The fourth freedom is rephrased as the "right to collaborate". Yeah, right, but the official "definition" is pretty clear to me so I don't see the point in re-writing it.
Moreover, but in my own opinion, nowadays free software is more about the obligation to share the modifications. I gather that many know what open source is, but that obligation, is crucially important and embodies the political statement of the FSF. It's what puts free software truly apart (and makes it great, IMHO). So I'd move the copyleft explanation higher in the list, closer to the 4 freedoms.
You'd be surprised how many people in the software engineering tutoring/mentoring group at my university -- the students who are more interested than anyone else in writing, sharing, and reading code -- who know no more about FOSS at all than that "open source" means you can download it for free, or do whatever you want with it. I'm not simplifying; many have not learned of the concept of software licenses in the way many FOSS contributors have.
It seems to me that the website is taking the position of explaining free software from first principles, rather than as this branch or strain of open source that you already have to understand to really get. I think that's a good thing.
> I'm not simplifying; many have not learned of the concept of software licenses in the way many FOSS contributors have.
Just give it sometime and you'll see many examples of that in this very thread. Everytime the topic of opensource license comes up, there will surely be a comment that conflates "source available" with "opensource" or even worse insist that "source available" also qualifies to be "opensource". It is like words don't mean anything anymore. It is like apple is apple and pear is also apple because it looks round and we can eat it. stops ranting. Coming back to this post, I think it is valuable to explain opensource from first principles like this.
This is becoming so common that I am starting to wonder if it's intentional, to erode and dilute the actual meaning of "open source."
There have been a spate of startup products that advertise themselves as "open source" to capitalize on the marketing/SEO advantages of doing so. And when called out on this in comments or whatever, the founder eventually presents some kind of argument along the lines of, "well if anyone can read the code then it is open!" or even the old "language can evolve" canard that people trot out when they want the meaning of a commonly-understood word or phrase inverted to fit their specific business or political agenda.
People "conflate" the two because that's the logical implication of the phrase when one does not have the historical context of living through and being aware of the early days of the open source movement. Words do have meaning. You're just adding additional meaning to them that is not commonly understood.
Attach the same adjective to "park" or "road," and you'll understand my frustration with people claiming dominion over the word, "open." You don't have an obligation to cut the grass or fill a pot hole.
If you want people to understand what you mean, use more specific language like, "copy-left." That was a fabrication of the open source movement and doesn't clash with commonly understood language. If they don't already understand the jargon, at least it will be obvious, and they can seek out the definition.
No, I don't care what some open source foundation defines it as. I don't recognize their authority to extend the definition of the word, "open."
Exactly. AT&T Unix was the quintessential example of source-available software. Companies and institutions could get licenses for the Unix source code, but they weren’t allowed to redistribute it, and it certainly wasn’t free. If I recall correctly, while it was originally only $99 for academic institutions, the price went up by orders of magnitude after the breakup of AT&T (before the breakup it was prohibited from monetizing Unix due to an agreement made with the US government when AT&T was involved in an anti-trust case).
We wouldn’t have our modern BSDs if it weren’t for people being frustrated with source-available Unix. BSD initially started out as essentially a collection of additions to AT&T Unix. Eventually after the divestiture of AT&T and the resulting dramatic price hikes for Unix source licenses, the Berkeley developers embarked on a campaign to replace AT&T source code to develop a fully open source Unix. They beat GNU to the punch (see GNU Hurd), only to be stymied by lawsuits from AT&T.
But copyleft already means something different from open source. It's more like what people unfamiliar with open source software think that open source software is.
Your analogy actually shows why “open” is appropriate in this context. It’s true, you don’t have to cut the grass or fill a pothole in an open park or an open road. Similarly, you don’t have to fix any bugs in an open source software either. But if you do, you cannot stop others from using the fix. In your analogy, if you do fix a pothole, you cannot prevent other road users from using the patched portion of the road.
