Hacker News Reader

Comment by captainmuon | Hacker News Reader

This is another case of tragedy of the commons. Individually, it doesn't make sense for a company to donate. I can't even resent them, since it puts the company at a (slight) disadvantage. Unless I was flush with cash, I probably wouldn't donate from company funds if I were an entrepreneur. But collectively, we would all benefit to fund something like this. (Edit: I like culturestate's response of budgeting a small amount for this. Wish more people would do that!)

For me, another data point why our current economic system is not optimal, and we need at least more coordination, or maybe a better system.

View on HN

> Unless I was flush with cash, I probably wouldn't donate from company funds if I were an entrepreneur.

I am a solo entrepreneur with a self-funded business and I set aside a percentage of revenue to be paid in developer sponsorships, through Github sponsors, Clojurists Together, and directly. I sort the libraries that my code uses by "indispensability" and fund the top 8 or so. It's not much, but if everybody did this, the ecosystem would be healthy.

I'm happy to see that at least the Clojure community is very mature and this approach is rather common.

It is in my best interest and it does make sense: I rely on those libraries, I want them to be maintained. And if they get abandoned, I will bear the very significant cost of rewrites.

The problem is if you're competing with someone else, all things being equal, that invests the money you give away they now have a better shot at beating you. From a purely selfish perspective it's better to allow people like you to fund it while everyone else gets it for free.

Bear in mind I do agree with you, it would be better if we supported these libraries properly. But it seems like the majority are willing to allow a few to carry them. It makes the idea of running these projects on sponsorships seem unsustainable.

First, this isn't a zero-sum game.

And second, seriously? I should worry that my competitors will be meaningfully boosted ahead (to the point of having a better shot a beating me) by me supporting a bunch of libraries that I use?

The world would be so much better if we all just did the right thing. It really isn't difficult to know what the right thing is at any given moment.

I agree in general with your comment, but even in your example, is donating to exactly the top 8 clearly the right thing? Why not 5, or 13, or numbers 6-15 (figuring that 1-5 are adequately funded by others)?

How many do you donate to?

I give what I can, because contributing any amount to any libs is better than just talking and contributing zero.

> Unless I was flush with cash, I probably wouldn't donate from company funds if I were an entrepreneur.

I specifically budget a small amount (~$100/mo) of company cash for this purpose. Some of it's donated to OSS, and some if it goes to pay for low-cost freemium products where we don't necessarily need the premium features but want to support the team[1].

I stole this idea from someone else so I know I'm not the only one, but more entrepreneurs really should do it.

1. https://mailsac.com/ is a good example where our usage would fit fine in the free tier but we subscribe anyway.

I’d classify it as a public goods problem. The difference is rivalry in consumption. You compete over fishing grounds but you don’t compete over use of Babel. It makes a difference in user strategy too: with the tragedy of the commons, your personal marginal gain is higher than the individual share of the social cost. This leads to overconsumption. In a public goods problem, you fully benefit without contributing. As a result, you maximize utility by not contributing. There is also an incentive to understate the the utility you derive from the public good in order to avoid being forced to pay, making it hard to estimate the utility generated by the public good. It’s also called the freerider problem.

Thanks for your analysis but isn't the share of the social cost the fact that Babel will stop updating when these people can't support themselves? It's just delayed in this case, no?

When something is too obviously a good thing, there is a game of 'chicken' where no one feels the need to contribute because the item is so obviously advantageous they are sure someone else will.

At the extreme, this is the opposite of a normal public good, which are characterised by a cost above any individual's marginal utility. Here, there are any number of organisations who probably have marginal utility from Babel above the cost of provision. See the snowdrift dilemma.

Prisoner's dilemma or tragedy of commons has a dominant strategy (defect), but chicken does not - your best strategy is to do the opposite to the other.

The first woman to win a Nobel Prize in economics worked on studying systems where the tragedy of the commons does not hold and determining under which conditions[1] such arrangements do work. Adopting open source to that system seems like a very difficult task to me though.

1. https://en.wikipedia.org/wiki/Elinor_Ostrom#Design_principle...

While I agree with your general point, I still think crowd sourcing open source projects is still doable. All we need is to centralize it.

I work for a very small startup that uses hundreds of big or small open source projects. And we never pay a penny for any of them.

And I don't think I can go to my boss and ask for a meaningful contributions to any of these. We are not profitable and there's no reward for that contribution.

However, if there was a way for me to subscribe to micro payments for a few projects, given some easy trivial incentive, we definitely would've.

I can easily subscribe to $5/m payments to like 5-10 projects that we use. Would've meant nothing for us. But economics of scale would've kicked in for them.

