Comment by mkleczek 2 months ago
Isn't it just a matter of choosing the license properly? To be honest, I find it contradictory:
On one hand OSS authors select permissive license exactly because they want big corps to use their software.
On the other hand - OSS authors are unhappy about big corps using their software in a way that license allows them to.
Just stick to GPLv3 / AGPL.
> I don't understand why an OSS author would select anything more permissive than that:
Honestly? Because many things are too small to be worth paying a lawyer to write a letter to license violators.
If I'm not willing to sue someone for disobeying the license on a library or tool, I just use a maximally permissive license. And at a good lawyer's hourly rate, it has to be a pretty big project before I'd even care.
So for minor projects, MIT or Apache it is. Or I just CC0 it. I wrote the code because I found it useful, and I decided that it wasn't worth the often heartbreaking effort of building a sustainable business around it. So if someone else finds my code useful, that's great!
(EDIT: See discussion below for why I don't bother with a license I'm not willing to enforce.)
Respectfully, I think you have it completely wrong. Copyleft licences don't force you to sue anyone.
> Because many things are too small to be worth paying a lawyer to write a letter to license violators.
Then don't! I would advice you never even consider it, that would be absolutely crazy!
But what you miss is that big companies using your library have a lot of money, and therefore it is a risk for them to not comply with your licence. What does that mean in practice?
* Big companies will favour permissive licences. If you publish your library under a permissive licence, you screw the other small libraries that have a copyleft licence (that would be an argument for a permissive licence). * Big companies can't use strong copyleft libraries in proprietary software, so that would be a no-go. * Big companies can TOTALLY use weak copyleft libraries in their proprietary software. But if they do, they have to distribute their changes. Because not doing it would be a risk they are not willing to take.
> Respectfully, I think you have it completely wrong.
Whenever someone starts a sentence with "respectfully", it's usually because they're about to say something disrespectful. Definitely the case here.
GP doesn't have it completely wrong. There is no wrong or right. The person who is doing the building gets to decide the license, and any reason or motivation they have for picking a particular license is by definition correct, because that's what they want to do.
> If you publish your library under a permissive licence, you screw the other small libraries that have a copyleft licence (that would be an argument for a permissive licence)
No one releasing something under an open source license has any sort of responsibility or obligation to some random other open source project they may not even know about.
> There is no wrong or right.
Only alternate facts?
I answered to a comment that said "I don't use copyleft because I don't want to have to pay a lawyer", which to me is a completely invalid point. It's like saying "I don't use copyleft because I don't want to eat bananas". You can use copyleft and not sue, just like you can use copyleft and not eat bananas.
> No one releasing something under an open source license has any sort of responsibility or obligation to some random other open source project they may not even know about.
I meant that permissively-licensed libraries tend to be favored, so that's an advantage when competing with similar libraries. If you care enough to choose a license but you are not trying to compete with other open source alternatives, then it would be nicer to go for copyleft.
I know permissively-licensed projects that were inferior to copyleft alternatives but got more traction because... well because companies could freeride on the permissive license. In that case those who chose the permissive license knew exactly what they were doing and decided to compete this way instead of technical merit.
> Big companies can't use strong copyleft libraries in proprietary software, so that would be a no-go.
Just to clarify: I have actually been involved in open source license enforcement efforts, because big companies do violate strong copyleft licenses regularly. Or their subcontractors do.
Usually, if you're willing to pay a good lawyer to yell at them, you can get the domestic ones to stop. And I know of some enforcement organizations that manage to get voluntary settlements on the order US$15,000, or who get the violators to appoint a "compliance officer."
For a big, commercially important project? This is often worth it. For a modest CLI tool, or something which I put a few weeks of full-time effort into? It is not remotely worth my time to enforce a strict license. This is a purely personal tradeoff. But if it's not worth $400 to me to have a lawyer write a letter, then I'm going to use a permissive license.
> But if it's not worth $400 to me to have a lawyer write a letter, then I'm going to use a permissive license.
Again: if it's not worth $400 to have a lawyer write a letter, just don't have a lawyer write the letter! But you can still have the copyleft license... for the situation where it may be worth a lot more in the future. Or just to give leverage to fellow developers: if I use a copyleft library at work, then I can tell my managers that I must contribute my changes upstream, during my working hours.
> Honestly? Because many things are too small to be worth paying a lawyer to write a letter to license violators
That’s a false alternative. You can chose copyleft license and not sue.
This is _especially_ useful for things I don’t plan to monetize. It allows free use but also protects the society from free riders.
> It allows free use but also protects the society from free riders.
That's a value judgment that others may not share. I'm not (quite) one of them, but there are people who believe that bits on a hard drive shouldn't be owned, by anyone. I think it's a perfectly reasonable stance to argue that releasing something as a public good under the least restrictive terms possible is a moral and correct thing to do.
Personally I prefer strong copyleft for most projects that I really care about. But that's my preference, and I don't kid myself into believing that I know what's right or wrong for society in this area.
I agree with you and I am a fan of EUPL, but even if it is pretty much the same as MPL 2.0, Google, for example, has an explicit ban on it just like it has on AGPL3.
Probably there's no real reason apart from legal saying "meh, we can't be bothered about reviewing this"
As I don't care about Google using my code or not, I choose EUPL, but it's worth mentioning that some companies only accept permissive licenses that grant them the right to do basically whatever they want with your code.
Free software has almost lost, now that MIT has become the default, pushed hard by corporations and its employees.
> Free software has almost lost
Not sure what you mean. GPL-style licensing is still very popular among open source projects. Sure, there are perhaps a lot more projects that are MIT-licensed (and similar), but that doesn't detract from the body of thriving GPL-licensed software.
> On one hand OSS authors select permissive license exactly because they want big corps to use their software.
I think you're reading too much into people's motivations. When I release something under a permissive license, it's because I don't care about it enough to license it under something like the GPL. Or I just don't want to deal with the possibility of there being license violations that I'd feel responsible for dealing with (if I'm not going to, then why bother licensing under a copyleft license?).
For the most part I don't really care who (if anyone) uses the stuff I release. Building a community around an open source project that I started could be fun and rewarding, sure. But honestly I'm not sure I'd even want big corporate users, since they're likely to expect things from me that I'm not willing to provide.
> On the other hand - OSS authors are unhappy about big corps using their software in a way that license allows them to.
Regardless, you're painting all of us with a very broad brush. Please don't assume anyone's motivations or licensing decisions fall under some simple, one-size-fits-all rubric that you have in your head.
There is a whole spectrum between AGPL and permissive. MPLv2 or the EUPL [1] basically say "you can use it as a library (and link it without the concerns of LGPL) but if you modify/extend the library, you have to distribute the code of the modified/extended library".
I don't understand why an OSS author would select anything more permissive than that: big corps can use MPLv2/EUPL libraries in proprietary software just fine.
[1]: https://joinup.ec.europa.eu/collection/eupl/how-use-eupl