With a recent change that removed dialog in favor of bsddialog, FreeBSD has now retired the last piece of GPL-licensed code in its base system. Although the impact of this change will be minor for users, it does highlight once again the Open Source divide that has split developers since the 1990s, when [Linus Torvalds] attached a scribbled set of notes to the v0.01 Linux kernel release that would later be replaced with the very similar GPL license and its derivatives.
This history and its impacts are also the subject of a recent video by [Brodie Robertson]. For the FreeBSD project the biggest change here is probably that the entire GNU subtree in the codebase is now gone, As detailed in the video, this is the end of a very long-running project, whereby FreeBSD in its early days incorporated GNU code for userland tools. These components and its replacements are detailed in the FreeBSD wiki.
Naturally, this change has upset some people for reasons best known to themselves, but from a project management perspective this makes a lot of sense. Not having BSD-incompatible licenses like GPL in your source code massively simplifies matters. The main difference being that the GPL is reciprocal, requiring that derivative works also be released under the GPL — a feature that is often at odds with commercial projects. Meanwhile, the BSD licenses merely require the use of BSD-licensed code to be declared.
The upshot of the BSD-license in the case of FreeBSD is that it has found it and its components used in many commercial products, including MacOS/OS X, the PlayStation’s Orbis OS, and of course the BSD networking stack is happily used in Windows, macOS and just about anywhere else. Meanwhile the GPL forced Linksys to open the firmware to its WRT54G series of routers, ultimately leading to the development of OpenWrt and similar projects.
Although the OSS licensing flamewars will likely never end, it’s hard to disagree that FreeBSD is pretty healthy at over thirty years old, even if using it as a desktop OS comes with a few asterisks.
I’ve tried FreeBSD as a desktop. It was FAST on my ancient old laptop. But.. too hard to watch Netflix so it had to go. I love my OPNSense router though which is based on FreeBSD.
I guess it’s cool that Apple stuff is a form of Unix and without BSD maybe it wouldn’t be. But iPhones are so d4mned locked down, might as well be any OS under the hood, one can’t really make use of it.
I would never release any source code under a BSD license. If you want to make use of my work… cool. I insist you reciprocate with the same permissiveness. But if you want to make money in the process.. No problem! So long as you share your mods. What’s far worse than a BSD like license? A non-comm license. What’s the point?
everything’s fast if you don’t run systemd/gnome/kde/browser :)
Systemd is modular, it doesn’t add overhead by itself.
systemd has tendrils into everything. it doesn’t work by itself. that’s its upside and its downside but there’s no ignoring that’s what it does.
just an anecdatum that i recently ran into…my raspberry pi mp3 player uses init and has a 580kB tmpfs /run partition. my rockpro64 mp3 player uses systemd and has a 37MB tmpfs /run partition! by default it keeps 8 days of logfiles, and it generates 4-5MB per day.
surely, it’s possible to work around this, but every time i try i meet walls. people like systemd because they don’t “have to” configure it, and if you don’t configure it, it is incredibly wasteful. and by incredibly, i mean, i wouldn’t believe it if i didn’t see it with my own eyes.
5MB per day of pure waste for no purpose is a horror story i tell my kids about what will happen to them when they grow up and leave my house and meet Windows for the first time.
I just see the license terms as a different form of payment. It gets added to any monetary payment to give a cost. Everything is a cost/benefit comparison. Usually GPL as the default license is a good start, and the offer of dual licensing tends to take care of most other issues.
However sometimes I just don’t care. Something I wrote years ago I want to share with a small group? Or a single individual?
It is bad form to just hand it out without any license, because technically they can’t legally use it at all. You add something, anything, license wise to give others that peace of mind. I will never spend a minute considering enforcing my rights to it.
In this case it is far simpler to throw a BSD license in with it and call it a day.
Sure you can say the same for the GPL but to what end? If I won’t enforce it, and someone goes about violating it, isn’t that teaching them they can get away with that kind of crap against other GPL projects that do care?
I think using GPL when you don’t intend to “use” and enforce any license harms others more than myself, while using BSD can really only harm me if someone was to take credit for it.
