Do we need to rethink what free software is?
mjg59.dreamwidth.org
mjg59.dreamwidth.org
Trying to license "good" and "evil" boils down to the same slippery slope that the TOR Project dealt with when they had to reject requests to add censorship mechanisms. "Can't we just censor hate speech?"
Software is a tool. How that tool is used is at the discretion of the user. If a developer is concerned that a project could be used for evil, then he or she should consider whether releasing the software publicly is a morally good choice in the first place.
So if you for example say that everybody except military, secret services and weapons manufacturers are allowed to use your software, that might not stop them from using it, but it makes other devs consider similar clauses and at one point it might be widespread enough to show some sort of effect (e.g. less people willing to work for the goons).
A license requiring a user to adhere to the software developers' moral code is the opposite of a free software license.
In your example regarding "weapons manufacturers;" in the United States the right to keep and bear arms is enshrined in the Constitution and interpreted by the courts as an inalienable human right. Your proposed software license would thus restrict the freedom of the user and is therefore non-free.
Yes, of course.
The question is whether we want to change that. Of course our current understanding of Free Software makes such a license outside its bounds. Of course the Four Freedoms and the OSD both prevent restrictions on use. But do the Four Freedoms and the OSD accurately capture what we really want out of free software?
To be clear, I personally lean towards "yes, they still do," but I think it's absolutely healthy and important to talk about why they do. The OSD wasn't handed down by the gods on a mountaintop; it's a set of rules that we believe facilitate a broader social purpose, and we should be always willing to ask whether we still agree with that social purpose and whether we still think the rules get us there.
(Also, it's true that "free software" might end up being the wrong term for a changed definition, but it still seems important that the free software community, including the FSF, participate in these discussions.)
I highlighted right above how one of the desired use cases of this proposed "ethical" licensing model is to restrict a fundamental human right of the users of the software.
In the best case, what we are talking about is not free software. In the worst case, there may be actors involved in these discussions hoping to use outrage mobs to destabilize the free software movement, in particular application code still licensed under GPLv2.
Arguing that free software as conventionally defined is better than what is being proposed is 100% within the mission of the Free Software Foundation. The FSF tells you how bad Windows is, even though Windows isn't free software. The FSF has a list of popular non-free licenses and tells you why they aren't free.
Arguing that members of the free software community who are interested in other licenses should continue to write free software - and why - is also firmly within the mission of the FSF. If they have concerns over free software, responding to those concerns in good faith is the job of the FSF. I'm no less of a member of the free software community simply because I write proprietary code from 9 to 5 on weekdays. (If I were, there would be very little free software community, and it would be mostly driven by well-funded corporations and not by individuals.)
Also, I'm of the opinion that you can accomplish the same goals by means other than licenses (e.g., better rules on who can participate in development), and I'm sympathetic to the goals. Shouldn't the Free Software Foundation be saying, if you want to achieve these goals, here is concretely how you can accomplish them using free software? If people are saying "free software cannot do X, we must abandon free software to do X," shouldn't it be an FSF priority to demonstrate that you can do X with free software?
> In the worst case, there may be actors involved in these discussions hoping to use outrage mobs to destabilize the free software movement, in particular application code still licensed under GPLv2.
I think if we're going to spin conspiracy theories, let's talk more about whether the NSA supported Richard Stallman as the spokesperson of the free software movement to ensure that it never gained mainstream acceptance by making sure the popular image of free software was some weirdo with bad personal hygiene, and how the community failure to get rid of him played right into government interests for decades. Sounds equally plausible.
I believe the right to bear arms is at a similar crossroads with those who want the right to feel safe from the government and those who want to feel safe from others. Who's freedom is more important?
Although I think these arguments are against what free software is, maybe it's a discussion worth having, maybe there is something that was missed.