Allow $1-10 monthly subscriptions, give the backers a badge or something they can use when hiring. Maybe show a badge next to their usernames on github or whatever. Maybe show a badge to the issues the backers create so they get a sense of priority. slowly you can build a small habit of people/companies subscribing small amounts to projects they use.

If there was such small incentives I would've immediately subscribed $5/m to following project:

* Nodejs

* Postgres

* Node-Pg

* Rabbit MQ

* Grapesjs

* Esbuild

* Webpack (Which I badly need to improve it's killing us everyday)

* Dokku

* Metabase

* MJML

It's in my company's benefit for each of these to survive/thrive and keep improving. I am afraid that once/if some of them stop that'll hurt us. $5/m and potentially getting some attention from maintainers (in smaller ones) if we had issues would be enough of an incentive for these all.

The incentives I suggest:

* Badges on Issue tracking on the issues created by backers

* Badges next to usernames on Slack/Discord/Irc/Whatever

* Badges on Github profiles for backers

An example: Docker. They started rate limiting their free tier, asking for $5/m. I don't care about the $5/m at all. My biggest problem with that change was not the $5/m, but the time I had to spend on our different tools setting up auth environment variables.

> If there was such small incentives I would've immediately subscribed $5/m to following project: Nodejs, Postgres, Node-Pg, [...]

Who would receive that money? Only the core developers? Anybody with "commit access"? In which share? Will it detract voluntary contributions? How should other companies working on the same code-base be treated? Are they allowed to contribute at all? Do they get some shares of the funds?

All these questions (and many others) that have been discussed in the Debian project for a long time and nobody has yet found an answer that is acceptable to everybody.

Some sources:

* "Debian + Dunc-Tank.org = dissent" (2006) https://lwn.net/Articles/201488/

* "Paying (some) Debian developers" (2019) https://lwn.net/Articles/790954/

* "DPL: Debian project has plenty of money but not enough developers" (2020) https://www.theregister.com/2020/09/10/debian_project_addres...

Debian's case is so complicated. Many of these complexities and dynamics don't exist in other software projects.

Most of the project's I'd support have central figures that can use the money and don't have such difficulties. Babel could've used my $5/m donations. Debian's theoretical and moral dilemmas are not universal.

I used to be involved in a similar project for KDE [0] and it was KDE e.V that received the money. How they allocated their money is an entirely different question that depends on each one.

[0] https://relate.kde.org/

> * Badges on Issue tracking on the issues created by backers

This can be a slippery slope. It could easily be mis-interpreted as "Hey, I payed $3 last month so I want you to prioritize my bugs". Generally I think for these kind of sums it should be clear that you donate and pay as a thank-you for what's already there, not for what (in your opinion) should come. Using it to highlight issues would work against that IMHO.

Your last sentence is worth emphasizing. Many people misunderstand the tragedy of the commons as an argument why a commons can't work, but the only possible conclusion is that the commons can't work under the current economic system (and specifically the dominant school of economic thought of the past half century).

The usual example given is a pasture shared by shepherds who are each going to optimize their use of the pasture by getting more sheep even if it means exceeding the sustainable capacity of the pasture, eventually killing it for everyone. Yet the real world shows that communal pastures have existed for centuries in rural communities just fine. The difference is that these pastures aren't used "optimally" and the shepherds aren't economically "rational actors" within the current economic system, though they are increasingly being pressured into that by the dominant system surrounding them.

The tragedy of the commons is not a warning about the unsustainability of a commons, it's a warning about the logical consequence of our dominant economic system.

That's mistaken in two ways.

First, the tragedy of the commons - aka the public goods problem - is an argument why a commons can't work, full stop. It's precisely mathematically specified. If the conclusion doesn't hold, that's because the assumptions didn't hold. You can certainly argue that the assumptions don't hold in many places, e.g. among Maine lobster gangs or in the woods around Japanese villages, and indeed that's what various authors have done. They have not argued - and it would make little sense - that a different economic system would magically make public goods games have a different equilibrium.

Second, you're mistaken in assuming that actors in successfully maintained commons are economically irrational. One simple interpretation of these commons is that they are repeated games. In such games it is rational to cooperate to preserve the commons, since defection leads to others defecting in future. That is standard economic rationality. Elinor Ostrom put forward that interpretation sometimes. She also emphasized that communities used things like punishments and fines to sanction uncooperative behaviour - again, these mechanisms work for standard homo economicus reasons.

The economic system is literally part of the assumptions.

The tragedy of the commons holds up under laissez-faire capitalism. It can be somewhat mitigated under other forms of capitalism but there are examples for commons that pre-date capitalism itself and also specifically enclosure (the formal privatization of common land), which itself pre-dates capitalism.

