[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Sun's Elliptic Curve Technology Contribution to the OpenSSL Project



Dragos Ruiu <dr_(_at_)_kyx_(_dot_)_net> had written:
> Second point.  There is nothing except copyright law to impede
> publishing the information contained in a patent or even a new
> blueprint for an implementation of a patent. (I haven't checked
> but I don't believe any patent offices do not waive copyright or
> litigate it.)  That is the point of a patent: to allow publication of
...

At least in the US, the patent itself is explicitly *not* copyrighted,
either by the applicant or anybody else, and may be freely distributed
without further ado.  I believe this is true even if the patent is
rejected.  Essentially it becomes part of the public record created by
the federal govt, which is not allowed to copyright its own material
(which includes patent applications, court transcripts, congressional
proceedings, etc.)  This does work differently from other countries,
and despite recent changes in copyright and patent law I think this is
unlikely to change - this is wired pretty deeply into US constitution law.

The original intent of patent law in the US was to promote the
publication of information about new ideas, so the patent document was
supposed to explain enough about the device to enable others to
understand it, explore it, &etc.  Also, the patent holder can't stop
discussion or experimental use of his algorithm.  How well this all
works in practice is of course another question.  There's not much
evidence the patent office is even competent to decide if a patent
actually does explain the device sufficiently well.  The patent holder
does have some incentive to explain it, as patents that are *not*
sufficiently complete and functional can be broken in court.  They
don't have any incentive to be clear and easy to understand though, so
most patents are completely turgid, except for the usual amount of
obscure opaque fantasy in the claims.

Copyright covers expression, patent covers the idea.  An independent
re-implementation of an idea is a new expression, so is not covered by
copyright but may still be liable to patent.  I don't know what patents
sun claims to hold regarding ECC, but if those patents are an integral
part of the operation of their ECC library, then any interoperating
reimplementation of the library is equally liable to the same patent
problems as use of sun's library.  To avoid risk, the only way to
distribute such a library would be outside of the reach of the US legal
system regarding patents on the use of mathematical algorithms.  There
is an interesting fine point here regarding the mere *distribution* of
the description of a patent (for the purpose of discussion and
experimental use), but I sure wouldn't want to be the one trying to
insist in court that all the pre-compiled binaries on the openbsd CDs I
was distributing were purely for experimental use and discussion, and
that therefore I wasn't infringing on any patents.

The sad thing about this all is that Sun and their lawyers probably have
no idea or understanding what their paranoia has done to the utility of
their efforts to participate in open-source projects.  The lawyers are
just doing what they've been trained to do: protect the potential business
interests of their client.  Sun's management clearly no longer understands
the utility of open source (in 1988, they were able to release
software under a useful license, as in SUN RPC, so they must have
understood it back then.)

					-Marcus Watts



Visit your host, monkey.org