Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Bug: Not GPL. :-) #834

Closed
pnathan opened this issue Mar 22, 2016 · 68 comments
Closed

Bug: Not GPL. :-) #834

pnathan opened this issue Mar 22, 2016 · 68 comments
Assignees

Comments

@pnathan
Copy link

pnathan commented Mar 22, 2016

Hi,

I love this project, but infrastructural software really should be under the GPL. It's very important that distributed software can be fixed by the end user, which the MIT licenses & friends doesn't require.

Please consider releasing as GPL.

Regards,
Paul

@bencook
Copy link

bencook commented Mar 22, 2016

More importantly, as derivative works of the original C coreutils, this must be licensed under the GPL as well.

@ghost
Copy link

ghost commented Mar 22, 2016

@bencook Why v2 but not v3?

@bencook
Copy link

bencook commented Mar 22, 2016

@ivegotasthma It would be whichever the original program is licensed under.

@willneumob
Copy link

It's a complete rewrite of, a) some POSIX commands, and b) other utilities commonly used on *nix systems. There are already BSD-licensed implementations of most of these tools on FreeBSD. Please don't license this as GPL; GPL is significantly less "free".

@ghost ghost mentioned this issue Mar 22, 2016
@cnd
Copy link
Contributor

cnd commented Mar 22, 2016

people please, don't start this holy war here, it's not constructive, first listen what owner think about this, if he don't mind then voting from contributors - there is just no other way

@pnathan
Copy link
Author

pnathan commented Mar 22, 2016

@bencook - I would suggest demonstrating the "derivative work" claim - that is a legally binding claim placing requirements on the derivators. I support moving to the GPL but such a claim should be demonstrated. :-)

@willneumob LGPL is ideal for libraries, but I don't think it affects programs like this: the mandate would apply, and people aren't generally going to be using a libcoreutil (I think... I've never done so or seen it done in any sense, and I've been floating around the scene for > 10 years... if that means anything. ;) ).

@johnstorey
Copy link

For what it's worth, the first thing I checked was the license. Then I moved the project off my list of potential things to work on.

@willneumob
Copy link

@pnathan Well, for a shell script, they are similar to libraries. If I write an extensive shellscript making use of these programs, might I be forced to release my works? I'm not suggesting people would want to actually link it against their code as a library. L is for Lesser, not Library.

@cnd
Copy link
Contributor

cnd commented Mar 22, 2016

@willneumob so people currently share all scripts which are using coreutils?

@willneumob
Copy link

@johnstorey Because of GPL or MIT? You're always free to fork the MIT version if that's what you wanted (as far as I understand the law)...

@Heather Well, no, but the coreutils project might technically be within their right to pursue action against someone who does not? I mean, what's the difference between a C developer using a library's API and linking against that library, and a shell script running these commands? It might be fine, and IANAL; just food for thought. I don't feel that strongly about it. But LGPL might allow more flexibility, while still requiring changes to this codebase itself to be shared.

@ghost
Copy link

ghost commented Mar 22, 2016

@willneumob

Well, for a shell script, they are similar to libraries.

Uhh...no

If I write an extensive shellscript making use of these programs, might I be forced to release my works?

Read up on what the license says.

I'm not suggesting people would want to actually link it against their code as a library. L is for Lesser, not Library.

Please read the license and the wiki page for it.

@willneumob
Copy link

I guess LGPL is fairly specific about using shared libraries, so, you're right, it might not apply. I still stand by shell utilities being abstractly similar to libraries, as they are external binaries and the "API" is the command line interface and arguments to the program. But I give up :) It's really not my thing to have an opinion about as I haven't contributed code to this project. Much respect to those who have, and it's their opinions that should matter in this.

@yincrash
Copy link

I apologize for any inconvenience posting to HN has caused. I hope that the incoming interest and possible future contributions will outweigh any distress.

@christianbundy
Copy link

Those who optimize for the freedom of software will have a copyleft fork, those who optimize for a free public will have a permissive fork. If you aren't happy with the current license, fork. The problem solves itself -- there's no need to harass @Seldaek about their obviously intentional decision.

Just a quick note: it looks like there's an existing GPL fork, which contains a license violation. You're welcome to impose a copyleft license on your software, but that doesn't mean that you can remove the author's license just because you find it inconvenient. @ivegotasthma please re-add the original license to your GPL fork as soon as practical. Thanks.

@ghost
Copy link

ghost commented Mar 22, 2016

@christianbundy What's the violation? The MIT is there until a certain point, after that everything is GPL. Isn't that how it works?

@adelarsq
Copy link

@ivegotasthma need something like neovim did (more than one license).

@Dreae
Copy link

Dreae commented Mar 22, 2016

@pnathan I don't understand the argument that it's important for the end user to be able to fix the software. The MIT license certainly doesn't prevent that, if it doesn't work right just fork it, fix it, and rebuild it.

Unless you're worried about someone down the line creating a non-FOSS fork of this and distributing it, in which case just don't use their software.

@jahewson
Copy link

@ivegotasthma, the violation is that in your GPL'd fork you've removed the copyright notice which says, amongst other things:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

Obviously, you can't do that. You need to keep the original notice intact and add to it, by saying something such as:

This software incorporates portions under the following license:

  [original license]

The remainder of this software is licensed under the GPL:

  [GPL]

Of course until you actually write some code, you don't have any remaining portions to which the GPL applies. What you can not do is claim that the original code is under the GPL, because it isn't. You can incorporate it into a GPL'd work (and have all your subsequent changes be under the GPL), but you can't change the license of the original work.

@christianbundy
Copy link

@ivegotasthma Here's the text that you'll required to include (emphasis mine):

Copyright (c) Jordi Boggiano

Permission is hereby granted, free of charge, to any person obtaining a copy of
this software and associated documentation files (the "Software"), to deal in
the Software without restriction, including without limitation the rights to
use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of
the Software, and to permit persons to whom the Software is furnished to do so,
subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

I'd recommend including the original license above the GPL (like @jahewson's example) so it's chronological, but I'm not a lawyer and Neovim's method is probably fine, albeit confusing.

@jahewson
Copy link

I'm not going to weigh-in on the GPL vs. MIT debate, but I'd like to debunk the following remark from @bencook:

More importantly, as derivative works of the original C coreutils, this must be licensed under the GPL as well.

Writing a compatible piece of software is not copying and does not constitute a "derivative work" because it is not an adaptation of the original work. It is its own unique work which, while deliberately designed to expose the same interface does not copy the internal workings of the original. In fact, many aspects of source code are not eligible for copyright protection at all, while others are but may be freely copied as "fair use".

It's popular to cite methodologies such as clean room design as being somehow necessary for developing software clones which are untainted by copyright claims, but as even the Wikipedia article points out: "Clean room design is usually employed as best practice, but not strictly required by law.". This is perhaps easier to understand intuitively when one realises that software is protected as a "literary work", under the same rules which apply to books and magazines: one does not expect the author of a fantasy novel to have never read a fantasy novel. Indeed, an author who had never read a fantasy novel is unlikely to write a fantasy novel at all.

Just as the original GNU coreutils did not infringe upon the copyright of the UNIX coreutils, neither does this project, as far as I can see. The suggestion that it somehow automatically does is worrying because it brings with it the idea that one does not have the freedom to read a piece of software and then write something similar, which is at best untrue, and at worst an impediment to freedom.

