The original sin of Free Software was calling it “free software”. “Software” is an analogy to wares, as in “warehouse”, and “free wares” clearly means “gratis”.
This was compounded by “FOSS”, where newbies will obviously, naturally interpret as gratis - “Free” is often capitalized to avoid confusion with “gratis”, but FOSS strips that hint by capitalizing every word (because it’s an acronym) and leaves users with just “software that is 1) open source, and 2) free”. The obvious conclusion is not that #2 means “open source (again) but with a slightly different underlying intentions”, it appears obviously meant to imply that it’s zero-cost.
The problem with de-emphasizing money among the community is that open-source brings in corporate money, which ends up being most of the money, and leads to all the most well-funded projects being thoroughly dependent on corporations, who (even with the best of intentions) have priorities that differ from actual people. Like, home server operators need software that’s low-maintenance and simple, whereas e.g. Google might prioritize software that scales well across three continents.
Money matters because a well-funded project is almost always way better. Firefox has an amazingly well-polished UI, and the reason it’s polished is because Firefox spends millions of dollars of income from Google on paying UI designers to polish it. MS Office has a very well-polished interface, and that’s because Microsoft spent millions of dollars of income on paying for the UI designers and UI research to polish it. I’m sure there are exceptions, but that’s missing the forest for the trees - the trend is very clear. Solid funding gets results!
In the long run, the libre software community needs to ween itself off corporate funding and get consistent income directly from it’s users if it ever wants 100% libre software stacks to outcompete proprietary software.
Good luck getting that kind of income when you’re giving the code away for free. You can’t have it both ways.
Or rather, you can do this in some domains where service contracts can pull in enough income, assuming you have all the expert developers of that software on your team so you have a major comparative advantage. That’s how companies like MongoDB and Couchbase started. But eventually outside groups can study the code and build up enough knowledge that your advantage evaporates. At that point you have to give up on open source (as they did) or close up shop.
In which case, good luck convincing people to pay for something they can get for free. Human nature just doesn’t work that way.
Either you have a tip jar and hope a few people donate (as today), or you give paying customers extras (like support, as with corporate users), or you somehow upend the whole economic system so developers are paid through some kind of public taxation (which is rather leftward of where any non-Communist country has ever gone, AFAIK.)
For backend stuff selling b2b to tech companies this might be true. For selling into enterprise and government it’s much less true. For selling to consumer it’s basically untrue.
I read this post with interest but it does not align with how I think about free software. The idea of the post is that the licenses we use when sharing software should do a better job of enforcing collective code-contribution structure that benefits free software and its authors. (It also points out that the current licenses and their description embody a certain set of values. This is interesting and independent and maybe there is something worth changing there.)
Basically I write free software because I think of my software release as a gift – growing a commons. I want people to be able to do this without sacrificing themselves, by being part of a society that recognizes and supports this work – just like we try to make sure that artists have the means to live and produce work, even when they are not commercially successful. It is not clear to me that software licenses are the right vehicle to enforce this; I would rather rely on a wider organization of society that values these gift-producing activities and supports them. (We have public-funded artistic residencies, and universities and research institutions basically provide a similar service for many type of programming contributions. There are other ways to support this that exist today or that should be made to exist.)
In fact we do have experience on the fact that licenses are not necessarily the right vehicle. Whenever we use more restrictive licenses, it often ends up being a barrier to code reuse for people who work in industry. Some people who would gladly use our software and contribute back (if only with usage feedback, bug reports etc.) will not be able to if their employer has lawyers that are afraid of license X and Y and we use it. Personally I am a hereditary-license person at heart, I would use GPLv3 for new projects, but if someone tells me that the license is a barrier for them I am generally happy to downgrade to MIT/BSD just to keep things simple. Yes, having companies give back to open source is really nice, I’m happy to think of ways to make that happen, but talking to lawyers is awful and it is a net improvement in my life to simplify that away with less restrictive licenses.
I read this post with interest but it does not align with how I think about free software
It’s also missing RMS’ key motivation: it was nothing to do with software developers, it was about the rights of users. This has nothing to do with anarcho-capitalism it is about preventing the authors of software from taking away rights from the users. The important historical context is that most software actually met his definition at the time that it was written. COTS software was a small part of the industry, the vast majority of software was written specifically for a single customer, who then held the rights to do whatever they wanted with it. The Free Software movement began in part because of the transition away from this. It was already common on early consumer microcomputers to have purely off-the-shelf software and RMS encountered the problem in a printer driver.
