Subject: RE: Patent-based dual-licensing open source business model
From: "Lawrence Rosen" <lrosen@rosenlaw.com>
Date: Wed, 13 Sep 2006 14:36:12 -0700

> I don't agree that a license that
> 
>    (1) allows me to offer a hardware product and a software product
>        that runs on the hardware, separately, but
> 
>    (2) does not allow me to offer the software copied onto the hard
>        drive of the hardware bit-for-bit,
> 
> where in both cases the software product includes full machine-
> readable source, is a free software license.  I suspect you'll find
> few who do agree.

The Covenant is not a license. 

What is the difference between buying a Dell computer and then downloading
Linux, as opposed to buying a Dell computer preconfigured with Linux? The
honest answer is: The second alternative offers us an opportunity for a
commercial license under patent law, without worrying about bit-for-bit
analyses or derivative work complexities. Dell is making money on that deal;
in our view, so should the author of the software and the inventors of its
embodied patents. :-)

If the economics of that trade-off drive customers to download and install
their own free software, we don't care. We like people who download their
own software. If enough people do so, we might be encouraged to drop our
license fees for commercial products.

But there is a more subtle legal issue implied in your question. If a
company tells its customers to buy its hardware, then tells them to
separately download specific patented open source software in order to avoid
paying royalties, I think the hardware vendor would leave itself open to
charges of contributory or inducing infringement. Our Covenant doesn't
protect hardware vendors from this.

> ... The problem is
> the boundary between the cases that obviously infringe, but you
> promise not to assert your rights, and those cases that obviously
> infringe, but I'd better watch my butt.  Where is that boundary?  The
> Covenant is not yet very clear, to me, anyway.

I'm trying in these discussions to make it clearer. The input you and others
are providing is helpful. I do suggest that (1) open source projects ought
not to worry their pretty little butts about such things, and (2) the
companies that make commercial products have plenty of lawyers to analyze
such questions. I can assure you that these are much easier questions than
those associated with derivative works and the GPL. EVERY reputable company
that sells such products in commerce now knows how to deal with patents;
they do it every day.

> If the software *is* open source, then it's irrelevant on what medium
> it's delivered: CD-ROM, EEPROM, hard drive, or a fully functional and
> configured workstation.  A user has the right to copy, modify, and
> redistribute it, period (along with access to source).  However, you
> do not give me the right to copy it onto a hard drive and redistribute
> it, if that hard drive "just happens" to be installed in a workstation.

Correct. You are distributing a commercial product consisting of hardware
and software. I'm not talking about a delivery medium. Copyright law deals
with "distribution" of software; patent law deals with making, using,
selling, etc.... Your example here deals with selling a product.

> If exercise of any of the four freedoms is impaired, you're not going
> to be able to call it free.  It looks to me like both use and
> distribution freedoms are restricted as the Covenant is currently written.

I don't call our Covenant free, since I see no reason to adopt a free
software definition designed for copyright law that fails to account for the
value of patented software in commercial products. We adopt the existing
definitions only to the extent that you can implement our patents for free
in free [i.e., open source] software.

Here's how we define "open source software":

   The term "open source software" in this covenant shall mean 
   software actually distributed to the public under software 
   licenses that have been expressly approved by Open Source 
   Initiative or the Free Software Foundation as of the date 
   of this covenant, and that provide that every licensee is 
   free to make copies of the software or derivative works 
   thereof, to distribute them without payment of royalties 
   or other consideration, and to access and use the complete 
   source code of the software.

That may not agree completely with others' definitions (although, except for
the phrase "actually distributed to the public" it probably does agree), or
their interpretations of their own definitions, nor are their licenses
necessarily consistent with our Covenant, but we have to build a profitable
business here consistent with our philosophical goals of promoting the
development of technical innovation. Please note from the paper I posted
that our objectives are much broader than just free software. Giving patents
away for free for commercial uses is no way to start that off.