I think you are missing the big picture. UNIX itself was meant just to be shared in Bell Labs. The original intent matters less than possibilities. What if Microsoft made millions on your code that you never thought would go beyond a few friends? I think everyone should know the story about Symbolics Inc. and Stallman. Maybe more should know about the fate of Maxima/Macsyma. Symbolics took the math software(Macsyma) out of MIT labs, and it eventually grew into a separate company and ported it to Windows. It dies a slow death, but The Department of Energy also managed to get a licensing agreement from MIT, and that became the Open Sourced Maxima, same code base that evolved a bit differently. If it didn’t than that code and functionality would be lost, as Macsyma re-united with Symbolics before going under. How in the world doe GPL harm others??? Stop seeing copyright as the norm. Imagine GPL is the natural, normal way. Maybe there’s some proprietary code where putting copyrights makes sense, In a talk I attended around ’99, Eric Raymond(ESR) mentioned a little piece of code a lumber company had to maximize the number of usable pieces of lumber that could be cut from any given tree truck. He told them there was no reason for them to open source, or “free” the code. Eventually someone would work out a similar solution that would be open sourced. But that was the exceptional case. The copyright is the weird solution, where enforcement becomes an issue. It requires active legal defense, if a copyright holder doesn’t sue violation that come to its attention, it can lose the copyright. That’s when the possibility of harming oneself or others really comes into play. HTF you can believe that GPL will harm others, especially in comparison to the other licencing structures? GPL is the answer to harming others the least. Don’t just imagine the possible consequences in your head from limited knowledge and FUD. Read a book on the subject, or at least get it AI to summarize it for you. Or read how ordinary copyright works in the real world.
It certainly harmed you. Look how mad you got at the thought I’d just let microsoft do that and not care. Angry to the point of personal attacks and insults to my intelligence over my opinion which license I get to choose. Then describe attributes of patent laws falsely applied to copyright.
You are not entitled to anything I create, and are not entitled to dictate what license I must use.
They explained why you were incorrect about the nature of three license. You are actually the angry one here, choom. The only one insulting you is yourself.
He didn’t insult your intelligence, he insulted your knowledge which are two completely separate things. Now I on the other hand am perfectly willing to say your intelligence is rather low, but that’s an informed opinion based upon seeing your lack of critical thinking skills and defensiveness over such a thing.
If you don’t care about helping others with your code then so be it, but don’t pretend you’re on some moral crusade by licensing in BSD or whatever when you don’t even know the definition of morality.
Are you confusing BSD with “Public Domain”???
If you truly don’t care then why not just spell it out by releasing it “Public Domain” or if you prefer “CC0”? That’s not “handing it out without a license” if you actually specify in a comment, a text file or whatever that is what you are doing.
BSD still has requirements to use. It’s just something along the lines of requiring that the copyright and non-warranty messages must be included. But those ARE requirements. Your same argument regarding not enforcing the GPL would apply to not enforcing BSD if you really don’t care.
Personally, if I release something small enough that I ‘just don’t care about it’ then that’s not going to hurt the GPL. If I was correct to not care, because it really is that small. Then so what? If someone violates the GPL on that it will not weaken the GPL overall because no one will ever hear or know about it. OTOH, if I was wrong, and it turns out to be an important part of the infrastructure of some big thing. Well, in that case I will start caring. I’ll insist that they respect the license they agreed to by using it and release that project OR rip my code out.
Public domain isn’t a license, it’s the state of the work after copyright protection expires.
In the US since the ’78 copyright changes, works are automatically under copyright with no act/choice on your part, and we can’t “renounce” that protection. It enters public domain once it expires. The closest thing is to license it with as few restrictions and burdens as possible and not actively enforce those, hoping the govt won’t either.
As for CC0 or MIT or Apache licenses, yes they too are comparable. I’m sure quite a few CC* licenses would be just as fine.
I in no way meant to imply BSD is the only way to license something to keep your name attached and nothing more.
You absolutely can renounce all personal copyright to a given software in the USA, the Unlicense is a premier example of such, and functions as a hyper-permissive license in countries like Germany where they enforce copyright regardless. It also functions as a liability waiver, which is why most people use it.