In my mind, the biggest failing of the free software movement has been that it has lost this user focus. MS Office introduced VBA, Excel’s macro language, and later PowerShell. Apple introduced AppleScript across their entire OS and application suite. These provided end users with something a lot closer to the FSF’s goals than access to source code: they provided extension points where end users (who were not professional programmers) could adapt the tools to their needs. Doing the same with most GNU projects required learning C or, if you’re lucky, Scheme. From an end user perspective, having access to source code and the legal rights to modify it is of no use if the code is a huge pile of poorly documented spaghetti written in a low-level language.
This is also, I believe, why corporations have taken such a big role. If you’re Microsoft, Amazon, or even a smaller company like Netflix, you can easily hire developers who can take advantage of the four freedoms. At the same time, you have lawyers who care about risk and will insist on a second source for critical infrastructure: something that you get automatically with F/OSS if you are able to hire developers (in house or externally) who will add features and fix bugs that you care about. If you take a F/OSS stack, you can quantify the cost of maintaining a fork of it if it decides to go in a different direction to your requirements, and you can reason about risk. With proprietary software you need to either be a sufficiently large customer that you can negotiate a custom support contract or you just need to accept that, at some point, it may stop doing what you want.
Basically I write free software because I think of my software release as a gift – growing a commons
I write free software because I can achieve more in collaboration with others than independently. One of the Sun founders (I forget which one) said ‘remember that more smart people don’t work for you than do’. As an individual, you are in the limit case of this: all of the smart people except one don’t work for you. Contributing to an existing project is far easier than duplicating all of its effort from scratch.
I also view F/OSS as reflecting economic realities: writing software is expensive (time, effort), whereas copying software is free. If someone makes a copy of something I’ve written, I’ve lost nothing. If they use it to get rich, I’ve lost nothing. If they send me a simple bug fix, I’ve gained something. If there’s a non-zero chance that other people will contribute to something I release, that’s a net win. This is why I use GitHub after years of refusing on ideological grounds: I want to make it easy for other people to contribute to my projects (I would love to see a publicly funded alternative to GitHub).
I would love to see a publicly funded alternative to GitHub
Isn’t this effectively what Codeberg is doing?
There’s also the models like Sourcehut (& future Smederee by the looks?) that make everyone pay a small amount to the privately-held service rather than having some loss-leading free tier for growth to go against that VC funded model.
In my mind, the biggest failing of the free software movement has been that it has lost this user focus. MS Office introduced VBA, Excel’s macro language, and later PowerShell. Apple introduced AppleScript across their entire OS and application suite. These provided end users with something a lot closer to the FSF’s goals than access to source code: they provided extension points where end users (who were not professional programmers) could adapt the tools to their needs.
One of my favorite quotes by RMS, that I just cannot find right now, is that he imagined that (executive’s) secretaries would use the capabilities of Emacs to ease their work. He had the right idea, but maybe elisp was the wrong execution.
Basically I write free software because I think of my software release as a gift – growing a commons.
Yeah. I write free software because I am free. I am free to use my time writing stuff that interests me, and I am free to take my toolbox at work from one job to the next, and I am free to share my work with others.
I don’t have answer how else this should work, but realize that writing software as a gift, while you still have to earn for a living, means that you need to be in a good position of earning income and have enough spare time to contribute to the given-away-for-free software (writing FOSS for a living is generally hard, and mostly relies on subsidizing FOSS work from a related commercial side business).
You do extra work in your spare time, while corporations benefit from permissive OSS for free, are not obligated to give anything back, or help you or FOSS in any way. You are doing free labor for somebody’s else’s business, and this isn’t controversial, only because there’s a layer of indirection between you and the business, or you treat it as fun, or never had hopes of being compensated for this work in the first place.
Imagine if OSS-using corp came to you before you wrote the software. “Hi! We’d like you to write a library for turbofoowidget. How much we pay? 0. We’ll send you complaints about defects so that you put more work into it, also for free.”