Sorry to be all legal and boring but it's important to get these things right. 😄

@adelarsq
Copy link

@johnstorey same as me. It's so hard to work with GPL projects.
@christianbundy since there is an explanation I don't think that really matters. But chronological order is better to follow.

@christianbundy
Copy link

@jahewson You're absolutely right. It's even been argued in court, like that one time that Oracle tried to sue Google for writing OpenJDK, because it used the Java API, just like this project uses the GNU coreutils API. It's possible that other countries may reach a different verdict, but all the evidence points to this being totally fine.

@jahewson
Copy link

@christianbundy Yes, that particular case is due to be heard all over again, this time regarding fair use. All bets point to Google winning and the world will certainly turn upside down if they don't. Notably, the FSF backs Google's defence of what they call a user's freedom to use APIs.

@ddevault
Copy link

I wanted to correct this:

More importantly, as derivative works of the original C coreutils, this must be licensed under the GPL as well.

So far as I can tell, the original C code has never made an appearance in this project. GPL is not required, and I'm supportive of the MIT choice.

@bencook
Copy link

bencook commented Mar 23, 2016

@jahewson What evidence do you have that this reimplementation was written in a "clean room" manner to avoid the possibility of being seen as a derivative work?

@Arcterus
Copy link
Collaborator

So far as I can tell, the original C code has never made an appearance in this project. GPL is not required, and I'm supportive of the MIT choice.

This is correct.

@Arcterus
Copy link
Collaborator

@yincrash It's not a problem. Thanks for posting. :)

@ddevault
Copy link

@jahewson What evidence do you have that this reimplementation was written in a "clean room" manner to avoid the possibility of being seen as a derivative work?

This doesn't need to be written in a clean room setting. GNU coreutils does not own the idea of the core unix tools. These are standardized tools specified by POSIX and implemented many times over. GNU wasn't even the first to implement these tools.

@Marqin
Copy link

Marqin commented Mar 23, 2016

For those waiting for creator's optinion - he's in travel.

@ghost
Copy link

ghost commented Mar 23, 2016

And the person who opened the issue doesn't post anymore. Well, this end the troll.

@pnathan
Copy link
Author

pnathan commented Mar 23, 2016

Hi, as the filer of this bug, I would ask that personal comments be not present. @ivegotasthma @StefanCristian - looking at you. (Personally, I'd appreciate it if you edited out those comments from the thread).

Seems like there's a lot of confusion in the thread about what this really means. The crux of the decision is whether the creators would want a company to be able to package these tools into a non open source system. The answer for these sorts of tools is usually, no. Would the creators of this software be okay with, as a hypothetical example, Apple replacing their BSD tools with these on OSX , making significant enhancements to them, and not having to share those enhancements?

This is precisely what I am asking: Is this ok? I don't want to contribute to systems that other people can grab without sharing my contributions & theirs out. If, however, uutils goes GPL, I am very likely to contribute to it.

@pnathan
Copy link
Author

pnathan commented Mar 23, 2016

@Annwenn Perhaps it's better to have the community discuss things, rather than stirring the pot. (And, oh, I have a job, and it was night. :-p).

@jahewson
Copy link

@jahewson: I am allowed to sublicense it, which technically means I can put it under GPL.

Yes, you are allowed to sublicense but you're not allowed to relicense. So you can add the GPL to the existing MIT license (it's compatible). But you can't remove the existing license, because that existing license is the only thing which grants you any right to copy or distribute the original source code. By default copyright law gives you zero rights to someone else's work unless they grant you a license to some of those rights. Remove or violate the license and you loose those granted rights.

Technically, yes you can add the GPL to the existing license, thus making the entire project GPL'd. But in practice, the original portions of the code are still effectively just under the MIT license, because you haven't modified them, even though, yes, technically they are under the GPL now.

So my 'fork' doesn't infringe on anything.

Right now it infringes on the original MIT license, the rights of the original author to not have their license removed, and the moral right or the original author to not have their name and copyright notice removed from their work. It's also probably violating the GPL, because by breaking the license terms of the original work, you loose the right to sublicense it under the GPL.

It's best not to be gung-ho about copyright, especially where the rights of others is concerned. The GNU project provides plenty of high-quality information explaining how to use their licenses. You should read and follow their advice:

How to use GNU licenses for your own software.

If you have copied code from other programs covered by the same license, copy their copyright notices too. Put all the copyright notices together, right near the top of each file.

@Seldaek
Copy link
Collaborator

Seldaek commented Mar 24, 2016

My two cents as original creator of the project: GPL vs MIT/BSD-style is an endless argument, and I don't think we can please both sides, so I'd rather leave it as is.

That said, I have been quite out of touch here, so if the people currently maintaining things feel strongly one way or another, feel free to do the change.

@Seldaek
Copy link
Collaborator

Seldaek commented Mar 24, 2016

@Heather with GPL AFAIK you lose the freedom to do commercial things in a closed source manner. You could sell the project copyright I suppose, although I am not sure how that flies when you have X contributors. However the evil corporation buying it would still have to release the modified source if they make any modification.

That's the main conceptual difference between those two licenses, MIT lets you do whatever you want with the code (except remove the copyright/license), GPL forces you to release any changes you do as GPL again.

GPL is also viral which means that if you embed the code in something it usually "spreads" to the surrounding project and everything should then be released as GPL. This isn't as important for this project though as they are binaries and not really meant to be interacted with as libraries.

I tend to prefer MIT because it offers more freedom, but my understanding is that many fear it as it enables bad actors to run away with open source code and not contribute back to it. You can't really reconcile those, which is why we have endless discussions with about 50/50 of 👍 and 👎 above.

@cnd
Copy link
Contributor

cnd commented Mar 24, 2016

@Seldaek honestly yes... I can't be sure whether it's possible or not to switch from GPL even when full core team is agreed to be sold but at least it's much more complicated

@iyra
Copy link

iyra commented Mar 24, 2016

@Seldaek

I tend to prefer MIT because it offers more freedom

This is based on a very limited definition of "freedom" which is "the ability to do whatever you want". Yet everyone would seem to agree with limiting some freedom - such as forcing the inclusion of the original copyright notice when one re-distributes, as the MIT license requires us to do.

As such, if we can call the MIT license as enabling freedom and the GPL license not enabling freedom, we have a problem - where do we draw the line at what is free and what is not?

When one looks at a group of people instead of merely one person forking and redistributing a project, we can see that the GPL consistently "guarantees" freedom by legal methods (except in the case of violation) of a given piece of code. This freedom comes in the form of the four freedoms.

On the other hand, when we look at non-copyleft licenses, we see that no freedom is guaranteed. The only freedom guaranteed is the initial freedom to fork from the original project - in this case, the Rust coreutils. In this way, it is clear that the GPL encourages software freedom and can in fact be said to be more "free" as it has the capacity to provide freedom to more people and it prevents what we have seen happening in the case where people will take the code and give nothing back other than a little copyright notice hidden away in the documentation that very few people, if anyone, reads.

As the OpenBSD project writes,

In fact, it is stunning how little support of any kind comes from companies that depend upon OpenBSD (or OpenSSH) for their products, but there is no requirement of compensation.

