13 ms·

Why is a few snippets of code so valuable anyway? We don't GPL scientific research, you expect people to cite you, and then build off your work, the licensing i

by foxes 5y ago

Why is a few snippets of code so valuable anyway? We don't GPL scientific research, you expect people to cite you, and then build off your work, the licensing in a sense is far more permissive. You don't GPL maths text books and then require proofs to be licensed back. Other knowledge doesn't seem to propagate in society like this. The human element of the programmer who can understand, adapt and develop some program is more valuable. The valuable thing is their time and overall knowledge, not the end piece of content necessarily. I think you do want to make sure people who have good ideas get recognised, (but this is like GitHub, employment record etc) but truly it's more beneficial to society if those good ideas propagate freely and get reused. I think the copilot thing really brings this to light. I believe BSD is probably the best option. Valuing the programmer more, caring more about who develops your app is more important than merely the code at the end of the day. Eg is this a reputable team I want to support? The copyright/license system seems out dated in light of new information. I think the citation model works better. If its truly well established knowledge then yes maybe a formal reference exists somewhere, but they sort of fade into the background.

wildmanx 5y ago
I think the key property of the GPL is its virality. It's part of the vision that if I can use some software, I can also change it, to adapt it to my needs, to learn from it, to build something on top of it. I can disassemble my bike, re-assemble it, change it, build on top of it. I want to do that with my browser, my OS kernel and my image viewer too. That this is connected to how "valuable" code is, is in a sense a happy coincidence. The conditions are attached and entangled because this forces people to listen and pay attention. A more "permissive" license (like BSD) is weakening that point. It allows you to reuse code but does not require the re-user to allow the above mentioned openness. So is kind of missing the point, from that perspective.
- foxes 5y ago
  Well I agree in that regard, but I think in the ideal world everything would just be open, and you would want to incentivise openness so it ends up as the natural state. In a sense by decoupling these issues. Eg Microsoft Windows OS could just be open source. There is some value to having windows itself, but the more valuable thing is the horde of programmers, and all the other resources behind it.
  - npteljes 5y ago
    This sounds good. How do you imagine getting there? GPL people think that contributing to this ideal is only worth it if you also require all participants to take part in the ideal. Thereby granting that at least the GPL software stays in that ideal world.
    - scrollaway 5y ago
      It's definitely a tough call. Maybe this is an industry scale problem. The GPL existing hasn't stopped walled gardens from developing, and open source is falling behind in many regards. Maybe the license isn't the correct place to address this. I'm not sure what the solution is, though...
      - npteljes 5y ago
        Proprietary is, I think, the "default" way how people always did things. Keeping the recipe secret is one of the moats that help the enterprise stay relevant. So naturally, "it works" and I think it always will. IMO, this is why it's important to keep the open source stuff open. Closed doesn't usually turn into open, so why should open turn into closed? I think the legal system is exactly the place to address this, in fact, I think the governments themselves should drive the development, as it's their own interest to really own their software and hardware stack.
        scrollaway 5y ago
        I don't remember who but someone here on HN once compared closed source vs open source to old school secretive alchemy guilds vs real science. I really liked that comparison and I've thought about it a lot since. And here it's worth asking, how did we go from alchemy to science? How did we end up figuring out that openness and transparency is beneficial? I know one thing, it wasn't with licenses. But I'm really curious, because I want to live in the world where open source is the default. I just don't know how we get there.
        5y ago
- bluGill 5y ago
  > It allows you to reuse code but does not require the re-user to allow the above mentioned openness That is the dogma, but in practice it is false. Khtml was GPL, that didn't stop Apple from forking it in such a way that their changes couldn't be reused. In practice many who change BSD software find it worth while to contribute their changes back - so they can continue to use the latest version instead of maintaining their own fork.
