I realise this isn’t going to be a popular opinion here, but I think it’s important to point out. These kinds of posts are popular, and they completely miss the point. “Open Source Burnout” shouldn’t even be a thing we’re talking about IMO.
You created something, chose a license, gave it away for free. And then whine about people using the thing for free, under the terms you gave it to them under.
When you give something away under a permissive license, and people make demands of you, you can simply say “No.” No is a complete sentence. If they’re being rude, tell them to go fuck themselves. If you want different outcomes, choose a different license. Your license specifically told them “I don’t owe you shit”, and also said “You don’t owe me shit.” Why do people then feel like they’re owed shit?
Yeah, it sucks when you volunteer a bunch of time and energy, make something great, and some mega corporations profits wildly from your work. But complaining because some mega corporation is taking the thing you created, under the terms you chose, and start using it without paying you is an entirely predictable outcome. It’s literally the outcome you chose. If you don’t want to work for free anymore, don’t.
I agree with what you’re saying, in that I don’t disagree with anything you said. But you’re being correct while completely sidestepping the real point. I think the real point is how we can make an open ecosystem work sensibly from a financial/value distribution perspective. Yes, sure, people should stop complaining in the way you mentioned, but what then? Are you eager to go back to the days when a 200-line library would come with an EULA? You would “install” DLLs with a wizard and enter license keys into text inputs, is that the solution?
I would be very happy to part with a non-trivial chunk of my income today if you found a fair way to redistribute value back to open source developers and if you could also bring other developers on board. We all owe our livelihoods to this ecosystem, it makes life better for us and for humanity at large, but we genuinely don’t know how to support this ecosystem within our economic system.
There are plenty of things we can do. But what we definitely can’t do is expect the current system to work.
We have the tools already. Copyleft with exception licensing is one way, within the bounds of what we already have, to accomplish this. I think there are better solutions out there, but this might be a good start.
(To answer your direct question, I wouldn’t install a 200-line library with an EULA. But I also don’t absorb 200-line libraries like they’re free drugs like many others do anyhow.)
But what we definitely can’t do is expect the current system to work.
Depends your definition of work.
I maintain a whole lot of open source projects, and contribute to many more. Every single minute I ever spent contributing to open source code was because, either:
My employer depend on it, so I’m paid.
I did it for myself or my own intellectual curiosity. Might as well share it.
Someone filed an issue and I felt like helping them.
There’s some major ecosystems that are thriving just on the above. By all accounts it “works”.
Part of me feels like there might be a glorious future out there, in which triumphant individualists hacking on ubiquitous core libraries get magically paid with no bureaucracy and no strings attached by “generous” corporations.
Part of me is like quit yer whining lol
:-)
There are plenty of things we can do. But what we definitely can’t do is expect the current system to work.
Yes! From the post: “Open Source is designed for a different world. It presents the first real opportunity in our society to begin to develop a post-scarcity economic model at scale. That likely wants unpacking in a future post.” I am tracking that next post here.
Copyleft with exception licensing is one way, within the bounds of what we already have, to accomplish this. I think there are better solutions out there, but this might be a good start.
Part of me feels like there might be a glorious future out there, in which triumphant individualists hacking on ubiquitous core libraries get magically paid with no bureaucracy and no strings attached by “generous” corporations.
I grew up thinking that this might actually happen in the future, and it would have been glorious. Back in the late 90s, early 2000s, it kinda did. Companies hired the people who created the things they depended upon, and paid well for the privilege. These were often called “Fellowship” positions. Those positions have, sadly, mostly disappeared.
It would be nice if these still existed, but I’m not holding my breath. :(
Companies hired the people who created the things they depended upon, and paid well for the privilege. […] It would be nice if these still existed, but I’m not holding my breath. :(
They still very much exist, but the key word is “depend”. The project in question need to be important enough for the company to justify it. It doesn’t apply to any random dependency that’s easily replaced.
e.g. if you are a Linux maintainer, you probably won’t have much trouble finding such position at a company with a vested interest in Linux development.
But if you maintain, say a linter, loads of companies will surely use it if it become a community standard, but it will never be important enough to the company to justify a fully loaded salary to fund that linter’s maintenance.
However many companies will happily let their developers contribute to it as part of their work without any red tape.
That’s where I generally don’t subscribe to such Open Source sustainability posts. Just because a company use some code and made money doesn’t mean they’d have paid for it if it was private software. The reality most of the time is that they’d either have done without or would have developed a (probably shittier) alternative in-house in a few days.
Foundations are picking up some of this. I love it when they do (vs. spending their budget on … marketing? unclear). Django and PHP are two that stand out.
I am surprised the PHP foundation has 8 people there on the payroll, maybe times are changing.
But still, looking back at my time in the project that’s still a bit winner-takes-it-all. 8 people being paid compared to all the volunteers, it’s just not a realistic goal to strive for (instead of getting any other job), it’s more of a serendipitous event if you are chosen. And this is PHP we are talking about. 28 years old, boring, default - other foundations often pay one person, half of their time.
You would “install” DLLs with a wizard and enter license keys into text inputs, is that the solution?
Such a solution is great because there is a dollar value attached to pissing me, your user, off. Similarly, if I don’t pay you a lot of money for something, I know I’m gonna need to spend engineer time on it.
If I want better support, I pay for it. If I can’t afford it, maybe I don’t build the thing or I learn to workaround the bugs.
Anytime you price something at zero and give it away, the cost discovery function of the market stops working–and somebody, somewhere, always incurs the cost.
I have nothing against pricing software, even a 2-line leftpad package. As I’ve mentioned, I’m willing to pay, I want to pay. But I don’t want to enter licence keys for the 230 packages I depend on, or make 230 sales calls to explain my commercial project and get quotes. How would that even work? What if I depend on a package directly and also transitively, will there be a different price? I also want to be able to fork my dependencies and make little fixes or customisations.