The GPL would be an advantage to the project as it will allow the project to re-use code that is included in forks made by other people.

It's a shame that so many people term this discussion on a "immediate freedom" basis rather than thinking in terms of everyone who uses software now and in the future. It would be a real shame to see this code used in some popular operating system with substantial changes and not being able to see how the changes were made, or only be able to run the program in certain circumstances. For example, the integration of MIT/BSD/ISC/etc.-licensed code under an EULA with ridiculous conditions such as this:

You may make one copy of the Apple Software in machine-readable form for backup purposes only.

or even preventing you from engaging in a class-action lawsuit:

Any Dispute Resolution Proceedings, whether in arbitration or court, will be conducted only on an individual basis and not in a class or representative action…

@christianbundy
Copy link

@projectilemotion I understand that you're trying to be persuasive, but it's intellectually dishonest to pretend that people who choose permissive licenses aren't "thinking in terms of everyone who uses software now and in the future". There are two viewpoints, with both advantages and disadvantages, and if it wasn't a gray area, we wouldn't be having this discussion. We'd all love to see a world full of open-source software, and we decide how we want to do that with the license we choose. Ultimately, there are two sides of the continuum: fascism and anarchism.

  • A copyleft license is a threat to leverage the judicial system to punish people who use your ideas in a way you don't like. It does not create keep your software free, it keeps your software from being used as an asset in proprietary software. The notion that it keeps your software free is, again, intellectually dishonest. Copyleft licenses are a threat to sue someone if they use your ideas in a way that you don't approve of, and is inherently fascist. This option guarantees that the author can sue anyone that uses the author's work in any form of propriety, and optimizes for "free" software.
  • A permissive license is an offer to stand on the shoulders of giants, and doesn't discriminate against people based on how they want to use your ideas. Instead of threatening to sue people, permissive licenses create a laissez-faire economy of ideas with no (or few) requirements. Permissive licenses are inherently anarchistic, which allows for a community fueled by ideas rather than another walled garden. This option guarantees that anyone can use the author's ideas, without any sort of discrimination, and optimizes for a free public.

It's a decision between fighting proprietary software creators with government violence, or letting people do whatever they want (with the knowledge that you may disagree with them). I'm biased, but I prefer the Free Public License: it's more permissive than the MIT license and doesn't require attribution/etc. Just my two cents.

@pnathan
Copy link
Author

pnathan commented Mar 24, 2016

@christianbundy

It would probably be in the best interests of a civil discussion if you
removed the "f----ism" word, as that's a great way to attract misguided
rants. It's plausible the "a-----ism" word should also be removed too, for
that identical reason.

You are correct that one relies on the legal system to guarantee
libre-ness, and one relies on the goodwill of others. My opinion, having
worked in major corporations, is that goodwill quickly vanishes when large
piles of money enter the picture; we therefore must choose a system to
enforce good behavior (you are likely to understand that this is a classic
reasoning against "a---ism"). As an example, Amazon Web Services has
profited billions off of the hard work of people who have never received a
single red cent, or public recognition, or other compensation for the
vast piles of cash. I don't think that's moral; at the least, I think AWS
should contribute significantly back to the community. A license such as
AGPL would protect against this rapacious behavior (Concrete action: I have
asked AWS recruiters to not contact me until this behavior changes); the
code would be required to be released. There are a lot of nifty things
that AWS does codewise... be a global improvement if the world could use
them too.

On Thu, Mar 24, 2016 at 10:39 AM, Christian Bundy notifications@github.com
wrote:

@projectilemotion https://github.com/projectilemotion I understand that
you're trying to be persuasive, but it's intellectually dishonest to
pretend that people who choose permissive licenses aren't "thinking in
terms of everyone who uses software now and in the future". There are two
viewpoints, with both advantages and disadvantages, and if it wasn't a gray
area, we wouldn't be having this discussion. We'd all love to see a world
full of open-source software, and we decide how we want to do that with
the license we choose. Ultimately, there are two sides of the continuum:
fascism and anarchism.

A copyleft license is a threat to leverage the judicial system to
punish people who use your ideas in a way you don't like. It does not
create keep your software free, it keeps your software from being used as
an asset in proprietary software. The notion that it keeps your software
free is, again, intellectually dishonest. Copyleft licenses are a threat to
sue someone if they use your ideas in a way that you don't approve of, and
is inherently fascist. This option guarantees that the author can sue
anyone that uses the author's work in any form of propriety, and optimizes
for "free" software.

A permissive license is an offer to stand on the shoulders of giants,
and doesn't discriminate against people based on how they want to use your
ideas. Instead of threatening to sue people, permissive licenses create a
laissez-faire economy of ideas with no (or few) requirements. Permissive
licenses are inherently anarchistic, which allows for a community fueled by
ideas rather than another walled garden. This option guarantees that
anyone can use the author's ideas, without any sort of discrimination, and
optimizes for a free public.

It's a decision between fighting proprietary software creators with
government violence, or letting people do whatever they want (with the
knowledge that you may disagree with them). I'm biased, but I prefer the Free
Public License https://opensource.org/licenses/FPL-1.0.0, more
permissive than the MIT license, and doesn't require attribution/etc. Just
my two cents.


You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub
#834 (comment)

@jahewson
Copy link

@projectilemotion, what you say captures nicely the philosophical difference which underpins this debate:

As such, if we can call the MIT license as enabling freedom and the GPL license not enabling freedom, we have a problem - where do we draw the line at what is free and what is not?

When one looks at a group of people instead of merely one person forking and redistributing a project, we can see that the GPL consistently "guarantees" freedom by legal methods

Yet the GPL does so by limiting the freedom of those very people, to build upon and redistribute the work in the manner that they wish. So there's a philosophical contradiction - hence this discussion.

On the other hand, when we look at non-copyleft licenses, we see that no freedom is guaranteed. The only freedom guaranteed is the initial freedom to fork from the original project - in this case, the Rust coreutils.

Yes and no. The MIT license provides the same four freedoms as the GPL, it just doesn't force them onto others. It's up to other developers to continue to provide those freedoms. But it's not like they don't exist: For any MIT licensed software where the developer has made the source available I can run the program (freedom 0), change how it works (freedom 1), redistribute copies (freedom 2), and distribute my modified version (freedom 3). What I can't do is force others to reciprocate.

In this way, it is clear that the GPL encourages software freedom and can in fact be said to be more "free" as it has the capacity to provide freedom to more people and it prevents what we have seen happening in the case where people will take the code and give nothing back other than a little copyright notice hidden away in the documentation that very few people, if anyone, reads.

I can't agree with this paragraph. The GPL provides more of the "you're not allowed to do this" kind of freedom at the expense of the "you're allowed to do as you like" kind. But it can't be said to be providing more freedom overall, because it's taking away an awful lot of freedom from software developers. Indeed this distinction is known to political philosophers as positive vs. negative liberty and it's a balancing act with no right answer. In many way, that's what this discussion boils down to.

Indeed, by your logic the end users themselves are people who "give nothing back", yet you would seek to ensure their freedom at the expense of alienating developers [1, 2, 3, 4, 5, 6] who otherwise might be more than willing to give something back. Clearly "giving back" can't be the yardstick against which the right to freedom is measured, so that particular line of reasoning doesn't add up. Which brings me back to this quote:

people will take the code and give nothing back other than a little copyright notice hidden away in the documentation that very few people, if anyone, reads.