  - temac 5y ago
    khtml was and is still LGPLv2.1, webkit still is, and contains major work from Apple that obviously could and has been reused...
darthrupert 5y ago
"Few snippets of code" can be pretty large. I see a lot of value in not pointlessly reinventing the wheel. GPL encourages/enforces people who fix things to give those changes back to upstream where they may benefit everyone. So ideally only one set of fixes per problem has to be globally implemented and maintained. The implications for global wealth are non-trivial. How much Linux alone has done, I wonder?
- diegoperini 5y ago
  > I see a lot of value in not pointlessly reinventing the wheel. GPL encourages/enforces people who fix things to give those changes back to upstream where they may benefit everyone. One can argue GPL encourages reinventing the wheel by not being permissive enough.
sangnoir 5y ago
> The valuable thing is their time and overall knowledge, not the end piece of content necessarily. The GPL doesn't give primacy to the thing that's built - but to the user of the thing. The GPL implicitly values the user's freedom over developer's convenience - the opposite is true for BSD-like licenses. Generally, I feel like the affinity most people have to one license or another entirely depends on how they more closely identify as a user of. software, vs as a developer
- alex_smart 5y ago
  >The GPL values the user's freedom, That assumes a very specific kind of user. The kind that has the time and energy, not to mention the ability, to read and understand source code and modify it according to his needs. Basically the RMS kind of user. For most users, that freedom is not much of a concern at all. For them, the most important freedom is: the freedom to enjoy the fruits of labour of a large horde of programmers who can make a fruitful living out of creating apps for said users. They enjoy this freedom a lot more in what GNU calls non-free platforms like Windows and Mac and iPhone, which is why they keep flocking to them.
  - marcodiego 5y ago
    > That assumes a very specific kind of user. The kind that has the time and energy, not to mention the ability, to read and understand source code and modify it according to his needs. Actually, no. End users also benefit when code is contributed back to the software they use.
    - alex_smart 5y ago
      Good point. Of course open source software has definitely benefitted users in a wide variety of ways: for instance by providing open source tools to help build said software (can't even imagine a world without free compilers these days). My intent was only to help the GPL and free software fanatics see an important freedom provided by non-free platforms, not to belittle the achievements of open source community.
xunn0026 5y ago
Have you heard of... patents? Why is an idea written on paper so valuable? Why can't people just use it (or, as it often happens, independently rediscover it) and just cite the patent if recognition is required? As for scientific research, most of the software behind scientific paper is basically closed source. Which is another reason why there's a whole reproducibility problem with science papers: it's quite possible many papers are just bad interpretation of data due to bad software. Also note how much software / research / data is proprietary. How can one quote that? I wouldn't mind using only BSD in the future if I get to read the Apple iOS source code on GitHub.
- foxes 5y ago
  I agree scientific papers should be open source. It is not acceptable there is a proliferation of bad practice. That seems a different issue however.
  - xunn0026 5y ago
    It's not a different issue. You used academic papers as the standard we could use for open source and it's provably worse. Even copyright... isn't Elsevier a major copyright holder for all academic papers and thus causing all sorts of problems. It's not like it's all kumbaya out there. The academic world is much worse than GPL and open source on multiple dimensions.
    - jsjohnst 5y ago
      > isn't Elsevier a major copyright holder for all academic papers and thus causing all sorts of problems No, I could be wrong, but it’s my understanding that the author of the academic paper still holds the copyright, it’s just the author licensed sole (or near exclusive) distribution rights to Elsevier in exchange for publication in their journals. It’s basically the same in the publishing world, I own the copyright on the book I coauthored, but WROX Press owns the distribution rights as we gave those up in exchange for them publishing.
    - fastaguy88 5y ago
npteljes 5y ago
If what you say stands, they should also include all of Microsoft's proprietary code in the learning process. Do you see that happening?
drran 5y ago
