Hacker News Reader

Comment by joshlemer | Hacker News Reader

Some other commenters in this thread have pointed out that nobody is forcing open source developers to work on the contributions that they make, and that is strictly speaking true. But in the culture of software development, there does seem to be an ambient message often repeated or hinted at, that it is good to "give back" to open source by contributing and that it is virtuous to do so.

I now am starting to rethink this sentiment because the vast majority of the benefit of open source contributions on github, be they to languages, runtimes, application frameworks, databases, etc, go towards increasing the bottom line of for-profit companies, not to developers. And the vast majority of the beneficiaries of open source will never support the project even when they are fortune 50 companies saving millions of dollars by using the work of one volunteer.

There is also the idea that contributing will be great for your career development. I have found that not to be the case at all, I think that no potential or current employer has ever given a rat's ass about open source contributions, and do not consider that work as valuable when making hiring decisions. The work you are paid to do is the only thing anyone cares about. I'm not saying that that shouldn't be the case, but just that that is the case.

Given then, that there's very little upside to doing open source, and most/all the benefits go to profit-making corporations, it is puzzling why do we even push for greater involvement in open source at all? It seems we shouldn't be, we should be warning people who want to contribute to Open Source that they should probably spend their time doing their own studying and personal/skill development which will allow them to succeed in the roles that they have with their current role or a role they'd like to obtain one day, for money.

View on HN

How is software unique in this? Tons of museums around the world are visited by the kids of people who make 100k a year even though the average worker there is an unpaid intern. Do you read any news? Listen to NPR podcasts? I think NPR does pay interns, but they don't pay them millions of listeners * 50min a week of listener time * $60 an hour of average listener earnings.

What about science? If you ever benefitted from a genetic screening for any disease, your doctor made anywhere between $100 and $400 an hour, and the graduate student who discovered the function of that gene $10 an hour if you're accounting for his tuition credit. Who tells these students to stop spending 60 hours a week at the lab and start using their skills doing something people will pay for, like develop a no-palm-oil no-aspartame but-still-addictive Mars bar for 25c unit cost?

People will accept less money in exchange for doing something meaningful, like sitting in their basement painting stuff rich people will speculate on for millions of dollars after their death, or campaigning for a rich lawyer who already makes 200k and will get a good bonus post-election because he has the right party colors.

If you could literally sell meaning to people for less friction than those alternatives they'd eat it up. And they'd spend their time making money to spend it on your meaning-making money-sucking machine.

Because that's all money is for: be exchanged for things that make your life not suck. And this exchange tends to bear transaction costs.

> If you could literally sell meaning to people for less friction than those alternatives they'd eat it up.

"If you want to get rich, start a religion" - L. Ron Hubbard

prediction: 2020's will be a decade of hyperactive cult growth, accelerated by proliferation of private alternative currencies, loneliness and the decline in traditional religious affiliations. Patronage of the ambitiously famous will be direct and recurring, without need of branded material maguffins to mediate the transaction loop.

https://en.wikipedia.org/wiki/Megachurch

> Who tells these students to stop spending 60 hours a week at the lab and start using their skills doing something people will pay for, like develop a no-palm-oil no-aspartame but-still-addictive Mars bar for 25c unit cost?

Almost everybody, whether in or out of academic science, tells them that.

Traditionally, OSS contributions are selfless labors of love and curiosity. These people are flexing their intellect and creativity doing something they thoroughly enjoy: solving problems in the most elegant way possible. Code is shared with an air of positivity, camaraderie, and we're-all-building-this-together. At the risk of sounding kitschy, it's this sense of collaboration that enables the open source community to produce robust software proudly capable of competing with proprietary alternatives.

If this culture is stifled the name of depriving corporations 'free labor' through OSS, the creative hacker community isn't elevated - everyone is brought down. IMHO anyone seeking to build financial value for developers or pad their resume is better off working on a side project/startup.

It's generally not a problem until the big companies start asking for features, asking for bugs to be fixed, etc., expecting anything more than what the maintainers choose to donate. If someone from Amazon reports a bug it's kind of scary, you feel like you've broken Amazon. You might even try to set aside time you weren't planning to set aside. It's hard to remember/enforce that you don't owe them that time (especially if they're nice).

If Amazon doesn't pay you, then why are you scared of breaking Amazon?

I understand why somebody would be scared of that. They're big and important to the tune of many many billions of dollars, while you or I are not. But you shouldn't let that get to you. You should make a conscious effort to not let that get to you. If they want you to take their problems personally, let them hire you. If they can be broken by you but are unwilling to pay you, then fuck them. Put them lower in your priority queue than private individuals.

https://youtu.be/3XGAmPRxV48

Mike Monteiro's version runs slightly longer:

https://invidio.us/watch?v=jVkLVRt6c1U

(Fuck you, pay me.)

Well that is indeed what I said, you just have to get used to telling people that. That's the big problem. That's the reason this thread on HN exists. People trying to figure out the right path, how to ask for money on their little OSS projects, etc.

I would love to have created something that was so critical it could break amazon. That sounds like a ticket to $5k a day consulting gigs

From the article:

> It’s still not clear that companies would be interested in such a contract. When Stenberg asked the company that needed him to fly to a different country to troubleshoot their problem to pay for one, they refused.

If one of your primary motivations for working on FOSS is to improve the world (which for many of us is the case), then it's very easy to feel an obligation to un-break Amazon because it is a quick way to improve the world for millions of people.

That, on its own, is fine. It brings you satisfaction rather than financial support. What starts being fishy is when companies like Amazon who can readily afford to offer financial support instead rely on and exploit that sense of obligation to get work done for free.

I realise this sounds whiny and that's not the point. The point is that by withholding FOSS contributors financial support, these companies are creating an environment in which the best and most dedicated contributors run the risk of burning themselves out, making everyone worse for it.

