Subject: Re: on software freedom and patents
From: Thomas Lord <lord@emf.net>
Date: Sat, 02 Dec 2006 11:18:40 -0800


    Stephen> It might be nice if patents worked as you propose,
    Stephen> but they don't.

 That is misleading because we are talking about areas that are
 not settled law.


    Stephen> It's possible that *in practice* things *could*
    Stephen> work out that way, at least for some free software
    Stephen> licenses.  This is the claim that the FSF makes
    Stephen> vis-a-vis the GPL v.2.  However, this is based on
    Stephen> the implicit patent license doctrine,


  This has nothing to do with the "implicit patent license
  doctrine".

  Very briefly: patent law (at least in the US) does not
  *create* new exclusive rights.  Rather, it *grants*
  some exclusive rights that would otherwise exist (for someone)
  to the inventor.

  For example, whenever a consumer buys a lightbulb, a unique
  contract is executed: that customer's money in exchange for
  that lightbulb.  That is an exclusive contract: only one
  seller gets to sell that particular lightbulb.  Holding the
  contract to sell that bulb is therefore an exclusive right.
  That right fits the descriptions "making" and "selling" and so
  it is among the rights granted to a patentee.  Patent law
  didn't create the exclusive right to sell that light bulb:
  patent law merely assigned ownership of an exclusive right
  that already existed.

  On the other hand, when a person downloads a free software
  program, while there are contracts that pertain to the
  bandwidth and so forth, there is no contract required for the
  person make that copy of the program (or to subsequently use
  it).  Obtaining and using copies is not an exclusive right.
  There is nothing about copying that program for Congress to
  assign to an inventor -- patent law is moot because the
  program is publicly licensed.   No patent licenses are required.

  For tangible inventions, public licensing of texts, under
  copyright law, is unimportant.  A description of a lightbulb
  doesn't give off much light.  Software, it simply happens to
  be the case, is different: there is no real distinction
  between a complete and accurate description of a software
  invention and a useful embodiment of the invention.
  Therefore, for software, patent disclosure requirements have a
  special consequence: in order to obtain a software patent, a
  useful embodiement of the invention must be propogated to the
  public.   The cat is out of the bag, so to speak.   Freedoms
  such as the freedom of speech ensure the public's right to
  propogate descriptions of the invention so long as they do
  not enter into exclusive arrangements of making, selling, or
  using the invention.

  This most certainly does *NOT* mean either that:

                           Not True:

    (a) all GPL programs are assuredly safe
            from patent infringement (even under this theory)

    (b) software patents are worthless

  GPL may not qualify as a public license under this theory
  because if a person violates the GPL they loose all
  permissions ordinarilly granted by the GPL.   Thus, the GPL
  does not convey a public right: it conveys a right to those
  people who never violate the GPL.   A less retributive
  copyleft license would be less problematic.

  Software patents are not worthless because the analysis
  presented here relies on the assumption of public licensing.
  In this analysis, software patents apply, in all customary
  ways, to proprietary software.


-t