It's ironic that the only instance of someone taking from this project and giving nothing back, not even the legally required copyright notice, is the GPL'd fork. Especially, as due to the licensing incompatibility the authors of that fork will never give back anything to this project.

I'm really struggling with the phrase "what we have seen happening", as if there's some sort of ongoing problem with MIT licensed software where people "take the code and give nothing back". What we've seen as a general trend is a huge growth in liberally licensed open-source projects in recent years, at the expense of GPL licensed projects [link], and it's been a boom time for open source. There's more participation in open source by big business than there's ever been, and it's being fuelled by liberal licenses, not the business-hostile GPL. If anything we're seeing the opposite effect from the one you claim, that businesses are willing to open and collaborate on liberally licensed codebases [link], but they won't touch a GPL'd codebase, and were that the only option they would instead keep their code private (which is worse for everybody). What you're saying reads more like a conspiracy theory than an appraisal of the current state of open source development, where the trend towards liberal licenses is resulting in more open source overall, instead of a relatively small number of copyleft projects.

What you have is a fairly muddled and inconsistent argument which says "people are bad and can't be trusted" but "we know better so we'll force them to do as we say", which is best described as paternalism. I'm not saying you can't argue in favour of the GPL (you can, see my comment on positive vs. negative liberty) but this particular argument isn't going to cut it.

@iyra
Copy link

iyra commented Mar 24, 2016

@christianbundy

Ultimately, there are two sides of the continuum: fascism and anarchism.

I don't think it's fascism to tell people that you would like your code used in a particular way. Wouldn't you say that it's even more "fascistic", by your own definition, to allow the code to be used in software covered by EULAs made by the likes of Microsoft, Apple and Sony, which prevent class-action lawsuits and in fact limit for what purpose you can run the software? If one considers that the code can be used in these circumstances, it is in fact enabling "fascistic" use of your code. Please tell me how that "optimizes for a free public".

It does not create keep your software free

It guarantees, provided that all parties act legally, that the four freedoms are conveyed. Software that has the four freedoms I call "free", as per the definition of "free software". It can also be called "open source". It goes by the OSI's and FSF's definitions, so I call it "free". The GPL does keep your software free.

it keeps your software from being used as an asset in proprietary software

This is true. But to say that it can't both keep your software free and prevent it from being used as an asset in proprietary software is a false dichotomy. How's that for intellectual dishonesty?

Copyleft licenses are a threat to sue someone

No they're not. They are a set of conditions for which I allow my software to be used. We might as well say that the MIT license is a threat to sue someone if you do not include the copyright notice. The threat to sue can be made by choice of the copyright owner.

if they use your ideas

Now this is outright false. A piece of software is not an idea, it is a document. An algorithm is more like an idea. As you know, copyright does not cover ideas but rather it covers works.

is an offer to stand on the shoulders of giants

You're only allowed to stand on my shoulders if you promise not to jump up and down on them. Isn't that reasonable?

and optimizes for a free public.

Are we really free when we can allow people to take away our freedoms?

rather than another walled garden

I don't think you know what a walled garden is.

with government violence

The government has no part in civil cases such as suing for copyright infringement.

@jahewson

Yet the GPL does so by limiting the freedom of those very people, to build upon and redistribute the work in the manner that they wish.

We limit freedom in the "real" world - mostly, as it should be, the freedom to take away freedoms from others and to harm others. The GPL does the same in the software world by taking away your freedom to remove the freedoms that others would have had.

it just doesn't force them onto others

I never said it did. I said that the MIT license guarantees the freedom to re-distribute. The provision of the other freedoms is upon the re-distributor.

The GPL provides more of the "you're not allowed to do this" kind of freedom at the expense of the "you're allowed to do as you like" kind.

What exactly is the difference between these "kinds of freedom"? It is irrelevant, because both licenses give the four freedoms upon the initial publishing as you have previously mentioned. What freedoms, other than the freedom to take away the freedoms from others, are "not allowed" by the GPL?

Indeed this distinction is known to political philosophers as positive vs. negative liberty and it's a balancing act with no right answer.

These ideas are irrelevant to the discussion. The MIT license gives a specific list of what you are allowed to do. The GPL does the same, but in more words. Neither license says "do whatever you want except for these things...".

the end users themselves are people who "give nothing back", yet you would seek to ensure their freedom at the expense of alienating developers who otherwise might be more than willing to give something back

There is no obligation, in any free software license, for the users to "give something back". I think the concept of "giving back" only applies when one has "taken" - and as such, the one who is taking is the person who re-distributes the software. The developers have the same freedoms as the users, and there is no expense involved. The GPL guarantees that there is no trade-off has to be made between the freedom of the developer and the freedom of the user - something which the MIT license would in fact allow. The allowing of the developer to restrict freedoms. Now we have developer freedom at the expense of user freedom. Speaking in terms of the four freedoms, the GPL makes sure, legally, that every party has them. The freedom you don't have is the freedom to take away the freedom of others.

Clearly "giving back" can't be the yardstick against which the right to freedom is measured

It's not.

It's ironic that the only instance of someone taking from this project and giving nothing back, not even the legally required copyright notice, is the GPL'd fork.

This is none of my business and this is only one data point. Those who act in such a morally wrong fashion have no sympathy from me. If the license wasn't the GPL, would you be bringing this issue up?

will never give back anything to this project.

If there are changes made and re-distributed, then in fact they are giving back. You just don't want to take it because you chose the MIT license. But don't complain about people not giving back when you just don't want to take it. It's right there for you to use.

What we've seen as a general trend is a huge growth in liberally licensed open-source projects in recent years

Unfortunately true.

and it's been a boom time for open source

I never said it hasn't been. My point was that now we have a boom in open source (whether or not you want to attribute this to permissive licenses or not is none of my business) but we have a corresponding boom in the number of projects available for someone to grab and integrate into a proprietary product or project - and people have been doing just that. Sony with the operating system that runs on their consoles, for example.

I'd also take a bit of an issue with this "boom for open source" idea. Though there's more open source being written and used than there ever has been, there's less open source that guarantees the freedoms for the next person, and one of the largest markets in computing, the "app" market, is chock full with closed source software, a lot of which I have a hunch relies on permissively licensed libraries.

We're seeing just as much of a "proprietary boom" as we are an "open source boom". Do you want to attribute that to the MIT license too?

There's more participation in open source by big business than there's ever been

Whether or not you want to attribute this to permissive licenses or not is none of my business.

not the business-hostile GPL

The GPL is not necessarily "business hostile".

they would instead keep their code private (which is worse for everybody)

Like the MIT license allows people to do.

@christianbundy
Copy link

It would probably be in the best interests of a civil discussion if you
removed the "f----ism" word, as that's a great way to attract misguided
rants. It's plausible the "a-----ism" word should also be removed too, for
that identical reason.

@pnathan I think that asking me to censor myself is inappropriate, but if my words make you uncomfortable I'd be happy to use generalized words like "authoritarian" and "libertarian" instead. In the future, I'd recommend against trying to censor people who criticize your philosophies for being authoritarian -- a civil discussion of why you disagree with my terms would've been much more respectful, but that's neither here nor there.

