The Value of Open Source Software(papers.ssrn.com)
papers.ssrn.com
The Value of Open Source Software
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4693148
25 comments
We then calculate the demand side value based on a replacement value for each firm that uses the software and would need to build it internally if OSS did not exist.
Presumably in the absence of OSS, companies would write far simpler software than the OSS they presently use. What fraction of the code in the Linux kernel do most users ever run?
Presumably in the absence of OSS, companies would write far simpler software than the OSS they presently use. What fraction of the code in the Linux kernel do most users ever run?
What leads you to presume company devs would be able to pull this off? Go write your own queuing solution from the ground up in real life. You'll learn it's non-trivial to cover all necessary edge cases while simultaneously avoiding tight/brittle coupling.
Insane amount of work compared to dropping in battle-tested (or even modestly used) OSS, with stackoverflow/gpt solutions at your fingertips for nearly anything imaginable.
Most companies core competency is in delivering business value, which means shipping useable products and features. This most often boils down to building atop a thick stack of open-source technologies (e.g. Dropwizard, Kafka, Postgres, MySQL, Opensearch/License, the vast Java and Golang ecosystems, Python, Django, RoR, PHP, and on and on). Without that, devs would need to spend $$$ licensing privately owned solutions, or do a ton of work developing homegrown spaghetti frameworks and platforms from scratch before even being able to start considering addressing core functional use cases. This would be very bad for startups with limited runway to get off the ground.
What can be in days to weeks through OSS resources would take months to years, or costly licensing.
Most companies core competency is in delivering business value, which means shipping useable products and features. This most often boils down to building atop a thick stack of open-source technologies (e.g. Dropwizard, Kafka, Postgres, MySQL, Opensearch/License, the vast Java and Golang ecosystems, Python, Django, RoR, PHP, and on and on). Without that, devs would need to spend $$$ licensing privately owned solutions, or do a ton of work developing homegrown spaghetti frameworks and platforms from scratch before even being able to start considering addressing core functional use cases. This would be very bad for startups with limited runway to get off the ground.
What can be in days to weeks through OSS resources would take months to years, or costly licensing.
> Presumably in the absence of OSS, companies would write far simpler software than the OSS they presently use.
Simpler, but also worse. Part of the beauty of sharing common code is that you only need to prevent a problem once and then it can be fixed for everyone. If you write it in-house, it's easy to say "oh it's fine, nobody will ever attack this, it'll only be used under boring conditions, we'll never hit those edge cases" which is great right up until the day when it suddenly becomes very painfully untrue.
Simpler, but also worse. Part of the beauty of sharing common code is that you only need to prevent a problem once and then it can be fixed for everyone. If you write it in-house, it's easy to say "oh it's fine, nobody will ever attack this, it'll only be used under boring conditions, we'll never hit those edge cases" which is great right up until the day when it suddenly becomes very painfully untrue.
Simpler isn't necessarily worse. Having fewer features tends to make software more reliable and secure.
I don't think you can generalize it like that. It really depends whether the complexity is necessary or not, and that's context-dependent.
If the implementation is too simple, it may have issues that impact its reliability. For example, code may be using simple linear search instead of fancier data structures/algorithms, and have issues with handling larger inputs or have accidentally-quadratic code paths. Or may updating data files without the complexity of atomic writes or inter-process locking, and corrupt its data if it's unexpectedly killed/restarted.
Implementations that are too simple may be less secure, e.g. lack features for precise access controls, or sandboxing. Simple parsers can become a vulnerability (e.g. request smuggling attacks from underestimating complexity of edge cases in HTTP and URLs).
Too-simple tools may just push complexity and security risks elsewhere. There's an endless supply of simple templating engines that make it too easy to introduce XSS vulnerabilities.
If the implementation is too simple, it may have issues that impact its reliability. For example, code may be using simple linear search instead of fancier data structures/algorithms, and have issues with handling larger inputs or have accidentally-quadratic code paths. Or may updating data files without the complexity of atomic writes or inter-process locking, and corrupt its data if it's unexpectedly killed/restarted.
Implementations that are too simple may be less secure, e.g. lack features for precise access controls, or sandboxing. Simple parsers can become a vulnerability (e.g. request smuggling attacks from underestimating complexity of edge cases in HTTP and URLs).
Too-simple tools may just push complexity and security risks elsewhere. There's an endless supply of simple templating engines that make it too easy to introduce XSS vulnerabilities.
