivalent of a BSD license though (CC BY 3.0), if only because I want other people writing books with permissive CC licenses to be able to incorporate my material. Also, for a closed textbook, a publisher might want to take a couple chapters out of my book, or 10 of the images, and I'm not okay with limiting the amount of quoting.
If you're approaching this from an "ethics of sharing" stance, then what what you're actually restricting is somebody else's use of their own code. Your code never disappears/changes/gets relicensed, it's their code that you're staking a claim to.
Now, if you're approaching this from a stance of withholding certain valuable rights to your code, such as the ability to be the only person that licenses closed source usage, then the argument is a bit more logical.
> They are free to open up their code to me if they want to use mine.
I do. Under the MIT license.
This is some very bizarre reasoning. They still have their code. They are still free to do whatever they want with it. If they want my contributions in their product, they made a serious error in their license choice. Maybe they should have used the GPL.
But do you know what? They can use my code. They are completely free to use my code. As long as they only use it in GPL software.
"I do. Under the MIT license."
By saying this, you are saying to me that I am free to re-release your code under the GPL and add my own contributions. If you are not happy with this arrangement, you have made an error in your license choice.
I have not made any errors in my license choice. It behaves exactly in the manner I want it to.
There's no way to resolve the conflict. We just have incompatible values and understandings about what sharing means. Our values are more liberal (the FSF uses the word permissive), so that means that you're free to exist in our world. We'd like to be part of your world as developers, but we aren't willing to embrace your restrictions.
You wish I'd chosen differently. I also wish you had chosen differently, as currently from my point of view, you're helping the propagation and spread of closed source software.
Simply: I don't want my code to be used in closed source software, so I can't use your license. That choice is simply not open to me.
No, I'll remain content with my license agreement, I'll just think that you're a dick.
That's OK, though -- some people behaving like dicks is simply the cost of collaborating openly.
If you don't want people to do something particular with your code, yet you choose a license that permits it, then by definition, you chose the wrong license.
I'm going to chalk your childish, immature and unprofessional insults down to frustration. And to prevent further noise, I am not going to reply to any further comments made by you.
1) Rude.
2) Simply the cost of doing business under the BSD license.
I don't write (or publicly release) much code but I truly want anyone to be able to do whatever they wish with it (whether the result is open or closed). Unfortunately, the GPL doesn't allow that.
I think a World full of closed source software would be dramatically worse than a World full of open source software. This is why I love the viral nature of the GPL.
Of course, the chain can't end with me and I'll have to publish my own changes. But it's worth it, I'm not a proprietary software asshole, and in the end I'll benefit from other people's changes and insert it back in my own code and improve my sales that way. You, me, the other programmers, Stallman, the users, everybody wins!
That's how I see it anyway. In the end, you have your own priorities and if you like to give without always receiving, that's understandable too. Ah, tradeoffs...
Permission to use, copy, modify, and/or distribute this software for any purpose with or without fee is hereby granted, provided that the above copyright notice and this permission notice appear in all copies.
Then that annoying all caps standard 'I AM NOT HELD LIABLE' sentence:
THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
Done :D
It's just about as permissive as you can get and it's more professional that the WTFPL license.
MIT license do not fulfill, as it require copyright notice and permission notice.
WTFPL do not fulfill the requirement, as it require edited software to change name.
Unlicense do fulfill the OPs requirements.
It really deserve to be upvoted above the MIT suggestion.
If you're afraid of offending people and therefore want to relicense under CC0 or MIT or whatever, you can, because WTFPL allows you to do WTF you like. Obviously.
For example, for LGPL, it doesn't mention you must allow customer modification and reverse engineering for the purpose of debugging those mods.
Even people who do everything else right, still mess this up.
Some licenses are just not very easily summarized, because they are complex and nuanced.
There needs to be a concise summary of how these licenses differ, one that they can comprehend without reading the whole license text. Website is a good idea, but IMO, needs a lot of work, and may even be harmful in it's current state.
I'm also not sure why they promote the GPLv2 (without + clause) over the GPLv3 either. That may convince someone that people will always share their code, and doesn't mention tivoization. Seems like a deliberate ploy from someone who dislikes the strong copyleft of GPLv3 family to prevent people using them.
I think an improvement would be to have multiple options and to filter down licenses until only a few remain: so eg, when I click "I care about sharing improvements" on the front page, it gives me another few options, like "I want people to share their modifications for hosted websites", or "I want people to be able to use in commercial apps" - which would filter to AGPL and LGPL respectively.
To distribute a product that uses GPL'd code, the rest of that product must also be under an open-source license, but there is no requirement that everything be licensed under the GPL.
For most discussions about the GPL, people call the GPL "viral" from a misconception that it can't be combined with or depended on by non-GPL code. This belief is obviously wrong, but it persists.
Both the GPL and LGPL can be combined with and depended on by non-GPL code, so if the LGPL is not "viral", then neither is the GPL.
As I gave as my example, I can use a library licensed under the LGPL, and link it with code licensed under the CDDL. I cannot do that with a library licensed under the GPL. With that GPLed library, I'd have to use a license that the GPL is compatible with. So, if I had a product that I wanted to link with both sides, I would need to license my code under a license that is mutually compatible, such as BSD or MIT; I would be restricted from using certain other copyleft licenses.
> Viral as in requiring my code to be compatible with the
> GPL ecosystem, and in restricting my choice of licenses.
> As I gave as my example, I can use a library licensed
> under the LGPL, and link it with code licensed under
> the CDDL. I cannot do that with a library licensed
> under the GPL.
> So, if I had a product that I wanted to link with both
> sides, I would need to license my code under a license
> that is mutually compatible
You're reinventing definitions to suit your argument. The BSD license creates no restrictions on what you can combine it with. The restrictions are defined by to the GPL.