You are correct that one relies on the legal system to guarantee
libre-ness, and one relies on the goodwill of others. My opinion, having
worked in major corporations, is that goodwill quickly vanishes when large
piles of money enter the picture; we therefore must choose a system to
enforce good behavior (you are likely to understand that this is a classic
reasoning against "a---ism").

As @jahewson mentioned above, this is incredibly paternalistic, and relies on the false notion that people owe me something for the ideas I've shared with them. I create open-source software because I care about the public, I couldn't care less about the software itself. The software I write is a means to an end, and I don't feel comfortable compromising the freedom of the public to guarantee the freedom of my software.

As an example, Amazon Web Services has
profited billions off of the hard work of people who have never received a
single red cent, or public recognition, or other compensation for the
vast piles of cash. I don't think that's moral; at the least, I think AWS
should contribute significantly back to the community. A license such as
AGPL would protect against this rapacious behavior (Concrete action: I have
asked AWS recruiters to not contact me until this behavior changes); the
code would be required to be released. There are a lot of nifty things
that AWS does codewise... be a global improvement if the world could use
them too.

Unfortunately, power imbalances are an intrinsic property of capitalism, and are outside the scope of this project. Intelligent people made the conscious decision to use permissive licenses, and treating them as victims of some proprietary corporate conspiracy doesn't solve the issue we're talking about. I'm not saying that you're wrong, because I'd obviously love for everyone to be perfectly compensated for everything that they give and take, but that's not how our economic system works (whether you're an open-source author, an AWS employee, or an AWS customer).

@pnathan
Copy link
Author

pnathan commented Mar 24, 2016

@christianbundy - I'm not an anarchist, nor do I subscribe to anarchist-syndicalist presuppositions, which your objections to the GPL are rooted in. If pushing to have a ground rule that improves society at large is paternalism/authoritarian (I don't agree that it is - I call that the fundamentals of a civil society), then so be it.


Here we have a chance to have a better core utils for GNU/Linux; let's take this opportunity and have a license that helps make society-at-large better through requiring spreading the source code via distribution. Lots of arguments for and against GPL have been made; I think that the primary talking points have all been hit.

I rest my case, and will not be replying further save to clarify points upon specific request to me, @pnathan. I ask, as the filer of this bug, to have a civil discussion without name calling or flamebait added. I also ask that unless something new can be added, to not have further comments, to avoid pointless bikeshedding.

I await the maintainers' decisions, myself.

@jahewson
Copy link

@projectilemotion

First of all, just to be clear, I'm not involved with this project.

The GPL does the same in the software world by taking away your freedom to remove the freedoms that others would have had.

You said it yourself, the GPL takes away your freedoms. Saying that's "as it should be" is an appeal to some sort of cosmic authority and not an argument in support of your position.

What exactly is the difference between these "kinds of freedom"? It is irrelevant

The freedom to do something and the freedom from something are two very different things. This is a major concept in the philosophy of political thought [link], it's generally called positive vs. negative liberty and it's generally seen as a dividing line between right and centrist politics vs. left wing politics. Any discussion on freedom needs to be based on the prevailing understanding of what freedom is and the forms it takes, otherwise you're just assigning arbitrary meaning to words.

What freedoms, other than the freedom to take away the freedoms from others, are "not allowed" by the GPL?

What you're doing here is dismissing a very important kind of freedom, the freedom for others to build upon your work in the manner of their choosing (and the freedom for the original author to allow them to do so). Your notion of "the freedom to take away the freedoms" is flawed because you're only looking at one kind of "freedom". Denying people the ability to build freely upon your work is not to give them freedom. In fact, an original author who does so is quite obviously exercising their freedom to take away the freedom of others, something which you claim should not be permitted.

Indeed this distinction is known to political philosophers as positive vs. negative liberty and it's a balancing act with no right answer.

These ideas are irrelevant to the discussion. The MIT license gives a specific list of what you are allowed to do. The GPL does the same, but in more words. Neither license says "do whatever you want except for these things...".

These ideas couldn't be more relevant to this discussion. They are the very essence of this discussion. Dismissing the entire body of philosophical thought on "freedom" is pretty much going to ensure that you're not making a valid or meaningful argument. The GPL does not have a monopoly on freedom and your dismissal or well-accepted thought on this topic shows a lack of awareness of the boarder context in which any discussion on "freedom" occurs.

If there are changes made and re-distributed, then in fact they are giving back. You just don't want to take it because you chose the MIT license. But don't complain about people not giving back when you just don't want to take it. It's right there for you to use.

No they're not. They're giving nothing back. What they're doing is taking the source hostage and saying "my way or the highway". That's not giving, it's taking.

but we have a corresponding boom in the number of projects available for someone to grab and integrate into a proprietary product or project - and people have been doing just that. Sony with the operating system that runs on their consoles, for example.

Why is this a bad thing? It's great that people can share software with others who use it to create products that people love. Do you really think Sony have never made any commits back to any of the open source projects they use? All I see here is a kind of resentment and jealousy that others are able to use their resourcefulness to create a successful business which utilises open source software, as if they have somehow cheated the original authors. This is such an uncharitable and petty outlook.

Though there's more open source being written and used than there ever has been, there's less open source that guarantees the freedoms for the next person, and one of the largest markets in computing, the "app" market, is chock full with closed source softwat re, a lot of which I have a hunch relies on permissively licensed libraries.

Again, you're only looking at one kind of freedom and ignoring the other. The original author of a GPL'd project has the freedom to build a proprietary product from that code, yet they deny that freedom to others. It's great that the app market is able to make such good use of open source code, I'm sure the authors of those projects would be thrilled to know that their code is being used to power real products in the real world. Why would they not?

We're seeing just as much of a "proprietary boom" as we are an "open source boom". Do you want to attribute that to the MIT license too?

Are we? Do you have any citations to back up this claim? If proprietary software fuels the release of more MIT licensed code then it could well be a good thing. Were GPL the only option instead of MIT, all of that proprietary code would certainly be kept private. You would rather it were that way?

The GPL is not necessarily "business hostile".

True, but it almost always is.

they would instead keep their code private (which is worse for everybody)

Like the MIT license allows people to do.

While the MIT license allows them to do so, the GPL license forces them to do so, because GPL is not a viable option for them. You're very much ignoring incentives here. Proprietary users of an MIT licensed code are incentivised to contribute back to the code, because doing so improves the code overall, which makes it more likely that others will start using the code and contribute too, and the business is able to reap the benefits of not just their own contribution but everyone else's too. Whereas for GPL'd code, proprietary users are simply unable to use the code, they won't be able to contribute back to the project, and will instead devote their (limited) resources to developing a proprietary alternative.

So the MIT license encourages everyone to share and collaborate together, and the incentives are naturally there to contribute back. Whereas the GPL forces everyone to into a system of sharing against their will, which as a matter of course generally removes their freedom to build a viable business which makes use of open source code. A business which would put them in a great position to devote resources to contributing back to the projects they make use of.

Again, I don't see anything more than paternalism as the basis of your argument, with a good dose of paranoia and mistrust of others. Not exactly positive thinking.

@iyra
Copy link

iyra commented Mar 24, 2016

@jahewson

The freedom to do something and the freedom from something are two very different things.

