[kwlug-disc] Permissive vs copyleft licenses

Mikalai Birukou mb at 3nsoft.com
Thu Dec 10 11:26:44 EST 2020


>> Can I connect this to Monday's talk about voting?
>>
>> "If you don't pay attention to politics, time comes when politics comes
>> after you."
>>
>> Choice of licensing is a political choice. You set the policy around
>> your software.
>>
>> In my talk I've referenced Belarus. I think it is possible to find
>> RMS-like figure in Belarus politics. Someone saying right things for
>> decades, scorned for sticking to principles. Now population is waking up
>> to a nightmare they've helped create. Ironic? Is it the right word?
> Personally, I have already made a political commitment to the FOSS movement. I use FOSS software, I contribute to FOSS projects, and I run my own FOSS project.
>
> I don't think that the choice of FOSS licence is as important as people are claiming here. I am disappointed at the level of sectarianism and polarization within the FOSS community about the choice of licence.

"Paying attention" in politics means spending of mental effort to 
elaborate about particular details at hand. When we don't go into 
details, we abstract into some concept. And we have observed that any 
non-trivial abstraction leaks.

>   I call bullshit on claims of hellfire, damnation and apocalypse if I don't use the GPL and choose a more permissive licence instead. The comparison of using a permissive licence with the terrible situation in Belarus is very extreme and also inaccurate.

I am trying to find what moves people. I am not talking about current 
terrifying event in that country. I am comparing a place were majority 
have forgotten and didn't care about actual important details, while 
there are figures that were blubing about fundamentals, seemingly 
trivial low level, "why should I care?", nerd-y details, meticulous 
repetition, ... like RMS?

Its a story about paying attention to details. That's it. If there are 
other possible suggestions, I didn't mean them. Comparison from reality 
have this tendency.

> The GPL doesn't have the magic powers that people attribute to it.
GPL captures reciprocity, and captures it strongly. And reciprocity in a 
form "my default action is do good to this new person I just met". (RE: 
Richard Dawkins, first computer models, results of competition.) This 
touches deep.
>   MongoDB and Redis have both discovered that the AGPL 3 does not, in fact, protect them from competition from cloud providers who steal their code and use it to compete against them. As a result, MongoDB and Redis have both relicensed their code under non-FOSS licences.

Let's get down to details. There is a project with some initial license. 
Contributors give their labor under an implicit understanding that this 
legal setup is what there is. Then implicit understanding is broken. 
Individual contributors feel cheated.

Is this an "understandable outrage" or a "religious zeal"? Is it a 
"hellfire"? And should we blame victims for their reaction?

Something tells me that if aforementioned examples didn't have 
"community of contributors" there would be no news around re-licensing 
event.

>   Those are two famous examples, but the same story is playing out with other companies as well, who are taking their code proprietary in the face of competition from cloud providers. [Eg, see this recent post: https://joemorrison.medium.com/death-of-an-open-source-business-model-62bc227a7e9b]

When GPL was articulated? Were then compute clouds around? What was the 
age of GPL, when compute clouds started to appear?

In this historic perspective, I'd say GPL is a success. Yes, life is not 
static, life is moving on. But it worked, didn't it? Even if I start to 
use now FreeBSD desktop, I understand that it would be mostly 
GPL-licensed code, from projects that used GPL license to inspire 
growth. Right?

<plug>Let me plug in an advert for January talk. We'll discuss 
architectures. Just licensing trick is no longer enough. Some 
architectures are simple, but have not been invested into, not been 
explored, cause they are resistant to monopolization.</plug>

> Just as the GPL is less effective than claimed at protecting you from bad actors stealing your code and competing against you, the GPL is also less effective than claimed at "forcing" companies to contribute source code back to the community. The reality is that modifications to GPL'ed code that you extract from a company by threat of lawsuit over licence violation don't have a lot of value. The situations where actual value is created is where the company *wants* to contribute back to the FOSS community, because now they are actively working with open source teams, and often also employing team members, contributing changes back in a form that the open source team can use. And in the situations where companies do that, the software is more likely to have a permissive licence than not. Of course I know a lot of people are paid by their companies to work on the Linux kernel, which is GPL'ed. But my previous employer employed one of the FreeBSD developers to speed up the TCP/IP network stack, contributing those changes back to FreeBSD. The reason for choosing FreeBSD was the permissive licence: my employer wanted to use this code in their own proprietary kernel. I've derived much benefit from the LLVM project, which Apple has poured a ton of money into. Again, Apple only did this because of LLVM's permissive licence. Apple would not have funded the GCC team for similar work.
And we may bore ourselves also with questions about business model at 
every particular instance. Details, details. ... :) ... I have code to 
write. A-a-a-a-a-a => /dev/null
> So the GPL is just another FOSS licence. There is no evidence-based moral reason for choosing it over another FOSS licence, it's a personal preference. By and large, corporations contribute to the FOSS community not because they are forced to by the GPL, but because they are founded by and employ engineers who have made the same political commitment to FOSS that I have. Projects with permissive licences are more likely to get corporate support and contributions, and may get more contributions from the developer community due to the perceived downsides of GPL code (ie, I'm contributing to this project but I can't use it at work).
>
> The GPL is important because it inspires religious zeal in people, due to its powerful narrative about making people free from the power of evil corporations. Important software has been built by people inspired by this ideology.

Let's encourage "paying attention" to details, talking details, avoiding 
generalizations that hide details.

That Google and Apple stance against GPL is a pressure to "not pay 
attention" to details, to not be "political"(TM). When a person trashes 
GPL without going into specifics, can't it be described as "religious"? 
When an office of lawyers produces a talking point that is not caring 
about details that matter to you, is it a "zeal" to say "hey!". Lawyers 
should know better. Is their omission intentional?

>   Permissive licences are also important to the health of the FOSS community; they are more attractive to libertarians and people who are skeptical of religious narratives. And permissive licences are responsible for a lot of money being pumped into FOSS software development by corporations, which gives us more choice of FOSS software and higher quality FOSS software to choose from. It takes all types of people to make a healthy FOSS community, so let's try to be tolerant.
As a random example of permissive: LGPL was specifically created to put 
devs on non-free side onto a drug of freedom. Cute little detail :)





More information about the kwlug-disc mailing list