Subject: Re: Re: Question Regarding GPL
From: Rick Moen <rick@linuxmafia.com>
Date: Fri, 20 Jan 2006 22:31:44 -0800

Quoting Mahesh T. Pai (paivakil@yahoo.co.in):
> Rick Moen said on Fri, Jan 20, 2006 at 03:58:42PM -0800,:
> 
>  > To my  knowledge, his most  recent statement was on  2002-10-17, as
>  > follows (in part):
> 
> Do have a look at
> http://people.redhat.com/arjanv/COPYING.modules

Thank you, Mahesh.  I will be incorporating those posts into my own archive.

> Also,  http://www.wasabisystems.com/gpl/lkmquotes.html   might  be  of
> interest to most of us.  The ToC for the above is at 
> http://www.wasabisystems.com/gpl/ 

I would respect much more what our NetBSD-using friends at Wasabi
Systems (publishers of excellent NetBSD CD sets, by the way) have to say
on the subject if they would spend more time citing the exact wording of 
relevant parties in proper, full context and less time editorialising.

But at least a couple of the hyperlinks are useful.  

As to the quotation from Prof. Moglen:  While of course I respect his 
views and works highly, obviously his opinion on what is and is not a
derivative work (in relation to the kernel) is even more irrelevant than
is Linus Torvalds's.

And, I would (greatly) presume to say, clearly wrong in that instance.

Let's work with the sort of hypothetical edge case that Torvalds
mentioned in his 1998 Linux Gazette interview:  Company X, having more
money than brains, decides to port proprietary device driver foo.c with a
long previous history in some other *ix, and does so with proprietary,
written-from-scratch C header files, using none whatsoever from the
Linux kernel, nor any other source code from a GPLed codebase.

Company X now compiles binary foo.ko files for a number of
distributions, CPU architectures, and distro versions, in the manner
described by Greg K-H, and lists them for public ftp.  A kernel coder
then sues for copyright infringement -- but on what grounds?  Facts will
show that foo.c simply was not derived from Linux at all, and even the
new *.h files are new, independent works.

That was constructed deliberately to be an edge case:  Torvalds's point
in the 2002 and 2003 postings is that the matter always hinges on a factual
determination about derivation.  I've stacked my example's set of facts.

However, Prof. Moglen denies that such an example could exist:  I say
that's perhaps wishful thinking, but certainly nothing like reality.

-- 
Cheers,             
Rick Moen                 "Anger makes dull men witty, but it keeps them poor."
rick@linuxmafia.com                                   -- Elizabeth Tudor