Post-Open Source
boringcactus.com
boringcactus.com
Why is open source dead? Because mozilla layed of a bunch of people? If only their code where open source, so that everyone else could keep working on the projects they were working on. Because github makes money? And is closed source? The writing about that is particularly revealing. "if they believed open source was in principle better, they’d be open source themselves." See, there are two alternatives: Either they really believe in open source, or they see open source is something suckers do that they profit from. But in fact github open sourced plenty of important code, like the ruby git library. Thanks. Maybe it's not a dichotomy, not something that you "really believe in" or not?
It's possible today, both on desktop and on a smartphone, to run a fully open source OS. That's a huge victory by itself, and also makes those still using closed systems freer by putting a limit on what the corportaions running them can get away with.
The author dismisses any kind of open source that isn't on a sacred crusade against closed source software as corporate friendly, lacking idealism. But I think that even if there are several closed systems based on open source competing (like a lot of closed source browsers based on an open source rendering engine) that's much better then having just one closed source browser.
So yes, open source really does make people more free, partly by making it possible for dedicated people to run a fully open source system and browser, partly by increasing competition.
You can run parts of an open source kernel and maybe open source user space. But forget about fully open drivers (binary blobs rule), especially remember resident platform blobs like Intel ME. The remaining exceptions are librem and some hacked thinkpads or chromebooks equipped with coreboot.
And let's not ignore the fact that most cellphones are locked up or have hardware without sane drivers, only ones provided in binary form. Not to mention the kernel running on the DSP to handle modem. Or other things.
Even higher level, just Android, try running its applications without Google proprietary services. While there is a reverse engineered version, it always plays catch up...
/e/ (e.foundation) delivers a polished google free experience that anyone can use.
so yes, it's possible.
It's rather the opposite. It's a slow grind, but slowly more free software is at the basis of things, due to its license nature and it growing more mature. At the same time I think the failure of open source software (if you can call it that ) was predictable and predicted long ago.
I am not sure you can call it a failure of open source, if open source never made sure, that some ideals are followed when using open source software. Perhaps we can talk about ideals and ideas in people's heads and what they thought open source was.
If a copyleft license is sufficiently strong that some people don't want to use it under those terms, some subset of those people may be willing to pay for an alternate license. (The set of people who wouldn't pay for an alternative to a copyleft license seem even less likely to consider any completely non-standard incompatible-with-the-universe license.) If you want to use a license that may potentially lose you a few over-cautious corporate users, but which protects a community of users and the possibility of revenue, a copyleft license like GPLv3 or GPLv2 may make sense.
Many of the people using permissive licenses don't work for companies, or otherwise don't have anything that might block them from using a copyleft license. Many people seem to worry about using a copyleft license because it might mean some people can't or won't use the software. I've even seen companies who would actually benefit from a copyleft license just not-think about the possibility of using one because they have a vague impression that Apache or MIT is what they should use. It wasn't that they considered and dismissed copyleft; they just didn't even think about the possibility and weigh it as a potential option.
(Also, last time I saw a survey of licenses on popular software repositories, GPLv3 was about as popular as GPLv2. LLVM vs GCC is a relevant and memorable problem, but those two projects hardly represent the whole landscape of software licensing. LLVM also gained popularity for reasons beyond just licensing, such as recognizing the value of extensibility and providing compiler framework libraries, much sooner than GCC did.)
I'm not going to say "use copyleft for everything!", and there are good reasons to use permissive licenses (such as propagating standards or releasing tiny projects). But next time you're releasing a project, give it some actual consideration. If you want to release something under a permissive license, and that's a deliberate decision, by all means do so; but don't just do so "by default" without thinking about it, and without thinking whether the permissive-by-default culture is actually in your best interests.
// Hacker News is bad.
if (document.referrer.startsWith("https://news.ycombinator.com")) {
document.location = "https://upload.wikimedia.org/wikipedia/commons/d/d4/Human_fart.wav"
}
It's literally being destroyed by Chrome and Chromium derived browsers, with Google now in almost the same position as Microsoft used to be, and Firefox market share falling to single digits. They can barely fund operations because of that.
It seems that FOSS ran by a small foundation cannot compete with open source ran by one of the biggest corporations in the world. Which has nothing to do with it being free software or not.
In the last couple of weeks I've installed an OS X machine, and a couple of windows machines, both of which insisted I provide phone numbers, emails, or credit cards in order to install just the basic OS.
Linux is already a clear win for privacy and the principle that my computer is, my computer, not a client terminal to a corporate "ecosystem" (with "users" obviously being the plants, or at best, herbivores).
Some public comments Linus has made about the choice to stick with GPLv2 over GPLv3: https://youtu.be/PaKIZ7gJlRU.
