Subject: Re: For Approval: BSD License, PostgreSQL Variant
From: "Chris Travers" <chris.travers@gmail.com>
Date: Fri, 12 Oct 2007 11:14:38 -0700

I am going to reply to Denial Wallace's point here primarily because
other GPL-supporters on this list seem to agree with him on the scope
of copyleft.

On 10/12/07, dlw <danw6144@insightbb.com> wrote:

>  See the "all third parties" who are the intended third party beneficiaries?
> They must be "caused" by the licensee to be bound
>  "under the terms of this License". That's the heart of the "copyleft
> theory" of Richard Stallman -- an unbounded recursive contract to make a
> contract to make a contract to make a contract make a contract. . . ad
> infinitum.
>
Not really.  You *may* have a case as regards the GPL v3 specifically,
but the GPL v2 is clear that it only applies to covered derivative and
compiled works, not to all pieces under the license.

This means that there is no problem with a BSDL project like
PostgreSQL offering GNU Readline support, but that commercial spinoff
projects may not be able offer that support, for example, without
running into the threat of lawsuit by those who think that dynamic
linking equals derivation (I am not convinced, but IANAL).

NOTHING in the GPL v2 states that you can only distribute your own
copyrighted elements under that license.  You can offer them under
whatever terms you want, but if you distribute them in a derivative
work containing copyrighted elements from another GPL'd program,
*then* the GPL kicks in.

Anyone can break the cycle simply by deciding to do offer their
specific changes under licenses which offer additional rights.  The
GPL v3 may even leave this loophole intact *if* section 7b legal
notices can be used to restrict the effects of permission removal
(under section 7, paragraph 2).

Best Wishes,
Chris Travers