[Patents] Re: overlapping legal regimes

[srctran@world.std.com (Gregory Aharonian)]

Paul writes:
>        Here's a thought.  Perhaps, a bare algorithm, with nothing but
>abstract variables, cannot constitute an invention.  Materialization, to
>use your notion, starts taking place at a semantic level, when the
>variables are made to refer to "things" in the real world ......

Which is the law of the land in the United States.  For example, as far
back as the early 1970s, the courts ruled (for example, in the case of
In re Christensen, 478 F.2d 1392) that you couldn't patent something if
the last step of the method was a mathematical equation to be solved,
even if the equation had real world inputs for the variables.  However,
as the dissent to the case argued, and later court cases supported, if
you associate the output with real world "things", then you are 
achieving a useful, specific technological process that is patentable.

For example, suppose I developed an elaborate, but not too elaborate
formula, for digging an oil well to a specific depth.  Oil well digging
is about as real world physical as you can get.  Consider three cases:
1) I implement the formula mechanical using gears, cam shafts, etc. An
awkward way to do so, but I can directly connect the mechanical device
to the oil drilling rig.  The combination of my mechanical invention
and existing oil drilling equipment, is a novel mechanical physical
invention that manipulates the physical.  It would seem to me to be
completely patentable, according to the arguments posted here about the
physical and mechanical.

2) I implement the formula with analog/digital circuit components which
I connect to a mechanical controller which is connected to the drilling
rig.  Again, this seems a completely patentable invention, because of
its use and integration of the types of equipment that have long been
patentable around the world for over 100 years - discrete mechanical,
electromechanical and electronic components.

3) I use a digital signal processor and some software to replace the
analog/digital circuit components of solution 2).  Or, I write some
software, and transform the software into FPGA patterns to be blasted
into a FPGA that is connected to the mechanical controller.  Or I get
a hardware/software codesign tool to accept as input my formula, and
generate some combination of the electronics in 2) and 3) to control
the drilling rig.  So why shouldn't the technologies of 3) be as
patentable as those of 1) and 2).

It doesn't matter to society which of 1), 2) and/or 3) I am doing - all
society wants is for me to drill oil wells more efficiently - that is
the invention to be disclosed to and protected by society.  So as long
as the invention starts in the real world, and ends in the real world,
it should not matter what goes on in between, even if software is being
used.  That is what a large body of US court rulings pretty much argue,
and is consistent with European rulings.

So when In re Christensen says the last step of the claims can't be the
solution of a mathematical problem, it is in agreement with 52.2(C) that
computer programs, as such, can't be patentable, the "as such" implying
no connection to the real world, not specificity in its application.
But add the real world connection, such as the above oil drilling control,
and the invention, in any form, including software, is patentable.

So the large number of software patents EPO is issuing (and much like the
USPTO, they need to control patent quality much better) is consistent
with 52.2.C - what EPO is doing is not illegal, as some have recklessly
hurled out here.  Sloppy, since EPO has the same prior art problems with
software that the US PTO has, but not illegal.

So by having an EPO rule that applicants must connect to the real world
with their software patents, or no patent is issued, then 52.2.C becomes
moot and might as well be dropped to eliminate unproductive discussions
as well as be consistent with US patent laws and court precedents that
are longer, deeper and as well thought out as European court rulings.
At the same time, this also solves the problem of people avoiding the
infringement of a hardware patent by using the technique in software
for which patent protection is denied.

That's my prediction on how Europe will ultimately decide, especially as
they see the open sourcers in Europe and the US start suing each other
using IP laws.  For example, there is a growing squabble (with payments
and lawsuits) between European and American backers of MySQL, a classic
open source development.  Where is all of the peace, love and love of
the common good there?  As open sources start realizing that the few of
them actually doing something new and useful that is starting to make the
them money, they will start invoking IP laws, including the patent laws
which are so disdained.  European politicians sympathetic to open source
will soon be not so sympathetic and find it easier to support the above
mainstream arguments for patents on software applications in the real
world.

Greg Aharonian