[linux-elitists] The breadth of SCO's claims

Karsten M. Self kmself@ix.netcom.com
Fri Jun 27 07:50:51 PDT 2003

Please do not CC me on list mail.

on Fri, Jun 27, 2003 at 09:34:58AM -0400, Aaron Sherman (ajs@ajs.com) wrote:
> On Thu, 2003-06-26 at 12:58, Karsten M. Self wrote:
> > Your copyright license interest in the combined work, under the BSD
> > license, is 'A', because your terms of licensing state that you have no
> > further interest in a combined work.
> I agreed with you up to here. The license specifically states
> restrictions that must be placed on derived works ("...as long as you
> retain..."). Since waiving a right (or rights in this case) cannot be
> done implicitly, you would have to say, "other than these restrictions,
> the copyright holder of this work disclaims all other rights to derived
> works". At least that's my understanding of the law, but I'm not trained
> in such things.
> Think of it this way: If another BSD-licensed project absorbed your
> code, and then a proprietary software company were to absorb THAT code
> and distribute it wihtout retaining your copyright notice, you would
> have the right to sue that third party without having to go through the
> second party. Why? Because they have violated the terms of the license
> on a jointly copyrighted work to which you were a party. *Your* license
> no logner applies, that was a license between you and the second party.
> The *new* license is crafted by the second party, but because of the
> terms of your original license, you must be a party to this new license
> ("source or binary form, as long as..."). 

The new work is still A + B.

17 USC 106(2) provides as an exclusive right of authors the right "to
prepare derivative works based upon the copyrighted work".  Copyright
interest is not lost through the fact that a derived work is now being
prepared at a distance.  Internationally, details differ but are
generally similar for Berne signatories (virtually all nations).

The rationale for "license transmissibility" is debatable.  In the case
of the GNU GPL, this is stated explicitly, that "the recipient
automatically receives a license from the original licensor".  In the
case of the BSD license, we have the case that:

  - There is no explicit transmission of license in the license text

  - The derived work is governed by copyright.

  - Absent an explicit grant of license, neither the 2nd party
    distributor, nor the third party recipient, have rights in the work.

  - A reasonable interpretation of intent of the BSD license is that it
    transmits such right.

So the portion of the derived work which is 'A' is now governed by the
original BSD license.  Plus whatever additional terms have been imposed
by the second party.

> This new license may have simple BSD terms, in which case your
> situation has not changed much.

See above.

> However, if the new license has more restrictive terms, you are now a
> party to those more restrictive terms, and you certainly have the
> right to enforce them, no?

I would disagree. 

These additional terms were not imposed by you, and were not part of
your original grant.  The original BSD work's authorship interest is
governed by the BSD license.

> As far as I know, you cannot distribute a work under two licenses at the
> same time without (as Perl does) allowing the recipient to choose
> between them (either circumstantially or by preference). 

This is grossly incorrect.

As a trivial example, virtually all proprietary software is governed by
a highly complex web of inter-relating licenses.  The recipient has no
option to chose which terms s/he wishes to apply.

The option to choose among licenses, as practiced in FSF Free Software
and OSI Open Source licensing, is strictly a _convenience_ to the
recipient.  It allows for more flexibility in downstream application of
a work, as well as compatibility with existing restrictive licenses
(most often the GNU GPL).  By dual (or multi-) licensing a work, the
recipient is governed by _the union_ of rights granted under _any_ of
the individual licenses, _if_ these apply to the work as a whole.  This
is as distinguished from, say, a combined work under a license allowing
partial coverage, say, the MozPL.  In the case of the GNU LGPL, the
option to relicense under the GNU GPL allows for one-way transitivity
from the GNU LGPL to the GNU GPL for an LGPL'd work, when combined with
one licensed under the GPL but _not_ the LGPL.

The usual redistribution option is then to continue offering the
multiple choice of license, preserving this compatibility, so long as
the original multiple licensed work is not being combined with a work
under terms conflicting with one or more of the multiple licenses.

By licensing a work, such as Mozilla or OpenOffice.org, to take to
actual examples, under both the MozPL and the GNU GPL, the option is
left open to take the original work and combine it on a file-by-file
basis with a proprietary or otherwise restrictive license (applying the
MozPL), *OR* with a GPLd work (applying the GPL).  It's an approach to
flexibility in licensing.

The offer of "at your option" is *strictly* a convenience to the
recipient of the work.

> Thus, you must view the BSD and Windows licenses as a collective
> license over the Windows code, no?


I view these as a mess.  You'd need a lawyer untangle what specificly
applies to which pieces of the work.



Karsten M. Self <kmself@ix.netcom.com>        http://kmself.home.netcom.com/
 What Part of "Gestalt" don't you understand?
    Windows Refund Day II:  fight for your right to refund
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://allium.zgp.org/pipermail/linux-elitists/attachments/20030627/5a014df8/attachment.pgp 

More information about the linux-elitists mailing list