Skip to main content
Open Source

Return to Answer

added 51 characters in body
Source Link
Max Morehead
Max Morehead
  • 31
  • 1
  • 1
  • 5

The two topics are entirely orthogonal, in my understanding. The debate about whether APIs can be copyrighted is about whether program interfaces are copyrightable even without the implementation of those interfaces. In the case of Oracle v. Google, the question was whether Google could duplicate class and method names without infringing Oracle's copyright. For example, is a method signature like

void <E> quicksort(List<E> items);

along with many other similar method signatures protected under copyright?

The issue with ZFS and Linux is unrelated as far as I can tell. In that case, all parties agree that Linux is copyrightable and ZFS is copyrightable because they are more than mere APIs. Instead, Linux and ZFS are implementations of APIs, along with the specification of the APIs.

To continue this point, as an example, I think people who would argue API are not copyrightable would say it would be okay for someone (e.g. Canonical) to completely reimplement ZFS with the same interface and release the reimplementation under any license. If that license were GPLv2 compatible, then it could be released together with Linux with no issue.

Whether you agree with Kuhn's characterization of "ZFS+Linux" as a combined, derivative work, rather than an aggregation of two separate programs is a separate question.

The two topics are entirely orthogonal, in my understanding. The debate about whether APIs can be copyrighted is about whether program interfaces are copyrightable even without the implementation of those interfaces. In the case of Oracle v. Google, the question was whether Google could duplicate class and method names without infringing Oracle's copyright. For example, is a method signature like

void <E> quicksort(List<E> items);

along with many other similar method signatures protected under copyright?

The issue with ZFS and Linux is unrelated as far as I can tell. In that case, all parties agree that Linux is copyrightable and ZFS is copyrightable because they are more than mere APIs. Instead, Linux and ZFS are implementations

To continue this point, as an example, I think people who would argue API are not copyrightable would say it would be okay for someone (e.g. Canonical) to completely reimplement ZFS with the same interface and release the reimplementation under any license. If that license were GPLv2 compatible, then it could be released together with Linux with no issue.

Whether you agree with Kuhn's characterization of "ZFS+Linux" as a combined, derivative work, rather than an aggregation of two separate programs is a separate question.

The two topics are entirely orthogonal, in my understanding. The debate about whether APIs can be copyrighted is about whether program interfaces are copyrightable even without the implementation of those interfaces. In the case of Oracle v. Google, the question was whether Google could duplicate class and method names without infringing Oracle's copyright. For example, is a method signature like

void <E> quicksort(List<E> items);

along with many other similar method signatures protected under copyright?

The issue with ZFS and Linux is unrelated as far as I can tell. In that case, all parties agree that Linux is copyrightable and ZFS is copyrightable because they are more than mere APIs. Instead, Linux and ZFS are implementations of APIs, along with the specification of the APIs.

To continue this point, as an example, I think people who would argue API are not copyrightable would say it would be okay for someone (e.g. Canonical) to completely reimplement ZFS with the same interface and release the reimplementation under any license. If that license were GPLv2 compatible, then it could be released together with Linux with no issue.

Whether you agree with Kuhn's characterization of "ZFS+Linux" as a combined, derivative work, rather than an aggregation of two separate programs is a separate question.

Source Link
Max Morehead
Max Morehead
  • 31
  • 1
  • 1
  • 5

The two topics are entirely orthogonal, in my understanding. The debate about whether APIs can be copyrighted is about whether program interfaces are copyrightable even without the implementation of those interfaces. In the case of Oracle v. Google, the question was whether Google could duplicate class and method names without infringing Oracle's copyright. For example, is a method signature like

void <E> quicksort(List<E> items);

along with many other similar method signatures protected under copyright?

The issue with ZFS and Linux is unrelated as far as I can tell. In that case, all parties agree that Linux is copyrightable and ZFS is copyrightable because they are more than mere APIs. Instead, Linux and ZFS are implementations

To continue this point, as an example, I think people who would argue API are not copyrightable would say it would be okay for someone (e.g. Canonical) to completely reimplement ZFS with the same interface and release the reimplementation under any license. If that license were GPLv2 compatible, then it could be released together with Linux with no issue.

Whether you agree with Kuhn's characterization of "ZFS+Linux" as a combined, derivative work, rather than an aggregation of two separate programs is a separate question.