> If someone from Amazon reports a bug it's kind of scary, you feel like you've broken Amazon.

Really? If Amazon filed a bug on one of my repos I'd probably see dollar signs flashing.

How would you go about getting paid for that? I doubt it would be possible, even if the people that reported it really want to and need it, they could never get approval from the bureaucrats.

Eh - I work for a big SV company in a group that provides key software and hardware. If we have a schedule risk that can be mitigated by paying a 3rd party engineer a reasonable sum we could probably get it approved in a week. It would require at least exploring alternatives and quantifying the risk, but it's really not hard to get (reasonable amounts of prudently spent) money when your project is in the profit pipeline. I doubt it would have to go as high as a VP for approval.

Some companies do have systems set up to pay OSS projects, but it's hard to navigate. I think a lot of people enjoy donating their time to OSS because it's fairly low-stress (until the above situation happens). I don't want to figure out billing and stuff when companies want my support, it's not a company, and that's why the article here exists. That's why this discussion is happening.

That's their problem, not yours.

https://invidio.us/watch?v=jVkLVRt6c1U

If a project starts having large companies relying on it, traditionally there have been 2 mechanisms to address this:

- start a company around providing enterprise support services;

- ask for the large companies to provide some resources such as a part (or all) of an employee's time to work on the project, for the parts that are more dear to that company.

If a company is unwilling to pay, either in the form of a support contract or employee's time, it should look at another solution. Not just because they are being disrespectful/annoying/distracting to the maintainers of the OSS project but also because it means they don't have a good contingency plan for when things go wrong.

For a larger business, the value of using OSS is not to save money, it's to get the right level of customization for specific needs.

You are forgetting about a third tool in the toolbox. Licenses. If you don't want your open source work to be used for profit select a license that is restrictive.

Ok but sometimes you don't care or encourage this as long as you don't have to support it for free. And you even expect companies not to be assholes taking but not giving back - not enforcing but if it's the nth ticket or bug from mega Corp X that profit from it then let them solve it themselves.

Or ask them to fix it and submit a PR. If their edge case broke it, then they should know best how to fix it.

"Fix it yourself" is not a full solution to this problem. It takes time to review and merge pull requests. Especially if some well staffed team is dropping a large chunk of code for a relatively esoteric use case. It will inevitably take a lot of back and forth to work it into the right form for the project's primary goals, and it may never fit. If a company really depends on a project, it behooves them to contribute financially through a license or employment, or to maintain a fork.

RedHat basically showed how you can setup an Enterprise OSS support business.

There's also a multitude of businesses selling support licenses for plugins/modules that extend OSS applications.

Try laughing at the absurdity of that until the nervous edge blunts. You should be feeling some empowerment once the existential terror has worn off.

You know it's gone when you're ready to send them a cost estimate at twice your hourly rate.

One of the main benefits of releasing free software (and IMHO a glaring omission from the article) is to get bug reports before the bug affects you. If Amazon wants to pay somebody to send you such bug reports, so much the better.

That is a great problem to have -- it is how commercial wings of open source projects are formed. The commercial wing becomes a for-profit mechanism to help champion/grease bugs/features through the update process, as well as providing PS and indemnity/support services.

You can politely ask these nice people to contribute funds towards your project that is so important to them. Since they are so nice, they will probably nicely send some money towards the oss developer.

That is the subject of the entire "oss funding" problem. Precisely that many companies are not willing to contribute funds even though the project is important to them. That's why this thread on HN exists.

If they are not willing to contribute funding, this might mean that the project is not much important to them?

Anyway, the developer does not need to stress out too much about Amazon employees having problems with their code...

"I'm sorry, I'm busy having a beer and playing a game right now. But I can offer you a full refund?"

Why can't the person from Amazon reporting the bug just submit a pull request?

Based on FOSS stats, majority of open source contributions is paid for. Either by companies or universities.

Enjoying to solve problems, yes. And curiosity too! Both are great motivators.

But selflessness? That's the lie we tell ourselves when the true driver is something else, like narcissism.

It's not a bad lie. The outcome is pretty much the same while it lasts. But altruism is not a reliable motivator, and certainly not a requirement for volunteer work. A requirement is to have basic human needs met with some minimal security. Like, not having to judge everything you do by how much money it will earn you.

Perhaps I'm in my own echo chamber, but I definitely view open source contributions on a resume as a huge positive when I'm reviewing resumes (as a software dev my boss trusts to evaluate applicants), and I had thought that if that wasn't at least the norm that it would be pretty common among tech companies / companies where developers are part of the evaluation process.

edit: Also, all of the open source I do or that I know of my peers doing is 100% for the love of programming and open source culture, without a bunch of hidden motives. Open source was definitely key to me becoming a programmer when I was a kid - without it I would never have gone as deep as I did into coding (linux, gcc, irrlicht), 3D modeling (blender), etc.

Have you encountered applicants with OSS on their GitHub profiles just to show off to HR? ResumeWare if you will?

I've seen a few trying to apply to my company. They had a few projects in the languages they had on their resume, but upon closer inspection you could see they were just project templates they had gotten from some tutorial, made one or two insubstantial changes to and then uploaded to github.

Ahh. Irrlicht. I miss the era of small-scale game engines and everyone working on their own in OpenGL/D3D. DevMaster website even had a list of engines, with at least 300 positions on it, sortable by language and other technologies. Unfortunately that era has faded away because Unity and Unreal Engine cornered the market and there aren't many people left working on their own engines.

But they are not working for free, they're doing this work during work hours. Individual volunteer maintainers this article is about are a minority. The majority of contributions to the open source projects that "the internet runs on" are from businesses that use the software and maintain or improve it.