You can already attempt to convince people to not do that though, on a massive scale, even. You do not need licenses for that. Licenses are a legal matter. Working for the military or secret services is completely legal, but many would say that it is unethical in some or all cases. Why would you need licenses to achieve the goals you mentioned, or how are licenses of any use to achieve these goals, or what is it about licenses that inherently make them more effective in doing that?
It's certainly possible to restrict usage based around a moral code. Look at the JSON License: https://www.json.org/license.html
> The Software shall be used for Good, not Evil.
So, yes, it would be possible to restrict software licensing based on a loosely-defined moral code that would constantly need updates to keep pace with new social injustices, but this would not be free software.
I take issue with this. First of all, this “well-defined” qualifier is ironically poorly defined. What does that mean? It either is or is not derived from some principle or set of principles; it is ultimately an example of ethics being imposed as a restriction.
Secondly, you don’t actually have “complete freedom”. A recent example I encountered was the utility github.com/gojp/goreportcard which still relies on github.com/alecthomas/gometalinter despite gometalinter being deprecated in favor of the much faster github.com/golangci/golangci-lint. Because golangci-lint is licensed under GPL and goreportcard is licensed under Apache 2.0 it would violate GPL to integrate golangci-lint into goreportcard. So either goreportcard has to adopt a new license and ensure they’re compliant with that license or suffer relying on subpar software that has been abandoned by its creator and community.
Now perhaps I’m ignorant (something I often find myself guilty of when speaking on such lofty topics, please forgive and correct me if this is the case) to what you mean by “complete freedom” but that doesn’t seem like “complete freedom” to me, it sounds like two incompatible sets of ethics.
For your second question, you are right that you don't have 'complete freedom,' which another commenter pointed out cannot even theoretically exist. The closest you really get is public domain, in which you're not even restricted like the MIT; you don't even have to list the author! (IANAL, so this is not legal advice, but just my understanding.) However, the GPL is not interested in guaranteeing complete freedom, in the style of very permissive licenses. Instead, it's interested in maximizing end-user freedom. An MIT library can immediately be placed in a proprietary program, and now the end-user has none of the Four Freedoms, though the author of the program still has all of them. So, the GPL puts restrictions on developers and how they can use the software ('limiting freedom', if you want, though it's the freedom to limit others' Four Freedoms that is being limited), to guarantee that the Four Freedoms reach the end-user. This does lead to incompatibility as you've outlined, and so goreportcard would have to adopt the GPL to use the library and remain compliant (again, IANAL), but from the perspective of the GNU foundation, this maximizes the total number of freedoms retained because it continues to protect end-users. I hope this helps, I wanted to expand on what the goals of the GPL actually are, since you seemed to earnestly be asking.
Distributors of software, on the other hand, are prohibited from licensing GPL-derived software in a way that restricts end-user freedom. (And they're also prohibited from licensing GPL-software in a way that allows others to license it in a way that restricts end-user freedom, which is why you can't incorporate GPL code into an Apache-licensed project.)
On a related note, the person behind the Hippocratic License (HL) which they define as "a modified MIT license that specifically prohibits the use of open source software to harm others" was recently asked by OSI [1] to modify the language in the original document since it might've led people to wrongly believe that the aforementioned license was Open Source Software and that software distributed under it was Open Source Software.
This sounds next to impossible, unless the entity is a war criminal,in which case a civil suit would be entirely pointless.
Or do you even think there would - as an example - be a US court prohibit the use of some software for ICE? Such a court ruling would imply the US government acts against UN human rights. Which court would open this box for a software licence?
https://mobile.twitter.com/CoralineAda/status/11769406055257...
Not to mention how reasonable people can disagree on where to draw the line. If you think the Israeli government violates human rights, does that justify banning all Israelis from using the software? What about Intel with its major facility in Kiryat Gat? Should you then ban the software from being run on any Intel processors, whether or not they were made in Israel?
I sympathize with the movement to hold coders and users accountable, but I just don't think this is workable.
People already have a hard time understanding what free software exactly is/requires/provides. Changing the definition at this point would only create confusion, especially if the new one is even more ethically grounded.