> and also most of the people using Linux right now are using it by accident, distributed as ChromeOS or Android
Also, those kernels are now built with LLVM. https://clangbuiltlinux.github.io/ (not all devices in existence, but all new ones for the past few releases of both distros and many old chromebooks that received recent kernel upgrades).
Which is essentially the same as tivoization...
Immediately nvidia-driver and radeon drivers would be illegal.
Those drivers do work better than their free/open-source counterpart but are hard to install (especially if secure boot is enabled) and do not play nice with the whole GNU/Linux ecosystem.
Radeon drivers specifically play nice with the ecosystem, but they feature crypto firmware blobs.
Network drivers tends to be good neighbors too.
Nvidia drivers cause more of a problem, as do ugly smartphone GLES drivers. (Not to mention very custom network chipsets there.)
Can you describe "their EEE strategy" in more detail?
What license update? And hinder it how?
The topic of this article is very close to me but the style of writing makes it too hard to read.
The easy fix is to start using the F in F/OSS more (AGPL, etc), and force anyone who wants to use and extend your software for commercial use to pay up. Take that money, and invest it back in F/OSS software if you want -- use it to buy treats for your dog if you want. If you want wider reach and use for your software (which in a sense is even more "free" for some meaning of the word), then use the GPLv2/MIT license or other such permissive licenses.
The hard fix is trying to win over hearts and minds when it comes to OSS-washing and the marketing pushes/strategies companies employ these days to get people to write/maintain their software for them. The problem here is that some people are absolutely fine with the current situation -- many get absurdly cheap (essentially free) meaningfully complex software even if it's basically shareware from corporations, and are fine with that. Many also are fine creating software that is permissively licensed because people write software for all sorts of reasons, fame and recognition being one.
It does make me glad to see this narrative coming up on HN though -- the corporatization of OSS has been happening for a long time now and needs more attention drawn to it. Every company has a right to try to make money with software, but some are more well-intentioned than others (and they usually suffer from a harder route to market viability).
Mozilla is in an odd position, with most of their revenue coming from their largest competitor. I think that's pretty unique.
FWIW, I think the 4 freedoms reinforce what end users want, or maybe used to want: compatibility and interoperability. The big companies somehow managed to convince users against having those two things in the name of a few conveniences as the author himself pointed out. Gatekeepers can only be stopped if users reject them, and the GPL isn't so bad a tool for doing exactly that.
What a horribly short sighted view...
What happens when the company stops developing the software for your device?
Symbian^3 was moved to a free and open licence (Eclipse Public License) for some time. The ecosystem died anyway.
Making a non-trivial change to an OS and deploying it successfully to a device on my own would be a challenge, though. Once Nokia stopped paying the internal developers and Symbian programming forum was closed, there were no gurus capable of answering questions.
And in case of a very complicated OS whose internal workings were very different from other OSes out there, having no vibrant community of experts = death, unfortunately.
This is only half the story.
The other half is that GCC is a monolithic app, built in a style that makes it hard to contribute to. GCC was (maybe it still is), ironically, defective by design in order to force companies to extend GCC, instead of building plugins that could be distributed as proprietary.
Either way, LLVM's license is still Free Software, still Open Source. And LLVM's permissive license is actually proof that you don't necessarily need a copyleft license to get companies to contribute. And we've got a whole ecosystem of such projects built with permissive licenses, hello BSD/OS.
Free Software and Open Source never meant just copyleft. That's a misconception.
---
> "what really killed free software? ... in a word, capitalism."
It's funny that the author says that, because it is capitalism that enabled Free Software to exist in the first place.
Copyleft wouldn't be possible without the existing copyright law.
Also the biggest contributors to FOSS projects have been companies. Not individuals in their spare time, not the state, but companies motivated by profit. Just like how technological innovations are often driven by companies motivated by profit. And yes, we could have a discussion about perverse incentives, and how we could avoid that.
But I was born in communism, I know how not living in capitalism is actually like, and unless people can suggest actual solutions, instead of a rant inspired by Marx & friends, we've got nothing to talk about. I wish articles with such bad takes would start with the conclusion, so I wouldn't waste my time.
This is more twenty-something edgelord 'life isn't fair' communist who clearly has never lived, or known anyone who lives(d) under communisim.
No, it did not. If you knew the history of FOSS, then you would know that for several decades, every major tech company strongly opposed it and was terrified of it. GNU and Linux were started by programmers and academics who wanted to work on software with other people and share the benefits in the open. They were not started by individuals interested in making a profit or getting an IPO. FOSS was not started by entrepreneurs or any company; it was a cooperative, not a competitive, effort by people who wanted something better and who realized that cooperation/sharing work would be helpful in that goal. You could say their goal was to communally create software ;)
> Copyleft wouldn't be possible without the existing copyright law.