You really should start doublechecking your sources, because copyright abandonment has been a legal doctrine for several centuries and has only become stronger in recent decades.
Do the people you share software with actually care about this? I’ve never associated with anyone who worried about the licensing of the software they use. Licenses only matter if you run a business and the lawyers get involved. Otherwise, just pirate everything.
It doesn’t matter if they care about the license or not because it applies you might as well head off the knock-on effects by just doing the correct thing in the first place.
I have been ignoring licenses for my entire life, and the only knock-on effect is peace of mind, because I get to avoid having lawyer-ese babble inside my brain. The “right thing” is to share knowledge freely, regardless of who supposedly “owns” it.
Out of sheer curiosity, what source code have you released under not-BSD licenses?
I release all of my code under the MIT license. It says, basically, “do what you want with it and don’t bother me”.
(https://github.com/ToolChainGang if anyone is interested.)
I think a better question is: has anyone, ever, been contacted by someone who wants to pay you to use your code commercially?
I don’t think that has ever happened. I think the license restrictions don’t make the authors any money, they’re only used to slap down others from making money.
I had someone contact me way back when asking to use my code in a dedicated video playback box when those were still new. After they told me a bit about their use i was able to point them towards better solutions already present in the python standard library. They’d been authorized to offer $2k for signing their license agreement, but as far as i could see their text would have required me to cease publishing the already GPL’d version, and possibly not allowed me to use it in my systems.
first of all, I’m sure it’s happened. Especially in the heady days before GPL was drafted Maybe you worded it poorly. Just offhand I’m thinking of Microsoft asking Tim Paterson at Seattle Computer Products for 86-DOS. But perhaps you meant something else entirely. What about Linus Torvalds and Linux itself? I think that’s an ongoing case of “we want to pay you to use your code commercially”. Any coder who wants to make money just copyrights the code from the start. Open Source not making creators money- that’s not a bug, it’s a feature. I’m not sure what your point is, why you have a “better question”.
Stallman got angry at Symbolics for “stealing” code from MIT computer labs. The whole point of FOSS is to stop this kind of behavior. He didn’t want money for himself, it very much was slapping down others from making money from code that is not theirs. Do I detect a note of superiority in that you slap everything with MIT licence? Maybe that part is worded in a way you didn’t intended, also
Yup, agree. As a commercial user of code in different companies, if it’s GPL it’s not worth the hassle of asking for a commercial licence and dealing with someone who probably thinks they want t3h codez 2b free!!1 and won’t licence it anyway, and quite possibly doesn’t understand what licences should cost (“company X makes millions so should give me millions for this library”).
GPL has become like copyright and patent law – yet another tool which was intended to protect creators but has ended up stifling creativity.
You’re essentially saying that using someone’s code in your commercial setting is somehow “creativity” and that GPL is stifling that “creativity”. Surely that is not what you intended to say? Cause it sounds like GPL is protecting the wishes of the code’s creator, and that bothers you enough to type in L33T.
Should the first person to write a sentence get to dictate how that sentence is written by others, for eternity? No? Then why have this requirement for code, other than narcissism?
That makes no sense at all. Licenses are permission to use someone’s copyright. Copyright does not work they way you are suggesting. Don’t be so melodramatic and certainly stop making bad-faith arguments.
So if someone writes code and “copyrights” it, I can write that same code, letter-for-letter, and use it freely? Or can they legally stop me just because they wrote it first? It’s exactly as straightforward as I stated, you just don’t like it because it makes copyright seem silly (because it is).
The freedom to copy is a matter of bodily autonomy. Copying is human nature; restricting it is anti-human.
I agree with your humanist leanings, but on copyright and code in particular, you’re very lost.
These pages explain it rather well, but the US page is lawyer speak. Basically, not every line of code is copyrightable. The copyright is for the entire work, yes, but some of the individual lines are considered too generic or foundational to the language to be copyrighted. If someone were to never see your code and made a program that does more or less the same thing, even if some lines end up the same, the second work does not infringe on the copyright of the first work.
