Subject: Re: "Derivative Work" for Software Defined
From: Ian Lance Taylor <ian@airs.com>
Date: 06 Jan 2003 08:11:41 -0800

andre@linux-ide.org writes:

> One of the questions about "Derivative Work" as it relates to binary
> only loadable objects, is the creation of a boundary layer of execution.
> Specifically, the design and publishing an API which properly glues into
> an open source gpl program or kernel(ie loadable modules services) designed
> to provide an execution layer between the GPL and Commerial private code.
> Where as no GPL code in any form is allowed to touch the Commerial code.
> The converse is true, obviously.  The execution layer or boundary.
> Now using this reference from 1995, many companies have gotten legal
> positions about binary modules.
> 
> http://groups.google.com/groups?as_umsgid=4b0rbb%245iu%40klaava.helsinki.fi

What Linus says presumably is valid for Linux.  RMS agrees with that
in the message you forwarded.  It doesn't necessarily apply to any
program other than Linux.  Note in particular the last paragraph in
Linus's message.

> Now what is position can be derived legally in a position or brief to
> prevent binary modules period.  I am against denying anyone the freedom
> of choice to use closed source objects or only open source ones.

This sort of question depends upon the legal definition of ``derived
work.''  Unfortunately, that term is not defined by the law or the
courts for this sort of thing.  Personally I suspect that weight would
be given to explicit statements like those of Linus and RMS--i.e.,
Linux permits binary modules, other GPL programs do not (unless the
author gives explicit permission).

> I ship and sell binary only products, so I have an interest in not
> restricting people.

Other than your customers, presumably.  Restrictions cut both ways.

Ian
--
license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3