I'm not talking about your small hobby project with a few dozen users - as good as it may be, the internet doesn't run on it. cURL is really an exception, not the rule.

And openssl?

(It is possible for there to be multiple anecdotal exceptions.)

The whole thread is anecdotal. I'm not aware of any quantitative study in the matter. There are also the distributions like Debian which is essentially run by volunteers.

I know there are volunteers that do a lot of Debian work, but that is not to say Debian is purely volunteer work. Obviously there is plenty of sponsored work. https://www.debian.org/partners/

I’d argue without trying to classify and quantify Debian contributions, there is no way to know who is doing work on Debian on their employer’s dime.

I also try to contribute to various projects in spare time as well. My contributions don’t really add up to much overall, though.

Sure, it’s all anecdotal. But otoh, the notion that FOSS is mostly volunteer work is probably worth challenging. Most non-huge projects I’ve come across only have 4 or 5 key contributors. Open source is at least, not as distributed as believed for most projects. On the other hand, companies using FOSS have incentives to contribute, and sometimes will even hire contributors on important projects.

I didn't do any quantitative study, but I'd be really surprised if most maintainers in the project are not volunteers. Those contributors, though usually individually small, tend to add up. It's also probably difficult to justify Debian package maintenance to your employer, unless you work for canonical. Even products that distribute a .deb tend to do it outside of the distribution, rather than really contributing.

I haven't done a quantitative study either, but here are two examples of extremely influential open source projects that the internet runs on:

The spring framework has contributions from many volunteers, but most big contributions and all of the integration work come from Pivotal.

The contributions to the linux kernel are graphed here:

https://www.extremetech.com/computing/175919-who-actually-de...

80% of new contributions are from companies. That shouldn't be surprising at all, because these companies are also the biggest users of the project.

As for myself, in the rare occasions I contributed a patch to an OSS project, most of the time it was to solve a problem I encountered at work, so I have anecdotal evidence that at least some from the category 'volunteers' also got paid to write that code.

Actually, there's certainly of reasons you might do maintenance work on company time. At Google for example, the gLinux distribution is heavily based on Debian[1]. Why would individuals do it on company time otherwise? One example is to solve a business problem. I have definitely contributed upstream to open source projects to solve problems at work.

[1]: https://en.wikipedia.org/wiki/gLinux

OpenSSL has hundreds of contributors, not just one guy, and the 'heartbleed bug' guy they were referring to wrote the code during his PhD, so it wasn't a volunteer donating his spare time.

And that's why we have things like GPL and free software, which ensure, that contributions are brought back to the community. If you want to have contributions brought back to the communities and not only rest inside for profit organizations, then perhaps choose a license to enforce that. This is what many open source people do not get: Many "don't care, it's open source"-licenses (MIT license for example) lack the backed in ethics of the free software licenses, which in contrast have been _designed_ to force contributions to be shared with the community. When someone chooses an open source and not free software license, they can't complain afterwards, that modifications are not brought back as contributions for the community. That's just being uninformed.

> vast majority of the benefit of open source contributions on github, be they to languages, runtimes, application frameworks, databases, etc, go towards increasing the bottom line of for-profit companies

This is nowhere close to true, unless you define “benefit” as “cash profits”. Even then, it’s a tough argument.

There have been extraordinary benefits to most people in the world from open-source contributions. Without a wide range of open source code projects there would be no Wikipedia, no SciHub, no web forums, no search engines, no web maps, no movie/book/restaurant review sites, no craigslist, no online dating, no news websites, no web education platforms (MOOCs, Open Courseware, Khan Academy, ...), no technical Q&A sites, no video sharing sites, no social media, no e-commerce, no online banking, no ...

The great majority of the benefits of all of these software products and projects accrue to their many users, not to their authors or corporate hosts.

(Of course technology is not all positive, and many people have also been harmed by technology. Open source also enables surveillance, stalking, bullying, new kinds of fraud, new venues for propaganda, new legal risks, new avenues for social and political control by self-interested and unaccountable people and institutions, etc.)

Open source or not, the big companies will always win. The only real difference is that as a whole society would slow down and be worse off without open source. Either way, it'll be equally unfair.

> Given then, that there's very little upside to doing open source, and most/all the benefits go to profit-making corporations

I partially disagree. Yes, the benefit will not be fairly shared. However, the volunteers' works accelerate entreprises' works, which can then provide new services and new products faster.

Imagine some ML algorithms could find the cure to aging. You have two options : either wait that some company develops it, but you might be dead by then, or accerelate the process by open sourcing your code and hopefully get the companies closer to this goal. Now, that doesn't mean you will be able to access that cure if a company finally finds it and decides to sell it, but you increase the odds that something might be discovered.

It's like giving $1000 to charity, but with the scale of software, your coded contribution can have way more impact than $1000 on the world.

Thanks for pointing this out. The quoted statement is only true if "benefit" is restricted to meaning "financial benefit." In whatever people do, not just in software, most of the financial benefit goes to profit-making companies because everything they do is directed toward making profits. But what about the unquantifiable benefits that come from improving people's lives?