I agree.

the freedom for others to build upon your work in the manner of their choosing

I contend that this freedom can be broken down into three options:

  • the option to build upon the work as part of a proprietary program
  • the option to not build upon the work as part of a proprietary program
  • the option build upon the work and not distribute it.

Outside of these there are no options. This freedom would be partially taken away (i.e the first option is disallowed) but you can still choose between the others.

Your notion of "the freedom to take away the freedoms" is flawed because you're only looking at one kind of "freedom".

How so? You alluded to two types of freedom: the freedom to do something and the freedom from something. By using the GPL, one is taking away the freedom to make the code proprietary and re-distribute it (i.e the freedom to do something). Does the second kind of freedom, the freedom from something apply here?

Denying people the ability to build freely upon your work is not to give them freedom.

True. But I'm giving them the four freedoms that makes the work qualify as free ("open source") software.

In fact, an original author who does so is quite obviously exercising their freedom to take away the freedom of others, something which you claim should not be permitted.

The freedoms I talk about are the essence of the OSI and FSF's definition of what free and open source means. That is, the freedoms that would be given to you if you use an OSI certified license, or the four freedoms. These are equivalent in spirit. The freedom to distribute binaries without obligation to provide source is not one of the freedoms.

Throughout the whole of my discussion, I have been using "freedoms" to mean the four freedoms, or what would qualify under the OSI definition. These are all that matter to me, which in my original posting led me to the conclusion that the GPL would guarantee maximum freedom - that is, going by the four freedoms. Sorry for any confusion on this matter.

What they're doing is taking the source hostage and saying "my way or the highway".

Why is this unreasonable? Unless you're releasing your code into the public domain, this is what you are saying regardless of what license you use. With the MIT license, "my way" is including attribution. With the GPL, "my way" is that you're not allowed to take away the freedoms of users.

It's great that people can share software with others who use it to create products that people love.

It's unfortunate that people don't love freedom as much as the products they use.

has the freedom to build a proprietary product from that code

If you were the copyright owner then you could do that too. This is a moot point because it is impossible to take away your own freedoms guaranteed to you by copyright law, unless you assign the copyright to someone else.

It's great that the app market is able to make such good use of open source code

I disagree, but it's useless arguing about opinions.

Why would they not?

I don't really care for what they think. I care for the users of the software and prospective developers who don't get their four freedoms.

Are we?

Yes. Though, you are free to disagree. My view is that one of the largest software markets, mobile and console gaming, which has exploded in recent years, works as a "proprietary software boom", due to the fact that most of these games are closed source. On a side note, we'd be lucky to see id software releasing any future games as free software.

Do you have any citations to back up this claim?

No, it is only my view.

If proprietary software fuels the release of more MIT licensed code then it could well be a good thing.

I partially agree.

Were GPL the only option instead of MIT, all of that proprietary code would certainly be kept private.

What makes you think that proprietary software developers would rather keep their code private than release it under the GPL?

because GPL is not a viable option for them

Why not, exactly? If you are releasing your source code, why would it matter that you choose MIT over the GPL? From my perspective, it would make more sense to release the code under the GPL as this way you can contributions from other people's software back, meaning you can keep competition down. If the companies cared about free software as much as they claim, this wouldn't be problematic at all.

Proprietary users of an MIT licensed code are incentivised to contribute back to the code, because doing so improves the code overall

What exactly is the incentive here? Why would I want to improve the code overall when I can just improve my own?

which makes it more likely that others will start using the code and contribute too

How so?

and the business is able to reap the benefits of not just their own contribution but everyone else's too.

Like the GPL?

Whereas the GPL forces everyone to into a system of sharing against their will

If you were going to share anyway if the project were using MIT, you wouldn't be sharing against your will if the project was GPL'd.

which as a matter of course generally removes their freedom to build a viable business which makes use of open source code

No it doesn't. I don't really care about businesses either.

Again, I don't see anything more than paternalism as the basis of your argument

I don't think that's necessarily bad.

Not exactly positive thinking.

I like my freedoms. That's positive thinking for you.

We can agree that open source and free software is good. I can't understand why one wouldn't want to see more good in the world. If I could stop terrorism, I would - even if that meant taking away the freedom for people to use bombs against other people. This isn't a practical analogy, but I hope you get what I'm saying.

@jahewson
Copy link

@StefanCristian

Example: If you're allowed to smoke in bars, you either smoke in bars or not

Analogy is not a good form of debate, as it distorts meaning with false associations. There has been enough rich discussion on freedom and its nature over the years that we can surely discuss it at face value. I like that you've identified an example where there are two forms of freedom at work and they sit in contradiction to each other, because that aspect is analogous to the GPL vs MIT situation, which reinforces my point that there are two essentially political, opposing views of what freedom can be. But the similarities end there.

Almost the same case as MIT licenses and GPL licenses. Once you have the possibility to close source and make money out of your fork, of course you're going to contribute part of the money and technology back. That's natural and sometimes you want the community to do the work for you.

You have a double standard here. You consider someone who uses and contributes to GPL'd software to be well-intentioned, but someone who uses and contributes to MIT licensed software to "want the community to do the work for [them].". That's not fair.

But who's the one winning here? Who won in the 'OSX vs FreeBSD' battle? Of course, those that hold the power of money.

When did it become a competition? When did people giving away their work for free suddenly become hungry for power and money? I'm really lost at the idea of "winning" here, as if the originator of a free work is somehow loosing. This sounds more like a paranoid delusion of "us vs. them" than of the factors motivating those who choose to write and publish software for free.

Now, let's go further on the subject and exemplify some situations from other open projects. A lot of them are built around the GNU/Linux structure without any problems ( i.e. VMware, Oracle, Microsoft, RedHat, SUSE, nvidia, amd, programs and support for various programs ), contributing to the linux kernel, GNU programs and so forth.

I'm not sure that citing Oracle, Microsoft, and VMware as champions of the GPL is a great argument. Microsoft especially have recently been opening up a lot of code under the Apache license.

Now, what would happen if 2 milion dolars ( from the previous companies ) would go into the devs of this Coreutils MIT to contribute to it as it is right now, in the MIT form. What would happen? Very easy: their team would take the whole code, modify it at will and close-sourcing it. One year later (just an example): MS Linux closed-source. On freaking money. How would @Heather feel? Judging by the fact that each dev would receive part of the money, but still not enough as Microsoft would get from selling that damn closed-source Linux.

This is a pretty whacky idea which says more about your imagination than about this discussion. Any project under any license can be taken closed-source if the original authors agree to it. You have no idea how the original authors would respond in such a situation and I could equally argue the exact opposite hypothetical point.

Now, think of this as a future possibility, just one of them. Think at this that everytime a person contributes to MIT license, it contributes for other one to take away the profit, money and glory. The original devs will be like FreeBSD: forgotten and out of the history books. and with no big deal money. most probably employed by the company that has the closed-source buyable-only fork. with own license.
I would feel miserable.

Again I think this really betrays the negative mindset surrounding this kind of GPL advocacy, that someone building upon work given away freely by others will somehow "take away the profit, money and glory.". I don't know about you, but I publish open source software in two contexts: at work, so as to give back to the public, and at home, for self-fulfilment. Neither of these goals preclude other people from dedicating their own time and resources to build on top of my work, and I can't begin to imagine having such a jealous resentment of anyone who goes on to make that a business, or build something else successful on top of it.

