If you've been reading News Picks, you may have noticed that both the ACLU and Red Hat have filed amicus briefs in a patent case, In re Bilski and Warsaw. I thought you might like to understand what is going on, as best as I can explain it. Here's the ACLU brief [PDF].
Red Hat's announcement indicates that it is asking the United States Court of Appeals for the Federal Circuit to consider something dear to your hearts, whether or not, or to what degree, patents should be allowed on software, and its brief explains to the court the damage that software patents are doing to Open Source software. Here's the brief [PDF].
For background and context, here's a Law.com article that explains the fundamental issues in the Bilski case:
As most patent lawyers know, the U.S. Court of Appeals for the Federal Circuit has signaled that it is reconsidering the patentability of business methods in a rare en banc hearing on In re Bilski. It's the most visible sign yet that the Federal Circuit is ready to fall in line with a series of Supreme Court rulings limiting patent rights by revisiting its own 1998 State Street Bank v. Signature Financial Group decision.
Many may also know that Bilski is about a method of trading weather risk invented by two former energy company executives who are founders of Pittsburgh-based WeatherWise USA Inc.
So, it's an appeal from a failed attempt to get a business method patent. But it's an opportunity to raise issues with the court that I know most of you have been wishing someone would raise. Red Hat in particular is bringing to the court's attention the braking effect software patents are having on Open Source software, where instead of encouraging innovation, it's threatening it.
The inventors, shall we call them, are appealing their inability to obtain a patent on their idea, as Law.com reports:
Warsaw sees his failure to get a patent, like the Minnesota investigation, as an example of visionless government bureaucracies. "Both these events are stifling creativity and innovation in areas where they should be encouraging them," says Warsaw. "We've invested tens of millions of dollars in our company. IP is no longer gears, valves, pipes and wires."
In short, it's litigation on whether patents can be stretched to cover a disembodied idea, an abstract idea that has no machine implementing anything, no technology at all. The USPTO wouldn't issue the patent, and that is what is being appealed. Here's the decision [PDF] by the Board of Patents and Interferences back in September of 2006 that upheld the examiner's position refusing to grant the patent, and it explains what the patent application was for and why the examiner said no:
The invention relates to a method practiced by a commodity
provider for managing (i.e., hedging) the consumption risks
associated with a commodity sold at a fixed price. It is
disclosed that energy consumers face two kinds of risk: price
risk and consumption risk (specification, p.1). The
proliferation of price risk management tools over the last
5 years before the filing date allows easy management of price risk
(specification, p. 2).
However, consumption risk (e.g., the need
to use more or less energy than planned due to the weather) is said to be not currently managed in energy markets, which is the
problem addressed by the invention (specification, p. 2). ...
The examiner's position is summarized in the statement that,
"[regarding [claims 1-11, the invention is not implemented on
a specific apparatus and merely manipulates [an] abstract idea
and solves a purely mathematical problem without any limitation
to a practical application, therefore, the invention is not
directed to the technological arts" (FR4). That is, the examiner
states that the invention is an "abstract idea," and apparently a
"mathematical algorithm," and does not fall within the
"technological arts" according to In re Musgrave, 431 F.2d 882,
893, 167 USPQ 280, 289-90 (CCPA 1970), where the examiner states
"The definition of 'technology' is the 'application of
science and engineering to the development of machines and
procedures in order to enhance or improve human conditions, or at
least improve human efficiency in some respect.' (Computer
Dictionary 384 (Microsoft Press, 2d ed. 1994))."