How would you weigh the benefits of millions of people using and loving a piece of software, even if no profits are generated? What if they are able to free up time which they use to improve their communities, spend more time with their friends and family, maybe even make money for themselves?

in the same vein i disagree with people who complain about freeloaders who use Free Software and don't seem to give anything back. in reality we do not know how they give back to society. they might be doing many of those things that you mention that i as the developer of the software can not see.

all good points. i also think what really pushes a lot of developers to do open source, and bear with me for the unpopular opinion--is ego. they love having followers, and a voice on twitter with some open source badge on their profile page. it's in a way a status symbol of "i did this important thing that a lot of people rely on, im pretty important." i highly doubt that even 5% of contributors are doing it out of the kindness of their hearts or the community.

you can sort of see evidence of this within repositories filled with smug comments within the issues or pull requests

It basically boils down to whether you think people are really being altruistic, or whether it's simply reciprocal altruism.

sure-- but the main difference is that the community pretends to be super altruistic under the guise of doing god's work for the good of all developers. it comes off as super fake. at least thats sort of the vibe i get from it. i also dont have a huge sample size of repos so maybe im talking mostly out of my ass in this case

i would honestly rather it all be backed by corps and people get paid for it without any hints of "benevolent dictators"

I agree that it's a thing, aside from some examples I have in mind, like for example, the original authors of OpenSSL (originally SSLeay) didn't really ask for any credit as far as I know. I think you might be overestimating the number of people who do a lot of self-promotion simply because those people are going to be more visible. There are plenty of people who work on important things and don't really do a lot of marketing or self-promo.

> i would honestly rather it all be backed by corps and people get paid for it without any hints of "benevolent dictators"

Well don't use opensource software then, go buy it... It's your choice...

Just like it's the choice of a developer to give away something they made for free...

You mean backed by red-hat under the benevolent hand of Ulrich Drepper?

Maybe just community? Building something useful to oneself, releasing it to find it solves others' problems, seeing one's work have a broad benefit, collaborating with others to make it better, etc. Not sure if this is just a positive spin of your "ego" hypothesis, but my gut tells me that it's a categorically distinct motivation.

I don't think this is fundamentally different from being proud of your importance as measured by salary. Granted a salary might be a more accurate measurement.

Nonsense. Salary is an indicator of utilized leverage, not of quality. Plenty of software devs could easily put themselves out of business, and not be sure what to work on next, without capturing all of the value they generate.

For me, it's the somewhat selfish but defensible idea that open source is a good way to get experience with technologies that I don't get to use in my day job. I get some amount of pride in seeing my name in release notes etc but that is about it, I don't do Twitter, don't have followers, and am not high-profile in any way.

Ego strokes is just one form of non-monetary compensation. OSS is not alone in being the product of voluntary work. Lots of worthwhile civic, charitable, communal and religious endeavors rely on volunteer labor which is compensated through recognition, friendships, power trips, reciprocal benefits, satisfaction of giving back, etc. The power of non-monetary compensation should not be underestimated, even in the context of paid work.

On the other hand, software maintenance is a category of work that tends to be poorly compensated even in the case of proprietary corporate software. So it is no shock that OSS maintenance also is poorly compensated either monetarily or non-monetarily.

I’d say ego is less than 10%, and the quiet, achieving minority are not doing it out of the kindness of their hearts or love for the community, they are doing it because it’s a love of their craft.

Once you see something that you have built, running at thousands of operations a second for years, once it’s running smoothly and all the bugs and edge cases have been smoothed out. Watching the machine dance, visualising the data structures in your mind - and algorithms running on them. Thinking about state changes, loops and execution, and watching it all running in a beautiful symphony of logic : there’s really no other feeling of joy and appreciation of beauty. I love being a programmer, and I suspect the majority of open source contributors are like this.

The snide comments and unhelpful github messages are the minority in my experience, especially with more mature products (kernel development, sys ops tools etc)

This kind of ego satisfaction is not a necessarily bad force, and is responsible for many of the advancements in world history. Why did guys like Newton or Salk publish their work instead of profit from it in secret? Because there is more to life than money. Because there is the satisfaction of knowing tour work has changed the world. Because they wanted their names to ring out in the annals of history, be carved into the tops of marble columns, etc.

I don't think that's true. The software itself doesn't benefit their bottom line because it's free for everyone. You can't charge a premium because your business runs on Linux.

All open source software does is provide a base-line of technology that all companies and all individuals can use. You can't directly or even indirectly profit of it -- you can only profit on the value you can provide on top of it.

As technologists, that's what we want! We want companies and individuals to stand on the shoulders of giants and peek a little bit higher. We don't need anyone to re-invent the wheel over and over.

> I think that no potential or current employer has ever given a rat's ass about open source contributions, and do not consider that work as valuable when making hiring decisions.

That is not the case with the D programming language effort. Quite a number of strong D open-source contributors have been recruited into very well paying positions directly because of their contributions.

I suspect the key is contributing to a higher profile open source project. One that is high enough that one can make major contributions, and not so high that one has a hard time standing out amongst the other contributors.

Also Linux.

A couple of things I've been thinking of (speaking of GPL sofware)

- free software (GPL) can be used for any purpose, by the user. Yes, this means a corporation can use the software without paying. However if it tries to redistribute the software it has the obligation to distribute the source code and any improvements made to it.

- Large corporations are made of people. It is also of value that a person inside a large corporation can learn to use GPL tools, then leave the corporation and download use the same tools at home or another job and continue to use the skills learned.

- GPL software is also great for education - students can not only use tools, but can also pull them apart and see how they work

- GPL software can be a legacy if you're into that. You can write and distribute software under the GPL, and it can survive your current job, or corporation, or live on past your lifetime. GPL software has outlived most of the places I've worked. Meanwhile the non-free software those places created during my tenures have largely disappeared.

- GPL software is a great foil against our losing war with privacy. I can foresee a future where running software you can view and modify might be the only way to ensure you know you aren't being taken advantage of.

It’s part of being a professional. We even have a word for it: pro bono. Volunteering your skills doesn’t make you a sucker.

Companies may not care about volunteer work when making hiring decisions, but that should not be why you do it.

You say most open source software is used by companies to better their bottom line, but if you’re making six figures through skilled work and not giving anything back to your community, I fail to see how that’s any better.

I'm the one you're responding to and I'm not making six figures, you can't assume everyone lives in USA making amazing salaries and bonuses.