And here we see the political difference. The idea that nobody should be able to profit from the work of another is a pretty hard left-wing concept and a very strong opinion. It's certainly not one that can be casually declared as being the only true form of "freedom", nor the only way that things "should" be. I think that the GPL advocates in this thread need to realise that what they're arguing for is essentially a political position and not an unquestionable piece of arithmetic where "more freedom" always wins.

Personally, to add one more thing, I believe that core-programs, system programs like coreutils, gcc, glibc and kernel... the ones that compose the whole system, should be GPL. The rest? At will, doesn't matter.
But at least the base of everything, the foundation of the systems, should be GPL. At least that. Are we asking too much?

There may be a stronger argument for this point of view, but you'll need to give some reasons other than you saying that it should.

@Arcterus
Copy link
Collaborator

I personally don't like the GPL very much. I would much rather use either the MPL or an MIT/BSD-style license.

@jahewson
Copy link

@projectilemotion

I'm glad that you're seeing two kinds of freedom now. Because we don't have to agree about how to balance them - it's a matter of opinion.

Your notion of "the freedom to take away the freedoms" is flawed because you're only looking at one kind of "freedom".

How so? You alluded to two types of freedom: the freedom to do something and the freedom from something. By using the GPL, one is taking away the freedom to make the code proprietary and re-distribute it (i.e the freedom to do something). Does the second kind of freedom, the freedom from something apply here?

It depends how you frame it. The GPL takes away the freedom to make proprietary redistributions with the goal of giving the user freedom from proprietary software which they can't modify and the freedom from being beholden to the authors of such software. Freedom from tends to be associated more with left wing ideology, such as the FSF's notation that free software exists to promote a free society, a society free from what it sees as coercion. You can see how this gets political pretty quickly.

Throughout the whole of my discussion, I have been using "freedoms" to mean the four freedoms, or what would qualify under the OSI definition. These are all that matter to me, which in my original posting led me to the conclusion that the GPL would guarantee maximum freedom - that is, going by the four freedoms. Sorry for any confusion on this matter.

This is the classic polarised outlook that one expects from the different sides of the positive vs. negative freedom camps. That one side simply does not regard the other side's freedoms as being legitimate, or even existing at all. There are more than four freedoms to be had and restricting the definition of freedom to just those is obviously going to limit thought.

What they're doing is taking the source hostage and saying "my way or the highway".

Why is this unreasonable? Unless you're releasing your code into the public domain, this is what you are saying regardless of what license you use. With the MIT license, "my way" is including attribution. With the GPL, "my way" is that you're not allowed to take away the freedoms of users.

Because it's not their code. You already said that those who take from others without giving back are doing wrong. So by your standard, this is the wrong thing to do. They should write their own code.

To be fair the attribution in MIT is a legal necessity, without a license the work would remain under the sole copyright of its original author. The MIT licenses does the bare minimum necessary to allow the work to be freely reproduced.

It's unfortunate that people don't love freedom as much as the products they use.

Again, this works only for your particular narrow definition of "freedom".

has the freedom to build a proprietary product from that code

If you were the copyright owner then you could do that too. This is a moot point because it is impossible to take away your own freedoms guaranteed to you by copyright law, unless you assign the copyright to someone else.

The GPL could easily contain a legally binding pledge that the author of the original software will not reissue the original software under a proprietary license. Yet it does not. No, the author of the original software is in a position of unfair advantage when using the GPL, they restrict the freedom of others while doing nothing to limit their own.

It's great that the app market is able to make such good use of open source code

I disagree, but it's useless arguing about opinions.

Quite. But again we're seeing what is essentially a political discussion.

Why would they not?

I don't really care for what they think. I care for the users of the software and prospective developers who don't get their four freedoms.

What about five freedoms, or six, or seven? Limiting freedoms to just these precludes many other useful and positive kinds of freedom. And what of the freedoms of developers? Are we to ignore them?

My view is that one of the largest software markets, mobile and console gaming, which has exploded in recent years, works as a "proprietary software boom", due to the fact that most of these games are closed source.

What worries me is that you see that as a problem. Are people not free to develop proprietary software? Would you rather their businesses didn't exist? Why are you so keen to dictate what others should do?

What makes you think that proprietary software developers would rather keep their code private than release it under the GPL?

Ha, ok that's an easy one. Look at GitHub. We've now seen liberal licenses outnumber copyleft licenses, and the trend continues, so it's clear that developers of proprietary software are happy to open up their code under liberal licenses. The vast majority of significant open source code has been developed by businesses. Even Microsoft have proved willing to open up .NET under a liberal license. This would never happen under the GPL.

because GPL is not a viable option for them

Why not, exactly? If you are releasing your source code, why would it matter that you choose MIT over the GPL? From my perspective, it would make more sense to release the code under the GPL as this way you can contributions from other people's software back, meaning you can keep competition down. If the companies cared about free software as much as they claim, this wouldn't be problematic at all.

Because they can't accept contributions under the GPL, as they would not be able to include them in their existing proprietary distributions. I love the completely false idea that only the GPL allows people to contribute back - the MIT license obviously allows that too. For example, a company who open-sources a PDF viewer may have no commercial interest in that viewer, it may simply be a necessary but non-strategic component of their system. Were they to release it under the GPL then they'd loose the ability to include it as part of their proprietary system, perhaps with some minor customisation. By releasing it under the MIT license the entire community can benefit from something that otherwise would be kept private, and the company can benefit from contributions from others which they are free to include in their proprietary distribution.

Proprietary users of an MIT licensed code are incentivised to contribute back to the code, because doing so improves the code overall

What exactly is the incentive here? Why would I want to improve the code overall when I can just improve my own?

Well, you took that quote out of context:

Proprietary users of an MIT licensed code are incentivised to contribute back to the code, because doing so improves the code overall, which makes it more likely that others will start using the code and contribute too

There's your answer. More people contributing to your codebase means more commits that you didn't have to write.

and the business is able to reap the benefits of not just their own contribution but everyone else's too.

Like the GPL?

No, because the business can't reap the benefits from GPL licensed code, because they can't use it, as the GPL isn't compatible with most business models.

Whereas the GPL forces everyone to into a system of sharing against their will

If you were going to share anyway if the project were using MIT, you wouldn't be sharing against your will if the project was GPL'd.

Likewise, if you weren't going to share anyway, then the GPL does force you to share against your will. Where are we going with this?

which as a matter of course generally removes their freedom to build a viable business which makes use of open source code

No it doesn't. I don't really care about businesses either.

Erm, yes it does. It clearly does. You're tacitly admitting as much when you say "I don't really care about businesses" and again, showing quite a strong political position - why are you against business? Should we all inherit our wealth? Or perhaps live in communes?

Again, I don't see anything more than paternalism as the basis of your argument

I don't think that's necessarily bad.

Come on, back that up with something....

We can agree that open source and free software is good. I can't understand why one wouldn't want to see more good in the world. If I could stop terrorism, I would - even if that meant taking away the freedom for people to use bombs against other people. This isn't a practical analogy, but I hope you get what I'm saying.