It's not "in the technological arts," the appeal decision stated, because it isn't in any way necessary to perform the method on a computer or have any other technological means to carry out the invention. It was easy enough to recognize a cotton gin as being worthy of a patent, let's say. But when software patents began to be covered, things got squishy. Computers, after all, perform abstract ideas, so where do you draw the line? Everyone seems to have suddenly realized that unless someone figures that out, there will be endless litigation to try to draw a line somewhere, as the Board goes on to relate:
The "useful arts" in the Constitution, are implemented by
Congress in the statutory categories of eligible subject matter
"process, machine, manufacture, or
composition of matter, or any new and useful improvements
thereof." Machines, manufactures, and man-made compositions of
matter represent tangible physical things invented by man and
seldom raise a §101 issue, except for the special case of
claims to general purpose machines (usually computers) that
merely perform abstract ideas (e.g., mathematical algorithms),
where the fact that the claim is nominally directed to a
machine under §101 does not preclude it from being held
This appeal involves
claims, i.e., the claims do not recite how the steps are
implemented and are broad enough to read on performing the steps
without any machine or apparatus (although performing the steps
on a machine would, of course, infringe). The steps of claim
1: do not recite any specific way of implementing the steps; do not
expressly or impliedly recite any physical transformation of
physical subject matter, tangible or intangible, from one state
into another; do not recite any electrical, chemical, or
mechanical acts or results; do not directly or indirectly recite
transforming data by a mathematical or non-mathematical
algorithm; are not required to be performed on a machine, such as
a computer, either as claimed or disclosed; could be performed
entirely by human beings; and do not involve making or using a
machine, manufacture, or composition of matter, We do not
believe the outcome in this case is controlled by the Federal
Circuit decisions in State St. Bank & Trust Co. v. Siqnature Fin.
Group, Inc., 149 F. 3d 1368,
47 USPQ2d 1596 (Fed. Cir. 1998) and
AT&T Corp. v. Excel Communications, Inc., 172 F.3d 1352, 50 USPQ2d 1447 (Fed. Cir. 1999) because we interpret those cases
to involve the "special case" of transformation of data by a
The question of whether this type of non-machine-implemented
subject matter is patentable is a common and important one to the
U.S. Patent and Trademark Office (USPTO), as the bounds of
patentable subject matter are increasingly being tested. In
recent years, the USPTO has been flooded with claims to
"processes," many of which bear scant resemblance to classical
processes of manipulating or transforming compositions of matter
or forms of energy from one state to another. Many of these
applications are referred to as so-called "business methods," but
claims to methods of meditation, dating, physical sports moves,
etc., are also presented. "Business methods" have long been
considered statutory subject matter when performed by
a machine. ...
The USPTO no longer
rejects claims because the claimed subject matter does "business"
instead of something else. See State Street, 149 F.3d at
1377, 47 USPQ2d at 1600 (referring to Examination Guidelines,
61 Fed. Reg. 7478, 7479 (1996)). Nevertheless, many questions remain
about statutory subject matter and what the tests are for
determining statutory subject matter. State Street and AT&T,
often called "revolutionary,"
involved patented machines or
machine-implemented processes that examiners have for sometime
regarded as nonexceptional. Perhaps encouraged by certain
general language in these cases, however, a wide range of ever
more general claims to "processes" come before the Office
(although the present case predates both State Street and AT&T).
Many, like the claimed process in the present case, are not
limited to implementation via any particular technology or
machine. Are such "processes"
patentable because they are
"useful"? Other "process claims" involve what seem to be
insubstantial or incidental manipulations of physical subject
matter--e.g., the mere recording of a datum: are these patentable processes? Still other process claims involve human physical
activity--methods of throwing a ball or causing a fumble. Do
these process claims cover patentable subject matter? Must the
examiners analyze such claims for compliance with the written
description and enablement requirements, and search the prior art
for evidence of novelty and nonobviousness?
In short, patent examiners and everyone else find themselves in patent LaLaLand. If you listen to the court's October hearing (which you can listen to as MP3 here - [Update: -- not any more, as the court now renders that url as the later May 8, 2008 hearing; October's seems to have disappeared.]
[Update 2: It's back. If you search for keyword Bilski on the court's search page, you'll find it first on the list. Or just go here.]
And you can read a very complete report on oral arguments on AIPLA's website.), it's quite amusing to hear them try to figure out if a chiropractor's hands are a kind of patentable process, if the spine is altered. It's getting so goofy.