When we’re dealing with open source software, it is basically impossible to claim ignorance of the open source code. If you use it, you’re bound by the license attached by the copyright holder.
Read it or don’t. I don’t care. I’m not the one making wild claims of infinite length copyrights, etc.
US
https://law.justia.com/cases/federal/appellate-courts/F2/982/693/137252/
EU (easier to follow)
https://fsfe.org/news/2025/news-20250515-01.en.html
Having near universal standards expressed in code for EVERYONE is a good thing. Having near universal sharing of effort expressed as code is also useful.
Maybe there’s a benefit to having two different open source OS available so long term we can see whether the worst and best predictions come true.
tbh i personally don’t see much use for the BSD license. For my own projects, i usually release GPL but if it’s small i release public domain, just a comment at the top “Use this however you want – Greg”. (not that anyone uses my projects)
It’s my impression that a lot of BSD-licensed code (I’m throwing MIT license etc into the same bucket) winds up being treated as if it was public domain…the attribution is spotty at best. Same for GPLed code, truly. It doesn’t seem to be common today but years ago i saw variants on the BSD license that seemed to require like a splash screen (attribution within the binary, not just in documentation) and i’ve never seen anyone comply with that. And a lot of times if a project pulls in several BSD-licensed components, the boilerplate from one of them gets copied into the documentation but the other ones wind up un-attributed. shrug doesn’t seem to matter from my perspective
Obviously other people will care, but for my personal work i am happy to share the source and modifications even if the license doesn’t require it. And i am happy to see people ‘steal’ my work even if they don’t attribute or re-share.
“Obviously other people will care, but for my personal work i am happy to share the source and modifications even if the license doesn’t require it. And i am happy to see people ‘steal’ my work even if they don’t attribute or re-share”
Then don’t GPL it, or, dual license.
Personally I’m allergic to GPL 3; its existence alone is a problem.
yeah i don’t care if people violate my license but i always reach for GPL v2 :)
I thought I remembered that splash screen requirement!
The only true BSD is NetBSD. 🤣
Remember. We have WiFi drivers in Linux today because when Linksys used Linux in it’s early WRT routers they violated the GPL. They were taken in court and forced to release the driver modules they had compiled into the kernel.
If the GPL didn’t exist and BSD was used instead all that would have been released is a little attribution .txt file and we might still not have native drivers for WiFi cards.
there’s more than a couple stories like this. bash comes to mind. it still rubs me the wrong way, though, because the success of these product categories was dependent on their open source nature. bash would be a niche nothingburger and we would be using a different open source shell if bash had not been open sourced, for example. and no one gains more from open source device drivers than vendors do. being able to check the box for “linux driver provided” and then your chip can be used in phones, set-top boxes, industrial controllers, laptops, PCs, and so on…it seems like if it hadn’t happened the way it did, it would have had to happen another way.
but then, less philosophically and more concretely…it really hasn’t happened! isn’t that crazy! like, i’m using intel wifi right now on my linux laptop…i even hacked the source, how can i say that? but it really isn’t. the wifi driver is a closed source blob in /lib/firmware, and the public iwlwifi source is ‘just’ glue. as glue goes, it’s pretty fine-grained, cohesive, much better than (for example) the interface to the closed source firmware on raspberry pi. but fundamentally the driver is still closed. and it shows — for certain classes of bugs, i wind up just trying different versions of the firmware driver, blindly hoping one of them ‘just works’, like i was a damn windows user.
and there’s a cherry on top that really gets me. my laptop — perhaps because at some level it’s counterfeit — is missing a ‘mandatory’ part of the firmware, the region-specific permitted frequency table. It doesn’t need to be mandatory — the open source part of the kernel provides cfg80211’s regulatory.db that seamlessly makes up for the lack. But the iwlwifi maintainers (Intel employees) have decided to require that there is a redundant copy in the ROM. So in order to use my laptop, i had to disable that requirement, and it effortlessly fell through to regulatory.db. But the iwlwifi maintainers rejected my patch! On principle! They don’t think they have a responsibility to provide an open source driver for the users, they are explicitly working for Intel trying to use their hegemonic control of the kernel development process to ban knockoff or region-unlocked hardware.