Forget I mentioned money, then. That doesn't change anything else I wrote.

I guarantee I earned less money in the past year than anyone reading this (I don't work in tech any more), and it didn't stop me from volunteering my skills. Being rich is not a requirement for contributing.

Sure, point taken. I'm not saying don't give to charity or volunteer for a good cause, I just question if contributing to open source software is really an effective way to go about doing good. If the project you contribute to has some special positive impact on the world, then probably, but if you're just contributing to the general software ecosystem, then it's hard to see how that directly benefits anyone other than, at best, a bunch of well off software developers (though that's not clear to me either) and at worst, only benefitting profit seeking corporations which we would normally expect to fund work being done to run their business.

As someone making hiring decisions, I very much do visit the github accounts of my applicants. I consider the quality of their code and look favorably on them for making code available. I look even more favorably on those who bother to make licensing decisions for their code. Do you get an extra $5k for pushing 1k lines of GNU-licensed code? No. Did you get a callback on your job application? You bet. Do I take pride in the OSS work of my people? Yes. Does that factor into their performance review? Yes. Does that factor into their annual raise and prospects for promotion? Yes.

I have also hired people who don't have GitHub accounts. In those cases, they have some other proof of work. Papers, thesis, references (which are a drag, because then I have to also figure out how to vet the reference) and, particularly valuable, referrals from people I already know personally. This last one goes both ways: not only do I have high confidence in the good intentions of the person making the referral, I can then retrospectively assess this contact's ability to assess.

Does your company allow engineers to work on OSS while they're not busy? If not then how can you claim you're proud of their contributions, and don't you feel like you're coercing people into working overtime to get a better raise or a promotion?

My take: He can claim he is proud because he acknowledges their work and - as the rest of the work - it influences promotion and raises. He enjoys and rewards the positive impact this work has for the company image and the skills of the employees.

I can understand that, I just disagree that to be a better employee one should also devote their free time to an activity that directly influences one's performance review at work. This is definitely a good deal for people who enjoy doing OSS work on their own time-- but it feels wrong and not fair to those who don't want or don't have time to put the extra hours for that.

We're so remote the sun never sets on our group, so the definition of "workday" is pretty fluid, so yes, "when they're not busy" is a good partition. Otherwise, as the other child comment says, there is a pride in supporting effort. I don't necessarily benefit directly from the projects they may work on. Like others have said, I can't coerce them to do write open source code. But I definitely benefit from their improved or sustained skills. You have to find differentiators to make such decisions. Are you arguing that I should ignore signal or that open source contributions are an invalid signal, or the use of the signal itself is immoral, by some definition of "immoral"?

I don't think you can make a sweeping generalization like that, because "employer" is not some omnipotent figure sitting in an ivory tower. It's other developers, and managers who are (typically) technical and can appreciate individuals who contribute to open source.

If I see someone is an active contributor on Scikit-learn, numpy, etc. that is a certainly a strong signal for me. Yes I'm not making the hiring decision, but I can assure you the HM is going to be influenced by those who interviewed the candidate.

Citations needed. This is just a pessimistic speculation based on your experience, but my experience is the exact opposite. So who's right? We should leave this kind of stuff at the door.

Is there research comparing the current situation with a hypothetical world where OSS doesn't exist? How many less developers would have a job? How much less would the remaining be paid? I'm imagining it and that's not a world I want to be in.

To add a bit more to the "factualness" of the conversation, according to the Bureau of Labor Statistics Software Engineers in the U.S. took home $144 billion, which for comparison is more than the revenue of Alphabet or Microsoft. How much of that do you think can be attributed to OSS?

There's a problem with this perspective. I have an insight that would help you see the picture more accurately. I would share it, but here on HN people from profit-making corporations could also see it and benefit, so just ask me in person next time you see me.

So since a corporation might get value from your work then nobody should get value from your work?

That's not a super charitable interpretation of what I said. I'm saying that for a lot of open source software, virtually 100% of the beneficiaries are corporations, so we shouldn't fool ourselves into thinking we're doing god's work here, and my second point is that people probably overestimate the personal reputational/career gains to themselves from contributing to open source. So, a large portion of open source contributions do nothing other than provide free highly skilled labour to Coca Cola, Walmart, and McDonald's.

>I'm saying that for a lot of open source software, virtually 100% of the beneficiaries are corporations

And this is bullshit. Most popular open source projects see heavy usage in academia, personal projects, public projects, government projects, etc. Python is a massive reason any students know how to program at all.

>Python is a massive reason any students know how to program at all

At other points in our history that language would be BASIC or Pascal. Their initial implementations may not have been open source

This is bullshit. Corporations benefit from a larger pool of students and cheaper pythons developers, Postgresql and mysql are very popular is business. By comparison, how many people use Ubuntu instead of Linux?

> I'm saying that for a lot of open source software, virtually 100% of the beneficiaries are corporations

What open source projects specifically?

> So, a large portion of open source contributions do nothing other than provide free highly skilled labour to Coca Cola, Walmart, and McDonald's.

In my opinion, any open source development is valued WAY MORE than closed source development, as it can be shared and verified by others. If you have a project that's benefiting multi billion dollar companies, then you've created something useful that can be shared with small companies and individuals too.

Also at the end of the day it's the developers choice, they can stop maintaining it, but the point is; sharing code is better for everybody

> What open source projects specifically?

I have not contributed to any of these projects but for instance: Spark, Flink, Druid, Akka, Hadoop, Cassandra, Kafka, Java Spring, Hibernate, etc etc.

All those (mostly distributed) systems are not exclusively used by big corporations. They are also used in research and smaller startups too. Hell I know people who use the Java Spring Framework on personal projects.