Now, after recent decisions by the US Supreme Court in particular, the tide seems to be shifting somewhat. The appeals court is now asking, what exactly is patentable? If it is a mere abstract idea and it's allowed to be patented, what happens next? Can you, say, patent democracy? Men invented it, after all, it's under the sun, it's not a law of nature, and it's a method of getting something done. Could that be granted a business method patent? If abstract ideas, unrelated to any machinery, can be patented, then why *couldn't* you patent democracy? The USPTO is struggling with where to draw the line. So are the courts. No one knows precisely where the line should be, which is what happens, in my view, when you have already crossed the line into Sillyland.
The US Court of Appeals for the Federal Circuit noticed, I gather, that there is a need to draw a line somewhere, and so after the hearing last October, instead of issuing a decision on the case on its own initiative announced in a February 15, 2008 order that it would hold an en banc hearing and it asked the parties and amici to answer five questions [PDF]:
This case was argued before a panel of this court on October 1, 2007.
Thereafter, a poll of the judges in regular active service was conducted to determine
whether the appeal should be heard en banc.
Upon consideration thereof, IT IS ORDERED THAT:
The court by its own action grants a hearing en banc. The parties are requested
to file supplemental briefs that should address the following questions:
(1) Whether claim 1 of the 08/833,892 patent application claims patent-
eligible subject matter under 35 U.S.C. § 101?
(2) What standard should govern in determining whether a process is patent-eligible subject matter under section 101?
(3) Whether the claimed subject matter is not patent-eligible because it constitutes an abstract idea or mental process; when does a claim that contains both mental and physical steps create patent-eligible subject matter?
(4) Whether a method or process must result in a physical transformation of an article or be tied to a machine to be patent-eligible subject matter under section 101?
(5) Whether it is appropriate to reconsider State Street Bank & Trust Co. v.
Signature Financial Group, Inc., 149 F.3d 1368 (Fed. Cir. 1998), and
AT&T Corp. v. Excel Communications, Inc., 172 F.3d 1352 (Fed. Cir.
1999), in this case and, if so, whether those cases should be overruled in
This appeal will be heard en banc on the basis of the original briefs and
supplemental briefs addressing, inter alia, the issues set forth above.
You've probably heard of the State Street case, as it's a famous case, the one that stretched patents to cover software. AT&T was the one that did the same for business methods. You might be interested to know that the judge reversed in that latter case was Judge Sue Robinson, the judge in the stalled Red Hat v. SCO case. She tried to do the right thing, as we would view it, but she was reversed, and now the question is, was she right after all? Should those two cases be modified or overruled? In fact, in that opinion by the appeals court, the question that is now stage front and center was raised in a quotation from Justice Stevens in the case Diamond v. Diehr, 450 U.S. 175, 182 (1981):
Our analysis of whether a claim is directed to statutory subject matter begins with the language of 35 U.S.C. 101, which reads:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title....
In his dissent in Diehr, Justice Stevens noted two concerns regarding the 101 issue, and to which, in his view, federal judges have a duty to respond:
First, the cases considering the patentability of program-related inventions do not establish rules that enable a conscientious patent lawyer to determine with a fair degree of accuracy which, if any, program-related inventions will be patentable. Second, the inclusion of the ambiguous concept of an "algorithm" within the "law of nature" category of unpatentable subject matter has given rise to the concern that almost any process might be so described and therefore held unpatentable.
Diehr, 450 U.S. at 219 (Stevens, J., dissenting).
Despite the almost twenty years since Justice Stevens wrote, these concerns remain important. His solution was to declare all computer-based programming unpatentable. That has not been the course the law has taken. Rather, it is now clear that computer-based programming constitutes patentable subject matter so long as the basic requirements of 101 are met.
Now the question is, did the law go too far? Duh, you say. But this is what they are asking. If you listen to the MP3, you will hear them talk a lot about In re Comiskey and here it is [PDF], along with some information about it. They also mention Parker v. Flook, a Supreme Court decision.