The non commercial clause in the creative commons is also sth like this: free for non-commercial usage, all others have to ask.
I don’t see why authors shouldn’t be allowed to excert at least a little bit of control what their creations are used for, if they like to do so.
Perfectly fine. But then it's not free software, or in this case free creative work. It's as you said: it's free only for some which makes it unfree in general.
I think "social" or "ethical" software could be a reasonable term for the views proposed.
That being said, and IANAL, but I'm pretty sure that such ideas would not held in court, unless the "unethical behaviour" is illegal. But in this case, why should criminals stick to software licences?
Can you run it? Read its code? Redistribute it verbatim? Make modifications to the code for private use? Redistribute the modifications? Then it's free software.
You are free not to like some free software licenses though.
Your answer proves that GP has a point: Commons Clause (on top of Apache for instance) fits your requirements perfectly, yet it is not "free software" according to FSF/OSI definitions.
There are degrees to freedom, and while FOSS activists would like to paint issues black and white, they really are not. Which freedoms a particular user cares about is not at all given. For myself, I would gladly use Commons Clause software, because if I make money off software then it's only fair that I pay commercial license for it - after all, it's in my interest that the software is actively developed and maintained. And your questions are really all I care about, and this clause conforms to them.
But I think in the end it is up to each user to decide which set of freedoms they care about to the point of not using non-conforming software. The cynic in me would say that most users care only about free (as in beer) - at least judging by Android and plethora of its free-of-cost apps.
EDIT: I am not arguing that Commons Clause & co. should be accepted under the "free license" term - for better or worse, this term has some widely accepted definition that is not compatible with them. Maybe a new term, like "hybrid licenses" is needed? We'll see.
You can't run Commons Clause software if you're receiving payment for doing so. You can't redistribute Commons Clause software verbatim if you're receiving payment for doing so. (As written, it even appears to prohibit selling a CD of the software for the cost of the blank CD, a service that's otherwise widely accepted as a useful community contribution.)
Because the Commons Clause restricts the situations in which you can run and redistribute the software, software licensed under it isn't free software.
Is it free software? No. Does it allow many more freedoms than closed source? Hell yes. Is that enough? We'll see.
I doubt that "can you run it?" is meant to be interpreted as: "can you run it, sometimes?"
To GP:
> There are degrees to freedom, and while FOSS activists would like to paint issues black and white, they really are not
There are software that are free software, and software that are not free software. This is the whole point of a good definition. Actually assessing whether a program is free software might be hard, but at the end no software is between free software and non-free software.
Now we can argue on a level of freedom but we are not talking about this definition anymore.
MIT or BSD -> "Open"
MIT or BSD: free, open source
GPL: free, open source, copyleft
I.e. it will even further concentrate power, and accomplish the exact opposite of what free software set out to do.
Not to mention the ethical nature of the software could be interpreted by potentially hostile courts. "Spreading disharmony against the state" will fall under "evil", and 'ethical' software will be reduced to another tool of control and propaganda.
> As stewards of the free software definition, the Free Software Foundation should be taking the lead in ensuring that these issues are discussed. The priority of the board right now should be to restructure itself to ensure that it can legitimately claim to represent the community and play the leadership role it's been failing to in recent years, otherwise the opportunity will be lost and much of the activist energy that underpins free software will be spent elsewhere.
Translation: now that Stallman is mostly gone, we may finally change the definition of free software, thanks to a better "representativeness"
But this is extremely ridiculously naive to think that just because it would be coming from that Frankenstein FSF, it would not fragment the community. Debian would not follow. And tons of others.
Whatever you think of the current def, I think there is exactly zero chance of not fragmenting significantly by attempting to change it. I even think this is a huge misunderstanding of the very nature of free software to consider that free corporate usage and exploitation shall be prevented.
And if they can't change the norms of the community, fracturing it beyond repair is fine with them because it doesn't deserve to exist if it does not conform to their moral standards.
