John Mark's article about "Why Open Source Failed" is so extremely misguided. It lambastes the open source community for driving inequality in the world by making their work freely available. He doesn't entirely let the "Big 4" technology corporations off the hook for being evil, but, because of that conspicuous difference between open source and corporate development, he certainly blames open source for helping them get where they are today. To conceive this canard, you have to be willing to take on board some pretty big assumptions about the power dynamics between technology corporations and open source. I am not, primarily because bending over backwards to conclude that sharing non-rival resources leads to inequality is a distraction from very real drivers of inequality originating from corporations. It is not the fault of the open source experiments that those corporations have been spinning out of control for decades. The corporations' leaders chose that path for themselves. The engine of inequality Mark is trying and failing to address is corporate abuse of open source, and corporate abandonment of good business sense in pursuit of pure profit without consideration for the larger tech ecosystem of which they and open source project are parts.
Copyright and open source
The first thing we need to realize is that offering free alternatives is not new. It has been a thing for far longer than the open source's two decade vogue—consider the public domain, works not covered or no longer restricted by the copyright system. It is the destiny of all copyrightable works to enter the public domain, even if that inevitability seems imperceptibly distant today. After that, they are free as the air to be used by anybody, and software (for better and worse) is copyrightable, so it shares that destiny. Maybe the software will be outdated in 95+ years, when copyright may expire today, but that is an immaterial accident caused by grafting the strictest version of our literary system onto bits. One could easily imagine creating a different species of copyright for software, as the (poorly named) design patent serves the fashion industry. The main innovation of open source within this system is the abdication of copyright restriction and allowing that destiny to come early, within the lifetime of its original creator. (Which, really, was not much of an innovation because most things were never copyrighted before formalities were abolished. "I would like to continue the status quo after you've changed the rules, please.")
When talking about copyright policy, we too often neglect tech's role in delivering this sorry state of affairs, but it is illustrative to take a look at the most common copyright trolling boogeyman: Disney. A famous user of free works like Beauty and the Beast or Hunchback of Notre Dame, does Disney's use of public domain literary works drive inequality in film? No. Hard no. Those works are free, and Disney made use of them to make projects that people have enjoyed and the use of those free works has spurred much economic prosperity within and without the mouse company precisely because they are available to all of us. Open source software voluntarily places itself at this important crux to our system of economic development and, again, because software is copyrighted, public domain non-licensing is the fate of every shred of it anyway, however far off that day may seem.
In short, we don't and should not shame Disney for using the public domain. We shame Disney for denying the public domain to everyone else in their attempt to keep their work out of the public's hands, which is a much stronger driver of inequality than releasing some work and not charging rent. The analogy to software companies is direct: don't shame them for using open source. Shame them for attacking the copyright system and lobbying to restrict our rightful access to works held within that system. The corporations made the establishment of open source as a counterculture a necessity in the first place. Don't turn around and shake your ignorant finger at open source after that, with the obvious villain at your back.
Regardless, lobbying for worse copyright laws and spawning our free-culture alone was not necessarily bad business. Some people legitimately think that longer copyright terms are beneficial to society, although I do not. Straight-up bad business sense enters the picture when you start thinking about the second order effects of open source development and corporations' active decisions to poison their relationship with it. This is detrimental to everyone, including themselves. I'll start with funding, since it is the most obvious way that corporations have abandoned their community responsibilities.
Failure to fund open source infrastructure
One of the ways that the analogy between public domain literature works and open source software breaks down is when you consider the fact that old literature is just as useful as it ever was. Old software is almost immediately outdated upon release. To continue working, most software needs to be updated, and so it needs maintainers. This is not too concerning with small projects, but large projects like Python are full-time operations backed by foundations. You don't have to look too hard to find such large projects, critical to the infrastructure of our entire world, are strapped for funding or only have a handful of maintainers who often work on the project in addition to whatever corporate job they have.