I'm not sure why you think you need to defend the concept of homo economicus. You're literally refuting your own first point with the second: a commons can work, but where it has worked it has usually come with constraints that run contrary to laissez-faire/free-market capitalism (or the approximation thereof) and that don't necessarily translate to the dynamics between globe-spanning corporations compared to that between a much smaller set of individuals.

No, the logic of public goods does not require any assumption about the economic system. Any set of actors whose actions create a negative externality, and who cannot enforce agreements to cooperate (either through the courts or via repeated interactions) will overproduce the externality. So, you could say the assumptions include the legal system. There are numerous tragedies of the commons under socialist economic systems, including environmental degradation and the inefficiency of socialist firms. There are plenty under capitalism too. Berlin under Rhineland capitalism has plenty of litter, just as New York does under Anglo-American capitalism. Hunter-gatherering bands are also perfectly capable of overusing natural resources and failing to cooperate. When they succeed, it's because they get the incentives right - at least, there is a reasonable case for that.[1]

I defend homo economicus to dispel a false impression: that successfully sustained commons refute the idea that humans (a) are self-interested and (b) make optimal choices. You could say that laboratory public goods games do that, but even there you'd have to qualify your argument very carefully.

[1] Guala, Francesco, 2012. Reciprocity: Weak or strong? What punishment experiments do (and do not) demonstrate. Behavioral and brain sciences, 35(1), pp.1-15.

I'm not sure who you think you're arguing with but so far most of your rebuttals attack points I never made.

I never said that humans aren't "self-interested" or that humans, in the abstract, can be described as acting optimally within given constraints.

I said that the constraints influence behavior (which you agree with) and I said that the current economic system results in a tragedy of commons (which you also agree with). I also said that it is also possible to avoid the tragedy of the commons (at least locally, for a given commons), which is demonstrated by successful commons existing and having existed under different constraints.

None of that seems to be anything you disagree with, yet here we are.

Note that both examples you've given — rentier capitalism in Berlin and New York, and (presumably) Soviet-style bureaucratic socialism — suffer massively from alienation, which demonstrably leads to inefficiency and disregard (e.g. littering). This is well-understood enough that creating a feeling of "ownership" or "autonomy" (without introducing monetary benefits or democratic sharing of actual control over the company) has become a management strategy to motivate employees under capitalism.

As a counter-example to Berlin under capitalism in Germany, consider affluent rural areas where you see barely any littering. These regions however also have a high rate of home ownership and parallel social support networks, which I assume is what you're gesturing at as "repeated interactions" (but in this case often spanning literally generations). But these systems are unstable under laissez-faire capitalism as they present an inefficient resource use in the greater economic system and will over time erode as residents move away for job opportunities and properties are sold off and reused for more space-efficient apartment buildings.

To summarize, I'm not saying a tragedy of the commons can not exist under other economic systems, I'm saying there are systems under which a given tragedy of the commons can be averted and that this is not the case for the current one.

OK, fine. But this is also false. There are many tragedies of the commons that are successfully avoided under capitalism. As you point out, affluent rural areas are one, and the claim that in future these are all automatically going to die off is a post-dated cheque. (Villages in the UK stockbrocker belt have little litter, despite being composed of wealthy incomers.) Another example is cartels. In the end, what's left apart from the uninteresting claim that "constraints influence behaviour"?

> the claim that in future these are all automatically going to die off is a post-dated cheque

It's not, though. We've literally seen this development happen over and over again and it's a logical consequence of the system as is the increasing monopolization via acquisitions and buyouts.

We're not in purely laissez-faire capitalist system, certainly not globally, but trying to paint present day Berlin as an example of a social market economy in the "Rhine capitalism"[0] spirit shows a complete unfamiliarity with the practical evolution of the German economic system and welfare state during the later part of the 20th century and the early 2000s, which saw a massive shift towards free market capitalism, especially under (ironically) the SPD-led Schröder government.

> Villages in the UK stockbrocker belt have little litter, despite being composed of wealthy incomers.

This isn't a counter-example to what I said, this is an example of what I said. Income isn't the determining factor, ownership (and legacy) is. Additionally, even if they don't expect to live there forever because they're comfortable relocating to follow the money, they're surrounded by potential business partners. This again isn't about their level of income but it creates a social mechanism that incentivizes displaying "good behavior".

> Another example is cartels.

An example for something avoided under capitalism? Certainly not. This is like saying Western countries don't have corruption because money is used to exert influence by signing entirely legal backroom deals (campaign financing in the US, promises of zero-effort board positions and highly paid "speaker" engagements in Germany) instead of literally handing wads of cash to police officers.