The world of open source benefits the powerful entities, and that's OK that's what open source is about... Everyone gets the code for free... No matter who you are

> The world of open source benefits the powerful entities, and that's OK that's what open source is about... Everyone gets the code for free

The risk is that it devalues labor. And could build a culture where so much is expected for free that there is never enough financial support.

We should strive to make things easier. If open source packages can make software easier to write, that's a good thing and devalued labour is a sign that the labour is easier.

Much like the improvements in farming and agriculture have changed from manually reaping the fields to AI controlled harvesting equipment.

There will always be software engineering problems to solve, they just move up the stack...

> There will always be software engineering problems to solve, they just move up the stack...

When fewer and fewer people are needed to do the same work it doesn't necessarily make for a better or more sustainable market: socially or ecologically. Consider a world where one needs 40 years of education (instead of say 20) to begin to contribute in a way the capital owners are willing to pay for. And paying for the privilege of much of that education.

That's true, as technology advances, the knowledge needed to maintain it increases. In 50 years the software stack could require petabytes of memory to run complex simulations..

But to me that's a pessimistic way to look at it. It's an accomplishment to need less people to build amazing things.

We're already at a point where even the most experienced engineer can only understand a fraction of the stack they use. Compare that with the 1970's where a programmer might understand 90% of the mainframe they use... because it was simpler...

Basically you don't need to know as much knowledge as earlier generations to accomplish the same thing, but this devalues that thing (or makes it much harder to be good at) ... All the low hanging fruit is gone I guess

My larger point is that there could be fewer paying roles and less willingness to pay for anything. A society of a few owners, few paid workers, and everyone else without much/any influence is not a balanced or healthy one.

Isn't this kind of the problem the GPL is supposed to solve?

Sure, a company can build off an open source project to increase their bottom line, but they'll have to contribute back any changes they make. Which means their work is now benefitting other developers as well.

I feel like the GPL has somewhat fallen out of favor in recent years. To an extent I understand why, and yet...

Back when people were making a big deal about the apparent lack of women/minorities/etc in open source projects, I objected to trying to get them to give away their labor for free.

The response to that was that these contributions lead to jobs. But, that idea seems to have largely disappeared. There was a period where people were claiming your github account is your resume, but I haven't heard that sentiment in several years.

Funnily enough, a GitHub-to-resume tool recently hit the front page[0], but yes, that opinion seems to be in decline.

[0]: https://news.ycombinator.com/item?id=20995056

Originally git repos use to signal for passionate developers, the kind of people that you couldn't stop from coding. Before the 2008 boom there were far less people actively interested in software and the best were the ones that just loved it. The rest where people working in corporate farms that still hadn't got laid off from the dotcom burst.

Then the startup boom started again and a new round of devs came into the profession looking to make money and needed to prove themselves. They heard "good devs have great github profiles!" so they started creating github profiles and flooding them with code to get hired.

Now we're in the new wave of cargo cult white boarding/code challenges. You may hate them but at least the early FAANG challenges were interesting and challenging. But since every company has to have them you have devs that couldn't pass a google white board challenge trying to create their own version... which usually means they are poorly thought out.

A better way to look at OSS is like charity. Giving to charity is not something you do to enrich yourself, although you will likely enrich yourself in certain ways. You don't give to charity so you can get a better job in the future. And the results of giving to a charity might result in profits for corporations: suppose the charity makes a breakthrough in medicine and a large medical research company uses that to leapfrog their R&D. That's a feature, not a bug. We shouldn't discourage people from giving to charity because of these aspects.

It's worth noting that many of these companies give back to the OSS they leveraged, and are often the genesis of some of the best OSS software themselves.

> Some other commenters in this thread have pointed out that nobody is forcing open source developers to work on the contributions that they make, and that is strictly speaking true. But in the culture of software development, there does seem to be an ambient message often repeated or hinted at, that it is good to "give back" to open source by contributing and that it is virtuous to do so.

I can’t see anything wrong with that, it’s like people doing favors every day for other people without asking for money

> I now am starting to rethink this sentiment because the vast majority of the benefit of open source contributions on github, be they to languages, runtimes, application frameworks, databases, etc, go towards increasing the bottom line of for-profit companies, not to developers. And the vast majority of the beneficiaries of open source will never support the project even when they are fortune 50 companies saving millions of dollars by using the work of one volunteer.

Yes open-source improve computation as a whole, companies as well as individuals

> There is also the idea that contributing will be great for your career development. I have found that not to be the case at all, I think that no potential or current employer has ever given a rat's ass about open source contributions, and do not consider that work as valuable when making hiring decisions. The work you are paid to do is the only thing anyone cares about. I'm not saying that that shouldn't be the case, but just that that is the case.

Not my experience, have never stumbled upon a recruiter that doesn’t value OS contributions

> Given then, that there's very little upside to doing open source, and most/all the benefits go to profit-making corporations, it is puzzling why do we even push for greater involvement in open source at all? It seems we shouldn't be, we should be warning people who want to contribute to Open Source that they should probably spend their time doing their own studying and personal/skill development which will allow them to succeed in the roles that they have with their current role or a role they'd like to obtain one day, for money.

Since all previous assumptios was wrong, conclusion doesn’t follow. But I agree that people should not do Open Source for money

I agree with what you are saying here. The underlying problem here is that companies (being emotionless optimization machines) aren't willing to shell out for the benefiet they are getting from open source, because people are willing to give it out for free. The only way this will get resolved is if open source contributors as a whole stop contributing leaving companies to either have to support open source in some way, or go back to developing in house. Most likely the later as it has the benefiet of retaining IP and competitive edge. What this says is that most likely open source will start to go away in the years following if corporate entities don't start sponsoring it more. Maybe some type of subscription model like academic journals, but we all know how much people like that.