I really like that Richard Stallman analogy where he says it’s understandable for a space station to charge for the air you breathe, but it’s ridiculous to be forced to walk around with a heavy bubble in your head just so that they can accurately measure your air consumption.
Love the analogy :) This is why periodic/snapshot of your code works well. Chad from Sentry allocated $x for each headcount & then thanks.dev helped distribute funds based on that monthly snapshot. I guess it requires more people to be aware of the air they are breathing & knowing the money they are providing is going directly to putting more air in the space station.
You make this sound like it’s a problem of the individual, the maintainer. It’s a problem for you too, it’s a problem for the entire system.
The “of course you’re being exploited when you work in an exploitive system” is a true and valid take, but misses the point a bit. If the problem is the system, how can we meaningfully change the system? You mention copyleft, which is one system that has existed for a long time with some impact. Are there other things we could try?
What about an open source workers union? Not just for maintainers that make software, but all the rest of us that use it, too. Fair wages, fair conditions, fair compensation to the open source ecosystem.
I wrote a free tool to help people who want to contribute get started https://www.codetriage.com. I wrote a book for people who want more of a coached experience https://howtoopensource.dev. I’ve been researching this sustainability issue for a LONG time. I’m convinced that to have a meaningful change in the system requires condensing power back in the hands of the maintainers and users. The best tool for that job is a Union.
It is a problem of the individual, the maintainer. And it’s absolutely not a problem for me.
Nobody is being exploited here. People are giving things away for free, based on the licenses they choose. And then they’re bitching when they don’t get exactly what they asked for.
And I’m not shouting from the outside with no experience: I created one of the most popular libraries in the Ruby ecosystem. And when I chose the license, I weighed the pros and cons, and knew exactly what I was getting into.
Yeah, it would’ve been nice if all of these companies that use it got together and gave me a whack of money to continue working on it. But I made my choices deliberately, just like every other FOSS contributor, with full knowledge of what I was giving away, and what I could expect in return.
Nobody’s being tricked into this. They’re making choices.
Thanks for whatever you did in the Ruby ecosystem. I’m currently a core contributor with syntax_suggest as a default gem.
I feel like you didn’t read the rest of my comment where I advocate for forming a union. Which was the bulk of my reply.
Let’s rewind a bit. This is where I’m coming from:
Lessig (1999) identifies four elements that regulate behavior online: Laws, norms, markets, and technology
Code/architecture – the physical or technical constraints on activities (e.g. locks on doors or firewalls on the Internet)
Market – economic forces
Law – explicit mandates that can be enforced by the government
Norms – social conventions that one often feels compelled to follow
What you’re describing is focused only on one of those components, the law, which is what the license is. You’re discounting the architecture, market, and norms of the situation. I.e. the system.
Nobody’s being tricked into this. They’re making choices.
They’re making choices within a system. One that’s not ideal for them, or for the community. We also have the power to change the system, but licensing/laws are clearly not enough. They’re not enough for me anyway. I’m advocating for using those other facets of a community to unionize.
Great questions. I have a pre-baked idea of what this would look like, but not everyone does.
Who, exactly, is in the bargaining unit for the union?
Unions have elections and permanent administrative staff. It would work the same way a union works. Notice the first benefit I listed was wage negotiation.
Are there dues? To whom are they paid?
Yes, paid to the union. That’s how things like paying people’s salaries when they strike work. The rest of the money? It’s up to members to decide.
I’m imagining some open source developer in residence programs that would cover a full salary for a period of time.
What does it mean to join the union? Who can’t join?
I’ve never heard of someone paying union dues to a union they weren’t affiliated with.
Beyond that it’s an open question. Maybe a minimum number of hours per year logged working towards contributing towards projects or other activities like mentoring other members.
At the end of the day what does it mean to meaningfully support open source? I would want any requirements to be aligned with supporting individual workers and supporting the open source community.
What happens during strikes?
Is seniority important? Does it even matter, in open-source software?
This would be a workers union. If you are in the union and work for Google, then yes, Google will care if you stop doing your job.
Would the union be by project, by contributing org, or what? Is there like “ZFS-on-Linux local 42” or something?
I don’t think that split would be helpful. Members in various states and countries could likely benefit from having region specific connections and possibly access to a legal specialist with some knowledge of labor laws.
Do we not accept patches from non-union members? Members in bad standing?
I would frame the union to be about supporting and encouraging open source use and development. I think if someone offers something helpful to the community it’s in the interest of the community to take it.
Sorry, I don’t know where it got lost: I would love to see a union, or professional organisation like other trades have.
But I also worry about what that might look like, and what it might include, and what it might exclude.
I get the Norms. One of the reasons I put RSpec under MIT was because that was the community norm. When I finally decided to release it, I had mostly settled on LGPL. But everything around me was MIT, so I figured fuck it. I don’t think we lost much, given the kind of tool it is. But I do not prefer the MIT license.
@schneems & @srbaker we’d love to opportunity to facilitate forming this union. We (https://thanks.dev/) have always strongly advocated for “the code needs to be free” but maintainers time & support doesn’t need to be.
Is there a union I can join that would make its central goal killing Discord (perhaps by making contributions to the Matrix chat ecosystem)? Or at least ban its members from using Discord as a chat platform for its internal discussions?
What happens when the Open Source maintainer stops maintaining the software? Or maybe they don’t maintain it very well due to their own $day_job? What happens if users stop reporting bugs to an Open Source project? What happens when said Open Source software starts to break or become incompatible with newer versions of other dependencies? When things start breaking at your own $day_job, then it will become your problem. If you still don’t think maintaining Open Source isn’t also your (or collectively our) problem, then you can try explaining that to your boss the next time production breaks due to an upstream issue.
I agree with you 100%, it would be great if we could all have a place we could discuss this topic with weekly cadence & proper project planning. I’d love to opportunity to connect & invite others to also join the chat about forming a union https://discord.gg/4pNMGyEw
But complaining because some mega corporation is taking the thing you created, under the terms you chose, and start using it without paying you is an entirely predictable outcome. It’s literally the outcome you chose. If you don’t want to work for free anymore, don’t.