And this isn't even to cede that cartels can't happen under capitalism, let alone under laissez-faire capitalism: they're not prohibited by the market but by regulations (i.e. laws), and even then there are always loopholes to enable corporations to engage in functionally identical behavior as long as they follow a sufficient number of layers of indirection (or are comfortable to use a pile of NDAs to hide actual illegal behavior). A recent US example widely discussed on HN was the "no poaching" agreements between large industry players to put a cap on the salaries of highly desired employees.

[0]: The term is "Rhine capitalism" not "Rhineland capitalism", btw, and it's a gross oversimplification to use it to refer to the post-WW2 German model, the economy of France in the 1990s and the economy of Germany after the Agenda 2010 reforms as part of the EU-wide Lisbon agenda -- which was one of contributing factors to the EU debt crisis, drastically raised the poverty rate in Germany, drastically cut the income tax rate and wreaked havoc in the welfare system. It was a stretch when Albert popularized the term in the 1990s (after the German reunification) and it's frankly ridiculous to still use it today.

I think you might be interested in my comment sibling to yours about the work of Elinor Ostrom. She spent her life documenting those systems.

Capitalism represents a solution to the tragedy of the commons: Privatize it. You guys are interpreting this as a takedown of capitalism when it's the exact opposite. Capitalism represents a solution to this issue. You might not like that solution, but this circumstance just goes to show that naive attempts to reinvent capitalism are unlikely to work.

Contrarily to the common claim of capitalism, this wouldn’t be an optimal solution. Let’s say there are ten shepherds sharing the fields. You privatize it and in a short amount of time, there will be one shepherd left, ultra rich, and nine unemployed shepherds. So capitalism is a good system to optimize the value of the field, not its value for society.

If that actually happens, it will be because one of the shepherds is substantially better at flock management than the others, and their financial success allows them to buy fields from the other shepherds. The other shepherds will only sell their field if the good-flock-manager shepherd makes them an attractive offer (probably enough money such that they can live off the interest for the rest of their life). Not to mention, they very well might get a job working for the good-at-flock-management shepherd, possibly even at a raise since superior profits means there's more wealth to share. (Realistically there will probably be at least 2 sheep conglomerates competing on wages to hire the best shepherds. I'm not an ideologue, I'm cool with the government intervening in the economy a little bit to break up monopolies.)

> If that actually happens, it will be because one of the shepherds is substantially better at flock management than the others, and their financial success allows them to buy fields from the other shepherds.

IFF all the other shepherds even got a cut in the privatisation, then perhaps.

But there's no guarantee of that. That's certainly not how things went down in the last major privatisation the world has seen, that of post-Soviet Russia in the 1990s.

In some cases that applies, but I don't think a paid version of babel would be viable. Electronic goods resist commodifying. It only works sometimes (like in streaming music), but you need to put in a lot of technological and legal effort.

I believe that societies develop in lockstep with technology. The industrial revolution was only possible with the beginning civil society and capitalism; capitalism as we know it was only possible at the end of feudalism and absolutism. It is intertwined. New technologies tend to disrupt societies, you can go where the tech leads you (e.g. content can be copied without effort), or you can go the other way and adapt the tech to the society (say, not using machines in order to save jobs). That is why I think we are due for a big change in our political and economic system. What that is going to look like, and how much it will still look like capitalism, I don't know.

(Sorry for derailing the thread...)

>I don't think a paid version of babel would be viable.

It would if there were no open source alternatives. Bear in mind that most of the users are commercial organisations deriving value from it. They need something like it to solve their problems and if it isn't available for free then they will pay.

But you can't privatize the air?

Also, do you think the indian company was a perfect example of why capitalism should not involve itself with common goods? Unless you think it isnt a capitalist company despite presenting all aspect of it (including regulatory capture and threats of violence against slaves/employees).

Hm, a solution. They privatize it, and then companies will say to their webdevs, we don't want to pay extra for a profit of some capitalist, write the website without it.

So each webdev will end up with their own half-baked solution. Tragedy of the lack of commons!

I hear you cry, but that's irrational! Precisely, but so is not caring about the commons and letting them die.

I am not under an illusion that under capitalism, people cannot pass the buck (responsibility for maintenance) away too.

Companies make those sort of trade offs all the time, whether to buy or build. It doesn't take many to decide to buy to fund a small project like Babel. Not only that but you are likely to see competition driving the whole thing forward.

My point was that framing this as a tragedy of the commons problem, and here's the standard solution, can itself be disputed. The hidden assumption is that the capitalist will take the responsibility to take care of the thing, and provide it for a fair price, but it is by no means given. Certainly, there are examples of economic failures of this type as well. (Even in software, Unixes/Lisps..)

