1. 28
  1.  

  2. 7

    This is fud. I won’t speculate as to why OSIA is doing this.

    The relevant text (not having regard to the amendments in the CPTPP “wrapper” treaty, which I don’t think are relevant):

    Article 14.17: Source Code

    1. No Party shall require the transfer of, or access to, source code of software owned by a person of another Party, as a condition for the import, distribution, sale or use of such software, or of products containing such software, in its territory.
    2. For the purposes of this Article, software subject to paragraph 1 is limited to mass-market software or products containing such software and does not include software used for critical infrastructure.
    3. Nothing in this Article shall preclude: (a) the inclusion or implementation of terms and conditions related to the provision of source code in commercially negotiated contracts; or (b) a Party from requiring the modification of source code of software necessary for that software to comply with laws or regulations which are not inconsistent with this Agreement.
    4. This Article shall not be construed to affect requirements that relate to patent applications or granted patents, including any orders made by a judicial authority in relation to patent disputes, subject to safeguards against unauthorised disclosure under the law or practice of a Party.

    The real meat is 14.17.1: No Party [signatory state] may require the transfer of, or access to, source code of software owned by a person of another Party [person of another party = person domiciled (the actual nature of the connection may be different) in another signatory state] as a condition for the import, distribution, sale or use of such software, or of products containing such software, in its territory.

    That bolded phrase again: No […] access to[,] source code […] as a condition for the import, distribution, sale or use of such software, or of products containing such software.

    This is specifically a prohibition on requiring sellers of software from having to show anyone else the source in order to be allowed to “import” their software. It does not preclude any private party from including any license term, including one requiring the sharing of source code as a precondition for the distribution of derivative works (which is what the GPL seeks to protect). Even if this did somehow make the GPL unenforceable, I cannot see how it would in any way affect non-viral licenses like the Apache License.

    14.17.2 confirms that reading: states may impose such a condition (of releasing the source) on the importation of software “used for critical infrastructure” (whatever that may mean).

    Honestly, I think the only parties who should be worried about this provision are closed source makers of infrastructure items like routers or operating systems, as this specifically authorises the imposition of source disclosure requirements on such items.

    The trend with such arrangements is that signatory states tend to align under relatively common approach. Perhaps Windows will have to be open sourced in the TPP zone; or perhaps there will be a series of side letters and agreements that prevent any forced source disclosure except under very defined circumstances.

    The OSIA documents linked in the article do not include any legal analysis as to why they think 14.17 prohibits the enforcement of source code sharing provisions in licences. (http://osia.com.au/f/osia_sub_201805_sscfadt.pdf and http://osia.com.au/f/osia_cptpp_pr6a.pdf).

    Academic disclaimer: I have not performed a full analysis of the whole text

    Legal disclaimer: this is not legal advice, this is an off-hand academic analysis by an internet random. This does not indicate any course of action for any person, and is not in any way intended to apply to your, or anyone else’s situation.

    1. 3

      If a specific class of licenses becomes unenforceable, what does that actually mean to software currently under that license?
      Does the software, in that jurisdiction, simply behave as if it has no license?
      Or would the copyleft /portion/ of the license simply no longer apply (eg. people would now get it without those restrictions)?

      1. 3

        Those would have been good questions to answer, but alas. It seems like the “give away” parts are still in effect, but the “give back” condition is what is ruled invalid, but that’s a vague guess.

        1. 2

          Most open source licenses revert to plain old copyright if the license is broken.

        2. 2

          AFAIK no valid licence means “all rights reserved” all over the world.

          This means that if GPL is not valid in a country, no one can use that software in that courty.

          1. 1

            Not always. e.g. the creators of the EUPL argue that under EU law, GPL, LGPL, MPL, etc are all equivalent and compatible in any direction (because you can’t enforce copyleft on anything above file level)

            1. 1

              Can you share some legal sources to back this statement?

              I’ve never listen such argument, but I have listened an opposite one, stating that software copyright is based on literary one, so that you can translate a C GPL software into Python only under GPL, as a derived work (with a caveat I do not remember). That’s different from common wisdom, as such protections are usually demanded to patents not copyright.

              So I’m very curious about your sources.

              And BTW file level copyright doesn’t means that, without a license you can use that file…

              1. 2

                And BTW file level copyright doesn’t means that, without a license you can use that file…

                Correct, but it means that e.g. the viral part of the GPL is entirely gone, as it only applies to a single source file, not to any binary artifacts, intermediate compilation products, or other sources.

                See https://joinup.ec.europa.eu/collection/eupl/eupl-compatible-open-source-licences#section-3 on why clauses on linking, in the eyes of the authors of the EUPL, have no legal validity.

                1. 1

                  Thanks this has been a great read for several reasons. I’ll double check with an European lawyer.

                  If this interpretation is confirmed, I will probably write an new stronger copyleft for my code instead of using AGPLv3.

              2. 1

                We’ve had this discussion before, and I don’t think that your interpretation of the interpretation of the EUPL authors is correct.

                https://lobste.rs/s/oroz5k/google_loses_android_battle_could_owe#c_wxup7r

                1. 1

                  In the discussion you admitted that you interpret it the same way, as them saying that based on written law, this would be the case, but there is no case law.

                  Generally, under civil law (which is what’s used in the relevant EU countries), decisions from courts rarely rely on previous court rulings, only on the written law. That’s why the US-american fixation on case law is a bit weird (and not exactly helpful).

                  1. 1

                    GPLv3 was written with consideration to EU law. It is why several of the legal terms in it were changed, to fit a more global perspective than GPLv2. I think Eben and his team are fairly competent at international law. I am aware of some difference between the legal traditions around the world, and I still think it’s uncertain that your interpretation of their interpretation is accurate.

                    I don’t see consensus that copyleft is invalid in Europe. It remains, as far as I can tell, a minority opinion, and despite the different usage of jurisprudence in Europe, successful legal challenges against copyleft still would give some credence to the interpretation that strong copyleft is currently null in Europe (with the patronising insinuation that copyleft is just some silly colonial idea).

                    After all, the VMWare case was brought up in Germany and it wasn’t immediately thrown out because copyleft is considered null. The judge seemed to consider copyleft to be worthy of consideration, if Hellwig could demonstrate sufficient copyright ownership over his kernel code.

                    1. 1

                      (with the patronising insinuation that copyleft is just some silly colonial idea).

                      What do you mean?

                      Btw, the EUPL authors does not say that GPL be void. Just the reciprocity would be invalidated (what some people call “virality”).

                      1. 2

                        That’s why the US-american fixation on case law is a bit weird (and not exactly helpful).

                        This statement sounds a bit like a European doesn’t like that a foreign legal document with its silly foreign ideas is being discussed in Europe.

                        One aspect of the hereditary nature of the GPL is exactly what was being considered in the VMware case in Germany. The judge there didn’t seem to think that strong copyleft was unworthy of consideration.

            2. 1

              In early 2017, the US withdrew from the treaty, which its President had previously described as a “terrible deal”.

              In before: The FSF proclaims Donald Trump to be a supporter of free software.

              1. 1

                Maybe another good reason to be selling GPL software. If copyleft disappears, nothing changes for the customers that are using it without modifications. That will be most.