The opposite of OSS is not "write your own", it's "buy commercial".
And while (clearly) that would "cost more", the difference would be insignificant. In the sense that lots of companies already buy everything commercially, and the software budget is trivial compared to just about everything else. We buy a lot of software that we use, and our software budget is still smaller than our coffee budget.
OSS delivers many benefits, but I'd argue that "price" is not the primary one. (It just feels that way when you're a student.)
And while (clearly) that would "cost more", the difference would be insignificant. In the sense that lots of companies already buy everything commercially, and the software budget is trivial compared to just about everything else. We buy a lot of software that we use, and our software budget is still smaller than our coffee budget.
OSS delivers many benefits, but I'd argue that "price" is not the primary one. (It just feels that way when you're a student.)
The opposite of OSS is not "write your own", it's "buy commercial".
I agree, but that's not the counterfactual the authors considered. They asked how much it would cost for everyone to independently rewrite the same software. It's not surprising that such a ludicrous concept produced an absurd estimated value.
I agree, but that's not the counterfactual the authors considered. They asked how much it would cost for everyone to independently rewrite the same software. It's not surprising that such a ludicrous concept produced an absurd estimated value.
Why would you presume that?
There's plenty of proprietary software to test this assumption against: Microsoft sells it, Oracle sells it, IBM sells it.
Are their wares simple?
There's plenty of proprietary software to test this assumption against: Microsoft sells it, Oracle sells it, IBM sells it.
Are their wares simple?
There wares have to meet the needs of a diverse customer base. Custom software developed for in house use only needs to implement the features you need.
Though without OSS, I suspect a lot of companies would opt for purchasing off the shelf solutions, which would quite possibly be a worse fit than the OSS solutions.
Or, they may just go without.
Though without OSS, I suspect a lot of companies would opt for purchasing off the shelf solutions, which would quite possibly be a worse fit than the OSS solutions.
Or, they may just go without.
The authors are talking about a counterfactual of companies building software internally, not buying it from a publisher.
One thing we now had come up three times during the past year was commercial closed source software we had in our eye or were using, where the company was bought, went bust or whatever and took their product down, ceased to do updates etc.
With open source software you at least could pick up the torch yourself or pay someone who does. A particularly bad case involves the software running on our core switches, which we now need to swap out because there won't be any updates soon.
Which is why we want everything to be open source if possible.
With open source software you at least could pick up the torch yourself or pay someone who does. A particularly bad case involves the software running on our core switches, which we now need to swap out because there won't be any updates soon.
Which is why we want everything to be open source if possible.
Related blog post criticizing this study: https://news.ycombinator.com/item?id=39340146
Open source is the very worst thing currently going on because it is so incredibly exploitative, it's far more exploitative than any actual company is of the workers who work at the company.
Even the people who are getting paid in open source are getting massively underpaid to do it compared to how much the people who are using their code are making, it's nothing compared to the power that is accreted by the people who have co-opted that work thanks to the open source model. And then mark zuckerberg gets to define how the internet works despite having paid for almost none of the software that his company actually needed to make that work.
It's like feudalism or serfdom, these people did the work and got nothing for it. It's like you took the worst aspects of capitalism for workers and the worst aspects of socialism for workers and put them together, that's open source. You get no power and you get no money.
It's exploitative whether the people chose to be exploited, just because someone chooses to let you exploit them does not meant that you didn't exploit them. And for the record that's how most exploitation works; convincing people to do something that turns out to be very bad for them and very good for you, and that's exactly what the open source movement has turned out to be.
I really don't see the "we post stuff on github under a gpl2 or lgpl or apache or mit license", all that is to me now is just exploitation. You can say that there's solutions but until someone demonstrates that those solutions work, it's the standard "real communism has never been tried" argument. AGPL is the only thing that I've seen so far that's an attempt to fix these fundamentally unfair compensation practices.
Even the people who are getting paid in open source are getting massively underpaid to do it compared to how much the people who are using their code are making, it's nothing compared to the power that is accreted by the people who have co-opted that work thanks to the open source model. And then mark zuckerberg gets to define how the internet works despite having paid for almost none of the software that his company actually needed to make that work.
It's like feudalism or serfdom, these people did the work and got nothing for it. It's like you took the worst aspects of capitalism for workers and the worst aspects of socialism for workers and put them together, that's open source. You get no power and you get no money.