I hear you asking "Well, if the arrangement is voluntary, can't the maintainers just walk away if they don't like that?" Then the project dies. Then the projects that run on it die. Then the corporations that make the projects that run on it lose key infrastructure. If such a thing happened under a corporation's own umbrella, it would excoriate management for such bad business practice. It is still bad business when dealing with an open source party. The fact that these people are releasing their contributions for free use is not the driver of inequality, because their contributions are deserving and, in fact, demand support from users as part of the transaction. Corporations refusing to assist the groups on whose labors they build drives inequality—tithing some funds is but the most trivial of ways to rectify that situation. Additionally, calling for corporations to support open source with tithes or even paid worker labor is not a plea for altruism or trickle down economics. It is merely stating plainly what open source's place in the ecosystem actually is. We must address that real, critical role directly because corporations often ignore it, to their own detriment, while causing a lot of inequitable division of labor and wealth. And who do we blame for this? Seemingly never the corporation literally acting against its own best interests.
And as for Mark's idea that individual developers holding software patents and restricting the tools' use would address inequality, who willingly uses S-PLUS over R? MATLAB over GNU Octave? Don't deny the reality that free, generally available tools provided by open source communities are powering modern software development. Question whether or not we are properly adapting to that reality.
Outsourced training and discrimination
Nonetheless, the other side of that is open source projects voluntarily submitting to free riding. It's true that they do so, it is their right, and it is not necessarily insidious for a small company to use Python and to not tithe to it even if they have the resources. Failing to support their open source dependencies is not the only way corporations are abusing the open source paradigm, though. Tell me, how long ago did you last hear someone spread the advice that a new developer should work on open source to prove to businesses that they are capable of doing their job? Trick question: it's every damn day, including this one. Much like allowing defendants to testify at their own trials destroyed the presumption of innocence in our criminal courts, employers taking open source contributions into account has destroyed the presumption of competence and the pathway into this industry. People who would not get out of bed to work for less than $100,000 are demanding that people prove themselves "productive" and "passionate" by working for free outside (thus, in addition to) a 9-to-5, which is a betrayal of the industry in three key ways.
Firstly, this is inherently discriminatory against those of us who don't find satisfaction in open source programming on our own time: I am not less "passionate" or, more principally, a less skilled developer simply because I don't like to do it at home. Secondly, corporations have no right to devalue junior developers by not hiring them and outsourcing their training to open source maintainers without compensation. This prevents people from entering the industry even if they choose to go the traditional university route and get a degree. It, in turn, prevents people from gaining industry credentials, which prevents people from being able to satisfy demand for experienced programmers driving the vacuous perception that there is a shortage of software "talent" out there. Finally, it's also barely-veiled as discriminatory in favor of software development's white male majority because open source is even more white and male than the general software developer workforce. You can talk about the whys of that and how to fix that until you are blue in the face, but the bare bones reality is that if your hiring practice is to filter for open source contributors today, you are filtering out women and people of color and share some direct responsibility for their underrepresentation in the field.
None of this means open source is bad, or is even a reflection upon open source. Contrary to Mark's article, sharing free software is a public good. Corporations systemically requiring people to contribute to open source without adequately supporting that part of the ecosystem with funds, manpower, or a managed understanding of what its dependence on open source means for the industry at large is sickening. Corporate abuse of the open source paradigm is one among so many other ways that corporations are engines of inequality in the world. We should condemn that abuse and seek accountability and reform. Blaming the exploited in this part of the scenario—the open source maintainers and contributors, and even non-contributors via discrimination—and declaring that they have "failed" when they have so clearly succeeded will not get us anywhere in the pursuit of a more equitable world.
Top comments (36)
As is always the case here, reminder to keep this discussion civil. This is a subject that can fire people up.
You people keep talking about the majority of developers being white males as if that was a bad thing. But it's not being caused by discrimination, (there is none because open source is as close to a pure meritocracy as has ever existed in human experience) so who cares what color or gender they are?
A meritocracy is a mirrortocracy.
"You people" means Democrats.
The Open Source community is a meritocracy - the most pure one which has ever existed. No one cares what color you are and most would prefer not to even know.
I never said Silicon Valley was a meritocracy. It is filled witb "diversity" and nepotism.