Also an angle you are considering are open source projects like Spark maintained by a company (databricks) for the benefiet of being used as a sales tool. This whole class of open source exists and pulls unwitting contributors into helping out who don't realize that they are benefieting multiple layers of corporate entities not willing to pay for it. And they will not see any benefiet from it (unless they use it as a personal sales tool to get a job at those companies).

> In the culture of software development, there does seem to be an ambient message often repeated or hinted at, that it is good to "give back" to open source by contributing and that it is virtuous to do so.

I can't think of any civic institution I belong to that isn't constantly asking for volunteers, and riding the volunteers they have to do more. I'm not sure it's any different.

Society needs volunteers, and money isn't everything.

> Society needs volunteers, and money isn't everything.

This is a great point. Still, the local church meals do not enrich multi-billion dollar companies; unlike FOSS.

Though I now cringe at the angsty teen outrage dripping from every word of this post, it's hard to believe that I wrote this on AskSlashdot 19 years ago:

https://yro.slashdot.org/story/00/01/22/1843258/open-defensi...

and still do not know whether any formal method exists (apart from the choice of license model) which could inoculate "Open Source" from being tragically impressed into corporate rent-seeking service. Or at least, as much of the Open Source ethic as is encapsulated in the public works of ethically motivated open source contributors; people who are essentially investing their time, skills and resources in the improvement of a lightly defended public common that continues to be parceled out to private interests.

* rereading the above I don't think I've made much progress in my posting style. Maybe it's something about the apparent structural inevitability of this problem, which seems more a recapitulation of deeper conflicts in human dynamics than something specific to "open source".

I have found the exact opposite actually.

Firstly, the reason I open-source most of my libraries is because, when I was a kid and learning to code, I couldn't afford or convince my parents to buy the expensive proprietary software I wanted to play with. Open source let me have access to this technology ( and to inspect how it worked ) without having to buy anything.

This is simple stuff like having GCC to compile C programs, rather than buying the intel or MS compiler.

Those early experiences got me where I am today. I contribute open-source software so that other, more junior developers have the same opportunities I did.

Again, I've never experienced this. I got my first programming job right out of high school, and my resume was filled only with various (small) open-source contributions. Without those I'm not sure what inspectable experience I would have had.

Again.... the exact opposite experience. I released a moderately popular library, built a community around it, and this has opened up career opportunities for me. In the niche I work in, several interviewers had actually heard of me before I even interviewed. The scalability of open-source to distribute a positive first impression to potential employers should not be understated.

I do agree with you here:

> And the vast majority of the beneficiaries of open source will never support the project even when they are fortune 50 companies saving millions of dollars by using the work of one volunteer.

Lastly, I think the reason most people find open-source to not live up to its promises is that they do not take on leadership roles in open source organizations. Frankly, as a more senior developer now, simply being a casual contributor to a project would not add to my resume. However, being a project lead, or component lead, would. You need to target your open-source contributions to things your employers would find useful. For example, I work on mainly backend, low-level things in my current career. As much as I would like to contribute to GNOME or GTK or whatever (and I have the skills to do it), there is not enough reward to doing so, so I don't. Instead I contribute to projects that I know will ingratiate myself to future employers. Perhaps that's why our experiences are so different.

>Firstly, the reason I open-source most of my libraries is because, when I was a kid and learning to code, I couldn't afford or convince my parents to buy the expensive proprietary software I wanted to play with

Just saying, software can be free without getting open sourced.

I agree that OSS gives raw access to technology, but in larger projects, I would prioritize a good documentation/reference book over source code to understand it in a complete manner.

I don't disagree, but for young tathougies, having access to the source code was vital. It's like the difference between reading about scientific success via articles in Nature and doing the research yourself. The former is not a substitute for the latter, when it comes to learning.

Do you remember being a developer in the 90s? It was horrible and you were stuck in the confines of truly for-profit companies and for-profit platforms and paradigms.

One benefit of OSS is to lead the world to a place where you want the world to go.

Having started to code during the 80's, what was so horrible about the 90's?

I had access to Amiga and PC, bough my software tools that I actually cared about (with student discount), and did some programming gigs already before getting into university.

Back then, making a living from selling developer tools still looked like a possible way of life.

The upside is the satisfaction of solving sometimes interesting problems. Some people do crossword puzzles, some play video games. A rare few write operating systems used by perhaps billions of people.

> Given then, that there's very little upside to doing open source, and most/all the benefits go to profit-making corporations

Such a wild claim would require some evidence. But it's so patently false, I'll give you pass.

Linus has gotten extreme benefit from Linux and Git. And developers have massively, directly benefited and the world indirectly from Linux and Git.

Wikipedia, directly benefiting everyone, would not exist without open-source either its infrastructure or the code it runs on. The Internet would not exist without open-source.

> It seems we shouldn't be, we should be warning people who want to contribute to Open Source that they should probably spend their time doing their own studying and personal/skill development which will allow them to succeed in the roles that they have with their current role or a role they'd like to obtain one day, for money.

Money is only a fraction of motivation for many people. Many do open source to make the world a better or more efficient place and reduce barriers to entry.

>> the vast majority of the benefit of open source contributions on github, be they to languages, runtimes, application frameworks, databases, etc, go towards increasing the bottom line of for-profit companies, not to developers.

Which inturn creates jobs doe devlopers.

Let's face it. As a devloper. My day to day would be tough and slow without many of the free tools devlopers created for fee.

So maybe the pressure for me to contribute is just.

The problem I see is, that companies make money by using OS, but are often very hesitant to commit to the code and instead bug the OS developers to fix these problems. In some rare cases only a lead developer of an OS project can do that without a fork.

OS matured, contribution in 2019 does not necessarily mean you commit code. Donating some money is fair use in my point of view and is simply a sign of respect.

