When to evaluate dual licenses

Michael Poole mdpoole at troilus.org
Sat Dec 15 14:39:41 UTC 2007


Michael Tiemann writes:

> On Dec 13, 2007 6:17 PM, Matthew Flaschen <matthew.flaschen at gatech.edu> wrote:
>
>
>     But, again, "GPLv2 or later" is not a software license and is outside
>     OSI's scope.
>
>
> You make a fascinating point.  I would agree that "GPLv2 or later" is outside
> our scope, because we cannot predict whether future versions of the GPL will be
> OSD-compliant or not.  However, in a different case, such as "GPLv2 or any
> other OSI-approved, copyleft license", may well be within our scope.

I do not see a significant difference.  If A is an OSD-compliant
license, allowing distribution under terms "A or B" versus "A or C"
makes no difference: both are OSD-compliant.  (Similarly, in Boolean
logic, if A is true, "A or B" and "A or C" are both true, regardless
of B's value or C's value.)  It only becomes a problem when someone
purports to remove A from the set of license options.

For redistributors, I am not sure that excluding the first license is
permissible with that kind of wording, and it may be outside the scope
of OSD compliance: it would effectively be a sublicense rather than
distribution under the (OSD-compliant) original license.

Michael Poole



More information about the License-discuss mailing list