Err… let’s see: imagine Bob has 1h a day to spend on open source because he’s got a full time job, and Alice spends 10h a day working on the same project because she really enjoys it and doesn’t need to work; she inherited a large fortune from her grandmother. Alice quickly rises as the tech lead of the project because of the work she was able to pour in it. How is any of that meritocratic? (And I won’t get into the origins of the term meritocracy.)
How much time you have to dedicate to any given endeavor is part of the merit of your qualifications for the endeavor.
The only difference in qualification between Bob and Alice in my example is that Alice has inherited money while Bob hasn’t. Are you suggesting Alice has more merit because she inherited money? That doesn’t fit a definition of meritocracy that I’m wiling to consider in a conversation.
I'm not a Democrat.
A study of gender bias in open source found that women who contributed openly got their pull requests merged less often than when they hid their gender. It's as drastic as 58% accepted when open and 78.7% accepted when hidden.
A "pure" meritocracy wouldn't consider a factor completely unconnected to the work in determining merit.
Also, if fewer women's open source PRs are accepted, then fewer women become open source contributors, then a filter for open source developers in the hiring process would accept fewer women.
Here's a link to the study they're talking about.
Meritocracy doesn't care about why you have merit, only that you have merit or don't.
So women, whose contributions are demonstrably accepted less frequently because they are women, therefore demonstrably have less merit because they are women?
Oh, must be that you're not from the U.S.
This guy on reddit seems to have a valid criticism of the methodology of that study: reddit.com/r/programming/comments/...
For the purposes of getting something on your resume to get an entry level programming job, you don't have to get PRs accepted by prominent open source projects to get some open source development experience. You can just code something, anything. Fork somebody. Even make a game. It doesn't have to be big or important. The point is just to prove that you can code: that's all. It helps get you over the "no experience" gap at the beginning of your career. Some people don't even have to go that route because they built something worth showing off for their coursework in college. But employers want to see that you've done something, so at least you know the basics. Putting something up on Github can help bridge that gap a little.
When I said "Meritocracy doesn't care..." I was responding to this:
My response to that was, "Meritocracy doesn't care about why you have merit, only that you have merit or don't."
But the more I think about this, the more I find this argument really absurd. Imagine if we were talking about being a concert pianist, and we have Alice who inherited enough money to dedicate her life to constant practice and instruction with the best teachers in the world, and Bob who doesn't have enough time to practice because he has to work. Well guess what, Alice is extremely likely to have more merit as a concert pianist than Bob, no matter what caused Alice to get there.
Look, I'm not going to respond to this any more than to point out that your insistence that I am either a Democrat or not from America (I am) is as blatant a demonstration as I could imagine that you are the one being partisan. Take a step back for the day, come back to it later, and maybe you'll see how the consequences of your argument (women and working-class people have less merit than non-women and people who inherit money) are ridiculous.
“Meritocracy (merit, from Latin mereō, and -cracy, from Ancient Greek κράτος kratos "strength, power") is a political philosophy which holds that certain things, such as economic goods or power, should be vested in individuals on the basis of talent, effort, and achievement, rather than factors such as sexuality, race, gender, or wealth.” —en.m.wikipedia.org/wiki/Meritocracy
Merit isn't the same thing as personal worth. Merit is about how well you can do the job. If your wealth buys you the best education and the time to pursue excellence, then OF COURSE you are more likely to have more merit for the task you've educated and trained yourself for than someone who didn't have that advantage. But this is not inherent to having wealth. Economic determinism is false. Rich idiots don't have more merit. But rich experts do.
How does the top right make you feel?
Yes, I didn't understand that article.
OSS failed? I had the impression it had won.
Many other big corps have adopted OSS as part of their portfolio. I mean, even Microsoft did?!
Mark's point is that people creating/maintaining the OSS are not getting compensation from all the tech giants that use it. So big business now doesn't have to pay for the creation of that software. They save money and don't spread the wealth. Simple concept. And from what I gathered as to this article's point, is the author here doesn't actually disagree, just seems flared up about nuances in how or why. And also throws in some white male privilege of course? (not saying that isn't a thing, just seems a bit unrelated to the OSS issue...)
Mark said that open source drives inequality in the world, and I say it does not. It's more than a nuance because we're directing blame for these problems in opposite directions. We could be said to agree that open source licensing technically allows even the largest corporations to free ride and corporations have done bad things, but very little, if anything else on this issue.