With just open source someone can steal your house and lock you out. With copyleft they can still "steal" your house - but not lock you out, so you can still enjoy any improvements. Company owners hate GPL because they don't control the keys.
What does “steal your house” mean in this context?
I think you are just talking about a situation where someone forks an open source code into a private one. But the original author still has the code. If someone steals my house but I get to keep my copy… nobody has lost anything.
I don’t think analogies to physical objects work all that well for software.
Software is more then just code. There are communities and ecosystems around it. When you release software you gain goodwill which is worth money to the right buyer. And all your users are good upsale and marketing targets. An unscrupulous company can take your code, win users over by marketing their fork, and if there is no copyleft they can implement extra features that only exist in their version.
Copyleft licenses typically only force you to publish modifications if you distribute modified software. But more and more software just sits on company's servers responding to API calls, which hardly constitutes distribution or "conveying the work", in GPL terms.
Most free software is not copyleft though, and fewer and fewer new free software projects are choosing copyleft licenses.
Ultimately copyleft is mostly a bureaucratic exercise, and it ends up stifling collaboration in all but the most hostile environments (such as the one Linux faced in its early days).
This is pretty sad. I believe there needs to be a critical mass. If, say, 90% of all FOSS projects were copyleft, then it would be much easier to introduce copyleft-licensed FOSS in companies, since they'd already know how to deal with it and it would not be a big deal.
Today, many companies basically ban GPL-licensed or otherwise copyleft software projects, perhaps apart from a handpicked list of exceptions, because they don't want to deal with the hassle. MIT, fine, GPL, no. And that's pretty sad. Because it means that if write a new project you need to choose between "people will use it but it won't be copyleft" and "copyleft but a sizeable part of the community won't touch it with a 10ft pole".
> Today, many companies basically ban GPL-licensed or otherwise copyleft software projects
At the same time, there are shining examples of commercially used and developed copyleft software like Linux, Blender, OBS-Studio. I doubt there exist many companies that banned really all GPL-licensed software, including Linux.
This shows that the GPL itself is not necessarily a hinderance to commercial use or development. In the end many hardware vendors accepted the GPL as a fact of doing business and opened their drivers to add them to the Linux kernel. With NVidia being the most prominent counter example.
I don't see why it's sad. As long as the code isn't forked and maintained internally as part of some proprietary piece of code, why would you care whether it's technically legal to do so or not?
It's sad because it basically gives a huge drawback to picking a copyleft license that has little to do with the actual contents of the license but mostly with market share of the license.
(I won't start discussing the merits of copyleft itself, if that's what your "why do you care?" question was aiming at.)
Yes, I was refering to the merits of copyleft itself - that's what I was getting at with my original comment as well - I view it as mostly unnecessary. I understand that you don't want to discuss that further, not trying to prolong this, just clarifying my position.
> Ultimately copyleft is mostly a bureaucratic exercise
Exactly, it is a hack that weaponizes the very system of software copyright against iself. Companies want the power to restrict everyone except license payers to use their software? Fine, then we can restrict their right to restrict our rights.
I think any criticism of copyleft should also contain the exact same criticism od copyright itself. Especially since it's copyright that started stifling collaboration in the first place... Originally, all software was "free software", since the concept of software copyright didn't even exist.
As someone else is pointing out, software has always been copyright protectable, and there has always existed software which took advantage of those protections, so that part is surely false.
The way I see it, there are two positions on software, both valid:
1. You want to develop a piece of software for your own sake, and potentially sell it to others. You copyright it and keep it proprietary.
2. You want to develop a piece of software in collaboration with other people. You apply an open source license to it.
My point about copyleft adding unnecessary bureaucracy to collaboration only applies to case 2. In case 1, you don't want that collaboration, and the copyright system helps you achieve that.
I should also note that if the copyright system didn't exist, most software would still not respect the 4 freedoms. Code would still be kept secret by companies (and employees would be contractually prevented from leaking it), and they would still impose hardware + software restrictions on which software you could install on hardware they sell. Even worse, they could actually do this with Linux or GNU utils as well.