This post makes the same mistake that most professional programmers make when it comes to free software, assuming that the institution is there to protect programmers. Indeed, for all the leftist rhetoric in the article, it’s coming from a place that is clearly heavily influenced by the needs of people writing software for profit.
Free software is there to protect users, not developers. Users can always access the source code of the software they use and modify it; users can switch which developer is maintaining their software; users don’t have to do what the developers tell them to do.
The developer doesn’t automatically get access to changes other people have made to software they wrote, because developers already have that power in conventional software copyright law. Developers start with all the power. Free software is a way to move that to users.
The thing is that users aren’t protected anymore. Or rather, end users. Copyleft projects that are ostensibly part of the commons are taken by large cloud companies which then turn those projects into profitable products which are not source available. Large corporations do not need protection, they have armies of lawyers to do that for them. The law as written is largely written according to their benefit. We should be protecting people who don’t have access to those resources as well as public institutions like universities.
If users really cared about having the four freedoms, this wouldn’t have happened. The vast majority obviously doesn’t care. It seems to me that the (bogus) assumption underlying the four freedoms is that the users are either programmers themselves, or technical enough to understand what use source code is and that they could hire a programmer to change it for them. And even then, the majority of potential users who do fit this assumption routinely choose nonfree software.
As a result of the underlying assumptions and conceptions of the FOSS movement, there is no methodology to require the sharing of changes and improvments with the commons.
Of course there is. The GPL famously requires that any derivative works — changes and improvements — also be GPL-licensed, which requires publishing their source code. That’s the whole reason a lot of companies and developers avoid it, and why so much (most?) free software still uses less restrictive licenses like BSD and Apache.
Its entire business model is predicated on the willingness of creators to share the product of their labour with the wider world. These creators[7] offer the product of their labour to the site, solely in exchange for the chance of exposure and discovery. There is no contract with GitHub which can provide payment for this delivery, save the ephemeral back-slap of the “Star” and the prospect of collaboration with others.
Um, what? Why should GitHub pay for the free source code people upload to it? That would create a bizarre form of “free to everyone except GitHub” licensing. Moreover, GitHub is providing valuable free services to developers, like code hosting, Git VCS, issue tracking, project management, CI… You used to have to pay for that kind of software and that kind of hosting. So let’s say it balances out.
The author makes fun of American libertarians (heck, so do I) but their own political/economic ideas seem to come from an equally extreme place on the Left.
They may be complaining that we cannot coerce individuals into revealing their private modifications. Which is true under the GPL if they don’t give out the binaries. Because the alternative is not good.
Many good points, but I feel licensing is the wrong tool for the job (even though it’s kinda our only tool right now). Right now, huge swaths of the open source world are wholly dependent on the paltry resources that corporations decide to give. If the terms change, corporations are most likely to migrate to an alternative or make their own alternative (Elasticsearch/Opensearch being the prime example).
Instead, we need something like (a better version of) the laws protecting labor organizing, where corporations are forced to negotiate, and we make “strike-breaking” actions (like forking right before a license change) illegal as well. The analogy is very loose here, so don’t read too far into into. My point is just that I think the solution to the “original sin” needs to come from broader actions than just the choice of license on individual projects.
The only open source “commons” I’ve seen work are those where every player has an individual monetary interest in it which is not threatened by the commons.
It seems to just give people fuzzy feelings to think something is shared, with the same connotations as “home made” or “organic”. And this happens regardless of whether it’s actually cooperative or not.
I know this: OPs article is nowhere near cynical enough.
The author needs to explain which collective rights they feel trumps the individual’s general liberty to use software, and why this either is not a restriction on them or is a just means to a higher end.
Saying that the DFSG and its file-off-the-serial-numbers copy the OSD are “original” rather than flawed attempts to flesh out the FSD also does a lot of work here.
There is no methodology to demand sharing because this would be coercion. See first comment.
As for the usual complaint that we should restrict uses of software only to ones we (and not corporations!) like, this is very easy to request but much harder to apply and hardly grows the commons.
Software licenses are choke-points for software. That’s it. That’s the joke. If you wish to advance a politics other than the politics of software usability using them you are greatly over-estimating the potential of their direct effectiveness for that. Sulking licences are the best you can do.