The white male privilege portion came in because one of the ways corporations abuse open source is to treat it as a filter during the hiring process. If you treat something as a filter, and the people with that something are overwhelmingly of a certain background, then you are filtering for that background. That makes entering the industry more difficult for those being filtered out, and I criticized corporations for doing that. I didn't get into the OSS side of that issue because this article was about the corporate side.
Well, thanks for clarifying that Mike. I re-read the article with your points in mind. Can you show or give me some info to certain statements you claim?
Mark says -
(emphasis mine)
(emphasis mine)
He never said open source "drives inequality". It seems he's saying there are problems and the complacency of the "open source community" is an issue. Every single example of a problem he gives involves "corporate abuse". (And he has cool charts and graphs, +1 lol...). But he actually makes very few points in the entirety of his article. So you agree with his main point but you say that you agree with "very little, if anything else on this issue". But Mark doesn't really make a single other point other than it would be good to make intellectual property reforms (once again geared to reducing the stranglehold of monopolies). That is literally the only other point he makes.
His article can be summed up in 3 points.
Please note. I don't actually agree with any of this. (Him or you). I just truly can't see what you "don't agree" with him on. Is it the intellectual property reform? Or his blaming the problem on developer/societal "complacency". Because those are literally the only two other topics besides the corporate abuse/mis-distribution wealth. Which is 80% of the article...
I think this is a more appropriate quotation to sum up his point in his own words:
The first sentence is what latched into my mind and, upon finishing, lead me to write my piece. Saying that it is mere complicity elsewhere might be a better hedge of the point, but that point is still blaming open source technologies for wealth inequality and, as I summarize the article overall, driving said inequality by allowing the corporations to rake in huge profits without compelling them to pay license fees. Most of my article was about the second order effects of this arrangement, not the lack of direct fees.
The second sentence shows how we differ when we say open source is being abused. In his view, free licensing is equatable to free labor in service of these large companies. He demonizes open source as "the building blocks of modern capitalist behemoths" and says that corporations somehow hacked the intellectual monopoly system by starting new projects with its tools. This is not a hack, allowing them to use the software is not "complacency," and open source developers are not idiots. Open source software is intended to be used as material to build on by everyone. Letting the blame for wealth inequality spread from the corporations to the open source tools is just wrong. No, corporations are abusing open source in two big classes of ways: 1) by not supporting the projects materially, whether by tithe or workhours, 2) by altering their expectations of open source developers in ways that harm the software industry at large. The first makes large projects less sustainable even as they become more important to the industry. They need not support the projects out of altruism, but they have every practical reason to do so because, without that support, the project may die. The second hurts basically every developer, especially ones with less professional experience, by making the industry more insular. That snowballs into a shortage of experienced developers later.
And Mark's remedies for this situation amount to making open source software more closed or appealing to some sense of fairness in the idea that a developer should be entitled to patent restrictions of work they made for hire. To be too brief, making software more closed isn't an actual remedy when the market is essentially demanding free software, so patents and restrictions don't address this.
Also, he makes a lot of other direct statements in the latter half of the piece that I disagree with like "Using or developing more open source software is not going to improve anyone’s lives." It certainly has. "Developing open source software is not a public good." It is, but the larger projects need support. In my view, the easiest way to allocate resources to these projects is based on who's using them and how much, which is something that can be regulated via norms although those norms have not exactly materialized. How exactly to get us there is an open question, but one way we can start is by not doing the things described in my article.
Mike -
Mark -
Mike -
Mark -
Mike -
Mark -
Not trying to be a jerk. The opposite actually. But it really seems like this guy is on the same side as you. Most examples you give of "disagreement" are actually small statements taken out of context. He is NOT anti-open source and he in no way "let's the corporations off the hook" (he does the opposite actually, his point is THEY won't stop unless you make them).
Once again. I do not agree with either of you. BUT... This guy should be your ally because you are both on the same anti-big bad corporations team. But because you can't let go of a few small statements or have different ideas to handle it you went and wrote an entire post against it. WTH? You agree with 80% of what he says!
