Subject: Re: Patent-based dual-licensing open source business model
From: simo <s@ssimo.org>
Date: Wed, 13 Sep 2006 08:34:18 -0400

On Wed, 2006-09-13 at 13:37 +0900, stephen@xemacs.org wrote:
> On Tue, 2006-09-12 at 07:01 -0700, Lawrence Rosen wrote:
>  > I'm interested in the reactions of this group to a new patent-based
>  > dual-licensing open source business model adopted by International
>  > Characters.
>  > 
>  > www.rosenlaw.com/IC-Business-Model.pdf. 
> 
> The phrase "license approved by the FSF" is insufficiently precise.
> The FSF is quite capable of approving and advocating licenses that are
> not free software licenses, even by their own definition.  You must
> specify "free software license approved by the FSF", or someone could
> put the US Constitution in a string as an invariant section of the
> program and license it under the FDL ;-).
> 
> Note that this Covenant doesn't address the issue of "submarine
> patents".  There's no requirement that the patent or the Convenant be
> mentioned in the software's documentation.  So somebody who receives
> "open source software" exempted under the Covenant with an academic
> license will naturally assume he's free to convert it to a proprietary
> product, or even with copyleft use a modified version internally, etc.
> 
> simo writes:
> 
>  > The use of this kind of licensing has always interested me.
>  > Your license, it seems to me, make me, personally, impossible to use any
>  > software I modify (even just to fix a bug, or to change a preference
>  > that make sense only to me) without releasing the source code in the
>  > public, and stay legal.
> 
> That's exactly what they want.  However, you haven't said what source
> code.

Well the problem is that not only applies to a company, but even to an
individual that made a trivial change for his own private use. It seem a
bit too much to require individuals to publish (it means you must set up
a website or something and maybe pay for the service, communicate with
IC etc..) the code for every trivial modification they want to do.

> I suspect that if "the Software Practicing the Patent" is incorporated
> in a library, and you do not modify the library, you're OK.  Unless my
> understanding of Larry's book on open source licensing is mistaken or
> he's changed his mind since he wrote his book, Larry doesn't believe
> in the "linking == deriving" theory.  If you fix a bug, or change a
> preference in the *library*, and you care enough about that not to
> publish it, then I expect you also care enough to pay licensing fees.

If you are a company I may agree with you, otherwise for personal use I
don't.

> You can probably make this fly even in a strong copyleft environment
> simply by creating a brute-force library that doesn't use the patented
> technology but performs the same function and exports the same API.
> Put both on your web site under an open source license, or certified
> mail them to International Characters with an open source license, and
> you're done.
> 
> It would also be costly for them to get past a claim that your
> modifications are "experimental".  "I was planning to post it on my
> web site next week, your Honor, really I was.  But it's not ready
> yet."

If I remember the paper correctly this is a non issue, as IC let you do
all the research you want, it is when you put the software in production
internally that you need to release the code. It's very difficult to
claim it is not ready when your 259 desktops use it all the day.

>  > It also strangles any little service company that want to try to
>  > help their customers by selling pre-configured hardware without
>  > making any relevant modification to the software except customizing
>  > the configuration.
> 
> It's "worse" than that.  Under your interpretation, they can't
> distribute pre*installed* hardware at all without risking legal action
> by International Characters.  However, if the software in flash ROM or
> whatever is bit for bit identical with something available on your
> website, then the preinstalled hardware, however complex and
> multifunctional, should be considered a distribution medium---not a
> combination---it seems to me.

Patent law is not copyright law, it cover this situation, and IC made it
clear that if the "medium" is actually hardware (computer+memory) then
you need a license.

> An important question here is whether receipt by *one member* of the
> public constitutes distribution "to the public".  If so, as long as
> you provide source for your configuration to that customer, you're
> clear.  Presumably that's not what they have in mind; if they did, it
> would just say "distribution."  But it is not at all obvious to me
> what "distribution to the public" means.  The commentary uses the
> phrase "public disclosure" instead.  Larry?

Distribution to the public has very clear meaning. Put the software in a
place where anybody can reach it.

The real problem is that it doesn't say for how long you have to keep it
public, and at which conditions, Larry?

Can I ask a fee for distributing the source code? :-P

>  > While I can understand and probably even favor the idea of making
>  > royalties against proprietary code, I think that blocking use of
>  > internally modified software or software completely public but
>  > distributed with hardware violates the very nature of free
>  > software.
> 
> All you need do is distribute the open source software and hardware as
> separate products, as far as I can see, and offer a bundled discount.

So you really think anybody would buy a router and the software
separately and then buy a prom writer to burn the software in? (Yes
stretching a bit :-)

Simo.