5 ms·

How would this work? Any such law has limited jurisdiction. There are already many examples of software development moving to avoid problematic laws. There ar

How would this work? Any such law has limited jurisdiction. There are already many examples of software development moving to avoid problematic laws. There are circumstances where further maintenance is all but impossible that happen in the real world. Open sourcing software cannot always grant all rights required to use said software legally, many of which may not be owned by the company. Some software has dependencies on third-parties that can void the relationship and for which no alternative exists, in some cases because the capabilities are unique and trade secrets. This either creates a de facto requirement to support all released versions of software indefinitely or a giant and easily exploitable loophole that renders it useless. How does removal of features, common in long-lived software, work under this regime? It would be nearly impossible to define "maintenance" in law such that it doesn't lead to de facto non-maintenance that also doesn't create unlimited liability that courts are unlikely to enforce. Furthermore, this would violate laws that expressly prohibit disclosure of code in some cases, that are unlikely to be repealed anytime soon. Any such exceptions would be widely exploited. This policy would create some perverse incentives and a massive number of loopholes.

jedberg 5y ago
I'm not sure why any of that stuff should be protected in the first place. 100 years ago, a company could build a machine. It could be full of "trade secrets". But at the end of the day, anyone could tear it apart and figure out how it was made and how to repair it. If you were willing to fix it, you could use that machine forever. Why should a modern "machine" be any different? Why should how the machine works be legally protected after the company no longer makes it or supports it? This is basically just an extension of the right to repair argument.
WorldMaker 5y ago
It could work like intellectual property law was meant to work in the first place? Patents are a limited monopoly granted in exchange for all of the technical details of the invention so that the invention can be built by others after the monopoly period expires. "All of the technical details" for a software patent should be expanded explicitly require "all of the relevant source code" and for good measure "relevant" should be described widely and probably "virally" (ala "copyleft"; maybe requiring a "copyleft" license?). Copyright is a limited monopoly granted in exchange for helping to insure that the product eventually is available to use by the public domain. For many years the US had a legal requirement for active copyright that a copy of every book be sent to the Library of Congress for cataloging. (This is no longer a requirement, though some US publishers still do it as a courtesy.) A requirement could easily be made that for software copyrights to be enforced in a court of law that same software must have all source code registered with a US Software Source Code Depository. We collectively could do a lot of interesting things to build automated tools for that. Back in the day when copyright required published books to be sent to the Library of Congress not every publisher followed the law and so it was exploited some, but if you get caught (exploiting it) not registering your source code and then get taken to court over it: you lose your copyright immediately, the court has the power to order you to hand over your source code, and at the very least makes it fair game as a part of the public domain to reverse engineer your software if your source code can't be found. On the surface it sounds like a great idea. It's also a very unlikely idea because it would require new legislation and would be opposed by many lobbies of corporate lawyers.
- jandrewrogers 5y ago
  This presumes that the software companies own all of their intellectual property. Most software companies don't own all the intellectual property that goes into their software or even all of the source code. Open source without the dependent intellectual property is a lot less useful and may be subject to incompatible legal regimes.
  - WorldMaker 5y ago
    It doesn't presume anything of the sort. The dependent intellectual property is still intellectual property and would be subject to the exact same considerations. (Per the proposal: If copyright, code for that dependent IP would need to be registered with a source code depository or be subject to forfeiture of said copyright in a court of law; if patented the code would need to be open source and maybe even copyleft to begin with.) The public may need to play whack-a-mole to access the full source code of larger projects, bring dependents and then their dependents and so on to court until the entire thing is open source, but that's still a better deal than the current situation where the public domain is owed nothing ever. (ETA: And "incompatibility between legal regimes" is mostly moot due to the Berne Convention anyway, at least per my lay understanding of it. That said, the proposal is entirely hypothetical and no such legislation would likely pass in any country so worry about international complications is mostly cart-before-the-horse.)

jedberg 5y ago

I'm not sure why any of that stuff should be protected in the first place. 100 years ago, a company could build a machine. It could be full of "trade secrets". But at the end of the day, anyone could tear it apart and figure out how it was made and how to repair it. If you were willing to fix it, you could use that machine forever. Why should a modern "machine" be any different? Why should how the machine works be legally protected after the company no longer makes it or supports it? This is basically just an extension of the right to repair argument.

WorldMaker 5y ago

It could work like intellectual property law was meant to work in the first place? Patents are a limited monopoly granted in exchange for all of the technical details of the invention so that the invention can be built by others after the monopoly period expires. "All of the technical details" for a software patent should be expanded explicitly require "all of the relevant source code" and for good measure "relevant" should be described widely and probably "virally" (ala "copyleft"; maybe requiring a "copyleft" license?). Copyright is a limited monopoly granted in exchange for helping to insure that the product eventually is available to use by the public domain. For many years the US had a legal requirement for active copyright that a copy of every book be sent to the Library of Congress for cataloging. (This is no longer a requirement, though some US publishers still do it as a courtesy.) A requirement could easily be made that for software copyrights to be enforced in a court of law that same software must have all source code registered with a US Software Source Code Depository. We collectively could do a lot of interesting things to build automated tools for that. Back in the day when copyright required published books to be sent to the Library of Congress not every publisher followed the law and so it was exploited some, but if you get caught (exploiting it) not registering your source code and then get taken to court over it: you lose your copyright immediately, the court has the power to order you to hand over your source code, and at the very least makes it fair game as a part of the public domain to reverse engineer your software if your source code can't be found. On the surface it sounds like a great idea. It's also a very unlikely idea because it would require new legislation and would be opposed by many lobbies of corporate lawyers.

jandrewrogers 5y ago
This presumes that the software companies own all of their intellectual property. Most software companies don't own all the intellectual property that goes into their software or even all of the source code. Open source without the dependent intellectual property is a lot less useful and may be subject to incompatible legal regimes.
- WorldMaker 5y ago
  It doesn't presume anything of the sort. The dependent intellectual property is still intellectual property and would be subject to the exact same considerations. (Per the proposal: If copyright, code for that dependent IP would need to be registered with a source code depository or be subject to forfeiture of said copyright in a court of law; if patented the code would need to be open source and maybe even copyleft to begin with.) The public may need to play whack-a-mole to access the full source code of larger projects, bring dependents and then their dependents and so on to court until the entire thing is open source, but that's still a better deal than the current situation where the public domain is owed nothing ever. (ETA: And "incompatibility between legal regimes" is mostly moot due to the Berne Convention anyway, at least per my lay understanding of it. That said, the proposal is entirely hypothetical and no such legislation would likely pass in any country so worry about international complications is mostly cart-before-the-horse.)