1) Your analogy is wrong, because none of the software licenses forces a user to have a license for his work. Moreover, a math book has strong copyright protection. You cannot just copy the math book and sell it as your own. GPL allows that. 2) Yep, the whole idea of GPL is to allow those good sources to propagate freely and get reused. Read the license, please. 3) Citation model is the core of BSD license, so use it for your own code.
belorn 5y ago
Scientific research is patented and exclusively copyrighted under publisher. If all scientific research were public and accessible we would not have sci-hub, and they would not currently be in court. We would also not live in a world where scientist has to work around patents and in some cases chose alternative research areas just because they can't afford (or use the result) because of patents. In addition there is a huge controversy in many Universities that patent from research is owned by the university, and if the researcher want to use technology that they themselves discover they need to pay the university. Removing patents and copyright from public funded universities is an on-going struggle that has yet to be fixed. We have yet to start to make all scientific research free from patents and copyright. We should strive towards the utopia where all that was expected is that people will cite the source, and then everyone could build off the work of everyone else.
- tzs 5y ago
  > Scientific research is patented and exclusively copyrighted under publisher. Research is not copyrighted. Particular write-ups of the research are copyrighted.
  - belorn 5y ago
    I am not sure I understand your point of distinction. Yes, Die Gedanken Sind Frei, the knowledge that exist in the researcher head is not copyrighted. The knowledge written down in the research paper by the researcher is. The knowledge in the head can however be patented, and so rendered unusable until the property owner has gotten paid.
    - tzs 5y ago
      A researcher discovers, say, a particular way to make a room temperature superconductor, and writes a paper describing how to make it, how they tested it, and giving a theoretical explanation of how it works. That paper would be subject to copyright. However, if I go read the paper at the library and then use the knowledge I gained from it to make the superconductor and test it and applied the theory I learned from the paper to other superconductors, that would not violate the copyright of the paper. If I wanted to then tell you how to make the superconductor, I would not be able to make a copy of the original paper and give you that copy. That would be violating the copyright on the paper. But I could write up my own description of how to make it in my own words, using what I learned from the paper, and give you that description. That would be fine from a copyright point of view.
swiley 5y ago
> We don't GPL scientific research Because people don't hide their equations and ask you to pay them money to evaluate them given a set of variables. GPL wouldn't be necessary if sharing code were already the default. It's an ugly hack to solve a really stupid artificial problem.
xdennis 5y ago
> Why is a few snippets of code so valuable anyway? Microsoft Copilot uses billions and billions of lines of code. It charges for the service. If Microsoft had to license the code, could it build as good a service? If not, then that proves that the (trillions of) snippets have value. Microsoft only says it's valueless because it can get away with stealing it. It's the same way water is free at the river. But that doesn't mean it has no value. Water has price tags in a supermarket. When Google photographs billions of streets it at least provides a public service with it and a way to remove yourself. Centuries ago, petrol was considered useless and a nuisance. Now it's very valuable because it has uses. Microsoft is trying to convince people that petrol has no value exactly because it does. Just because every line of code is worth millionths of a penny, doesn't mean it's valueless when Microsoft will make billions of dollars off of the service.
x4e 5y ago
Your comment gives the impression that GPL is a set of restrictions that developers are unfairly enforcing upon others. In fact it is the complete opposite: it is a set of permissions granted by the developer which otherwise would probably be forbidden by law.
- Aachen 5y ago
  There is no "probably" about it. Law definitely forbids basically everything that GPL allows, that's why GPL exists: to grant those rights. And of course one would then expect the thankful user to at least honor the same terms. Nothing odd about that to me.
  - x4e 5y ago
    I was allowing for cases of different legal jurisdictions or future changes to copyright (which I hope will happen).