So, i mean, it’s a gray area…i’m lucky i was able to hack the driver at all. But it’s not quite as simple as “thanks to the GPL, we have freedom today.”
All valid points.
Do you remember when the only way to get WiFi in Linux was ndiswrapper?
oh i thought of another example, which i think more properly shows the feedback relationship that you have to take into account when imagining alternate futures :)
in the late 1990s, there was a market for a common interface. for ISA ethernet, it was ne2000. i could buy any card so long as it said “ne2000 compatible.” for pci, it was rtl8139. no one ever advertised “linux compatible”, but “realtek compatible” was all over.
One of the reasons that never happened for wifi is that we have so many vendor-supported drivers officially in the kernel. So i just buy whatever wifi, and then i find out after it gets here whether it works with linux (it does). In fact, when i order wifi, i always get something different than i ordered, so there isn’t even an opportunity to shop for the one with a linux driver. I’m way more worried about matching the wrong kind of M2 slot than Linux driver support.
Could argue either way about which is the tail and which is the dog, right? But if we didn’t have nearly every single vendor contributing Linux drivers, then inevitably there would be one model that somehow had drivers (perhaps reverse-engineered) and we’d all look for it and it would become a kind of defacto standard. The fact that no defacto wifi device even exists in our mind is a result of all these ‘open source’ drivers.
I would be curious to see if Microsoft and Apple actually credited FreeBSD and other BSDs for software they use.
Apple certainly did for Darwin. https://developer.apple.com/library/archive/documentation/Darwin/Conceptual/KernelProgramming/BSD/BSD.html
I’m not sure if they still have any BSD code in current versions of macOS.
one of the many copies of the bsd attributio at microsoft: https://support.microsoft.com/en-us/bing/third-party-software-notices-and-information
android phones generally have the bsd attribution somewhere in settings -> about phone too
Apple specifically prioritized BSD code for inclusion until it could be replaced, which they have mostly done now.
The thing I like most about GPL is how much it bothers people who feel like they get to decide what others do with their own personal time and creations. If I make something for my personal use or for education purposes, I will happily use GPL code. If I make something for commercial reasons, GPL is not even a consideration. I know that I have to build what I need. I certainly don’t go around complaining about GPL code because I can’t use it to make a profit. For commercial purposes, GPL code effectively doesn’t exist, so don’t fret over it. For non-commercial purposes, I’m happy GPL does exist, and I’m happy we can all benefit from it equally. GPL isn’t a middle finger to the world. It’s someone rightfully deciding how their creation gets used by others who most likely are benefiting from that creation at no cost. Long live GPL.
License only matters if you can afford the lawyers to back it up or go after people that steal it from you.
There are organizations that might help you, at least regarding enforcing the GPL.
example of fortuitous GPL violation, since it’s something people tend to gesture around…
ipcameras. a ton of them are based on ‘anyka’ platform, which is uboot, linux, busybox, and whatever else was handy. because it’s open source, anyka can send each of their customers the complete source to a reference implementation. And the customers can customize it for their proprietary android app. And then none of them re-share their modifications! There have been a handful of what appear to be accidental or thoughtless open source releases over the years, out of hundreds of vendors. It’s a clear GPL violation, and a consequential one in the case of uboot and the kernel.
But, say i’m buying a $10 camera designed to lock me into a subscription service…and i hack into it. once i break through the wall, would i rather find a GPL violation or something truly proprietary?
I’d rather people obey, but even when they violate, it still does me a huge favor compared to the realistic alternative!
My 2 cents:
It’s hard to predict accurately what requirements will be a significant burden to someone else; any choice is a trade-off when it comes to licensing.
By default, copyrighted material (eventually) enters the public domain. I would to see a “ShareAlike”-type requirement with a time limit on it: I imagine it could keep developers in a legal “safe zone” (w.r.t. copyright) as long as they are actively improving the software. OTOH, what do I know? Software projects making heavy use of “permissive” licenses seem to get by without anything like this.
remember Faker.js? “I released under the licene the is explenatly so that it”s users don’t have to contribute anything but. not I’m mad that they aren’t contributing anything back.”