True, but you are not digging into the core issue here. If there was no capitalism (i.e. no companies seeking to maximize profits/valuation), why would there need to be copyright law in its present form? Copyright law exists because companies want to minimize costs, thus leading to incentives to steal other people’s IP for their own ends. Copyleft is a solution within this system; if no profit motive existed, then there would be no need for all of this defensive licensing.
> Also the biggest contributors to FOSS projects have been companies
As stated before, not for the first couple decades of FOSS. Do you know anything about Microsoft’s historical relationship/commentary about Linux? In the immediate present, your statement is true. That’s why it is described in the article. This is the problem. Corporations have co-opted the FOSS movement and used it to build closed source projects/SaaS products that they alone control and benefit from. The goal of free software was co-opted into free software on unfree platforms (e.g. AWS).
> But I was born in communism, I know how not living in capitalism is actually like
This is a classic false dichotomy. There are more options than simply capitalism and the 20th century implementations of communism. Even if you reject all forms of communism, communism != !capitalism. That is why ‘anti-capitalism’ is a separate word. Capitalism can be critiqued or replaced; it is not inevitable or unchangeable.
It's also enables proprietary extensions, easily enabling companies to make compilers for their proprietary languages, see CUDA. That seems pretty contrary to the mission of free software.
Also, just curious, how many patches have FreeBSD and co. actually received back from Sony and Apple in the past few years? They still seem to be struggling to add features to the point of begging for 802.11ac support.
If only if they looked at OSI's open source definition, they would have realized the same 4 freedoms are held in it.
Its intention is to be an internationally-recognized license equivalent to public domain
What makes it really difficult for me is my eyes can't seem to anchor anywhere when I read. It's like reading a big phrase, or like speaking without taking a breath.
For this article where Capitalization was missing, it only worked when using the neural network parser :) which sadly isn't fast enough :( if you don't have a decent GPU.
Chrome : https://chrome.google.com/webstore/detail/colorify/pipnfjhpb... Firefox : https://addons.mozilla.org/en-US/firefox/addon/colorify/
I sincerely hope no new project ever uses that license.
> Legally-conscious objectors aren’t betraying their own dastardly motivations; they’re refusing to enter into an ambiguous contract. Put differently: they’re not saying, “I’m an evildoer,” they’re saying, “I don’t understand what you want.”
(Please do not cite Artifex v. Hancom. Special conditions prevailed in that case that do not apply, in general, to all software under a nonexclusive open source license.)
Generally speaking, under common law, in order for an agreement to have the force of a contract, there must be an offer, acceptance, and consideration. Since most people acquire open source through a download, there is no consideration (monetary value) being remitted to the author of the software, and hence the license does not have the force of a contract binding on the author. The license is thus a bare license, and the author has the right to revoke it at any time. Unfortunately that's how contract law works in common-law jurisdictions, and we are in for some nasty surprises once this actually does get tested in court.
Civil-law jurisdictions, for example Germany, do not necessarily require consideration in order for a contract to be binding. Open source licenses such as the GPL can be binding contracts in Germany.
It gets even worse if the "schlub" is an entity like Google -- and you appear on their radar and they decide you're a competitor.
Some Linux contributors threatened it after the CoC was instituted, but nothing came of it.
Do you have an example of somebody actually revoking copyright and successfully suing or settling?
This same guy actually did try revoking the license to his open source slot machine game from the GeekFeminism contributors, and the GF people were like "okay, yeah, sure, whatever, we promise not to touch your shitty game again".
So this remains untested in court.
But if you actually look at the relevant bits of contract law, you'll see... he actually has stopped-clock syndrome on this, and it really has dire consequences for how we do open source. At least in the USA, and admittedly, different views on copyrights and licensing prevail outside the USA, where much of the work on Linux is done.
I haven't the slightest clue what you're taking about. You can change your license on future versions, but you can't retroactively change the conditions of an offer that's already been accepted.
If you're talking about the ability to delete published packages from NPM, that's entirely unrelated to licensing.
https://lulz.com/linux-devs-threaten-killswitch-coc-controve...
Yes, you can.
In order to be a legally binding contract, three things are required: an offer, acceptance, and consideration. Consideration is the value gained by entrants into a contract.
In short, if you did not pay to license some piece of open source software, there is no consideration for the author of that software. Hence, the license attached to the software is NOT a contract binding on the author. It is a bare license -- simply the terms under which you are permitted to use the author's IP -- and the author is legally entitled to revoke it at any time.
You may not like it, and it has unfortunate implications for open source, but that's how contract law in common law jurisdictions like the USA works.