LGPL clarification

[kmself at ix.netcom.com]

on Wed, Nov 01, 2000 at 06:06:19PM +0100, Mark Wielaard (mark at klomp.org) wrote:
> Hi,
> 
> On Wed, Nov 01, 2000 at 11:13:38AM -0500, Bryan George wrote:
> 
> > - Assuming there are no such circumstances, are there any GPL-compatible
> > licenses that behave similarly to the LGPL, but without the "obnoxious
> > modification and reverse engineering clause"?
> 
> I am not sure about your interpretation of the LGPL. But you might be
> interested in the way the FSF distributes for example guile and classpath.
> They are distributed as GPL plus an exception when you link the whole
> library. It could be seen as a simplified version of the LGPL.

In addition to the two examples given, the Linux kernel itself contains
an exception to allow linking of proprietary drivers (in non-source
form) directly with the Linux kernel.  This sort of exception has been
used in a number of other contexts.

I'll let the lawyers apply the final gloss, but my understanding is
this:

  - The GPL claims rights of authorship under copyright law, then cedes
    some of these rights back to the recipient of a copy, provided rules
    are adhered to.

  - A special exception applies yet another level of ceding of rights
    back to recipients, under special conditions.  I believe this allows
    compatibility of the base source with GPLd code, while allowing for
    broader application in certain cases (though a codebase in which the
    exception has been exercised -- e.g.:  non-GPLd code is included,
    might not be GPL compatible itself).

Other options might include using an alternative license -- the Mozilla
style was designed to address issues similar to those you seem to be
tackling.  A BSD (w/o advertising clause) license might be good for
promoting a technology, though it does allow for a proprietary
derivative of the work to be created (feature/bug discussion omitted).



I'd be interested in seeing a discussion of the relative merits of a
BSD/GPL dual license for cases like this.  The rationale is as follows.

The GNU GPL is a license which has the advantage of denying any single
entity a control locus.  As Chris Rasch wrote recently on the FSB list,
it avoids a prisoners' dilemma -- neither side in a transaction can get
an upper hand.  Others have noted that the GPL avoids creating any
points of control.  This creates a credible trust zone, if you will.

However, one of the foundations on which free software is based is open
standards.  The BSD/MIT licenses are essentially standards-propogation
licenses.  While the GPL creates credibility, there may be aversions to 
adoption, as Bryan indicates.  Yoking the two licenses together might be
the sugar-coating that the GPL needs.  BSD means that the terms of the
GPL need not persist into proprietary implementations of the codebase.
However, the existence of a GPLd primary base means that any codefork
which is attempted under BSD or other terms must compete against the
credibility of assured public existence of the GPLd variant.

This is an idea I've been rolling around for a while, there are a couple
of possible holes in it, but I'd be interested in seeing a read from
others.

IANAL, this is not legal advice.

-- 
Karsten M. Self <kmself at ix.netcom.com>     http://www.netcom.com/~kmself
 Evangelist, Opensales, Inc.                    http://www.opensales.org
  What part of "Gestalt" don't you understand?      There is no K5 cabal
   http://gestalt-system.sourceforge.net/        http://www.kuro5hin.org
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 232 bytes
Desc: not available
URL: <http://lists.opensource.org/pipermail/license-discuss_lists.opensource.org/attachments/20001101/9af4c07c/attachment.sig>