Permissive licenses (or "licenses" like SQLite's Public Domain declaration) like BSD/MIT have lower barriers for use by many organizations, and many of them will contribute back because having a bigger community involved in maintenance benefits them, even without a GPL-shaped gun to their head.
That community effect can make BSD/MIT/etc. better for users.
The idea that GPL is "best for users" seems to rest on at least one two assumptions, both of which are flawed:
1. No one considers licenses when they select free software, so that someone who would use a piece of software under a permissive license will use it under a GPL-style license (discounting the fact that there are organizations that simply won't touch software with GPL-style terms, that would use it with BSD-style terms),
2. People always and only give back when legally required to do so.
There are lots of communities around GPL licensed software as well, so that's kinda a moot point, no?
The assumptions you cite are not the primary reason GPL software is considered "best for users" (by some). The second point is not even relevant for users but developers.
GPL software is guaranteed to stay free. That's what's relevant for the user.
(Personally I like both kinds of licenses, so I'm not arguing for/against either)
> There are lots of communities around GPL licensed software as well, so that's kinda a moot point, no?
No, the point wasn't "only permissively licensed software can have communities" (which would be ridiculous), it was "the fact that many organizations are less receptive to the GPL can result in larger potential contributing communities for permissively-licensed projects".
> GPL software is guaranteed to stay free.
GPL software is no more guaranteed to stay free than permissively licensed software is.
The closest thing to that that is a real difference between GPL and permissively-licensed software is that, assuming no radical change on the part of the FSF (or, alternatively, assuming the "or any later version" option isn't used), GPL software is guarantted to not have legal non-free derivative works not separately licensed by the copyright owner. Which is relevant to the degree of control that the copyright owner can exercise -- and why it is popular for commercial Open Core schemes -- but not particularly relevant, at least in a positive sense, for users.
> "the fact that many organizations are less receptive to the GPL can result in larger potential contributing communities for permissively-licensed projects".
"Can" result, perhaps, but there seems to be little real word data behind that claim. I've heard this a few times, but I've never seen any numbers to back it up. As a counter example: The Linux kernel seems to be the most commercially backed OSS project of all. If what you say would be true, why would the major players not flock to the BSD:s instead?
> not particularly relevant, at least in a positive sense, for users.
What you say about the GPL is only true for projects where copyright is assigned away from the contributor. And a lot of developers as well as users tend to shy away from those for just the same reasons you state.
> "Can" result, perhaps, but there seems to be little real word data behind that claim.
There's a reason I use "can" instead of "does". Its very hard to quantify the effects of licensing since you can't easily isolate it from other contributing factors.
> As a counter example: The Linux kernel seems to be the most commercially backed OSS project of all. If what you say would be true, why would the major players not flock to the BSD:s instead?
That's a very good point. OTOH, while there are widely used GPL RDBMS, which ones have the non-first-party commercial backing of SQLite or Postgres?
> What you say about the GPL is only true for projects where copyright is assigned away from the contributor.
From the point of view of a downstream developer, its more of a concern with assignment, true. In any case, its not a positive benefit for users.
> "Can" result, perhaps, but there seems to be little real word data behind that claim.
The wording of the claim is particularly relevant. You seemed to acknowledge it, but then ignored it. The claim doesn't require data; it's using a priori reasoning. In particular, it assumes that the GPL is one potential road block preventing certain entities from using software. If a roadblock is removed, then that can result in a larger community. (N.B. Removing a road block, in and of itself, doesn't imply that more people will come.)
This is specifically one practical reason why I use permissive licenses. I want to maximize the number of people using my software. If I use the GPL, due to its infectious nature, people tell me they can't use it because their employer won't allow it. This conflicts with my goal of maximizing the number of people who can use my software.
> As a counter example: The Linux kernel seems to be the most commercially backed OSS project of all. If what you say would be true, why would the major players not flock to the BSD:s instead?
That's not a counter example. The claim in question here does not imply that GPL projects cannot have large communities.
>In particular, it assumes that the GPL is one potential road block preventing certain entities from using software. If a roadblock is removed, then that can result in a larger community.
But on the other hand a permissive licence could just as well be a road block preventing 'certain entities' from contributing code, and if it had a copyleft licence it could result in a larger amount of contributions.
In short, in lack of any data to back either of these hypothesises means they are just that, hypothesises.
There are successful collaborative projects using both types of licences, andydroid pointed out Linux which is GPL licenced and also the largest collaboratively developed software project in the world, so certainly GPL is not a serious problem when it comes to getting contributions or use (Linux is practically everywhere).
