Announcement

Collapse
No announcement yet.

Clutter's Cogl Relicensed To Be More Permissive

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • #16
    Originally posted by GreatEmerald View Post
    The first point applies to copyleft software just as well. But in case of permissive licenses, corporations can and do choose to not disclose their code in order not to have it benefit their competitors. And that's a loss for everyone but the corporation (the corporation itself gets a lot of work done for free).
    The second point is questionable. Their choices should never change, because there is no reason for them to change. If they decide on a license given enough thought, they shouldn't need to reverse their decision and back out of it at any point in time. As for legal cost and risk, I see no risk anywhere, and there are still costs either way. The only difference is license length (which is needed for copyleft, since it relies on certain nuances of copyright law), and the GPLv3 is written well enough that it covers a lot of corner cases (which is another reason why it's long).
    You forgot to mention that copyleft licenses often introduce incompatibilities that aren't foreseen.
    The large swathe of GPLv2 only projects have effectively killed off the potential uptake of GPLv3 because code would be incompatible between the two.
    Personally, I think that the author shouldn't be allowed to bind by version, if they want to use someone else's open source license, they should also have to accept the terms of later licenses for compatibilities' sake.

    Comment


    • #17
      Originally posted by intellivision View Post
      You forgot to mention that copyleft licenses often introduce incompatibilities that aren't foreseen.
      The large swathe of GPLv2 only projects have effectively killed off the potential uptake of GPLv3 because code would be incompatible between the two.
      Personally, I think that the author shouldn't be allowed to bind by version, if they want to use someone else's open source license, they should also have to accept the terms of later licenses for compatibilities' sake.
      That's why you never use GPLv2 only or GPLv3 only. Because GPLv2+ and GPLv3+ is the same as GPLv2 only and GPLv3 only, except it allows linking to GPLvN++ software.

      Comment


      • #18
        Originally posted by GreatEmerald View Post
        The first point applies to copyleft software just as well. But in case of permissive licenses, corporations can and do choose to not disclose their code in order not to have it benefit their competitors. And that's a loss for everyone but the corporation (the corporation itself gets a lot of work done for free).
        The second point is questionable. Their choices should never change, because there is no reason for them to change. If they decide on a license given enough thought, they shouldn't need to reverse their decision and back out of it at any point in time. As for legal cost and risk, I see no risk anywhere, and there are still costs either way. The only difference is license length (which is needed for copyleft, since it relies on certain nuances of copyright law), and the GPLv3 is written well enough that it covers a lot of corner cases (which is another reason why it's long).
        You still got it wrong You see the loss in the code a company writes but does not contribute back. The actual loss is the code the company never writes and so never contributes, because it does not want to deal with GPLv3 code (their resentment appears more limited towards GPLv2). It is exactly the same as with piracy: pirated copy != lost sale. Yes, corporations keep proprietary patches on permissive projects. That does not mean they do not contribute, nor that they contribute less than either full proprietary or copyleft.
        Case in point: Cogl has been written almost exclusively by companies (Intel, Red Hat and Colabora), and the copyright owners still want to license the whole thing as permissive.

        Regarding your rebuttals of my point, yes, point 1 applies to copyleft and permissive, and point 2 to proprietary and permissive, hence two points
        Also, feel free to ask your upper management if they think decisions in a company are permanent, and then ask your legal team how they feel about dealing with GPLv3 code. Your mileage may vary, but they might disagree with you, and incidentally they might be the ones deciding on such matters.

        Comment


        • #19
          Originally posted by carewolf View Post
          Apparently not, if they did, LGPL would never ever be a problem. It is only a burden if companies DO NOT want to contribute back.
          Who contributes to LLVM, X11, wayland, Android Mono and AFS software then? No company contributed code in there?

          Comment


          • #20
            Originally posted by GreatEmerald View Post
            That's why you never use GPLv2 only or GPLv3 only. Because GPLv2+ and GPLv3+ is the same as GPLv2 only and GPLv3 only, except it allows linking to GPLvN++ software.
            This is an issue of trust. Even though I trust the FSF more than many other corps/orgs, it would be foolish of one to sign a paper you have not read (ie: the future license version). If GPLv4 changes in a direction you don't like, well tough titties if you agreed to it in advance.

            For the record I do think GPLv3 improved on v2, but there was no way to know that before it was released.

            Comment

            Working...
            X