This happens all the time (whether on purpose or simply because the maintainer loses interest or time), even way back in the beginning before open source was being widely used commercially. The real issue is that such software is now being used as essential components in commercial products. Now, the question is: how can this be done sustainably, so that businesses can keep doing that without randomly losing support when the maintainer doesn’t feel like it anymore, and how can maintainers justify continuation of such support, which will become an increased burden when the software becomes used more widely?
Obviously, businesses can’t hire each and every developer of each and every minor dependency they end up using somewhere. But developers won’t want to keep working for free to produce software that earns others lots of money, either.
Do businesses go into just slapping a load of dependencies into their project, without knowing where they came from, what the potential risks are, and without a plan for the future in case something winds up maintained? I suspect the answer is yes, but in that case: those businesses are getting exactly what they chose as well.
On the one hand, I agree that on an individual level, we should be careful not to feel entitled to compensation for things we gave away freely.
I do view these things as a gift, and likewise I’m grateful for Python and Linux and everything else I built my career on
But on the group level, there is probably a way to make open source software a lot better, and more of it
Also, I don’t use a lot of “open source” because it’s tied up with companies that write bad code, and are “interested”.
The terms are kind of messed up, because right now it seems the best use of open source is for COMPANIES to collaborate, not individuals. (e.g. see Terraform – I noticed a lot of people who care about that also own their own companies using Terraform. It’s not really about individuals and end users.)
The system is working in that sense, but not for individuals. Another common thing is people doing open source to differentiate themselves to get jobs at companies.
That is all fine, but I also think there’s room for simply making open source software because it’s better. Right now it lags behind the state of the art pretty severely.
That is all fine, but I also think there’s room for simply making open source software because it’s better.
When I initially started using open source, that was because it really was better (when I got into Linux, Windows 9x and classic MacOS were the shitty alternatives…) and free (I was a poor student back then). And I still find open source way better in the sense that it (typically) doesn’t track me or do other user-hostile things (and if it did, it’d be easy to rip out and it would get forked). It’s sad that when you pay, you often get treated worse than when you use a free alternative.
But open source is also a lot worse, because the UI often is pretty crappy, there are important features lacking or buggy, or it simply doesn’t support common formats or hardware. In the worst case, some software simply isn’t available as open source at all.
Yeah I think our current “open source” model conflates a few things:
Making source available to end users
Making source available that is “honestly” intended to be customized by end users
The creators of the software aren’t financially compensated for making it or maintaining it
There things are coupled right now – making source available generally means no compensation – but that’s not fundamental. It’s mainly a question of what we value as society.
I agree that the UI is a big problem, but my area of interest is that systems software design is also bad.
I always bring up the fact that Linux has a deep design debt to Unix, and Unix was made with a lot of money. Most OSes and programming languages are made by companies with monopolies.
I think that makes sense in at least some ways, but there should also be room for other models.
There things are coupled right now – making source available generally means no compensation – but that’s not fundamental.
It’s hard to prevent - once the code is out in the wild, it can be copied freely, so any type of compensation from users would have to be somewhat “voluntary”. That’s the nature of a digital medium.
Proprietary software had this problem too of course. They used to work around this problem by adding copy protection schemes, usually involving some physical thing that only the buyer would have (codes in manuals, hardware dongles, making the physical medium “unique” by doing non-standard things to the disks, that kind of stuff). All types of source-available software could do the same, but if you have the code it’s trivial to disable or bypass the copy protection routines.
Nowadays proprietary vendors just put everything behind a SaaS server which cannot be (easily) replicated, so you don’t even receive a copy of the code - it runs on other people’s machines. This is also fundamentally incompatible with source-available software.
But if you’re talking traditional run-it-yourself software, for businesses there’s the BSA which could come audit you and fine you when you were found to be infringing on copyright by using pirated software. This could be a valid approach for source-available software, too. All that’s needed is a license which explicitly requires software used in a business context to have been paid for. The BSA itself as it currently stands is probably not an option as the membership fee would be a problem for most developers, but one could imagine a similar organization for source-available software. I think most current FLOSS proponents would find this kind of forced compliance rather offensive, though (it sure smells like a protection racket to me). Perhaps it’s just that we’re all too brainwashed by the status quo to accept it.
The beauty of Open Source lies in its near-pure meritocratic essence: projects providing significant value thrive with downloads and contributions. As a community, we can also determine the acceptable support model. We can all agree that code is free but we can all also work towards building the solution “maintainers time is not free”. If anyone has any idea’s or suggestions about how we can build the solution, [we’d love to collaborate] (https://discord.gg/jHt4TBJ4).
I think open source is not sustainable because it is intrinsically at odds with the system it lives within. The article mentions it but doesn’t go deeper into that. Our current economic system rewards profit in the short term over wellbeing and long term sustainability.
Unless the incentives change, I don’t think this can improve for services offered freely. “The path lies through platforms” is patching the issue with charity, and I don’t think it will bring real long term change.
The path lies through an overhaul of our economic systems.
Open Source is designed for a different world. It presents the first real opportunity in our society to begin to develop a post-scarcity economic model at scale. That likely wants unpacking in a future post.
The programmer will be paid exactly as much (or slightly less) than the same programmer in the queue for this job wants for the same job.
For any popular open source project, it is true that the next in line is ready to do this job for “thank you”. Just give me the password for your repo, or archive it and put a link to the official continuation fork.
Hence, a simple and understandable conclusion is that all popular open source products are and will be developed for free. Unless, of course, you don’t have a very good tongue, and you haven’t managed to talk a few sponsors into it.
And even if you wanted to pay someone for maintaining open source software, why would you employ someone from a part of the world where $60k (as in the example) is merely a “livable wage”?