“Bruce Perens, father of the Debian Project, holds to the peculiar American cult of the Individualist Libertarian, as does Eric Raymond,”
close tab
This post is not about software, nor is it very good. I strongly disagree with the FSF and GNU on a lot of their core beliefs and their goals, and despise the GPL, but they’re not oppressing anyone, and to claim otherwise is farce.
This is what you get when the key to making it through tertiary education has become re-framing everything in terms of oppression and critical theory. Garbage in, garbage out.
The original sin of Free Software was calling it “free software”. “Software” is an analogy to wares, as in “warehouse”, and “free wares” clearly means “gratis”.
This was compounded by “FOSS”, where newbies will obviously, naturally interpret as gratis - “Free” is often capitalized to avoid confusion with “gratis”, but FOSS strips that hint by capitalizing every word (because it’s an acronym) and leaves users with just “software that is 1) open source, and 2) free”. The obvious conclusion is not that #2 means “open source (again) but with a slightly different underlying intentions”, it appears obviously meant to imply that it’s zero-cost.
The problem with de-emphasizing money among the community is that open-source brings in corporate money, which ends up being most of the money, and leads to all the most well-funded projects being thoroughly dependent on corporations, who (even with the best of intentions) have priorities that differ from actual people. Like, home server operators need software that’s low-maintenance and simple, whereas e.g. Google might prioritize software that scales well across three continents.
Money matters because a well-funded project is almost always way better. Firefox has an amazingly well-polished UI, and the reason it’s polished is because Firefox spends millions of dollars of income from Google on paying UI designers to polish it. MS Office has a very well-polished interface, and that’s because Microsoft spent millions of dollars of income on paying for the UI designers and UI research to polish it. I’m sure there are exceptions, but that’s missing the forest for the trees - the trend is very clear. Solid funding gets results!
In the long run, the libre software community needs to ween itself off corporate funding and get consistent income directly from it’s users if it ever wants 100% libre software stacks to outcompete proprietary software.
Good luck getting that kind of income when you’re giving the code away for free. You can’t have it both ways.
Or rather, you can do this in some domains where service contracts can pull in enough income, assuming you have all the expert developers of that software on your team so you have a major comparative advantage. That’s how companies like MongoDB and Couchbase started. But eventually outside groups can study the code and build up enough knowledge that your advantage evaporates. At that point you have to give up on open source (as they did) or close up shop.
This is my point: we need a culture of paying for libre software. Right now we don’t, and in fact have created the opposite, and so people don’t pay.
In which case, good luck convincing people to pay for something they can get for free. Human nature just doesn’t work that way.
Either you have a tip jar and hope a few people donate (as today), or you give paying customers extras (like support, as with corporate users), or you somehow upend the whole economic system so developers are paid through some kind of public taxation (which is rather leftward of where any non-Communist country has ever gone, AFAIK.)
For backend stuff selling b2b to tech companies this might be true. For selling into enterprise and government it’s much less true. For selling to consumer it’s basically untrue.
I read this post with interest but it does not align with how I think about free software. The idea of the post is that the licenses we use when sharing software should do a better job of enforcing collective code-contribution structure that benefits free software and its authors. (It also points out that the current licenses and their description embody a certain set of values. This is interesting and independent and maybe there is something worth changing there.)
Basically I write free software because I think of my software release as a gift – growing a commons. I want people to be able to do this without sacrificing themselves, by being part of a society that recognizes and supports this work – just like we try to make sure that artists have the means to live and produce work, even when they are not commercially successful. It is not clear to me that software licenses are the right vehicle to enforce this; I would rather rely on a wider organization of society that values these gift-producing activities and supports them. (We have public-funded artistic residencies, and universities and research institutions basically provide a similar service for many type of programming contributions. There are other ways to support this that exist today or that should be made to exist.)
In fact we do have experience on the fact that licenses are not necessarily the right vehicle. Whenever we use more restrictive licenses, it often ends up being a barrier to code reuse for people who work in industry. Some people who would gladly use our software and contribute back (if only with usage feedback, bug reports etc.) will not be able to if their employer has lawyers that are afraid of license X and Y and we use it. Personally I am a hereditary-license person at heart, I would use GPLv3 for new projects, but if someone tells me that the license is a barrier for them I am generally happy to downgrade to MIT/BSD just to keep things simple. Yes, having companies give back to open source is really nice, I’m happy to think of ways to make that happen, but talking to lawyers is awful and it is a net improvement in my life to simplify that away with less restrictive licenses.
