Subject: Re: New angle on the patent problem
From: "Karsten M. Self" <>
Date: Mon, 20 Sep 1999 10:07:32 +0000

Returning from vacation, apologies for the delayed response.

Assaf Arkin wrote:

> > > GNU covers a lot of software, but surprisingly I can get through the day
> > > without using any GNU code. Unfortunately, many of the companies we are
> > > afraid will come with a patent claim are not well known GNU users.
> >
> > These are examples.  A key IMO is finding the right class.
> I just realized that's impossible and sent a separate e-mail to explain
> why.

Seperate email received, but no particular explanation found.  The
quoted content here appears more complete, but not compelling.

> The patent law is supposed to give coverage to companies that would like
> to protect their assets. That is the spirit of the law and that is how a
> court would interpret it.
> Now, let's assume that most companies out there use Linux. Let's assume
> company X files patent suite against Emacs for violation of a patent it
> owns.
> In your proposal, the GPL would immediately terminate the right of
> company X to use Linux, which means they could not go about their
> business, which means they cannot for all intents and purposes ever
> claim patent infrignment, whether valid or not.

Not quite sure who you're replying to, but this wasn't quite my

1). License for use of Linux specifies that patent infringement action
against any of a class of software works will trigger the termination
clause.  The Emacs license cannot terminate the right granted by another
license to another work -- Jonathan and I hashed this out earlier.  Even
if the text of the Emacs and Linux licenses is identical, the Emacs
license pertains to Emacs and the Linux license to Linux.  My goal is a
license to _a specific work_ which has a termination trigger based on
actions to _a class of works_.

2). The company can certainly claim patent infringement, but they have
to accept the consequences of this claim.  The existence of a defense
doesn't make an action impossible, it merely attaches a cost to it.
> A court will recognize that as a limitation that goes against the spirit
> of the patent law, and the court will allow the company to file the
> suite by simply disregarding the clause in the license. Thus, the
> company will be able to both claim patent infrignment and contrinue
> using Emacs/Linux.

3). Conveniently ignoring the fact that the spirit of copyright is that
an author (or more specifically, copyright holder) has the right to
control the making of copies of his or her works, subject to certain
limitations which may be considered sufficiently strong to prevent
wholesale appropriation of a work as you suggest.  Not being intimately
familiar with law, I'm not sure exactly how a license can be made
seperable or inseperable (valid in part if a portion is found not
binding, or not valid in any part if a part is found not binding).  

However I believe Assaf's assertion is naive.  IMO a court would have to
balance the powers of copyright and patent, and might well find that the
two parties are at an impasse -- the copyright holder cannot practice
art without infringing the patent, and the patentholder cannot use the
software without violating copyright.  Copyright and patent statute,
much less the Constitution, will not be destroyed if this is the case.

Moreover, while I am familiar with instances in which the courts have
relaxed the power of a company to exercise its limited monopoly granted
by a patent (Xerox suffered this fate in the 1970s, possibly leading to
some of the circumstances giving rise to Apple, and WIMP interfaces due
to poor IP management), I don't know of any instances in which a
reasonable claim to copyright (significant work of original authorship
fixed in tangible media) was been rescinded by the courts.

> Now, you may claim that they accepted the license. That is true, if they
> accepted the license before learning that a patent is violated and if
> the license was reasonable.

4). I've not made this claim.  However the (c) vs. patent clause I'm
proposing is, IMO, much less starkly controlling than some of the
clauses currently contained in SW shrinkwrap|clickwrap licenses, to say
nothing of clauses which are made possible under UCITA.  The matter is
for the courts to decide.  However, if recent trends in copyright
licensing and contract law bear out, I suspect that a decision will
either sustain the (c) vs. patent clause I propose, or greatly weaken
UCITA.  Either option is a win, IMO.

> A license that prevents them from using Emacs is reasonable, a license
> that prevents them from using all GNU code, if GNU is a substantial code
> base, is unreasoable and opposing the rights granted in patent law.

5). I don't know where you're reading that one license closes all access
to all GNU works -- see my reiteration of the license terms above in

6). Substantiate your claim that the clause (as I've stated it, not as
you're trying to read it) would or would not be reasonable, citing case
law or other precedent. You've made a legal opinion, I'd like to see
supporting evidence.

Karsten M. Self (
    What part of "Gestalt" don't you understand?

SAS for Linux:
Mailing list:  "subscribe sas-linux" to    
  2:38am  up 54 min,  0 users,  load average: 0.28, 0.16, 0.12