Cost of living varies around the world. But it’s hard to avoid the fact that open source maintainers who want to get paid are competing with peers who (like anton_samokhvalov pointed out) are willing to do the work for free, or for a tenth of the wage. It’s hard to imagine sustainable, highly-paid open source development working without an enormous amount of goodwill.
This is a restatement of the iron law of wages, with the crucial difference is that software development requires a level a bit above subsistence. Also, someone who does not want to spend time on open source will probably not start digging ditches, but instead take a job as a paying software developer.
If enough open source developers do not find the resources to continue development, this falls back on the organizations depending on their labor. From having to idly check dependencies and readmes once in a while, they might have to hire a developer to patch that critical security feature in the software they depend on. This will increase their costs quite a bit, and if the developer they hire is understandably reluctant to share their changes in open source form, it will increase the costs for every other organization.
Open source sustainability is definitely in the labor extractive sector of the economy’s interests to guarantee. Organizations who depend on open source developers labor might believe there’s an inexhaustible pool of developers available. I doubt there is.
It would be nice if we had better words to define “source available” that allows folks to confidently describe their terms to users & still keep their business opportunities or prevent corporations from using the software without contributing back (be that with patches or finance). It’s really weird that AGPL is the only copyleft-with-string-attached options out there–and if it’s odd that attaching something like The Commons Clause (which lets you use code, but not directly compete with) is boo’d & ridiculed (see: Redis) as “if it’s not open source, it’s proprietary” while continuing this cycle armchair-complaining that Amazon & the like gobble up the money offering that service without compensation or indie maintainers aren’t getting the income equal to value of their labor as if nothing could have been done about it. A lot of this can also come down to the economic position society has put us in where providing such often value isn’t rewarded & you are the fool if you aren’t trying to monetize or capitalize on every project since independence if not survival is dependent on it.
We’re working on this. FSL is conceptually close to Commons Clause (and ofc Heather helped with both), but so far doesn’t seem to draw the same sort of ire. I guess the difference is meaningful: FSL (like BSL before it) reverts to another license, vs. presenting an alteration of the base license in perpetuity.
Since the early days of the Open Source movement, companies have experimented with finding a balance between granting their users the basic freedoms guaranteed by Open Source licenses while also capitalizing on their investments in software development.
I think the path lies through careful discussion, because the points raised against “open washing” are not without merit, and yet there are tensions here worth resolving. I think the implications go well beyond Open Source, narrowly considered (gets into future of work, post-scarcity economy, etc.).
Cool to see stuff like this out there testing the boundaries. The ‘Non-Commercial’ Clause in CC *-NC has some legal issues of ambiguity but there is something to the spirit where some folks really don’t want their creations to be used commercially–especially to compete against them (or stronger term about capital like The Anti-Capitalist Software License, but whether or not your political stance should be in the license is a different can of worms I’m not sure which side I’m on yet). I agree ‘open washing’ is a concern, but there still feel like there is an important addendum that needs to be added somewhere to the 4 freedoms somewhere, given the strife this is causing.
By definition of the sustainability problem open source is not meritocratic at all. The one that succeeds are the one that can afford the time to do it. Which is exactly the problem talked above :)
We estimate the supply-side value of widely-used OSS is $4.15 billion, but that the demand-side value is much larger at $8.8 trillion. We find that firms would need to spend 3.5 times more on software than they currently do if OSS did not exist. […] Further, 96% of the demand-side value is created by only 5% of OSS developers.
I’m glad to have an updated, more thorough resource to refer to. This sort of big picture thinking is foundational to sorting out what success looks like for Open Source sustainability.
I read the paper, and I think it is fundamentally flawed. :-(
firms would need to spend 3.5 times more on software
This assumes that every company would have to separately rewrite all of the OSS it uses, vs. buying equivalents from other companies. They consider the latter case in an appendix, and (if I’m reading correctly) considered that way there is effectively no impact. Here’s my full writeup.
To throw more toasters in the bathtub–though not really ones I think we’re equipped to discuss in a meaningful way here on Lobsters–I think it’s also worth considering a few adjacent questions about sustainability:
Which population(s) are disproportionately represented in the creation of the open-source software we all use?
Is there a plan to reduce the proportion of those populations in OSS? Are those plans working by any objective measure?
Are we currently in the process of demonizing those populations or discouraging their contributions?
What proportion of OSS is developed by commercial interests? Does that change with ZIRP alterations and macroeconomic conditions?
What proportion of OSS is developed by people who have the spare means to do so without compensation?
What proportion of OSS is developed by people who do not have the spare means?
Is it reasonable and fair to expect those without means to develop OSS?
Is it reasonable to expect quality OSS from those without means?
What proportion of OSS is developed by people who primarily want to solve problems and ship code?
How does OSS cater to that population, if at all?
Are those people put-off by the arrival of folks doing it for other reasons?
Bluntly: I think OSS is collateral damage by folks scrambling to grab one last ticket into the middle class and by well-meaning folks who are more interested in social justice and equity than running code.
The existing OSS system cannot solve for both (or perhaps even either!) of those things, but it is still tremendously useful to everyone in spite of that fact.
I love working on my opensourceprojects, but my two GH sponsors, and the grant I received last year, barely amount to minimum wage for the total time I spent doing FOSS.
I’m lucky that I’m not burnt out, since they’re low-maintenance, but by the same token, I doubt they’re popular enough, in absolute numbers, to be economically sustainable, so I’m contemplating moving on. I definitely can’t spend the same amount of time on them this year as last year.
No, I was encouraged to apply for it, and the Clojure community is fairly small, so I felt I had a good chance, even if I hadn’t actually received it.
The bigger issue was I felt honor-bound to complete what I said I’d do in my grant application (add HTTP/2 support to Aleph), and it took way more time than estimated (Mythical Man-Month and all that).
That’s on me, but one consequence is now I’m reluctant to apply again.
Heh, no I kept working after the grant ended, to ensure the goal was met. Just released the big final update in the last couple weeks, in fact.