It’s also missing RMS’ key motivation: it was nothing to do with software developers, it was about the rights of users. This has nothing to do with anarcho-capitalism it is about preventing the authors of software from taking away rights from the users. The important historical context is that most software actually met his definition at the time that it was written. COTS software was a small part of the industry, the vast majority of software was written specifically for a single customer, who then held the rights to do whatever they wanted with it. The Free Software movement began in part because of the transition away from this. It was already common on early consumer microcomputers to have purely off-the-shelf software and RMS encountered the problem in a printer driver.
In my mind, the biggest failing of the free software movement has been that it has lost this user focus. MS Office introduced VBA, Excel’s macro language, and later PowerShell. Apple introduced AppleScript across their entire OS and application suite. These provided end users with something a lot closer to the FSF’s goals than access to source code: they provided extension points where end users (who were not professional programmers) could adapt the tools to their needs. Doing the same with most GNU projects required learning C or, if you’re lucky, Scheme. From an end user perspective, having access to source code and the legal rights to modify it is of no use if the code is a huge pile of poorly documented spaghetti written in a low-level language.
This is also, I believe, why corporations have taken such a big role. If you’re Microsoft, Amazon, or even a smaller company like Netflix, you can easily hire developers who can take advantage of the four freedoms. At the same time, you have lawyers who care about risk and will insist on a second source for critical infrastructure: something that you get automatically with F/OSS if you are able to hire developers (in house or externally) who will add features and fix bugs that you care about. If you take a F/OSS stack, you can quantify the cost of maintaining a fork of it if it decides to go in a different direction to your requirements, and you can reason about risk. With proprietary software you need to either be a sufficiently large customer that you can negotiate a custom support contract or you just need to accept that, at some point, it may stop doing what you want.
I write free software because I can achieve more in collaboration with others than independently. One of the Sun founders (I forget which one) said ‘remember that more smart people don’t work for you than do’. As an individual, you are in the limit case of this: all of the smart people except one don’t work for you. Contributing to an existing project is far easier than duplicating all of its effort from scratch.
I also view F/OSS as reflecting economic realities: writing software is expensive (time, effort), whereas copying software is free. If someone makes a copy of something I’ve written, I’ve lost nothing. If they use it to get rich, I’ve lost nothing. If they send me a simple bug fix, I’ve gained something. If there’s a non-zero chance that other people will contribute to something I release, that’s a net win. This is why I use GitHub after years of refusing on ideological grounds: I want to make it easy for other people to contribute to my projects (I would love to see a publicly funded alternative to GitHub).
Isn’t this effectively what Codeberg is doing?
There’s also the models like Sourcehut (& future Smederee by the looks?) that make everyone pay a small amount to the privately-held service rather than having some loss-leading free tier for growth to go against that VC funded model.
One of my favorite quotes by RMS, that I just cannot find right now, is that he imagined that (executive’s) secretaries would use the capabilities of Emacs to ease their work. He had the right idea, but maybe elisp was the wrong execution.
Yeah. I write free software because I am free. I am free to use my time writing stuff that interests me, and I am free to take my toolbox at work from one job to the next, and I am free to share my work with others.
I don’t have answer how else this should work, but realize that writing software as a gift, while you still have to earn for a living, means that you need to be in a good position of earning income and have enough spare time to contribute to the given-away-for-free software (writing FOSS for a living is generally hard, and mostly relies on subsidizing FOSS work from a related commercial side business).
You do extra work in your spare time, while corporations benefit from permissive OSS for free, are not obligated to give anything back, or help you or FOSS in any way. You are doing free labor for somebody’s else’s business, and this isn’t controversial, only because there’s a layer of indirection between you and the business, or you treat it as fun, or never had hopes of being compensated for this work in the first place.
Imagine if OSS-using corp came to you before you wrote the software. “Hi! We’d like you to write a library for turbofoowidget. How much we pay? 0. We’ll send you complaints about defects so that you put more work into it, also for free.”
This post makes the same mistake that most professional programmers make when it comes to free software, assuming that the institution is there to protect programmers. Indeed, for all the leftist rhetoric in the article, it’s coming from a place that is clearly heavily influenced by the needs of people writing software for profit.