by foxes 5y ago

wildmanx 5y ago

I think the key property of the GPL is its virality. It's part of the vision that if I can use some software, I can also change it, to adapt it to my needs, to learn from it, to build something on top of it. I can disassemble my bike, re-assemble it, change it, build on top of it. I want to do that with my browser, my OS kernel and my image viewer too. That this is connected to how "valuable" code is, is in a sense a happy coincidence. The conditions are attached and entangled because this forces people to listen and pay attention. A more "permissive" license (like BSD) is weakening that point. It allows you to reuse code but does not require the re-user to allow the above mentioned openness. So is kind of missing the point, from that perspective.

foxes 5y ago
Well I agree in that regard, but I think in the ideal world everything would just be open, and you would want to incentivise openness so it ends up as the natural state. In a sense by decoupling these issues. Eg Microsoft Windows OS could just be open source. There is some value to having windows itself, but the more valuable thing is the horde of programmers, and all the other resources behind it.
- npteljes 5y ago
  This sounds good. How do you imagine getting there? GPL people think that contributing to this ideal is only worth it if you also require all participants to take part in the ideal. Thereby granting that at least the GPL software stays in that ideal world.
  - scrollaway 5y ago
    It's definitely a tough call. Maybe this is an industry scale problem. The GPL existing hasn't stopped walled gardens from developing, and open source is falling behind in many regards. Maybe the license isn't the correct place to address this. I'm not sure what the solution is, though...
    - npteljes 5y ago
      Proprietary is, I think, the "default" way how people always did things. Keeping the recipe secret is one of the moats that help the enterprise stay relevant. So naturally, "it works" and I think it always will. IMO, this is why it's important to keep the open source stuff open. Closed doesn't usually turn into open, so why should open turn into closed? I think the legal system is exactly the place to address this, in fact, I think the governments themselves should drive the development, as it's their own interest to really own their software and hardware stack.
      - scrollaway 5y ago
        I don't remember who but someone here on HN once compared closed source vs open source to old school secretive alchemy guilds vs real science. I really liked that comparison and I've thought about it a lot since. And here it's worth asking, how did we go from alchemy to science? How did we end up figuring out that openness and transparency is beneficial? I know one thing, it wasn't with licenses. But I'm really curious, because I want to live in the world where open source is the default. I just don't know how we get there.
        5y ago
bluGill 5y ago
> It allows you to reuse code but does not require the re-user to allow the above mentioned openness That is the dogma, but in practice it is false. Khtml was GPL, that didn't stop Apple from forking it in such a way that their changes couldn't be reused. In practice many who change BSD software find it worth while to contribute their changes back - so they can continue to use the latest version instead of maintaining their own fork.
- temac 5y ago
  khtml was and is still LGPLv2.1, webkit still is, and contains major work from Apple that obviously could and has been reused...

darthrupert 5y ago

"Few snippets of code" can be pretty large. I see a lot of value in not pointlessly reinventing the wheel. GPL encourages/enforces people who fix things to give those changes back to upstream where they may benefit everyone. So ideally only one set of fixes per problem has to be globally implemented and maintained. The implications for global wealth are non-trivial. How much Linux alone has done, I wonder?

diegoperini 5y ago
> I see a lot of value in not pointlessly reinventing the wheel. GPL encourages/enforces people who fix things to give those changes back to upstream where they may benefit everyone. One can argue GPL encourages reinventing the wheel by not being permissive enough.

sangnoir 5y ago

> The valuable thing is their time and overall knowledge, not the end piece of content necessarily. The GPL doesn't give primacy to the thing that's built - but to the user of the thing. The GPL implicitly values the user's freedom over developer's convenience - the opposite is true for BSD-like licenses. Generally, I feel like the affinity most people have to one license or another entirely depends on how they more closely identify as a user of. software, vs as a developer

alex_smart 5y ago
>The GPL values the user's freedom, That assumes a very specific kind of user. The kind that has the time and energy, not to mention the ability, to read and understand source code and modify it according to his needs. Basically the RMS kind of user. For most users, that freedom is not much of a concern at all. For them, the most important freedom is: the freedom to enjoy the fruits of labour of a large horde of programmers who can make a fruitful living out of creating apps for said users. They enjoy this freedom a lot more in what GNU calls non-free platforms like Windows and Mac and iPhone, which is why they keep flocking to them.
- marcodiego 5y ago
  > That assumes a very specific kind of user. The kind that has the time and energy, not to mention the ability, to read and understand source code and modify it according to his needs. Actually, no. End users also benefit when code is contributed back to the software they use.
  - alex_smart 5y ago
    Good point. Of course open source software has definitely benefitted users in a wide variety of ways: for instance by providing open source tools to help build said software (can't even imagine a world without free compilers these days). My intent was only to help the GPL and free software fanatics see an important freedom provided by non-free platforms, not to belittle the achievements of open source community.

xunn0026 5y ago

Have you heard of... patents? Why is an idea written on paper so valuable? Why can't people just use it (or, as it often happens, independently rediscover it) and just cite the patent if recognition is required? As for scientific research, most of the software behind scientific paper is basically closed source. Which is another reason why there's a whole reproducibility problem with science papers: it's quite possible many papers are just bad interpretation of data due to bad software. Also note how much software / research / data is proprietary. How can one quote that? I wouldn't mind using only BSD in the future if I get to read the Apple iOS source code on GitHub.