I’m still the lead maintainer for now, though I plan to step back from new feature development and minor bugs. I have a couple substantial collaborators, and I’d hoped they’d take on a bigger role (though one of them just announced he was busy with a new kid, so…)
Open Source sustainability is when any smart, motivated person can produce widely adopted Open Source software and get paid fairly without jumping through hoops.
I absolutely love that definition. Hoops clarified later:
Self-determined individuals freely produce Open Source software, and get paid in proportion to their productivity without having to also become tech influencers, or sell contracts, or any other shenanigans.
I realise this isn’t going to be a popular opinion here, but I think it’s important to point out. These kinds of posts are popular, and they completely miss the point. “Open Source Burnout” shouldn’t even be a thing we’re talking about IMO.
You created something, chose a license, gave it away for free. And then whine about people using the thing for free, under the terms you gave it to them under.
When you give something away under a permissive license, and people make demands of you, you can simply say “No.” No is a complete sentence. If they’re being rude, tell them to go fuck themselves. If you want different outcomes, choose a different license. Your license specifically told them “I don’t owe you shit”, and also said “You don’t owe me shit.” Why do people then feel like they’re owed shit?
Yeah, it sucks when you volunteer a bunch of time and energy, make something great, and some mega corporations profits wildly from your work. But complaining because some mega corporation is taking the thing you created, under the terms you chose, and start using it without paying you is an entirely predictable outcome. It’s literally the outcome you chose. If you don’t want to work for free anymore, don’t.
I agree with what you’re saying, in that I don’t disagree with anything you said. But you’re being correct while completely sidestepping the real point. I think the real point is how we can make an open ecosystem work sensibly from a financial/value distribution perspective. Yes, sure, people should stop complaining in the way you mentioned, but what then? Are you eager to go back to the days when a 200-line library would come with an EULA? You would “install” DLLs with a wizard and enter license keys into text inputs, is that the solution?
I would be very happy to part with a non-trivial chunk of my income today if you found a fair way to redistribute value back to open source developers and if you could also bring other developers on board. We all owe our livelihoods to this ecosystem, it makes life better for us and for humanity at large, but we genuinely don’t know how to support this ecosystem within our economic system.
There are plenty of things we can do. But what we definitely can’t do is expect the current system to work.
We have the tools already. Copyleft with exception licensing is one way, within the bounds of what we already have, to accomplish this. I think there are better solutions out there, but this might be a good start.
(To answer your direct question, I wouldn’t install a 200-line library with an EULA. But I also don’t absorb 200-line libraries like they’re free drugs like many others do anyhow.)
Depends your definition of work.
I maintain a whole lot of open source projects, and contribute to many more. Every single minute I ever spent contributing to open source code was because, either:
There’s some major ecosystems that are thriving just on the above. By all accounts it “works”.
To be clear, I was using a definition of “work” that is implied (and sometimes explicitly stated) in these kinds of posts.
I think the current system “works”. And does so exactly as designed.
I’m not not with you:
:-)
Yes! From the post: “Open Source is designed for a different world. It presents the first real opportunity in our society to begin to develop a post-scarcity economic model at scale. That likely wants unpacking in a future post.” I am tracking that next post here.
I think Governing the Commons gives us an as-yet-untried blueprint. More on that in another future post.
I grew up thinking that this might actually happen in the future, and it would have been glorious. Back in the late 90s, early 2000s, it kinda did. Companies hired the people who created the things they depended upon, and paid well for the privilege. These were often called “Fellowship” positions. Those positions have, sadly, mostly disappeared.
It would be nice if these still existed, but I’m not holding my breath. :(
They still very much exist, but the key word is “depend”. The project in question need to be important enough for the company to justify it. It doesn’t apply to any random dependency that’s easily replaced.
e.g. if you are a Linux maintainer, you probably won’t have much trouble finding such position at a company with a vested interest in Linux development.
But if you maintain, say a linter, loads of companies will surely use it if it become a community standard, but it will never be important enough to the company to justify a fully loaded salary to fund that linter’s maintenance.
However many companies will happily let their developers contribute to it as part of their work without any red tape.
That’s where I generally don’t subscribe to such Open Source sustainability posts. Just because a company use some code and made money doesn’t mean they’d have paid for it if it was private software. The reality most of the time is that they’d either have done without or would have developed a (probably shittier) alternative in-house in a few days.
Foundations are picking up some of this. I love it when they do (vs. spending their budget on … marketing? unclear). Django and PHP are two that stand out.
I am surprised the PHP foundation has 8 people there on the payroll, maybe times are changing. But still, looking back at my time in the project that’s still a bit winner-takes-it-all. 8 people being paid compared to all the volunteers, it’s just not a realistic goal to strive for (instead of getting any other job), it’s more of a serendipitous event if you are chosen. And this is PHP we are talking about. 28 years old, boring, default - other foundations often pay one person, half of their time.
Such a solution is great because there is a dollar value attached to pissing me, your user, off. Similarly, if I don’t pay you a lot of money for something, I know I’m gonna need to spend engineer time on it.
If I want better support, I pay for it. If I can’t afford it, maybe I don’t build the thing or I learn to workaround the bugs.
Anytime you price something at zero and give it away, the cost discovery function of the market stops working–and somebody, somewhere, always incurs the cost.
I have nothing against pricing software, even a 2-line leftpad package. As I’ve mentioned, I’m willing to pay, I want to pay. But I don’t want to enter licence keys for the 230 packages I depend on, or make 230 sales calls to explain my commercial project and get quotes. How would that even work? What if I depend on a package directly and also transitively, will there be a different price? I also want to be able to fork my dependencies and make little fixes or customisations.
I really like that Richard Stallman analogy where he says it’s understandable for a space station to charge for the air you breathe, but it’s ridiculous to be forced to walk around with a heavy bubble in your head just so that they can accurately measure your air consumption.
Love the analogy :) This is why periodic/snapshot of your code works well. Chad from Sentry allocated $x for each headcount & then thanks.dev helped distribute funds based on that monthly snapshot. I guess it requires more people to be aware of the air they are breathing & knowing the money they are providing is going directly to putting more air in the space station.