> I think that no potential or current employer has ever given a rat's ass about open source contributions, and do not consider that work as valuable when making hiring decisions.

We do care. We're hiring for backend engineers. https://news.ycombinator.com/item?id=20871478

I hope you take it with a grain of salt. I've developed huge systems for my employers that aren't open source and my own open source contributions are nothing like my work (by design).

What helped me understand this better was thinking about writing software as a subset of all types of writing. People do write for free because they enjoy doing so or because they feel like they can advance knowledge in a particular area. This is true for both prose and also for the software medium. In both software and prose, you can gain valuable contacts and social credit by sharing your writing with others even if you don't make as much direct income from it.

The publishing industry is much better than software at getting the money to flow towards the creator, but software jobs overall pay better than most writers because of the skill sets involved. Both publishing and software, however, do have multiple ways of going from amateur enthusiast to highly paid professional if you have the talent and connections. And open software, like contributing to small literary magazines, is one way that you can make that transition.

programmers love - and i do mean love - to program, so it's not surprising that they're willing to give away their labors of love for free.

hell, most programmers would probably take a pay cut to work in a "cool" language. i'm willing to wager that C/C++/COBOL/java mean salaries are a standard deviation higher than haskell/elixir/clojure salaries. why? many reasons but i bet that companies know that they can use programmers' passion for mastery against them. leave your boring C++ slog job to "hack in Haskell!"

i'm guessing many of these guys are just clueless when it comes to how potentially valuable they are as individuals and what that's worth, while they put up a $10/mo patreon that 6 people use. i'm guessing most engineers are severely underpaid when it comes to value added, they consistently fall for the okie-doke's of "we get to type stuff into a computer and learn and oh wow, i ought to be so grateful for my six figure salary", yet they're putting entire swathes of the economy out of work for their corporate overlords.

it's tough to convince management you're worth a lot of money if you're willing to give away your precious time and resources freely.

According to the Stack Overflow Developer Survey[0] your intuition is wrong. C, C++, Java devs are paid less than average and F#, Clojure, Elixir devs above average. COBOL and Haskell didn't rank in the survey.

I 100% agree most engineers are severely underpaid. A lot of my colleagues get paid more than anyone in their family ever has (sometimes whole family put together) and just like to ride it out.

[0] https://insights.stackoverflow.com/survey/2019#salary

fascinating, thanks for this link. i stand corrected.

Isn't there a disconnect here?

> "the benefit of open source contributions on github... go towards increasing the bottom line of for-profit companies, not to developers"

Don't those for-profit companies employ people -- including developers?

>"[developers] should... spend their time doing... skill development... to succeed in... current role or a role they'd like... for money"

Are those roles you mention positions in for-profit companies? In "for money", whose money?

I mean, they perhaps should focus instead on absolutely nailing it in their position, or acquiring skills to do so, or to study to find a better position, rather than trying to benefit random companies they have nothing to do with, because if you hit it out of the park at work that will actually be rewarded, generally, whereas there is no reason to expect open source work to pay off.

As much as I dislike licensing zealotry, this is where I think the BSD guys have it dead wrong. The BSD license strictly enables corporate parasitism by dis-incentivising upstream contributions and source-sharing. If the pride in suspecting that one's work is widely used is payment enough for a person, then so be it. But I've never understood why subsidizing corporations is a point of pride for those folks.

one can make an argument that it depends on who is the copyright holder. if it is an individual or a non-profit entity, then the GPL is beneficial. but if it is a corporation, then a BSD license is preferable for others because it allows competitors to that corporation to compete on equal terms.

so to turn the argument around BSD unix would not exist in its current form if the initial corporate developers had not released it under a BSD compatible license.

so it's not just about volunteer BSD developers subsidizing corporations. but there is a give and take in the BSD world too.

compare that to mariadb which contains most of the code that is owned and controlled by oracle now and which prevents mariadb to sell to clients that demand a custom license agreement.

now for a GPL fan, this is irrelevant because they shouldn't be selling such licenses anyways, but fact is that the owner of a GPL product is the only one who can sell such licenses, whereas anyone who uses a BSD product can sell such licenses.

in this sense, BSD code ownership can not be sold because actual ownership of the code is pretty much irrelevant, whereas GPL code ownership can be sold because the owner always has an advantage.

But then, to avoid that, you can embrace the concept of copyleft (free software) instead of open source.

So, you can make your software free as in 'free speech', and obligate other companies to keep the products they construct above yours still free.

Better than to make them free as in 'free beer' where they can just consume your work and not give back to the overall movement.

> the vast majority of the benefit ... go towards increasing the bottom line of for-profit companies, not to developers

The problem is in the licensing.

Large companies have become the primary user rather than the real end users and other developers... because the licenses make it possible.

It benefits other developers by making their lives easier: no management purchase decision hassle, can inspect source, fix bugs, customize, etc all the standard open-source stuff.

It's just not financially beneficial to developers.

>we should be warning people who want to contribute to Open Source that they should probably spend their time doing their own studying and personal/skill ...

Its Open Source that is chiefly responsible for every bit of that personal skill I have.

In this game, the "giant" corporations are the irrelevant little players that hardly matter. They don't contribute, so they may as well not exist. They can be safely ignored except for the rare occasion that they "fart in our elevator" with their patent nonsense.

A lot of the FAANG and tech companies contribute to open source projects.

Honestly not everything is business-centric—if I’m working on something for pure love of it, that’s going to turn up the best quality I can muster. Those big companies run by business people usually just thinking about dollar signs. Just another shitty side effect of capitalism.

this is the core reason for free software

This. I use permissive licenses mostly, but if you're disgruntled about companies ripping off your work without ever giving back, it could be that you've used the wrong license.

Reply on news.ycombinator.com