So there is quite a lot at stake, and the tide appears to be turning away from the profligate issuing of patents in such categories. At least the right questions are being asked. But to what extent will they reach a decision to modify or pull back? This case will help to decide. Of course, the Supreme Court is higher than this court, but having already given some direction, the Bilski case gives the appeals court the opportunity to undo what many view as damage by some decisions that opened the floodgates to all kinds of crazy patents. And that is why suddenly a large crowd of interested parties are submitting amicus briefs. The list at the court is impressive. Unfortunately, the court does not make the briefs digitally available, but Red Hat's announcement says it will be on that page soon. And thankfully, Patently-O provides some descriptions of positions taken and explains the process. More amicus briefs listed here.
ACLU argues that this case raises another question, a Constitutional one: if you can patent an abstract idea without any machine component, what about the First Amendment? Can the government grant a monopoly to someone on an abstract idea? The ACLU press release puts it like this:
Introducing a rare argument applying the First Amendment to patent law, the American Civil Liberties Union filed a friend of the court brief today urging a federal court to uphold the denial of a patent that would, if awarded, violate freedom of speech. In the brief, the ACLU argues that Bernard L. Bilski is seeking a patent for an abstract idea, and that abstract ideas are not patentable under the First Amendment.
"The court must ensure that any test it uses in determining whether to award a patent is in line with the Constitution," said Christopher Hansen, senior staff attorney with the ACLU First Amendment Working Group, who filed the brief. "If the government had the authority to grant exclusive rights to an idea, the fundamental purpose of the First Amendment - to protect an individual's right to thought and expression - would be rendered meaningless."
Here's the Red Hat statement:
Today, Red Hat took a public stand challenging the standards for patenting software. In the Bilski case that is now before the Federal Circuit Court of Appeals, this patent issue is ripe for consideration. In a friend of the court brief submitted to the Federal Circuit Court of Appeals in the Bilski case today, Red Hat describes the special problems that patents pose for open source and seeks modification of the standards for patentable subject matter that take open source into account. Here is a quick summary of our brief.
Open source software is one of the most dynamic, innovative sectors of the U.S. economy, but the U.S. patent system is a costly hindrance to open source innovation. We believe that although the patent system was created to foster innovation, it’s simply not an engine for innovation for open source. Software patents were barely recognized when open source began, and so the hope of obtaining a patent did not motivate the first developers. Those pioneers were generally opposed to software patents. The open, collaborative activity at the heart of open source is at odds with the patent system, which excludes the public from making, using or selling a patented invention. Open source developers seek to contribute code to the community – not to exclude others from using the code.
Today the patent system is, if anything, a hindrance to open source. Developers face the risk that the original code they have written in good faith could be deemed to infringe an existing software patent. It’s impossible to rule out this possibility, because there are now more than 200,000 software patents, and those patents cannot be efficiently searched. Software patents are difficult to interpret, even for experts in computer science and software engineering. Experts often disagree as to whether a particular patent claim covers a particular program. Thus , a risk of litigation exist for every open source project, and the potential cost of patent litigation runs into millions of dollars for a single case.
Given the litigation risk, some open source companies, including Red Hat, acquire patents for the sole purpose of asserting them defensively in the event they are faced with a future lawsuit. Red Hat also provides open source intellectual property protections through our Open Source Assurance Program that protects our customers and encourages them to deploy with confidence. Our strategy is a prudent one and mitigates the risk of patent lawsuits, but it would be unnecessary if the system itself were fixed.
Despite the hindrances of the patent system, open source continues to expand at an exponential rate. One can only imagine how that expansion would accelerate if the braking effect of the patent system were lifted.
In the Bilski case now pending in the Federal Circuit, the appeals court will address the issue of the boundaries of the subject matter that may be patented. The Supreme Court has repeatedly stated that abstract ideas and algorithms (which are the core building blocks of computer programs) are not patentable. Applying this principle consistently would greatly narrow the availability of software patents. As the Federal Circuit addresses the standard for patentable subject matter, we urge in our brief that it take account of the perverse effects of the patent system on open source and narrow the standard for patentability.