You make this sound like it’s a problem of the individual, the maintainer. It’s a problem for you too, it’s a problem for the entire system.
The “of course you’re being exploited when you work in an exploitive system” is a true and valid take, but misses the point a bit. If the problem is the system, how can we meaningfully change the system? You mention copyleft, which is one system that has existed for a long time with some impact. Are there other things we could try?
What about an open source workers union? Not just for maintainers that make software, but all the rest of us that use it, too. Fair wages, fair conditions, fair compensation to the open source ecosystem.
I wrote a free tool to help people who want to contribute get started https://www.codetriage.com. I wrote a book for people who want more of a coached experience https://howtoopensource.dev. I’ve been researching this sustainability issue for a LONG time. I’m convinced that to have a meaningful change in the system requires condensing power back in the hands of the maintainers and users. The best tool for that job is a Union.
It is a problem of the individual, the maintainer. And it’s absolutely not a problem for me.
Nobody is being exploited here. People are giving things away for free, based on the licenses they choose. And then they’re bitching when they don’t get exactly what they asked for.
And I’m not shouting from the outside with no experience: I created one of the most popular libraries in the Ruby ecosystem. And when I chose the license, I weighed the pros and cons, and knew exactly what I was getting into.
Yeah, it would’ve been nice if all of these companies that use it got together and gave me a whack of money to continue working on it. But I made my choices deliberately, just like every other FOSS contributor, with full knowledge of what I was giving away, and what I could expect in return.
Nobody’s being tricked into this. They’re making choices.
Thanks for whatever you did in the Ruby ecosystem. I’m currently a core contributor with syntax_suggest as a default gem.
I feel like you didn’t read the rest of my comment where I advocate for forming a union. Which was the bulk of my reply.
Let’s rewind a bit. This is where I’m coming from:
What you’re describing is focused only on one of those components, the law, which is what the license is. You’re discounting the architecture, market, and norms of the situation. I.e. the system.
They’re making choices within a system. One that’s not ideal for them, or for the community. We also have the power to change the system, but licensing/laws are clearly not enough. They’re not enough for me anyway. I’m advocating for using those other facets of a community to unionize.
“Let’s form a union” doesn’t really serve to solve these problems in any way that I can see, and introduces a whole new class of issues.
etc., etc.
All of this concern of not getting compensated is sidestepped, of course, by actually charging for software–
When we dig into it, the reason that people contribute to these projects is usually one of:
Great questions. I have a pre-baked idea of what this would look like, but not everyone does.
Unions have elections and permanent administrative staff. It would work the same way a union works. Notice the first benefit I listed was wage negotiation.
Yes, paid to the union. That’s how things like paying people’s salaries when they strike work. The rest of the money? It’s up to members to decide.
I’m imagining some open source developer in residence programs that would cover a full salary for a period of time.
I’ve never heard of someone paying union dues to a union they weren’t affiliated with.
Beyond that it’s an open question. Maybe a minimum number of hours per year logged working towards contributing towards projects or other activities like mentoring other members.
At the end of the day what does it mean to meaningfully support open source? I would want any requirements to be aligned with supporting individual workers and supporting the open source community.
This would be a workers union. If you are in the union and work for Google, then yes, Google will care if you stop doing your job.
I don’t think that split would be helpful. Members in various states and countries could likely benefit from having region specific connections and possibly access to a legal specialist with some knowledge of labor laws.
I would frame the union to be about supporting and encouraging open source use and development. I think if someone offers something helpful to the community it’s in the interest of the community to take it.
Ultimately it’s up to members.
Sorry, I don’t know where it got lost: I would love to see a union, or professional organisation like other trades have.
But I also worry about what that might look like, and what it might include, and what it might exclude.
I get the Norms. One of the reasons I put RSpec under MIT was because that was the community norm. When I finally decided to release it, I had mostly settled on LGPL. But everything around me was MIT, so I figured fuck it. I don’t think we lost much, given the kind of tool it is. But I do not prefer the MIT license.
@schneems & @srbaker we’d love to opportunity to facilitate forming this union. We (https://thanks.dev/) have always strongly advocated for “the code needs to be free” but maintainers time & support doesn’t need to be.
Is this something you’d be interested in discussing further? https://discord.gg/wwpK2Ec7
Is there a union I can join that would make its central goal killing Discord (perhaps by making contributions to the Matrix chat ecosystem)? Or at least ban its members from using Discord as a chat platform for its internal discussions?
What happens when the Open Source maintainer stops maintaining the software? Or maybe they don’t maintain it very well due to their own $day_job? What happens if users stop reporting bugs to an Open Source project? What happens when said Open Source software starts to break or become incompatible with newer versions of other dependencies? When things start breaking at your own $day_job, then it will become your problem. If you still don’t think maintaining Open Source isn’t also your (or collectively our) problem, then you can try explaining that to your boss the next time production breaks due to an upstream issue.
I agree with you 100%, it would be great if we could all have a place we could discuss this topic with weekly cadence & proper project planning. I’d love to opportunity to connect & invite others to also join the chat about forming a union https://discord.gg/4pNMGyEw
This happens all the time (whether on purpose or simply because the maintainer loses interest or time), even way back in the beginning before open source was being widely used commercially. The real issue is that such software is now being used as essential components in commercial products. Now, the question is: how can this be done sustainably, so that businesses can keep doing that without randomly losing support when the maintainer doesn’t feel like it anymore, and how can maintainers justify continuation of such support, which will become an increased burden when the software becomes used more widely?
Obviously, businesses can’t hire each and every developer of each and every minor dependency they end up using somewhere. But developers won’t want to keep working for free to produce software that earns others lots of money, either.
Eh. The businesses are making decisions too.
Do businesses go into just slapping a load of dependencies into their project, without knowing where they came from, what the potential risks are, and without a plan for the future in case something winds up maintained? I suspect the answer is yes, but in that case: those businesses are getting exactly what they chose as well.