And IMO most of what you state as not agreeing with are not actually where he is, but rather your perception of where he is because of how you interpret certain statements. (If that makes sense). Just food for thought. I'm done now. Feel free to respond but I probably won't (not in a disrespectful way, I just feel I've thoroughly made my point and there is nothing left to say, feel free to have the last word, I'll just listen and not speak)
Aren't many big OSS projects created AND maintained by corps?
VSCode, React, Gatsby, Expo, Apollo, Amplify, GraphQL, etc pp.
Yup. I don't have much of an opinion on the subject myself. I think open source is great. I also think that many times big business can make bad or even harmful choices. It seems both authors have some interesting points. I just thought it strange how flared up this article's author was when it seems to be mostly in agreement with what it is roasting. Mark point seemed to be that open source "failed" because it didn't protect itself from abuse. This article's author seemed hung up on the "Open source failed" click-bait title and missed the forest for the trees... (Just my opinion of course). But in all seriousness, after reading both articles it really doesn't seem that Mike and Mark disagree...
Mark's article picked on Microsoft specifically because of its big swing in favor of open source in the last few years, saying that open source allowing free riding enables their bad behavior. Basically, open source "failed" because it didn't prevent corporations from being bad actors.
Now, I certainly have qualms with how Microsoft has gone about joining the open source movement, such as immediately buying its way onto the Linux Foundation board in light of its previous "Linux is evil" stance or "Embrace, Extend, Extinguish." The Linux Foundation's questionable history with conflicts of interest doesn't help, either. But the general idea of acknowledging that open source software is foundational to modern software is just being realistic about where we're at.
I think part of the issue with the discourse is that open source is so complicated, so politicized from the beginning, so incredibly varied across projects.
Open source is pretty abstract, and therefore, arguments can often devolve to two sides making straw man points. It's easy to argue against the most extreme ideas on each side of certain debates.
As I ventured to become more familiar with open source I found that there was so much nuance that people attempted to label as the definitive answer. It seemed like people had claimed to have "solved" certain open source dilemmas which, to me, seem more like ongoing economic tradeoffs which can never be solved.
In your post, and in my reply, it's all very philosophical and I wouldn't want to think that anyone in this conversation "nailed it" with their argument. We're all just holding court and chatting.
It's definitely one of those subjects about which discussion tenaciously runs away to abstractions. :) That's good, in the sense that open source ideas are versatile and can be a lens on many things, but it also allows people to argue past concrete details and criticisms by appealing to higher principles. I tried to balance that by including both the more heady origins for it in corporation-driven changes to the copyright law and also the concrete ways in which corporations risk their operations and make the industry less healthy by failing to support these groups they depend on.
However, the result does also feel like two articles merged into one to me, and thereby more of an abstraction, haha.
One of the main reasons we open source is to help fix bugs by crowdsourcing our development. No matter how big the organization is it cannot have enough resources to tackle all issues with closed/ private repos and deliver value in time.
For many popular projects its a win-win situation as contributors see their name in there and organizations improve their solutions almost for free.
Everything else is politics.
I wonder how much Open Source has contributed to the massive decline in global poverty.
I suspect it's role is not at all trivial. The ethic of "let solve problems because they are worth solving" and "let's give powerful tools to the masses" Changed the game in exponential ways.
How many business decisions are made in the world today without the use of Linux, apache, mysql, php, or another free variant of such? Nearly none. We have cheap smartphones in the pockets of the poor and developing world powered by linux or a variant.
When I was a kid, It cost dollars to call a state away. I watched my kids play Magic the Gathering with over video chat with their friends in Australia for several hours.
There is a Libertarian socialism that has emerged without the need for government revolution or force. It is handing the means of production to the masses. Those tools being distributed allow communities much better ability to solve problems locally and to communicate and coordinate with others to obtain resources that are needed.
You get the benefits of socialism without the oppression. Still a ways to go, but the impact of technology seems to pay compound interest.
Open Source as such did not fail. We are having more sourcecode in the open than ever. What failed are the expectations people had about Open Source.
Back in the day, having proprietary source code was meant to be a competitive advantage. What was right: having knowledge others have not is a competitive advantage. Source code was thought to be part of that knowledge.