Our full brief will be posted online soon. Check back for the link.
Oral argument will happen May 8 at 2 PM in Courtroom 201. Here's a map showing all the circuits. The Federal Circuit is, you'll see, in Washington, DC.
I've always told you that in my opinion the best hope for fixing the software patent mess was in the courts, but I confess I was thinking of the Supreme Court, not the U.S. Court of Appeals for the Federal Circuit. But that is not a complaint. I'm stunned but happy it is even asking these questions.
Update: EFF has filed an amicus brief also:
In re Bilski is an appellate court case that provides an opportunity to eliminate business method patents and curtail efforts to claim monopolies on basic human skills, behaviors, and interactions. Bilski is challenging the rejection of his application for a patent on a method of managing the risk of bad weather through commodities trading. EFF submitted an amicus brief [PDF] (in conjunction with The Samuelson Law, Technology & Public Policy Clinic at UC Berkeley Law, Public Knowledge, and Consumers Union) supporting the rejection of Bilski's patent application and setting forth a framework for determining patentable subject matter that focuses on the use of technology in the claimed invention....
EFF's brief proposes that the patent office and the courts should determine whether an invention is technological before even considering it for patent protection and lays out a set of factors to help make that determination. This litmus test will help ensure that the patent system is not used to monopolize everyday interactions and other human behaviors. It will also provide a more intuitive demarcation of patentable subject matter so that innovators and the general public can avoid infringement....The outcome of this case could change the landscape of patent law.
FSF's End Software Patents has an announcement about its amicus brief:
End Software Patents (ESP) has filed an amicus curiae brief in the Court of Appeals for the Federal Circuit's (CAFC) rehearing of the In re Bilski case. The rehearing could lead to the elimination of patents on software. ESP executive director Ben Klemens said, "This is an historic opportunity to fix the US patent system, as the Bilski rehearing will directly address the boundaries of the subject matter of patents. In our brief, the End Software Patents project supports the Supreme Court's long-held position that computer software should not be patentable, and has highlighted to the Court the real economic harm software patents cause the US economy."
With the boundary to what can be patented effectively destroyed by previous Federal Circuit rulings, massive-scale liability has been created throughout the US economy. ESP's brief also demonstrates that this liability is not merely a theoretical prediction, but a real economic harm. Over the last few months alone, ESP has tallied over fifty non-software companies being sued for infringement regarding their web sites or other course-of-business software, including the Green Bay Packers, McDonald’s, Dole Foods, Kraft Foods, Caterpillar, J Crew, Burlington Coat Factory, Wal-Mart, and Tire Kingdom. The rest of this list can be found at http://endsoftpatents.org/a-litany-of-lawsuits.
Ironically, the Federal Circuit's own web site at http://www.cafc.uscourts.gov/ is produced using software that likely infringes some number of software patents. In fact, the last decade of software patents has brought about many onerous and frivolous lawsuits, inspiring Congressional action and causing many to question the entire patent system. Many patents famous as the rallying points for patent opposition, such as the "Blackberry patents" from NTP v. Research in Motion, Ltd., have been software patents.
ESP's brief points out that these patents centered on claims over pure information. Under US law, pure information is not patentable. Further, the Supreme Court ruled three times that pure information does not necessarily become patentable when recited in combination with a physical object, such as information written to paper or loaded into a computer's memory. However, the Court of Appeals for the Federal Circuit ignored the Supreme Court's repeated rulings, and began allowing patents on information plus any physical component: a formula, if saved to a computer's hard drive; a price list, if money is eventually moved; not a correlation, but the act of correlating. The ESP brief recommends re-establishing the Supreme Court's rule that information should not be patentable, even when claimed in tandem with a physical afterthought.
Here's the amicus brief [PDF].