No argument there! Such sustainable fundings might help reduce the risks on this style of working, though!
On the one hand, I agree that on an individual level, we should be careful not to feel entitled to compensation for things we gave away freely.
I do view these things as a gift, and likewise I’m grateful for Python and Linux and everything else I built my career on
But on the group level, there is probably a way to make open source software a lot better, and more of it
Also, I don’t use a lot of “open source” because it’s tied up with companies that write bad code, and are “interested”.
The terms are kind of messed up, because right now it seems the best use of open source is for COMPANIES to collaborate, not individuals. (e.g. see Terraform – I noticed a lot of people who care about that also own their own companies using Terraform. It’s not really about individuals and end users.)
The system is working in that sense, but not for individuals. Another common thing is people doing open source to differentiate themselves to get jobs at companies.
That is all fine, but I also think there’s room for simply making open source software because it’s better. Right now it lags behind the state of the art pretty severely.
When I initially started using open source, that was because it really was better (when I got into Linux, Windows 9x and classic MacOS were the shitty alternatives…) and free (I was a poor student back then). And I still find open source way better in the sense that it (typically) doesn’t track me or do other user-hostile things (and if it did, it’d be easy to rip out and it would get forked). It’s sad that when you pay, you often get treated worse than when you use a free alternative.
But open source is also a lot worse, because the UI often is pretty crappy, there are important features lacking or buggy, or it simply doesn’t support common formats or hardware. In the worst case, some software simply isn’t available as open source at all.
Yeah I think our current “open source” model conflates a few things:
There things are coupled right now – making source available generally means no compensation – but that’s not fundamental. It’s mainly a question of what we value as society.
Related thread about “good design costs money”
https://lobste.rs/s/vhi7yd/whatever_happened_programming_2010#c_k6nsay
I agree that the UI is a big problem, but my area of interest is that systems software design is also bad.
I always bring up the fact that Linux has a deep design debt to Unix, and Unix was made with a lot of money. Most OSes and programming languages are made by companies with monopolies.
I think that makes sense in at least some ways, but there should also be room for other models.
It’s hard to prevent - once the code is out in the wild, it can be copied freely, so any type of compensation from users would have to be somewhat “voluntary”. That’s the nature of a digital medium.
Proprietary software had this problem too of course. They used to work around this problem by adding copy protection schemes, usually involving some physical thing that only the buyer would have (codes in manuals, hardware dongles, making the physical medium “unique” by doing non-standard things to the disks, that kind of stuff). All types of source-available software could do the same, but if you have the code it’s trivial to disable or bypass the copy protection routines.
Nowadays proprietary vendors just put everything behind a SaaS server which cannot be (easily) replicated, so you don’t even receive a copy of the code - it runs on other people’s machines. This is also fundamentally incompatible with source-available software.
But if you’re talking traditional run-it-yourself software, for businesses there’s the BSA which could come audit you and fine you when you were found to be infringing on copyright by using pirated software. This could be a valid approach for source-available software, too. All that’s needed is a license which explicitly requires software used in a business context to have been paid for. The BSA itself as it currently stands is probably not an option as the membership fee would be a problem for most developers, but one could imagine a similar organization for source-available software. I think most current FLOSS proponents would find this kind of forced compliance rather offensive, though (it sure smells like a protection racket to me). Perhaps it’s just that we’re all too brainwashed by the status quo to accept it.
I will note that this only works as long as the system accept the liability disclaimer in the license.
We are big enough and infrastructure enough now that it is not going to work for a lot longer.
Secondly, it only works if you can sleep soundly after your code errors that you did not fix have killed people.
For a lot of us, we can’t
The beauty of Open Source lies in its near-pure meritocratic essence: projects providing significant value thrive with downloads and contributions. As a community, we can also determine the acceptable support model. We can all agree that code is free but we can all also work towards building the solution “maintainers time is not free”. If anyone has any idea’s or suggestions about how we can build the solution, [we’d love to collaborate] (https://discord.gg/jHt4TBJ4).
I think open source is not sustainable because it is intrinsically at odds with the system it lives within. The article mentions it but doesn’t go deeper into that. Our current economic system rewards profit in the short term over wellbeing and long term sustainability.
Unless the incentives change, I don’t think this can improve for services offered freely. “The path lies through platforms” is patching the issue with charity, and I don’t think it will bring real long term change.
The path lies through an overhaul of our economic systems.
Yes! As I say in the post:
I’ve stubbed out a GitHub issue for the future post. :)
I will express an unpopular opinion:
The programmer will be paid exactly as much (or slightly less) than the same programmer in the queue for this job wants for the same job.
For any popular open source project, it is true that the next in line is ready to do this job for “thank you”. Just give me the password for your repo, or archive it and put a link to the official continuation fork.
Hence, a simple and understandable conclusion is that all popular open source products are and will be developed for free. Unless, of course, you don’t have a very good tongue, and you haven’t managed to talk a few sponsors into it.
And even if you wanted to pay someone for maintaining open source software, why would you employ someone from a part of the world where $60k (as in the example) is merely a “livable wage”?
yeah, those examples were rather shocking. $100k for an “entry level position” WTH
Cost of living varies around the world. But it’s hard to avoid the fact that open source maintainers who want to get paid are competing with peers who (like anton_samokhvalov pointed out) are willing to do the work for free, or for a tenth of the wage. It’s hard to imagine sustainable, highly-paid open source development working without an enormous amount of goodwill.
This is a restatement of the iron law of wages, with the crucial difference is that software development requires a level a bit above subsistence. Also, someone who does not want to spend time on open source will probably not start digging ditches, but instead take a job as a paying software developer.
If enough open source developers do not find the resources to continue development, this falls back on the organizations depending on their labor. From having to idly check dependencies and readmes once in a while, they might have to hire a developer to patch that critical security feature in the software they depend on. This will increase their costs quite a bit, and if the developer they hire is understandably reluctant to share their changes in open source form, it will increase the costs for every other organization.