And certainly this is the case here: getting rid of freedom 0 is not exactly something I would casually suggest esp in the name of preventing fracturing; I would expect the effect to be precisely the opposite :/ And I'm not sure at all why anybody would expect otherwise.
If you decide to speak out against the Church of Scientology, a Scientologist may show up at your door telling you to stop. If you don't stop, your boss at work will start to get anonymous calls about you, exposing what a horrible person you are -- perhaps a sexual harasser, rapist, or child molester. Their story will be backed up -- by other Scientologists (who won't identify as such).
As for where Matthew is coming from, the details of that are not so important as that the Right People are in leadership of the community. If the community is not led by the Right People in service to the Right Ideas, then it is part of the system of White-Supremacist-Capitalist-Patriarchy, does not deserve to exist at all, and should be destroyed. And yes, Matthew "Ted Ts'o is a Rape Apologist" Garrett is one of the Right People (though thankfully, Ts'o seems to have a bit more spine than Stallman, so the smear didn't work nearly as well as was hoped).
And that's the problem. Who is trustworthy?
> "The priority of the board right now should be to restructure itself to ensure that it can legitimately claim to represent the community"
Who is this appeal to, and what is it for? Is it genuine or an attempt, in the current upheaval, to have the board restructured to make it less effective?
> "and play the leadership role it's been failing to in recent years"
Failed in what? In "Representing the community"? What community should be represented here? Is the timing accidental? What is being sold to whom, or who to what?
Meanwhile rms is looking for new accommodations[1]:
"There should not be ... a digital listening device such as Echo, Siri or "Hello Google", or a card lock that records who opens the door. I will ask about cameras."
If nothing else, rms is trustworthy when it comes to software freedom. I'm not so sure about any new pretenders.
BSD licensed software is fine as well, nobody would say it is unethical.
In the same way the GPL3 came about due to external circumstances, other forms of licensing should be able to be considered free (like the AGPL or maybe others) with regards to the current landscape and circumstances.
The problem of ethical use of something is not limited to software. When you get any tangible object, you don’t usually get usage restrictions from a license. So, if this is a problem to solve, I think that the solution should be sought more generally, not restricting the scope to software. (edit: and not to an individual software, which would be a weak way of "preventing" an unethical practice at the society level anyway).
Usage of tangible objects is restricted by law. I think it should be the same for software. And yes, law is not perfect, and not the same for different countries.
Everybody has their own, evolving, ethics. If everybody starts restricting usage for random things in licenses, we may be in for a big mess. And I say this as someone who has an opinion on many ethical matters.
A company can refuse to sell you a product if they don't agree with what you're going to do with it but they can't do anything at all if you somehow legally get ahold of it through secondary means.
Same with sold copyrighted works through the first sale doctrine (and this carries over to patents as well).
Software has seemed to have carved itself out a special niche where you don't actually own the product so the producers can dictate what you can and can not do with it after "first sale" since they don't really sell (or grant) you anything other than the right to use the product.
I don't see how you could claim that it wouldn't be influencial if this became standard practice.
I assume you're talking about a government or government contractor, because anybody else wouldn't care about "licensing" if they're building a system to kill people. In the former case, "pay a lot more money" means spending more of everyone else's tax dollars, which they definitely won't have a problem with.
I guess I'm operating under the assumptions that behavior under capitalism can be influenced by pricing, but government vendors may be a bad example.
That is why freedom 0 (the freedom to run the program as you wish, for any purpose) in free software definition is important, and rejection of enforcement of specific ethical standard was intention and not omission.
This reflects modern liberal society, which accepts diversity of ethical systems and positions, accents civil cooperation and enforces just common necessary rules through law.
It is true that copyleft licences like GPL enforce free software idea itself, and are ideological in that sense. But such ideology is restricted in scope, so people with diverse ideological positions (say RMS and ESR) can accept it as a common ground. And even with this there is fracturing on copyleft vs non-copyleft free licences.
If 'free' software licences commited to some wider range of ideological and ethical positions, then it would lead to fracturing of communities along those lines and to endless infighting.
I would expect that such change would be advantageous to people with mainstream ethical positions, but oppressive to people with heterodox ethical positions. It would be ironic if free software would be more oppresive that commercial one (who is at least limited by consumer protection laws).
In conclusion, seems to me that such change is not benefical to free software and its users as a whole, but may be benefical to people who want to use free software as a power lever to force their ethical positions on others.
(There are also technical problems, like such software would be incompatible with GPL, and if forced as a new version of GPL that could delegitimize 'any newer' ammendment in GPL-3+ licences.)
> Maybe you think that allowing people with a broad spectrum of opinions to work together on open source was only necessary in the early days when the movement was small and obscure, and now that it's popular and successful, those on the Right Side of History are justified in closing ranks and purging all who disagree because it's Their Movement. But I would say: why do you think the movement was successful in the first place? If you kill the common ground between hackers of all stripes, you're slaying the golden goose. The movement will not only resume languishing in obscurity, it will collapse in a series of petty squabbles over who has the correct exegesis of bell hooks's Ain't I A Woman with respect to routing underprivileged TCP/IP packets and who is obviously wrong and a cryptofascist and must therefore be expelled. Because that's where the movement will end up, once the current contingent of people pressing for power in FOSS get their way. (There is another theory which states that this has already happened; RMS is no right winger after all.)
> And that will be music to the ears of upper management of Intel, Microsoft, Oracle, etc. They will then drop their façade of loving open source. The peace floats will stop dead in their tracks and the tanks will roll out.
Even something simple like the JSON license is something that the FSF sees as fundamentally incompatible with their values.
Matthew Garrett certainly knows it and certainly wants that something is done on the matter. Which could be good. However, the title implies a possible evolution of the definition of free software, but I just can't see that as conceivable.
That would be akin to change the definition of the prime numbers to exclude 1, breaking many demonstrations in the process ⸮
But one isn't prime: https://primes.utm.edu/notes/faq/one.html
https://en.m.wikipedia.org/wiki/Prime_number#Primality_of_on...
It would be like the AMA deciding medical procedures couldn’t be performed on unethical people.
It’s a neat idea, but not sure why author thinks FSF would consider this as ver suggestion is not free.
Freedom 0 says: The freedom to run the program as you wish, for any purpose.
If free software moves away from this how do we handle the curation of acceptable and not acceptable uses? This is a can of worms I 'd rather be left unopened.
I agree that there is a problem about monetization of software and I welcome the discussion about it. Abolishing freedom 0 however is something I 'd be very hesitant.
They came to a consensus yeeeears ago about what can and cannot go into Free Software licenses.
I'm all for more fairness, but remember that for example the GPL is a giant legal hack, and maybe the next breakthrough will be an equally big legal and/of technical hack and not necessarily an attempt to amend other free licenses?
Edit: seeing how things play out in the US and Canada I'm even more sure than before that politics should be kept far out of tech.
Tax avoidance schemes are also legal hacks. If you can (ab)use something for bad things, then you should be able to (ab)use it for good things as well, yes?
EDITED
Maybe you are thinking of 'tax avoidance'?
With software specifically, those authors concerned about the ethical implications of the use of their software can add additional clauses to their license of choice, spelling out ethical constraints, but I fail to see how it can be reasonably enforced. I don't imagine for one moment that powerful corporations or governments are going to stand by and watch their empire crumble if the alternative would be to violate the ethical terms of a crucial piece of software that will give them an edge in their amoral powerplays.
True. Which then means free isn't entirely free. Those using such software are taking on risk - whether they care to admit it or not.
What's mind boggling is that this includes business entities large enough to mitigate that risk (i.e., contribute people and/or money). I'm being a bit cheeky but perhaps a license names "Use at your own risk" would be the way to go? That would be something upper level decision makers (and legal) would understand. As it is, many of those don't know GPL from MIT, etc.
I understand the motivation for other more specific licenses, but it is also an advantage to just have a small set of licenses that people are familiar with.
I like to release stuff under ‘use either Apache 2 or GPL’, your choice. However, I honestly don’t know what problems that might cause people.
If I had to choose one license for all my stuff I would choose LGPL (permit free commercial use, changes to library must be shared) but a lot of my open source code is examples for my books so I like to be broad as possible, license wise.
Thanks, I never knew this and needed a chuckle this morning.
In case you're not aware, the idea that there are no universal moral truths is not universally accepted. :)
(But, yes, it was a joke at George W. Bush's expense, per wikipedia).
Yes, I know, which further strengthens the point that this is an extremely subjective territory. :P
Most people today can't even control their television (they can select what to watch, but not decide not to be spied on by their own tv. This is on the frontpage right now: "Watching You Watch: The Tracking Ecosystem of Over-the-Top TV Streaming Devices" [pdf] (princeton.edu) https://news.ycombinator.com/item?id=21100404 )
These are social/political issues. I think the FSF here is overwhelmed by a tidal wave.
Yes we do need to do that.
Initial idea of FOSS was more about personal use of OS applications, so at times when "you can share your code with your neighbour without your neighbour being able to deny the same freedom to others".
But that model simply does not work (or at least is not fair) in cases when OS is used in paid services. And 90% of OS code now is in the form that is suitable only for services. And that is the problem - FOSS is not motivating fair use of our code. Sometimes - quite opposite.
This.
Free software was initially aimed at protecting both personal and commercial freedoms, not just personal ones.
Rather, the assertion that the license must not assert any restrictions on use.
Maybe there should be a standard license for non-commercial source code, but that could be called a "source available" license.
What I suspect will happen is that the FSF will keep their happy-naive view of what "Free Software" is, ignoring the commercial and viability aspects of software financing and their definition will become stale.
And we will see more software that while not being officially OpenSource™ will, for most practical purposes, be open source.
> If free software is going to maintain relevance, it needs to continue to explain how it interacts with contemporary social issues. If any organisation is going to claim to lead the community, it needs to be doing that.
True. Though I don't think it will be open source that will suffer, rather, the organizations that miss the trends.
20 years ago most of the relevant free software was GNU, Apache, heavily Linux centric, etc
Today, not so much.
There are a lot of reasons to like UBI as a funding channel for open source. Open source contributes immensely to people learning how to use computers and even the bad ideas can influence the future. There is all this value and ecosystem being excluded when we look at funding on a per-project basis.
This reminds me of US visa application forms. After you list all your relatives and past employers and places you've been in the last decade, there's about three pages of questions like "have you tortured kids or are you currently planning to", "have you committed genocide with biological weapons or are you planning to" and such. The idea is that genocide mongers will tick "yes", so the US can deny their visa.
You have failed to understand the system.
The idea is that they will tick “no”, and that then even if their crimes can't be proven to the criminal standard of proof (beyond a reasonable doubt) necessary to impose imprisonment, capital punishment, or other criminal sanction they will then be subject to civil sanction for immigration violations, including deportation and bar from future legal entry, on the lower standard applicable to those non-criminal sanctions.
Too bad, even if you were to disallow that in your license, governments won't care, those things often are often already illegal - courts will dismiss your cases about violation of copyright by saying it is "national security secret".
Additionally, licenses are based on copyright and cannot put additional restrictions. Consider GPL, it doesn't add restrictions, rather it allows you to do what copyright normally wouldn't allow. Programs are usable for any purpose by default, GPL doesn't change that. You can use EULA to add restrictions, however this requires an user to actually agree to those restrictions (by, say, clicking "I agree").
And let's be frank, app stores are a much better model for the non-techie end user than the chaos of Windows 9x.
Stallman's (deserved) fall from grace is leading to some soul-searching in the movement he founded and mismanaged for decades. Better late than never, but I'm afraid it's already too late.