It's exploitative whether the people chose to be exploited, just because someone chooses to let you exploit them does not meant that you didn't exploit them. And for the record that's how most exploitation works; convincing people to do something that turns out to be very bad for them and very good for you, and that's exactly what the open source movement has turned out to be.
I really don't see the "we post stuff on github under a gpl2 or lgpl or apache or mit license", all that is to me now is just exploitation. You can say that there's solutions but until someone demonstrates that those solutions work, it's the standard "real communism has never been tried" argument. AGPL is the only thing that I've seen so far that's an attempt to fix these fundamentally unfair compensation practices.
All software created by private companies should be created by worker owned enterprises so they aren't forced to hand over the lions share of profits to separate owners. Instead the employees are the owners. They decide how to use the profits collectively and democratically among themselves.
Nothing stops like-minded employees from banding together, and forming a collective. Indeed these exist.
As you feel strongly about this, might I suggest you post an ad to find other like-minded individuals to align with and start said collective. (Seriously, I'm not being sarcastic. Collectives exist and do well.)
On the other hand you gloss over the role that owners play - perhaps because you've not been an owner? The obvious one of course being capital - developers like to be paid -while- creating, and not just "maybe later if it sells".
What you call "profit", some owners might call "return on investment". If your collective is prepared to invest time, or money, then indeed they should share in the return (if there is one.)
For better or worse, owners also have responsibilities. As much fun as collective decision making is, decisions do need to be made. What to build. When to hire. When to fire. Who to sell to. How to reach them. How much to spend reaching them. It takes a lot of right decisions to make a success, and only a couple of bad ones to sink the whole enterprise.
But, and this is my point, there are lots of models to choose, and everyone should pick thd one that suits them best. Employer or employee or man-alone. Open Source or Commercial. Hierarchical business or collective.
With software (unlike say, building roads) anything goes. Your future is the one you choose.
As you feel strongly about this, might I suggest you post an ad to find other like-minded individuals to align with and start said collective. (Seriously, I'm not being sarcastic. Collectives exist and do well.)
On the other hand you gloss over the role that owners play - perhaps because you've not been an owner? The obvious one of course being capital - developers like to be paid -while- creating, and not just "maybe later if it sells".
What you call "profit", some owners might call "return on investment". If your collective is prepared to invest time, or money, then indeed they should share in the return (if there is one.)
For better or worse, owners also have responsibilities. As much fun as collective decision making is, decisions do need to be made. What to build. When to hire. When to fire. Who to sell to. How to reach them. How much to spend reaching them. It takes a lot of right decisions to make a success, and only a couple of bad ones to sink the whole enterprise.
But, and this is my point, there are lots of models to choose, and everyone should pick thd one that suits them best. Employer or employee or man-alone. Open Source or Commercial. Hierarchical business or collective.
With software (unlike say, building roads) anything goes. Your future is the one you choose.
Financing is a big barricade for creating a worker owned enterprise in the US.
Some countries, like Italy, have programs for workers who are laid off. If ten workers that were laid off at the same time agree then the state will give them all of their unemployment wages in a lump sum to start a worker owned company together. Some countries will give workers the Right to First Refusal if their employer decides thatvthey want to close the place of employment or sale it. And the state will provide low interest loans to the workers so that they can buy the company.
There are lots of solutions to the financing issue, but we need the states and federal government to be more supportive of the practical realities of non-wealthy workers starting worker owned enterprises and implement policies that create a level playing field. Banks in the US currently frown on this type of lending.
Some countries, like Italy, have programs for workers who are laid off. If ten workers that were laid off at the same time agree then the state will give them all of their unemployment wages in a lump sum to start a worker owned company together. Some countries will give workers the Right to First Refusal if their employer decides thatvthey want to close the place of employment or sale it. And the state will provide low interest loans to the workers so that they can buy the company.
There are lots of solutions to the financing issue, but we need the states and federal government to be more supportive of the practical realities of non-wealthy workers starting worker owned enterprises and implement policies that create a level playing field. Banks in the US currently frown on this type of lending.
>> Financing is a big barricade for creating a worker owned enterprise in the US.
I'd argue that even allowing for the VC approach, financing is a big barricade for creating any business.
>> Banks in the US currently frown on this type of lending.
Given that 90% oc businesses will fail in the first 5 years, its not surprising that lending money to a new business is considered, well, risky.
But both these points return us to the same starting point - which is the value of providing capital, and the value of taking risk. Since an investor has a choice, we attract these investors by making them "owners", with the promise of future profit.