Free software is there to protect users, not developers. Users can always access the source code of the software they use and modify it; users can switch which developer is maintaining their software; users don’t have to do what the developers tell them to do.
The developer doesn’t automatically get access to changes other people have made to software they wrote, because developers already have that power in conventional software copyright law. Developers start with all the power. Free software is a way to move that to users.
The thing is that users aren’t protected anymore. Or rather, end users. Copyleft projects that are ostensibly part of the commons are taken by large cloud companies which then turn those projects into profitable products which are not source available. Large corporations do not need protection, they have armies of lawyers to do that for them. The law as written is largely written according to their benefit. We should be protecting people who don’t have access to those resources as well as public institutions like universities.
Yes, so let’s protect users again by preventing this. This is the intent of things like the AGPL and http://catern.com/freedist.html
Or sometimes something a bit stronger like Parity License.
If users really cared about having the four freedoms, this wouldn’t have happened. The vast majority obviously doesn’t care. It seems to me that the (bogus) assumption underlying the four freedoms is that the users are either programmers themselves, or technical enough to understand what use source code is and that they could hire a programmer to change it for them. And even then, the majority of potential users who do fit this assumption routinely choose nonfree software.
Of course there is. The GPL famously requires that any derivative works — changes and improvements — also be GPL-licensed, which requires publishing their source code. That’s the whole reason a lot of companies and developers avoid it, and why so much (most?) free software still uses less restrictive licenses like BSD and Apache.
Um, what? Why should GitHub pay for the free source code people upload to it? That would create a bizarre form of “free to everyone except GitHub” licensing. Moreover, GitHub is providing valuable free services to developers, like code hosting, Git VCS, issue tracking, project management, CI… You used to have to pay for that kind of software and that kind of hosting. So let’s say it balances out.
The author makes fun of American libertarians (heck, so do I) but their own political/economic ideas seem to come from an equally extreme place on the Left.
Small not: the GPL does not require publishing code. It requires offering to anyone you distribute to that they may have the code if they ask for it.
They may be complaining that we cannot coerce individuals into revealing their private modifications. Which is true under the GPL if they don’t give out the binaries. Because the alternative is not good.
Many good points, but I feel licensing is the wrong tool for the job (even though it’s kinda our only tool right now). Right now, huge swaths of the open source world are wholly dependent on the paltry resources that corporations decide to give. If the terms change, corporations are most likely to migrate to an alternative or make their own alternative (Elasticsearch/Opensearch being the prime example).
Instead, we need something like (a better version of) the laws protecting labor organizing, where corporations are forced to negotiate, and we make “strike-breaking” actions (like forking right before a license change) illegal as well. The analogy is very loose here, so don’t read too far into into. My point is just that I think the solution to the “original sin” needs to come from broader actions than just the choice of license on individual projects.
The only open source “commons” I’ve seen work are those where every player has an individual monetary interest in it which is not threatened by the commons.
It seems to just give people fuzzy feelings to think something is shared, with the same connotations as “home made” or “organic”. And this happens regardless of whether it’s actually cooperative or not.
I know this: OPs article is nowhere near cynical enough.
The author needs to explain which collective rights they feel trumps the individual’s general liberty to use software, and why this either is not a restriction on them or is a just means to a higher end.
Saying that the DFSG and its file-off-the-serial-numbers copy the OSD are “original” rather than flawed attempts to flesh out the FSD also does a lot of work here.
There is no methodology to demand sharing because this would be coercion. See first comment.
As for the usual complaint that we should restrict uses of software only to ones we (and not corporations!) like, this is very easy to request but much harder to apply and hardly grows the commons.
Software licenses are choke-points for software. That’s it. That’s the joke. If you wish to advance a politics other than the politics of software usability using them you are greatly over-estimating the potential of their direct effectiveness for that. Sulking licences are the best you can do.
“Bruce Perens, father of the Debian Project, holds to the peculiar American cult of the Individualist Libertarian, as does Eric Raymond,”
close tab
This post is not about software, nor is it very good. I strongly disagree with the FSF and GNU on a lot of their core beliefs and their goals, and despise the GPL, but they’re not oppressing anyone, and to claim otherwise is farce.
This is what you get when the key to making it through tertiary education has become re-framing everything in terms of oppression and critical theory. Garbage in, garbage out.