How big of a problem is it to change from babel to some alternative? I'm thinking of a possibly bigger disadvantage should that need to happen. I'm genuinely asking, since I'm not familiar with js tool chains.

I successfully removed it years ago because Typescript replaced everything I was using Babel for, things even build way faster and I even got Intellisense autocompletion and refactoring as bonus.

I doubt typescript is faster than babel with all the static analysis it does. In fact it's typical for projects to use typescript as a check in CI and actually build with babel, for speed. Fast alternatives to babel are things like esbuild and swc

Typechecks are optional, just use transpile-only mode for the build: it merely removes the type annotations, it's fast.

You can have the best of both worlds: Intellisense based on types in the editor yet fast transpile in the build.

True, I guess if a big company used it they probably could fund it from pocket money, rather than spending more money on switching. Sad that people have to write a "we're running out of money" blog post to make that happen.

There is also a separate aspect, that this also has do with unawareness, and not just excess stinginess. I might speak for only myself, but to a much larger extent that I'd like to admit, I take work and maintenance of open source projects for granted.

Maybe it's the disconnect caused by large businesses properly funding some of them, with an erroneous assumption of the level of maintenance others require. If something is core to a major tool chain then I just, sad to say, assume it's well funded, either by larger companies or universities.

Trying to think of an example, LLVM seems integral to a lot of progress on web functionality, just EMSCripten alone makes that case. This in turn is sponsored by most of the major tech companies [1]. Whether the same foundation also funds emscripten, I couldn't find out.

https://foundation.llvm.org/docs/sponsors/

I think my point is, if anything, if you have a open source project that requires having several full times employees, it might be high time to create a non/profit and work on getting funding. The post they wrote, and the visibility it gets can be an example of a doing just that, albeit maybe just a temporary fix.

I was just thinking how messed up it is that basically companies can take those expenses off their taxes and they still don't contribute, but then I was thinking - do they think that if there was a healthy ecosystem then they would create problems to employ people because the healthy open source community is then another competitor for top quality talent.

The inventors of our economic system didn't conceive of people giving away valuable stuff for free.

Our economic systems weren’t invented. They’re all hacks on top of hacks on top of hacks. Every economic theorist of note is well aware of that and they always have been. Everyone who has tried to build a new complex system from scratch has failed. Usually they’ve failed completely but occasionally they’ve gotten something functional at the end, but not what they designed.

I’m reasonably confident every economist knows people give stuff away for free but it’s certainly been true since Adam Smith, whose Theory of the Moral Sentiments is chock full of psychological and philosophical insight.

> Everyone who has tried to build a new complex system from scratch has failed.

Has anyone tried to design an economic system recently, based on tools from the information age?

China would probably be the best case study available, but they are of course notoriously difficult to get reliable data about. DeFi (Ethereum Decentralized Finance) would be the next best case study but finding someone capable of wading through all of the scams to find the actual projects is going to be almost as hard as getting reliable data out of China.

In this case the problem would actually be solved if the project was a bit more capitalist and charged money for usage, no? This isn't capitalism failing, it's Burning Man gift economy type stuff failing.

I never understand the reluctance to charge for software amongst this community. Especially when you consider projects like Babel are mostly being used in commercial environments. Why do people expect that their employer should get this stuff for free or that the developer should have to beg for a few crumbs at the end of the month.

I really like the model where tools are free for educational, open-source, and other non-commercial uses. I've ended up paying for software libraries because I was first able to try them out for free in school and other personal projects.

One issue with fixed pricing or licensing is that it can lock out people in developing countries from participating. A few hundred dollars is nothing for a San Francisco based company but it might be a huge amount for someone in Latin America. How many great ideas fail to take off due to insufficient access to funding?

In recent years game engines have been doing an interesting job at exploring this space. You can use stuff like Unity or Unreal and the price depends on how much money you're making. I think something similar is done in the mobile app store ecosystems as well.

For many OSS is a breath of fresh air. A project where you're in charge, free to set your own goals and deadlines. Free to tinker and experiment. Free to say "do it yourself" when someone gets a wrong idea and goes against your creative vision. Money create exactly the kind of obligations you're running away from. Even if it's not demanded, you're still feel pressure to provide support, make your code fit for some purpose. And for what? Some chump change?

> Why do people expect that their employer should get this stuff for free

Because it's offered for free.

> developer should have to beg for a few crumbs

They shouldn't. Developers unhappy with their financial situation should pick up some contractual work or offer their product as SaaS.

Reply on news.ycombinator.com