Looking at the larger open source landscape, it's my impression that copyleft is predominantly used in larger, finished application type projects (of which Light Table is a perfect example), while permissive licencing dominates in component/framework style code.
At the end of the day it's up to the developer to choose the licence for _their_ code, both copyleft and permissive licencing fulfills a need, else they would not be so popular amongst developers.
>This is specifically one practical reason why I use permissive licenses. I want to maximize the number of people using my software.
Nothing wrong with that, but there's also nothing wrong with wanting to licence your code so that end users of your code and it's derivatives are given rights (which include the source).
> But on the other hand a permissive licence could just as well be a road block preventing 'certain entities' from contributing code
A permissive license, by definition, cannot prevent an entity from contributing to it. I've never heard anyone tell me they won't use a permissively licensed project for any reason related to its licensing.
> In short, in lack of any data to back either of these hypothesises means they are just that, hypothesises.
If you still think this claim requires data, then I'm afraid you're missing the point. The claim is using a priori reasoning, and makes an assumption about removing road blocks from using open source projects. The claim does not make any quantitative claims relating permissively licensed projects and GPL projects.
In any case, making a quantitative claim here is nearly impossible. There are too many confounding factors.
> Nothing wrong with that, but there's also nothing wrong with wanting to licence your code so that end users of your code and it's derivatives are given rights (which include the source).
I don't understand your point here. I didn't say there was anything wrong with that. I merely stated that the claim made by dragonwriter actually factors into my decision to use permissive licenses. i.e., Anecdotally, the GPL is enough of a barrier that I perceive a permissive license as better if I want to maximize the number of people using my software.
>A permissive license, by definition, cannot prevent an entity from contributing to it.
Nor can a copyleft licence 'prevent' an entity from contributing it. It is a choice, just as someone could choose to only contribute to permissive projects they could also choose to only contribute to copyleft licenced projects, so I fail to see what point you are trying to make here.
>I've never heard anyone tell me they won't use a permissively licensed project for any reason related to its licensing.
Now you bring up 'use' which is different from 'contributing'. And given that copyleft requires the source code to be open, it can't be used with proprietary projects (which of course permissively licenced code can), but that is by design as copyleft exist to give end users the rights which proprietary software typically removes.
>and makes an assumption about removing road blocks from using open source projects.
And I pointed out the assumption is flawed since he (dragonwriter) talked about 'contribution'.
Saying that a lot of 'entities' are happy to _use_ permissively licenced code is not the same as them _contributing_ code under permissive licences.
> I merely stated that the claim made by dragonwriter actually factors into my decision to use permissive licenses.
And I merely pointed out that there are other factors than 'maximising code use' which developers may consider when licencing their code.
Oh, and I'd like to take the opportunity to thank you for Wingo, really like it!
I never said the GPL was a roadblock for people contributing code. I said it was a roadblock for people using it.
Pure and simple: someone works for an employer whose policy is not to use GPL'd code. Therefore, they cannot use any code I publish under the GPL license.
No such policy exists (that I've heard of) for permissively licensed code.
> Now you bring up 'use' which is different from 'contributing'.
You're the one who said "contributing"! :-) I initially said "use" several comments ago: "I want to maximize the number of people using my software."
> Oh, and I'd like to take the opportunity to thank you for Wingo, really like it!
Thanks :-) Come hang out on IRC/FreeNode at `#wingo`. (Although it's kind of dead.)
Yes it is. Given the counter example one might just as well argue that the GPL can attract more commercial contributors than a permissively licensed project, or even that such licensing is a "roadblock" for some contributors.
You really find this sort of "a priori reasoning" meaningful?
A counter example demonstrates a claim as false by assuming it is true, and the demonstrating that, in reality, it is false.
Your counter example does no such thing. It merely shows that a GPL project can be popular. In particular, your example does not denote any relationship between GPL and permissively licensed projects.
> You really find this sort of "a priori reasoning" meaningful?
Yes. Re-read my last comment to you. I explicitly described how it was meaningful. I'd rather remove roadblocks to using my software than add them.
That community effect can make BSD/MIT/etc. better for users.
The idea that GPL is "best for users" seems to rest on at least one two assumptions, both of which are flawed:
1. No one considers licenses when they select free software, so that someone who would use a piece of software under a permissive license will use it under a GPL-style license (discounting the fact that there are organizations that simply won't touch software with GPL-style terms, that would use it with BSD-style terms),
2. People always and only give back when legally required to do so.