/Larry


> -----Original Message-----
> From: stephen@xemacs.org [mailto:stephen@xemacs.org]
> Sent: Wednesday, September 13, 2006 1:08 PM
> To: lrosen@rosenlaw.com
> Cc: 'Free Software for Business'
> Subject: RE: Patent-based dual-licensing open source business model
> 
> Lawrence Rosen writes:
> 
>   lr> My responses are inline....
> 
>   simo> It seem a bit too much to require individuals to publish (it
>   simo> means you must set up a website or something and maybe pay for
>   simo> the service, communicate with IC etc..) the code for every
>   simo> trivial modification they want to do.
> 
>   lr> Not only isn't that what we want, but that also isn't what the
>   lr> Covenant or patent law require.
> 
>   lr> Consider the "first sale" (or "exhaustion") doctrine in patent
>   lr> law. When you buy any product in commerce, you are allowed to
>   lr> repair it if necessary to use it, regardless of the patents.
>   lr> (You can't *change* it, but you can *fix* it; you can put new
>   lr> sparkplugs in your Chevrolet, but you can't turn it into a
>   lr> Cadillac, and you certainly can't start making and selling new
>   lr> cars on the model of the one you bought!
> 
> But as I read that Covenant, I think simo is right to worry.  Software
> does not wear out or even break under load; there *is no equivalent to
> changing sparkplugs.*  Rather, what's under discussion here is adding
> some trim and chrome wheels to turn a Ford into the equivalent Mercury
> model.
> 
> Maybe that's still trivial and OK, but that's not clear to this
> non-lawyer.
> 
>   lr> If software is preinstalled on hardware, then it is a
>   lr> combination. The issue is not whether it is "bit for bit
>   lr> identical" (that's a copyright issue) but whether the product
>   lr> embodies our patent claims.
> 
> >From a legal standpoint, of course you're right.  From the point of
> view of understanding whether this is a free software license (based
> on patent) or not, however, the "bit-for-bit identical" case is a
> salient example.
> 
> I don't agree that a license that
> 
>    (1) allows me to offer a hardware product and a software product
>        that runs on the hardware, separately, but
> 
>    (2) does not allow me to offer the software copied onto the hard
>        drive of the hardware bit-for-bit,
> 
> where in both cases the software product includes full machine-
> readable source, is a free software license.  I suspect you'll find
> few who do agree.
> 
>   lr> For us, we're glad to be out of the business of worrying whether
>   lr> a bug fix is a derivative work. Although the patent law may seem
>   lr> complicated, it is a well-understood process to lay a patent and
>   lr> a product on the table and determine whether the product
>   lr> infringes. Meanwhile, we've been arguing about derivative works
>   lr> for years and still don't have a good answer.
> 
> That *is* a real benefit to basing a license on a patent, but it
> doesn't address the question.  The problem is not the edge cases where
> it's not clear whether there's infringement or not.  The problem is
> the boundary between the cases that obviously infringe, but you
> promise not to assert your rights, and those cases that obviously
> infringe, but I'd better watch my butt.  Where is that boundary?  The
> Covenant is not yet very clear, to me, anyway.
> 
>   lr> Or that purchased hardware preloaded with software means the
>   lr> software is "free"?
> 
> If the software *is* open source, then it's irrelevant on what medium
> it's delivered: CD-ROM, EEPROM, hard drive, or a fully functional and
> configured workstation.  A user has the right to copy, modify, and
> redistribute it, period (along with access to source).  However, you
> do not give me the right to copy it onto a hard drive and redistribute
> it, if that hard drive "just happens" to be installed in a workstation.
> 
>   lr> can you see how this new model can lead to a new form of
>   lr> software freedom based on patent law and not just copyright law?
> 
> It's easy to see that it's a new model.
> 
> It's difficult to admit it as free software.  Free software
> definitions are basically intended to implement "no rights reserved".
> Your Covenant, however, is clearly based on "some valuable rights
> reserved".  I happen to like your choices as to which rights to
> reserve, but I'm reluctant to call the result a free software license.
> 
> If exercise of any of the four freedoms is impaired, you're not going
> to be able to call it free.  It looks to me like both use and
> distribution freedoms are restricted as the Covenant is currently written.
> 
> Regards,
> Steve