A government funded model, which provides "collective capital" for no equity sounds like a good plan, that can work well when applied to the right people.
Unfortunately it's likely that 90% of the money will be spent for no permanent job return, and that seems like it will become an issue once it spans multiple election periods.
Most successful small businesses are started with little capital, but a fair amount of time. I was fortunate that my wife's salary supported me in the years it took for mine to make money.
I'd argue that even allowing for the VC approach, financing is a big barricade for creating any business.
>> Banks in the US currently frown on this type of lending.
Given that 90% oc businesses will fail in the first 5 years, its not surprising that lending money to a new business is considered, well, risky.
But both these points return us to the same starting point - which is the value of providing capital, and the value of taking risk. Since an investor has a choice, we attract these investors by making them "owners", with the promise of future profit.
A government funded model, which provides "collective capital" for no equity sounds like a good plan, that can work well when applied to the right people.
Unfortunately it's likely that 90% of the money will be spent for no permanent job return, and that seems like it will become an issue once it spans multiple election periods.
Most successful small businesses are started with little capital, but a fair amount of time. I was fortunate that my wife's salary supported me in the years it took for mine to make money.
We should have more public banks like North Dakota does and which are found in many countries. There is clearly a vested interest in the US by the power centers to keeping public banks from spreading.
You are correct when taken your value definitions of compensation and power, that FOSS is exploitative but i think you miss some important aspects here.
https://news.ycombinator.com/item?id=33298158
This great article depicts some "loosers" as voluntary, that simply want to have a fulfilled life. FOSS is in itself a product of passion, no exploitive sociopath could ever bootstrap this. The free accessibility of software and knowledge is what enables the FOSS enthusiast and the profiteurs alike. Reducing FOSS to the latter is a valid perspective and your concerns are reasonalbe but even if there was no exploitation/corporate use of FOSS software, i assume, the maintainers would still do their unpaid thing. The expoitation comes second and is not inherent to FOSS. I wouldnt call it "worst thing on".
Also, steven balmer disagrees with you :)
Edit: Maybe some kind of GPL-like infectious form of enterprise could be a solution, where such a non-exploitive/worker owned company is bound to prefer business partners/suppliers that adhere to it too. Just a throw-in.
https://news.ycombinator.com/item?id=33298158
This great article depicts some "loosers" as voluntary, that simply want to have a fulfilled life. FOSS is in itself a product of passion, no exploitive sociopath could ever bootstrap this. The free accessibility of software and knowledge is what enables the FOSS enthusiast and the profiteurs alike. Reducing FOSS to the latter is a valid perspective and your concerns are reasonalbe but even if there was no exploitation/corporate use of FOSS software, i assume, the maintainers would still do their unpaid thing. The expoitation comes second and is not inherent to FOSS. I wouldnt call it "worst thing on".
Also, steven balmer disagrees with you :)
Edit: Maybe some kind of GPL-like infectious form of enterprise could be a solution, where such a non-exploitive/worker owned company is bound to prefer business partners/suppliers that adhere to it too. Just a throw-in.
> i assume, the maintainers would still do their unpaid thing.
That's a broad generalization. Sure, there are pure-passion maintainers doing voluntary work and can afford that by living on savings or alongside a paid job.
That's a broad generalization. Sure, there are pure-passion maintainers doing voluntary work and can afford that by living on savings or alongside a paid job.
Poor me, getting paid better than 99.9% of the world to provision public goods.
Why, I'm benefiting people at the 99.9999% level in the process! Clearly this is the very worst thing currently going on, because it is so incredibly exploitative.
Why, I'm benefiting people at the 99.9999% level in the process! Clearly this is the very worst thing currently going on, because it is so incredibly exploitative.
Much of this is inherently interesting self-directed development work, and there already seems to be an ample supply of people who want to do it. If you’re looking for a market that would clear better by increasing compensation, look to gaps like documentation and live support and security patching that go ignored because they aren’t fun (for nerds).
We then calculate the demand side value based on a replacement value for each firm that uses the software and would need to build it internally if OSS did not exist.
We estimate the supply-side value of widely-used OSS is $4.15 billion, but that the demand-side value is much larger at $8.8 trillion. We find that firms would need to spend 3.5 times more on software than they currently do if OSS did not exist."
If programming is "theory building" [1], then OSS is sharing our theories of how to program computers. It makes sense that sharing our ideas is more valuable to the aggregate economy than keeping them secret.
[1] https://pages.cs.wisc.edu/~remzi/Naur.pdf