Yes, sharing is good. Open source is great. To borrow your analogy, maybe you could stop terrorism by taking away encryption? Or maybe you'd end up creating a paternalistic world fuelled by mistrust and resentment, with a self-riteous elite forcing their version of acceptable thought on everyone else?

@christianbundy
Copy link

@jahewson Just a quick note on this:

To be fair the attribution in MIT is a legal necessity, without a license the work would remain under the sole copyright of its original author. The MIT licenses does the bare minimum necessary to allow the work to be freely reproduced.

This was true until recently, when the OSI approved the Free Public License. In the interest of full disclosure, I was part of the team that pushed applied for its approval, I thought you might like to know that "the bare minimum" has recently been lowered a bit.

@iyra
Copy link

iyra commented Mar 25, 2016

@Arcterus Why don't you like the GPL?

@jahewson

The idea that nobody should be able to profit from the work of another

This is a caricature of what the user was saying, and it's not very extreme at all in my opinion.

You can see how this gets political pretty quickly.

No, I don't see it that way. It can be seen that way, but I would disagree with such a framing. It is only political because we say that we can ascribe the two types of liberties to the left and right. I don't think it is necessarily so.

That one side simply does not regard the other side's freedoms as being legitimate, or even existing at all.

I do not think that creating proprietary software should be a freedom that is conferred at all.

The GPL could easily contain a legally binding pledge that the author of the original software will not reissue the original software under a proprietary license. Yet it does not.

This is news to me, and I doubt that anyone would select such a license that limits their own freedom anyway.

No, the author of the original software is in a position of unfair advantage when using the GPL, they restrict the freedom of others while doing nothing to limit their own.

It's similar to the disadvantage that end users have when they can't get the source at all. I'd rather have a small "unfair advantage" as you describe instead of the massive disadvantage of being subject to an EULA and no source code.

What about five freedoms, or six, or seven? Limiting freedoms to just these to restrict many other useful and positive kinds of freedom.

I don't think those other freedoms are as relevant when talking about free software. Just as if we were to talk about a society, the freedom of someone to commit theft is not usually taken into account. The only semi-relevant freedom I can imagine is the freedom to make the code proprietary, which I mentioned earlier as being a freedom I regard as immoral to give.

Are people not free to develop proprietary software?

They are not.

Would you rather their businesses didn't exist?

Yes. I'm also in opposition to the use of copyright today that gives people almost zero freedoms. I would rather that businesses that do not confer rights I regard as essential to their users not exist. It's a price I'd be willing to pay for freedom.

Why are you so keen to dictate what others should do?

Everyone with an opinion wants to dictate what others do. I find it hard to imagine a case in which one would not. Besides, I am petitioning for developers to change their mind on this issue - so I am not forcing anyone to do anything of course.

I want to see a world which I regard as better.

The vast majority of significant open source code has been developed by businesses.

I really wouldn't say the vast majority.

I love the completely false idea that only the GPL allows people to contribute back - the MIT license obviously allows that too.

The problem is further on down the chain where people don't have access to the source at all. If you care so much about allowing people to contribute back, it would be better to release it under the GPL.

@jahewson
Copy link

This was true until recently, when the OSI approved the Free Public License. In the interest of full disclosure, I was part of the team that pushed applied for its approval, I thought you might like to know that "the bare minimum" has recently been lowered a bit.

@christianbundy, nice work.

@jahewson
Copy link

@projectilemotion:

That one side simply does not regard the other side's freedoms as being legitimate, or even existing at all.

I do not think that creating proprietary software should be a freedom that is conferred at all.

Case in point.

It's similar to the disadvantage that end users have when they can't get the source at all. I'd rather have a small "unfair advantage" as you describe instead of the massive disadvantage of being subject to an EULA and no source code.

This "massive disadvantage" isn't created by the MIT license. It's created by people not releasing their source code. It's not fair to characterise the MIT license as denying source, it doesn't do that. A company who publishes its MIT licensed source provides no disadvantage to their users whatsoever. The GPL does not have a monopoly on openness.

What about five freedoms, or six, or seven? Limiting freedoms to just these to restrict many other useful and positive kinds of freedom.

I don't think those other freedoms are as relevant when talking about free software. Just as if we were to talk about a society, the freedom of someone to commit theft is not usually taken into account. The only semi-relevant freedom I can imagine is the freedom to make the code proprietary, which I mentioned earlier as being a freedom I regard as immoral to give.

Again, this is a classic political view to hold. You're dismissing the freedoms which don't align with your own world view. And worse still, labelling them as somehow immoral.

Are people not free to develop proprietary software?

They are not.

I think you'll find that they are. You might not want them to be, but they are.

Would you rather their businesses didn't exist?

Yes. I'm also in opposition to the use of copyright today that gives people almost zero freedoms. I would rather that businesses that do not confer rights I regard as essential to their users not exist. It's a price I'd be willing to pay for freedom.

Well, now that's a pretty extreme view isn't it? I admire how principled it is, but I get the feeling that it's fuelled by dreams of utopia where everyone follows your particular world view. We've seen political ideas like that before you know... and your definition of freedom involves a lot of telling other people what to do.

The vast majority of significant open source code has been developed by businesses.

I really wouldn't say the vast majority.

I wish I had some good statistics to back this up, because I'm certain it's true. .NET core alone is 2.6M lines of code - that's the equivalent of an awful lot of hobbyist code. Personally, most open source software which I use has been created during the course of business by other people, and most of the open source code I contribute has been created during the course of my business. Do you really think that the open-source output of hobbyists (who almost certainly work as developers by day) outstrips the open-source output of businesses (where code is often written all day long by those same developers)? Even the Linux kernel, the definitive piece of GPL'd software has an estimated 70% of contributions coming from companies [link]. So you're almost certainly wrong here: it's the vast majority.

The problem is further on down the chain where people don't have access to the source at all. If you care so much about allowing people to contribute back, it would be better to release it under the GPL.

Now that's a different problem. The MIT license didn't cause that problem. People caused that problem. The MIT licensed code is always there for people further down the chain to make use of. I highly doubt that such end-users are going to suddenly start contributing back. It's developers who contribute back, not end-users.

@yincrash
Copy link

I'd like to suggest closing this issue. It seems that most discussion is just a back & forth argument about each person's personal philosophy on what the word freedom is. I think a fork is sufficient if someone personally feels they need a GPL'd version of the project to contribute.

@Seldaek
Copy link
Collaborator

Seldaek commented Mar 25, 2016

Yup. Thanks everyone for keeping this a relatively sane discussion, but I think we've heard enough from both sides. If you want to continue please do so on another platform. I'll lock to contributors in case others here want to voice their opinions but for me I'd still proceed with MIT as we have now.

@uutils uutils locked and limited conversation to collaborators Mar 25, 2016
@cnd
Copy link
Contributor

cnd commented Mar 25, 2016

@yincrash honestly I liked this thread because I take it as experience understanding some Licensing cases.
Many contributors haven't shared opinions but from what I see currently majority of those who did don't liked the idea of switching from MIT.
Personally currently I see many negative aspects of GPL and generally people, e.g. different people have different understanding of different aspects there

@cnd cnd closed this as completed Jun 7, 2016
@rivy rivy self-assigned this Dec 18, 2021
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

No branches or pull requests