Open source sustainability is definitely in the labor extractive sector of the economy’s interests to guarantee. Organizations who depend on open source developers labor might believe there’s an inexhaustible pool of developers available. I doubt there is.
It would be nice if we had better words to define “source available” that allows folks to confidently describe their terms to users & still keep their business opportunities or prevent corporations from using the software without contributing back (be that with patches or finance). It’s really weird that AGPL is the only copyleft-with-string-attached options out there–and if it’s odd that attaching something like The Commons Clause (which lets you use code, but not directly compete with) is boo’d & ridiculed (see: Redis) as “if it’s not open source, it’s proprietary” while continuing this cycle armchair-complaining that Amazon & the like gobble up the money offering that service without compensation or indie maintainers aren’t getting the income equal to value of their labor as if nothing could have been done about it. A lot of this can also come down to the economic position society has put us in where providing such often value isn’t rewarded & you are the fool if you aren’t trying to monetize or capitalize on every project since independence if not survival is dependent on it.
We’re working on this. FSL is conceptually close to Commons Clause (and ofc Heather helped with both), but so far doesn’t seem to draw the same sort of ire. I guess the difference is meaningful: FSL (like BSL before it) reverts to another license, vs. presenting an alteration of the base license in perpetuity.
See also the report that OSI just put out (underwritten by yours truly) about “delayed Open Source publishing”:
I think the path lies through careful discussion, because the points raised against “open washing” are not without merit, and yet there are tensions here worth resolving. I think the implications go well beyond Open Source, narrowly considered (gets into future of work, post-scarcity economy, etc.).
Cool to see stuff like this out there testing the boundaries. The ‘Non-Commercial’ Clause in CC *-NC has some legal issues of ambiguity but there is something to the spirit where some folks really don’t want their creations to be used commercially–especially to compete against them (or stronger term about capital like The Anti-Capitalist Software License, but whether or not your political stance should be in the license is a different can of worms I’m not sure which side I’m on yet). I agree ‘open washing’ is a concern, but there still feel like there is an important addendum that needs to be added somewhere to the 4 freedoms somewhere, given the strife this is causing.
There is the EUPL
I wish the future followup to this was submitted. Instead, just as we were getting to the good part, the post ended.
This was basically just a definition of a term, some examples, and “next episode, on Open Source Sustainability…”.
Gotta start somewhere. :-)
I’ve stubbed out a GitHub issue for the next post.
Indeed! To be clear I’m glad to see this out there - my comment was about what I’d like to see on Lobsters, not saying the post itself was bad.
But it seems to have generated a big discussion anyway which I’m looking forward to reading, so I’m also happy to eat crow :D
Good point @strugee I know Chad is working on https://fossfunders.com/ & actively looking for other companies looking to join the mission. The beauty of Open Source lies in its near-pure meritocratic essence, let’s see who else can step up to be like @Sentry https://blog.sentry.io/we-just-gave-500-000-dollars-to-open-source-maintainers/ & @FrontendMasters https://frontendmasters.com/blog/how-were-supporting-open-source-maintainers/
By definition of the sustainability problem open source is not meritocratic at all. The one that succeeds are the one that can afford the time to do it. Which is exactly the problem talked above :)
In related news, here’s a recent Harvard Business School paper: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4693148
I haven’t read the whole paper yet, but the summary is not so far off from my napkin math ~7 years ago. :-)
I’m glad to have an updated, more thorough resource to refer to. This sort of big picture thinking is foundational to sorting out what success looks like for Open Source sustainability.
I read the paper, and I think it is fundamentally flawed. :-(
This assumes that every company would have to separately rewrite all of the OSS it uses, vs. buying equivalents from other companies. They consider the latter case in an appendix, and (if I’m reading correctly) considered that way there is effectively no impact. Here’s my full writeup.
To throw more toasters in the bathtub–though not really ones I think we’re equipped to discuss in a meaningful way here on Lobsters–I think it’s also worth considering a few adjacent questions about sustainability:
Bluntly: I think OSS is collateral damage by folks scrambling to grab one last ticket into the middle class and by well-meaning folks who are more interested in social justice and equity than running code.
The existing OSS system cannot solve for both (or perhaps even either!) of those things, but it is still tremendously useful to everyone in spite of that fact.
I love working on my open source projects, but my two GH sponsors, and the grant I received last year, barely amount to minimum wage for the total time I spent doing FOSS.
I’m lucky that I’m not burnt out, since they’re low-maintenance, but by the same token, I doubt they’re popular enough, in absolute numbers, to be economically sustainable, so I’m contemplating moving on. I definitely can’t spend the same amount of time on them this year as last year.
Nice work Matthew, How did you get those two sponsors? Did you ask them or was it something that just happened?
Just sort of happened. Collaborators I knew either passing along some of their own funding, or getting their company to contribute a bit.
Just sort of happened. Collaborators I knew either passing along some of their own funding, or getting their company to contribute a bit.
Do you feel having to apply for the grant was hoop-jumping, as per the article?
No, I was encouraged to apply for it, and the Clojure community is fairly small, so I felt I had a good chance, even if I hadn’t actually received it.
The bigger issue was I felt honor-bound to complete what I said I’d do in my grant application (add HTTP/2 support to Aleph), and it took way more time than estimated (Mythical Man-Month and all that).
That’s on me, but one consequence is now I’m reluctant to apply again.
thanks for sharing your experience @kingmob Do you still contribute to the project or did you stop when the grant ended?
Heh, no I kept working after the grant ended, to ensure the goal was met. Just released the big final update in the last couple weeks, in fact.
I’m still the lead maintainer for now, though I plan to step back from new feature development and minor bugs. I have a couple substantial collaborators, and I’d hoped they’d take on a bigger role (though one of them just announced he was busy with a new kid, so…)
I absolutely love that definition. Hoops clarified later: