Software Patents Post Bilski: A Look Ahead
Blog post by Michael A. Spiegel. Please email any comments on this entry to <mspiegel@softwarefreedom.org>.
In
the haze of confusion surrounding the Supreme Court’s recent
decision in Bilski v. Kappos,
the appeals board of the United States Patent and Trademark Office
issued a ruling last week that
takes a definitive stand against the worst kinds of patents that
threaten software developers every day.
Despite
the Court's failure to
provide much guidance or adopt a bright-line test for patentable
subject matter in Bilski,
the appeals board ruling
in Ex parte Proudler is a sign of the growing skepticism
towards software patents that continually test the boundary between
acceptable technological innovation and impermissible abstraction. In
Proudler, the Board rejected a
number of claims to a software invention, citing Bilski in
its reasoning.
This
rejection is particularly noteworthy as it dispenses with a number of
fig leaves that patent attorneys have been using for years to make
software inventions seem less like abstract ideas, and
therefore patentable. While
the USPTO has always held that software is unpatentable, patent
attorneys were usually able to get software patents granted
by adding such
seemingly magical phrases as “a
computer readable medium containing computer executable instructions”
to a series of data
processing steps, thus transforming
software into a patentable
physical component of a computer.
Taking a
page from the same playbook, the
Proudler application
claims a
“method of controlling processing of data in a computer apparatus”
and a “computer
program stored on computer readable media for instructing a
programmable computer to implement a method of controlling the
processing of data.” Since
the Federal Circuit’s infamous 1998 State Street
decision which opened the
floodgates to software and business method patents,
such claims have been deemed patentable without question. Although
the Board could have justified the July 7th decision on other
grounds, it took particular pains to reject the
Proudler application
for claiming
unpatentable software.
“A claim that
recites no more than software, logic or a data structure (i.e., an
abstraction) does not fall within any statutory category,” the
Board said in its rejection.
In its rejection,
the Board cited both recent
Supreme Court cases (including Bilski)
as well as pre-State Street decisions
from the Federal Circuit.
I
believe the Board correctly deduced two notable things from Bilski,
in an interpretation which
hopefully will take root in the
courts as well.
First,
the “machine-or-transformation test”, in which a patentable
process must either be closely tied to a particular machine or
transform a particular article into a different state of thing,
survives as a primary
test for routine use.
Although the Supreme
Court held that “the machine-or-transformation test is
not the sole test for deciding whether an invention is a
patent-eligible ‘process’,” it did characterize the test as “a
useful and important clue.” Without
any hint of what processes, if any, may fail this test while still
being eligible for a patent, the burden will
lie with
software patent applicants to prove that their processes are
patentable after failing the machine-or-transformation test.
Second, the Bilski
decision effectively kills the “useful, tangible, and concrete”
test long favored by patent attorneys who sought to overcome
rejections for ineligible subject matter. As the USPTO has tightened
its guidelines for computer-related inventions over the years, many
applicants cleverly decided to draft their software claims to appear
as if their process was tied to a particular machine. However, for
those patents which were granted on the basis that their inventions
were “useful, tangible, and concrete”?most notably software
patents granted immediately following State Street?the
Court’s decision in Bilski makes these patents especially
vulnerable.
Assuming
this characterization
of Bilski survives on
appeal, I think we will start
to see the USPTO take more aggressive steps to stanch the flow of the
worst types of software patents. One need only look at the specific
claims in the Proudler
application to see how
patents of this type stifle
innovation in computer technology. I cannot
think of any computer program that does not have “logically related
data items” processed by associated rules. The rest of the claim
language contain unintelligible nonsense—the perfect weapon to use
against small software developers who can’t afford to defend
themselves against patent aggression.
So what will the
status of software patents be going forward? Given the Supreme
Court’s reluctance to categorically exclude any type of invention
from patent-eligibility in the face of unforeseen developments in
technology, I believe that at least some forms of software will
remain patentable, barring any (unlikely) legislative response to the
software patent issue.
Expect to see
software patent claims for “computer readable mediums” encoded
with “processor executable instructions” challenged under the
machine-or-transformation test as not being tied to a specific
machine. These challenges may prove successful in invalidating the
most abstract of computer software patents, to which patent
applicants will respond by drafting process claims with ever
increasing references to specific computer hardware. I also expect to
see an increasing reliance on system claims with generic hardware
elements, such as a processor or computer network, configured to
perform steps performed in software. The patentability question for
computer software running on general-purpose computers will be
decided on the Court’s clear disfavor for “abstract ideas” as
seen in the context of the reaffirmed Benson-Flook-Diehr
trilogy of Supreme Court cases. The language in these cases which
deny patents to “algorithms” and “mental processes” may prove
useful in invalidating the most harmful of software patents.