foxes 5y ago
I agree scientific papers should be open source. It is not acceptable there is a proliferation of bad practice. That seems a different issue however.
- xunn0026 5y ago
  It's not a different issue. You used academic papers as the standard we could use for open source and it's provably worse. Even copyright... isn't Elsevier a major copyright holder for all academic papers and thus causing all sorts of problems. It's not like it's all kumbaya out there. The academic world is much worse than GPL and open source on multiple dimensions.
  - jsjohnst 5y ago
    > isn't Elsevier a major copyright holder for all academic papers and thus causing all sorts of problems No, I could be wrong, but it’s my understanding that the author of the academic paper still holds the copyright, it’s just the author licensed sole (or near exclusive) distribution rights to Elsevier in exchange for publication in their journals. It’s basically the same in the publishing world, I own the copyright on the book I coauthored, but WROX Press owns the distribution rights as we gave those up in exchange for them publishing.
  - fastaguy88 5y ago

npteljes 5y ago

If what you say stands, they should also include all of Microsoft's proprietary code in the learning process. Do you see that happening?

drran 5y ago

1) Your analogy is wrong, because none of the software licenses forces a user to have a license for his work. Moreover, a math book has strong copyright protection. You cannot just copy the math book and sell it as your own. GPL allows that. 2) Yep, the whole idea of GPL is to allow those good sources to propagate freely and get reused. Read the license, please. 3) Citation model is the core of BSD license, so use it for your own code.

belorn 5y ago

Scientific research is patented and exclusively copyrighted under publisher. If all scientific research were public and accessible we would not have sci-hub, and they would not currently be in court. We would also not live in a world where scientist has to work around patents and in some cases chose alternative research areas just because they can't afford (or use the result) because of patents. In addition there is a huge controversy in many Universities that patent from research is owned by the university, and if the researcher want to use technology that they themselves discover they need to pay the university. Removing patents and copyright from public funded universities is an on-going struggle that has yet to be fixed. We have yet to start to make all scientific research free from patents and copyright. We should strive towards the utopia where all that was expected is that people will cite the source, and then everyone could build off the work of everyone else.

tzs 5y ago
> Scientific research is patented and exclusively copyrighted under publisher. Research is not copyrighted. Particular write-ups of the research are copyrighted.
- belorn 5y ago
  I am not sure I understand your point of distinction. Yes, Die Gedanken Sind Frei, the knowledge that exist in the researcher head is not copyrighted. The knowledge written down in the research paper by the researcher is. The knowledge in the head can however be patented, and so rendered unusable until the property owner has gotten paid.
  - tzs 5y ago
    A researcher discovers, say, a particular way to make a room temperature superconductor, and writes a paper describing how to make it, how they tested it, and giving a theoretical explanation of how it works. That paper would be subject to copyright. However, if I go read the paper at the library and then use the knowledge I gained from it to make the superconductor and test it and applied the theory I learned from the paper to other superconductors, that would not violate the copyright of the paper. If I wanted to then tell you how to make the superconductor, I would not be able to make a copy of the original paper and give you that copy. That would be violating the copyright on the paper. But I could write up my own description of how to make it in my own words, using what I learned from the paper, and give you that description. That would be fine from a copyright point of view.

swiley 5y ago

> We don't GPL scientific research Because people don't hide their equations and ask you to pay them money to evaluate them given a set of variables. GPL wouldn't be necessary if sharing code were already the default. It's an ugly hack to solve a really stupid artificial problem.

xdennis 5y ago

> Why is a few snippets of code so valuable anyway? Microsoft Copilot uses billions and billions of lines of code. It charges for the service. If Microsoft had to license the code, could it build as good a service? If not, then that proves that the (trillions of) snippets have value. Microsoft only says it's valueless because it can get away with stealing it. It's the same way water is free at the river. But that doesn't mean it has no value. Water has price tags in a supermarket. When Google photographs billions of streets it at least provides a public service with it and a way to remove yourself. Centuries ago, petrol was considered useless and a nuisance. Now it's very valuable because it has uses. Microsoft is trying to convince people that petrol has no value exactly because it does. Just because every line of code is worth millionths of a penny, doesn't mean it's valueless when Microsoft will make billions of dollars off of the service.

x4e 5y ago

Your comment gives the impression that GPL is a set of restrictions that developers are unfairly enforcing upon others. In fact it is the complete opposite: it is a set of permissions granted by the developer which otherwise would probably be forbidden by law.

Aachen 5y ago
There is no "probably" about it. Law definitely forbids basically everything that GPL allows, that's why GPL exists: to grant those rights. And of course one would then expect the thankful user to at least honor the same terms. Nothing odd about that to me.
- x4e 5y ago
  I was allowing for cases of different legal jurisdictions or future changes to copyright (which I hope will happen).