If I released some code with a license that said:
This work may not be combined with any work which has a
license containing more than twenty capitalized letters.
So if you define "viral" to mean "a license which can be incompatible with other licenses", then the GPL would fit that definition, but so would many other licenses (including the BSD license). Thus, it is not a useful definition.
There is no useful definition of "viral" which covers the GPL but excludes the LGPL.
Debian obviously thinks this is a problem, or else they would not have any problems with the GPL incompatible OpenSSL license.
Source: https://lwn.net/Articles/428111/
For example, the OpenSSL license is also incompatible with the LGPL; does that mean the LGPL is "viral"?
I am not anti-GPL. My original and only point was that the distinction between GPL and LGPL is important to explain, and the situation with PostgreSQL at Debian shows there are real world implications of this.
> For example, the OpenSSL license is also incompatible with the LGPL; does that mean the LGPL is "viral"?
You have any source for this? LGPL was created to allow linking with any software. If you just copy paste source they are not compatible.
> My original and only point was that the distinction
> between GPL and LGPL is important to explain
This is most relevant when using static linking, but also applies if you're distributing your program with a bunch of included dynamic libraries. The LGPL is important for proprietary applications because it lets them (for example) bundle a version of GTK+ that they know will work properly, so they don't have to depend on whatever version is installed on the user's system.
> You have any source for this? LGPL was created to allow
> linking with any software. If you just copy paste
> source they are not compatible.
The LGPL offers a workaround, which is to separate out the LGPL'd code entirely into a separate library, and then set up your build system so users can relink against a modified version of the LGPL'd code. In this case, the OpenSSL code is being treated as if it were a proprietary blob. So it solves the problem of distribution, but with the hassle of maintaining that clean library separation.
For example, I could write an application that uses GNU Readline, and release the source under the BSD license, and that would be perfectly legal.
Claiming that dependent code must be GPL'd is outright FUD.
> You may convey a work based on the Program, or the modifications to produce it from the Program, in the form of source code under the terms of section 4, provided that you also meet all of these conditions:
> c) You must license the entire work, as a whole, under this License to anyone who comes into possession of a copy. This License will therefore apply, along with any applicable section 7 additional terms, to the whole of the work, and all its parts, regardless of how they are packaged. This License gives no permission to license the work in any other way, but it does not invalidate such permission if you have separately received it.
Also, read this in the FAQ: https://www.gnu.org/licenses/gpl-faq.html#IfLibraryIsGPL
The linked FAQ is:
> If a library is released under the GPL (not the LGPL),
> does that mean that any software which uses it has to
> be under the GPL or a GPL-compatible license?
>
> Yes, because the software as it is actually run
> includes the library.
For section 5, consider the implications of your quote, especially the final sentence:
> This License [...] does not invalidate such permission
> if you have separately received it.
This isn't wholly true - the compatibility is only one way, which was back to my original point. "Compatibility" with the GPL is, according to the FAQ:
> In order to combine two programs (or substantial parts of them) into a larger work, you need to have permission to use both programs in this way. If the two programs' licenses permit this, they are compatible. If there is no way to satisfy both licenses at once, they are incompatible.
A BSD work which depends on a GPL work therefore, cannot be "Compatible" with the GPL because it fails to meet the criteria previously quoted: "You must license the entire work, as a whole, under this License ...".
What you are trying to claim is that you can write BSD code which is API compatible with readline, but if you are not distributing readline with it, it is not a dependent work. (ie, there are other works which are API compatible with readline). The condition here is that APIs cannot be copyrighted. In that case, you can distribute only your code which uses the API under your license of choice, but you cannot distribute the work as a whole under anything but the GPL.
If your code makes calls to readline functions, and you distribute readline with it, it is undoubtedly a "covered work" as described by the GPL, and is subject to section 5.c of the license which states that the whole work must be GPL if distributed
So you can distribute a BSD project with readline, and recipients are free to choose the BSD license for the BSD part and then distribute a closed source binary linked against a compatible readline replacement. The GPL does not invalidate the permissions granted by the BSD license.
The only goal of the FSF here is that the source code for anything distributed linked with readline be made available under the GPL. They don't stipulate that it only be made available under the GPL and not any other license.
The GPL imposes major headaches / restrictions on potential adopters. This reduces uptake, which reduces contributions, thereby reducing sharing of improvements.
And your claim of restricting potential developers is entirely speculation. e.g, see Linux vs any of the BSDs.
There is something contradictory about requiring others to act "more freedom promoting."
Other more "permissive" licenses can be used to constrain anyone though, because the developer sees the license was only for him, and those permissions are not relevant for his end user.
I "care about patents" -- does that mean that I want users to not violated my patents, or that I want to grant my users a license to my patents?
As this is a github project, I'd definitely like to make it known that it would be really cool to have some integration with the license on github's side on a per-repo basis. Asking people to create a license when creating a public repos at the same time as a README could really help get people thinking about licensing.
Their new repo page points to choosealicense.com and lets you choose any of the major OSS licenses, though it does show the MIT, GPL3, and Apache ones in bold.
I'd definitely like the option to add a license to my repo after creation though, instead of having to manually add a license to the repo and then pushing it. A UI for asking collaborators if it's okay to retroactively open-source their work under a specific license would be useful too.
Additionally, If you are distributing any source material (images, source code and so on), then any modified file need to include a notices stating that the file has changed. You may also not remove copyright, patent, trademark, and attribution notices in said files.
If the Work includes a "NOTICE" text file (also any credit buttons with similar content), you may not remove it.
Thats about it.
I was somewhat disappointed by the dearth of selections presented by the site.