For Approval: Open Vendor Public License (OVPL) and Open Vendor Lesser Public License (OVLPL)

[Alex Bligh]

--On 30 June 2005 09:10 -0400 mdpoole <mdpoole at troilus.org> wrote:

> As Peter Muldave showed, add-ons to the GPL and LGPL are used by the
> FSF in several components of the GNU system.  For example, libstdc++

That's the library linking problem, which I think is different, and derives
solely from how the GPL is framed (in that linking non-GPL code with GPL
code is at the least a gray area). It's a problem which is (we think)
solved in the OVPL/OVLPL anyway. The problem here isn't about LINKING the
licensed software to something else, it's about use of the licensed
software producing something which in itself is (without a specific
exception being made) a derived work of the licensed work (because it
includes 'bits of the original code'), and thus may (unintentionally) fall
under the terms of the license. Examples would include the builtins in a C
compiler (I'm *not* talking about standard libraries).

I recognize that in practice Matthew Seth Flaschen's objection could be
addressed in the same way with the OVPL/OVLPL: i.e. just say "if you
are initial developer and your project produces code, remember to put
in a special exception in the form of an additional grant if you
want to make it clear that programs produces by running your project's
code are not meant to be subject to the license". Indeed, this is
doubly practical with the OVPL/OVLPL because the license-back provision
allows the ID to do this even for a project that has such code
added later (for instance an interpreted language which has a compiler
added). What I am trying to work out is whether there is a benefit of
making the exception explicit in the license itself. My two problems are

1. Might someone not want to make such an exception? What if the result
   of running the program was always 99% derived work, and 1% addition?
   For instance an executable that created a copy of itself by minor
   self-modification?

2. Is the potential exception sufficiently clear-cut that a general term
   can be inserted whose effect will always be certain?

Currently I am havering on the side of "if the ID wants an exception,
the best thing to do is for them to put their own exception in as
an additional grant", but I'm quite prepared to be persuaded.
Comments? Matthew?

Alex