That last assumption is now mostly rebutted - of course e.g. Google does not share the source of its Borg-platform but we get Kubernetes (yay!).
It is not the openess or closedness of software which determines economic success. It is the sheer amount of knowledge you have. And this includes (of course) the ability to (physically) store and leverage this knowledge for economic success.
In contrast to what was once thought a competitive disadvantage - opening up your source code - it was a business model of success. You open your software but you keep the knowledge silos closed or at least build access gateways. As an aside: Facebook itself recognized that a bit too late, giving external access to its family jewels. But after realizing that, they closed the access down again.
That said, it is quite obvious, why this belief - more open software equals a more open and more free society - must have faild: It was never about software, it was about knowledge and its concentration. Or to use a more hip term: it is about platform capitalism.
And Open Source played in as far a key role in that, that it allowed to collaborate on the software which makes it easier for everyone to built platforms, which in turn helped the platforms itself to improve their software to build a platform. Microsoft wants people to have a Linux version of VSCode? Let's do that whith the help of the community. A smart move, indeed. Everybody wins in some respect.
The problem is: what to make of it as a society? The question of how to deal with platform capitalism is a question, which can not be answered with software. It can only be answered in a free and public discourse.
As an afterthought:
From a technical point of view, it has never been easier to disrupt the industry with a new business model, since building the software itself is less of an obstacle to tackle; think of Alibaba, Wish, etc.
In this respect Open Source has helped shuffling the cards.
Nice, thorough argument Mike. I can see both sides, and it's definitely a mutli-faceted discussion and analysis.
We discuss these kinds of issues on The Changelog pretty frequently. You'd probably enjoy (but probably disagree with) Zed Shaw's take on the subject.
Zed Shaw has flipped his bozo bit with me for life.
An hour and a half of him is a big ask, so I read the PDF transcript of his Twitter comments that got him on the show. Because those comments were primarily motivated by his persecution complex (They blacklisted me just for a different opinion! (No. No, they didn't.)) when they weren't outright and laughably wrong (Patreon is money laundering), I find them irrelevant. He's most directly addressed in my article where I talk about maintenance needs differentiating open source software from literature. It's just not true that companies can continue to use the software indefinitely if that software has no ongoing maintenance, and that maintenance requires external support from at least somebody when that software project gets sufficiently large.
I'll definitely try to check out a different episode of your show, though. Always out to find new fellow podcasters! ^^
I think it's misleading to conclude that we must either fight for it with legal methods like viral copyright licenses or just accept that corporations will make a lot of money without paying. There are obvious reasons for corporations to tithe money to open source projects. And, from a risk management standpoint, it is increasingly beneficial for them to do so as a security measure to make sure that their important dependencies continue being developed.
On the other hand, the necessary increases in open source donations do not seem to have happened perfectly in a vacuum. I would first suggest social stigma as one factor to the solution, but corporations only seem to respond to that when the advocacy comes from within (see Google walking back its technical support of China's censorship). Failing that, legal remedies might work, but I'm eternally skeptical about the efficacy of enforcement and I'm sure new, completely free alternatives would turn up to replace any of our current open source dependencies that adopted a mandatory payment model. If the way the market moves is toward completely free, then the market must adapt to keep itself sustainable, even if that means the market needs to read an economics text more complicated than "Supply and Demand" or "Obey the invisible hand."
I'm not concerned with your race. I also did not say people were being racist. I said that a system that uses open source contributions as a filter for employment biases the hiring process in favor of hiring open source developers, the vast majority of whom are white and male. Any data scientist--or, hell, just an opinion pollster--would challenge the implicit bias of any similar filter in any other context. Unfortunately, reactionary defenses against imagined charges of racism like your comment throw out these considerations because "it's fine for me." What if it's not fine for someone else?
Also, ask your manager how important it is for them to prove that they can manage in an interview. Ask a mechanical engineer how important it is that they prove their engineering capabilities when they're not getting paid for it. Most other fields have much more generous assumptions of competence in recruiting than software development does. They do just fine without locking people out over this.
Some comments may only be visible to logged-in visitors. Sign in to view all comments.