The GPL isn't hacking the copyright system against itself, to "disable it" for software. It's hacking the copyright system to extend protections that would not be possible in a world without copyright.
> It's hacking the copyright system to extend protections that would not be possible in a world without copyright.
And if it weren't for software copyright, we wouldn't need that protection at all!
In a software-copyright-less world, reverse engineering wouldn't be a copyright violation, so the scenarios you mention would be very unlikely to happen.
Reverse engineering is already legal and is very explicitly not a copyright violation - that's why we have open source Unix today. You may be thinking of decompilation, but that can be defeated (at least partly) with various techniques.
Not to mention, with so much software moving to SaaS models, access to even the binaries to be able to do reverse engineer them is going to be extremely hard.
Also, embedded devices using GNU utils and Linux would have no reason to allow any access to the code running on them in a world without copyright.
> Originally, all software was "free software", since the concept of software copyright didn't even exist.
Take it from a graybeard, this is absolutely not true. "Software copyright" has always been a thing, from day 1. Early on, it didn't really affect things too much, but most software was under copyright.
> "Software copyright" has always been a thing, from day 1
What exactly do you mean by "day 1"? First algorithm written? First computer program written? First program shared between machines? I remember reading Richard Stallman's story about software sharing, and him explicitly saying how software was initially free "by default", and how it was companies that invented software copyright, which is the spirit my original comment was intended in.
I am not familiar with exact history of US software copyright, so if you have any hard info about that, I'd be happy to read it, and correct any of my future statements on the topic.
Well, "day 1" was an exaggeration, since the first program was arguably written in 1843. And it's not exactly clear when the first commercial computer was put on the market (it depends on how you count).
But, looking back on the earliest times that I was involved with computers -- which predates microcomputers entirely -- all of the software that came with the machines was copyrighted, as evidenced by the copious amount of copyright notices all over everything.
> Richard Stallman's story about software sharing, and him explicitly saying how software was initially free "by default", and how it was companies that invented software copyright, which is the spirit my original comment was intended in.
Taken literally, Richard Stallman is simply wrong here. But I doubt that he meant that all software was free of copyright (simply because it wasn't). In the US, before 1978(?), you had to register your copyright and it was legally possible to intentionally not copyright works. Lots of software did that, so there was certainly software that wasn't under copyright. I suspect that's what Stallman was talking about. The need to file to obtain copyright protection does mean that copyright didn't apply "by default" then. But that's not saying that there wasn't copyrighted software!
There's no such thing as "software copyright", by the way. It's just "copyright", and software doesn't have any sort of special status now, nor did it then.
As to hard info regarding the history of US copyright, there's tons of it available online. You could do worse than reading the various histories and issues the government has at www.copyright.gov.
I evolved over the decades. I started out releasing most of my projects under a copyleft license. Then I switched to an open source license. Now, I'm mostly back to copyleft.
I switched back largely because of the popular confusion about what "open source" actually is, and also because modern open source advocates have come to think that "open source" means that there's an overt collaboration aspect to it. In my projects, there is not. So it's easier to avoid the mess and just go with copyleft.
I don't think that's true anymore. Linux proved to many companies how beneficial collaborating on software can be, and I believe the vast majority took that lesson to heart. The problems that the GPL mitigates compared to more permissive licenses just don't happen in practice for many classes of software.
Unironically, AGPL3 is the best license. Even if you don't believe in the principals of free software, it has a lot going for it: Unlike MIT, you get full access to GPL3 and Apache licensed software, no need to worry about relicensing your code, no need to worry about derivative works or anything, just set it and forget it. In addition, corporations will avoid you like the plague. No need to worry about breaking anything, tinker away on your project freely without concern for backwards compatibility.
MIT is free as in free tech support, the sooner people accept that the better off we'll all be.
This just doesn’t make sense to me. The best license is the one that meets the constraints of your situation. Good luck getting *GPL software onto iOS, for example. And yes, that’s Apple / Apple users’ problem, but it also means that iOS just can never use your software, excluding it from tons of users.
Personally I don’t mind something like MIT / Apache. If you wanna hack it and not guarantee stability, just hack on and make it clear in your culture, and companies relying on it get the instability as their problem. How you support your software isn’t the same question as how you license it.
It's the best license because choosing anything else means enriching corporations for free. I've come to think of open source software as just a wealth transfer from developers to corporations at the former's expense. It just makes no rational sense to release code that isn't free as in freedom. It's either that or fully copyrighted, all rights reserved.
> I've come to think of open source software as just a wealth transfer from developers to corporations at the former's expense.
It's the reverse: corporations pay for most OSS software development. Where individual developers do it, it's voluntary. If your summary were true and complete, how would you explain either of those sentences?
Even if that's the case, it doesn't change anything. They should all be paying for free as in freedom software development instead. Like the Linux kernel where 80% of patches come from people hired to work on it.
> Where individual developers do it, it's voluntary.
I used to do it voluntarily too. Now I believe we're working against our own best interests and therefore I'll use AGPLv3 in all of my new projects.
I'm not blaming anyone. I'm recognizing the fact that permissive licensing benefits corporations disproportionately and destroys our leverage as free software developers. I'm advocating against their use and noting that I myself will no longer use them from now on.
He likely means that companies will use the software and create issues with their problems without contributing anything back = developer will be free support for companies making money with no benefit.
AGPL doesn't have such problems because it's more radioactive than anything for companies = will be used only by other free software developers
> Generally speaking, all open source software is free software, and vice versa.
(From the tooltip in https://writefreesoftware.org/learn/.) This seems to be the most objectionable statement made, though it is properly hedged. The description in the paragraph above it is pretty unobjectionable, too.
Under the normal OSI definition of open source that's pretty much true but I think what would clarify it is that when some people say free software they mean copyleft and you have to know from context.
It’s a bit disappointing to see MIT and BSD still being recommended in the “permissive license” category for new projects. Their shortcomings are pretty well known by this point, but this is a good summary: https://writing.kemitchell.com/2019/03/09/Deprecation-Notice...
The author is a copyright lawyer by trade as well a software developer so he understands both problem spaces; he identifies several glaring legal defects in the BSD and MIT licenses; and together with a group of other copyright lawyers developed a license that addresses all of those shortcomings. If I were in his shoes I would promote it too.
Of course he had. Check https://blueoakcouncil.org/about Kyle Mitchell who wrote that MIT license deprecation article is also the one who developed Blue Oak license.
No, MIT license is not deprecated. Ton of FOSS projects use it and will continue to use it.
It’s not “deprecated” in any official sense, but it has several legal defects that in 2023 ought to remove it from consideration for use as a permissive license in nearly all cases.
> Kyle Mitchell who wrote that MIT license deprecation article is also the one who developed Blue Oak license.
OK, well, if that's not good enough for you: The other two Blue Oak board members listed at that site — Heather Meeker and Luis Villa — both have serious reputations in this field.
(Disclosure: Kyle Mitchell is a friend, and both a good lawyer and a good guy.)
It is not more common because popular software don't use it and there is no need to. We don't need yet another FOSS license. There are plenty of them serving different kinds of needs. The license is developed by the author of the link in parent comment so it's no wonder that they are promoting their own license every chance they get.
Can you identify a single other license that addresses all the shortcomings in the MIT license in a better way than the Blue Oak Model license?
Saying “we don’t need another license” is like saying we don’t need more software, since there is already plenty of software “serving different kinds of needs.” We create new tools to address specific concerns in a better way. What’s the problem?
Honestly, I have a desktop app that I have grand designs to release, but even though it's a desktop app, I'm thinking of using AGPL for it.
Simply, there are services and facilities that are wrapped in a GUI that could readily be lifted and plopped into a "service", so just seems that AGPL would be better.
That blog post strongly implies (though doesn’t explicitly say) he’s the author, and links to a git repo for the site on SourceHut that is under his own account.
Wait…it’s his own blog on his own website… isn’t writing about your own projects kind of the point in that context? What about this is “self-serving” in a way that requires disclosure???
Not sure how this comes across as self-serving or what the nature of this protest is, but if you explain it in more detail perhaps I can make a correction.
Sorry if this came off as a protest. It is not a protest. Maybe I am overthinking this but creating this page (which is a great idea) and then have a link from it point back to sourcehut (which is an open source business of yours IIRC) seems a bit off. It wouldn't be a lot of trouble to disclose the connection between the author of this page and the owner of sourcehut, would it? I think it'd be nice to do it so that it does not look like you are using the agenda of promoting free software to also promote your own open source business.
I dunno, it's just a place to store the code, and SourceHut is itself free software. I could put it on Codeberg or something, I guess, but it doesn't really seem like a big deal. Never heard anyone criticize the FSF or GNU for their use of savannah.gnu.org.
just curious: what would be the answer from FSF/OSS people to , cheating and opensource games/libraries?. i mean if for example dota2 were open source it would be fairly trivial to make scripts to send commands on events (even closed source games get cheat mods!).
off topic : on the other side, i think is inevitable that at some point in the future players will have AI assistants looking at their screens , making decisions and inputs just like a real human, (heck! you can sort of do that now with linux/x11 , screen grab and proton very easily), i personally would love to see competitive gaming involving AI agents and assistance.
Cheat mitigation needs to be server-side. Anything coming from the client has to be fair game. No matter what you do, this is inevitable in the light of reverse engineering and AI progress.
Chess is already there. It's really easy to cheat. You can though analyze games after the fact and pretty reliably conclude whether somebody cheated. That's one example of server-side cheating mitigation. No matter what the client feeds you, cheaters are detected and eventually banned.
tl;dr: access to the source code will help people develop cheats but it's not really required, your game's gonna have cheaters unless you do server side cheat mitigations.
That covers things somewhat, but not entirely. In Dota (and other games), there are a lot of purely client side cheats, that are obtained by reverse engineering the closed source client. Some examples that can't be detected server side reliably.
* Keyboard macros. Actions that require multi keys pressed in sequence can be done with one key press.
* Bigger viewport. Your client has information about the entire visible map, but by default only shows a certain part on your screen. Client side hack zooms out the view and shows you a much larger area.
* Auto disable. A very strong cheat. If an enemy suddenly jumps on you from out of vision, immediately cast disable spell on them. Doesn't matter if you were looking somewhere else or doing something else.
Valve can't prevent people from creating these cheats, but they put code in other parts of the close source client that detects the usage of these cheats. Then they create a list of these cheaters and every few months have a ban wave. And change the code enough to break the cheat.
Its a team game. So you can see anything that is visible to your allies, and you can (and should be able to) pan your viewport around to see that stuff. So the client needs to know everything that is visible to your allies.
Plus there are things that happen outside your viewport that affect whats in your viewport, and trying to limit your client from knowing those things would cause all sorts of graphical artifacts.
Eg. there are some spells that affect an entire area (and have graphical affects associated with the whole area). Even if part of the spell area is in your Fog of War, your client needs to know about it completely, for the graphic drawing algorithms to run. The client draws the whole thing, and then obscures the part that should not be visible to you. Otherwise, the graphic algorithms would become get additional complexity that will possibly have performance impacts.
What do you mean? It's already pretty trivial to develop cheats for fully proprietary games. The only way to prevent cheating is to do it server side. You'd then be protected even if the server itself were also open source.
I'm tired of "source-available" or open-source but unorthodox licence software being mocked. Things like SQLite (which has a clause about Christianity and its moral values) or licences which prohibit specific sectors or use-cases like surveillance or dual-use technologies.
I suspect the reason behind that is lobbying by large companies who can't use software like that, so they push complete open source without restrictions so they can benefit from it.
The "clause about Christianity" is the Code of Ethics that they tacked on to check a box someone insisted on them checking. The source code is public domain and there is no clause about Christianity if you want to contribute.
From the Code of Ethics:
> This document continues to be used for its original purpose - providing a reference to fill in the "code of conduct" box on supplier registration forms.
and
> No one is required to follow The Rule, to know The Rule, or even to think that The Rule is a good idea.
Makes it pretty clear to any but the most conflict seeking people out there that it's just a box filler (for some contract purposes) and not an expectation for the community generally.
I agree, except for the SQLite thing; the code is fully public domain and the license is safely ignorable. The reason it exists in the first place is for corporations to have some formal license to buy if they have an internal policy requiring it. They wrote it mostly as a joke.
> ... the freedom to use, study, improve, and share your software.
Does this mean that the software released under an OSI license with the Commons Clause attached [1] is considered "free" by their definition? Because CC allows all of the above, but prohibits commercialization.
If not, then perhaps they should elaborate on that aspect.
Since your site quite explicitly aims to be an introduction to the F/OSS, it'd be prudent to explain that the "use" includes "resell". Seeing one's altruistic work meant for the common good being unceremoniously repackaged, sold and SEO'd above the original can be very discouraging.
It's nowhere near being explicit. "Commercial use" is really quite ambiguous.
In common parlance of people somewhat exposed to the software licensing it means "use for work", because of the proliferation of "personal use" and "commercial use" licensing duos.
Realizing that this clause also covers other people being allowed to sell your work is a non-trivial effort. I'd also argue that this would go very much against expectations of FOSS newcomers as it's pretty much counter-intuitive as it goes against the altruistic spirit of the FOSS. Hence the need to have it explained upfront and very clearly.
Well, sure, you and I know that. This website is aimed at the novices however.
One of them will read through the website and get excited to give FOSS a try. He will put in the work, release the project only to see it getting reskinned and put up for sale. This will be as unexpected as it will be demoralizing. So this eventuality needs to disclosed upfront and not hid behind amorphous "commercial use" and misleading "a fee [...] for instance to cover the costs of bandwidth". Just be frank with those whom you are trying to convert, especially about the caveats.
The lack of very obvious attribution or association in the website copy from @sircmpwn is a little odd. The very-subtle-is-it-even-clickable circle top-right is the only indicator.
Otherwise this looks pretty good (though my first impression was "who is this? why should I read it?" so it took me a while to get into it) - I'll definitely be sending it to people; it's clean, tidy & straightforward.
One nit-pick: the term "free software" is often understood to be narrowly defined (copyleft), so including permissive licences is tricky, though they're at least described AFTER the copyleft section, which is nice.
The term "free software" as defined by FSF always included permissive open source licenses. Do you have any link or evidence to present that supports your claim that FSF's free software definition excluded permissive licenses?
On the contrary. MPL protects source code files. If you copy the source code files, you copy the license. If you add new source to the same files, that code must share the same license. That's perfectly simple to understand.
Compare to the LGPL, which has different rules depending on the exact details of how the library is constructed and linked.
There is no "self destruction", I don't really know why you think there is. The license requires you to provide source code for the MPL-2.0 licensed source code files regardless of how your software is built (that's why it is in fact copyleft).
What it doesn't do is force you to license non-MPL-2.0 source code files under the MPL-2.0, even if you combine the two into a binary. It is explicitly non-viral, but still enforces copyleft protections onto the originally licensed code.
Moreover, but in my own opinion, nowadays free software is more about the obligation to share the modifications. I gather that many know what open source is, but that obligation, is crucially important and embodies the political statement of the FSF. It's what puts free software truly apart (and makes it great, IMHO). So I'd move the copyleft explanation higher in the list, closer to the 4 freedoms.