Patent Office Stifles Innovation

An appeals court ruling favoring the U.S. Patent Office puts in jeopardy tech companies' ability to get protection for further software advances

By John Kheit

Information Age innovators need not apply. At least that's the implied message being stretched like police tape across the door of the U.S. Patent & Trademark Office (USPTO). The agency seems fixated on eliminating the last, true, sustainable American advantage: our capacity to innovate.

Recent moves by the USPTO have resulted in a precedent-setting legal victory that now threatens software patents with extinction, putting companies like Apple and Google at risk along with the U.S. economy.

Software is key to modern innovation. But more important, it's among the last strongholds in American economic might. Open the box on a new Apple (AAPL) product, and you'll see a note saying "Designed by Apple in California." It leaves out the implicit "made in China." That's what America does. It innovates. It designs. It conceives and invents. Today, America makes ideas. And then we outsource the manufacture of our ideas, and sell those to the world. Indeed, in just a few decades, the U.S. has gone from a country that largely produced tangible goods to one that produces blueprints for intellectual property. And one of the greatest areas of intellectual property production in the U.S., as noted by Apple CEO Steve Jobs, is software.

Notably, during Apple's recent quarterly report, Jobs commented about the iPhone: "The traditional game in the phone market has been to produce a voice phone in 100 different varieties.… [W]e approach it as a software platform company, which is pretty different from most of our competitors."

Left in Limbo

Now, it appears the USPTO is hell-bent on destroying such intellectual property rights. The most recent evidence came from an Oct. 30 ruling by an appeals court that could narrow what's patentable and make it much more costly to obtain patents, particularly for newer companies. Some fear the decision also leaves many innovative software technologies in limbo.

The decision came in response to a request by the USPTO for clarification of what is and isn't patentable. In so doing, the USPTO opted to retry principles that had been settled more than a decade ago, putting all patents protecting business-method software at risk.

The USPTO kicked the question to the appeals court because it wants the power to pick and choose what is patentable subject matter—rather than merely performing its well-established mandate to examine patent applications. At best, this is plain lazy. At worst, this push is economically stupefying. How can the USPTO recognize the next Google—particularly if it won't even examine the invention?

Admittedly, it's always been difficult to distinguish "abstract" ideas, which are not entitled to patent protection, from "applied" ideas, which are. In short, E=mc2 is not patentable, but applying the formula within a machine (as software on a computer) to run a nuclear reactor more efficiently and safely is patentable.

Easier to Prejudge

Congress has long equipped the USPTO with legal tools to let the quality wheat advance ahead of the chaff. Legal doctrines known as "novelty" and "non-obviousness" help measure whether a purported advance is inventive. But examining patent claims to figure out if they are new and non-obvious is hard work. It's far easier for the USPTO to arbitrarily narrow or prejudge what it thinks should not be patentable subject matter, so it can ease or outright avoid doing its job.

That's exactly what the USPTO did in response to an application by Bernard Bilski for a financial-services "hedging" process. Bilski's patent application was chock full of hopes and dreams that amounted to the financial-services equivalent of the wheel. In other words, it was a poster child for bad patent claims.

The USPTO's decision to deny that application was intellectually consistent with centuries of patent jurisprudence. In 2006, when Bilski appealed the USPTO's denial, the agency could have prevailed simply by arguing that the description was too broad, vague, incomplete, lacking in novelty, and obvious. Instead, the USPTO used the case as a ruse to ask the appeals court to "clarify" what is patentable.

The court handed down its decision on Oct. 30, issuing a ruling that established a new test. For a software process to be patentable, it must either be bound to a particular machine or transformational. The court elaborated only on the "transformation" aspect of the test, noting that there must be a "transformation…central to the purpose of the claimed process." But the court punted on the machine portion of the test.

The bottom line to the above legal gobbledygook is that software and business methods are still patentable, though some fear the court has left many innovative software technologies in limbo by theoretically narrowing what is patentable. In reality, the ruling will just make getting such patents more costly. It will create more work for clever patent attorneys to tie software to physical sources, thereby making it less affordable for smaller companies (BusinessWeek.com, 11/14/07).

Out of Touch

The appeals court did concede that "the Supreme Court may ultimately decide to alter or perhaps even set aside this test to accommodate emerging technologies." Still, the court's decision is horribly out of touch. For example, it's now doubtful that protocol- or signal-based software technologies are patentable. So had this ruling been in effect when the technology that makes the Internet tick was being developed, it's conceivable the technology would have been deemed unpatentable subject matter unless it had been tied to some particular machine. The entire point of the Transmission Control Protocol/Internet Protocol is to send information over networks and not be tied to any particular machine.

It is astounding that the USPTO and appeals court could be such Luddites or show such hubris as to narrow the meaning of a patentable "process"—a decision that only Congress has the purview to make. There's a good chance the court overstepped its authority, intruding on the bounds of the legislature. It is further mind-boggling that the appeals court has effectively shifted the judgment of patentable subject matter down to the USPTO.

Yet, it is very likely the USPTO will not rest until it succeeds in narrowing patentable subject matter and killing off patents for software business methods. There is a good chance the Bilski case may be picked up by the Supreme Court. But if the high court takes a pass, it may take only one more bad appeals decision to kill software business-method patents. And the USPTO has pipelined many more cases to bait the court.

Spate of Knockoffs

If a court finds software and business methods are not subject matter that should be patentable, the rest of the world will feel free to knock off U.S. inventiveness without repercussion. Apple, which took a great risk in making its iPhone, is already contending with knockoffs by small companies. But in making its iPhone, Apple took on the goliaths of the cell-phone industry. It's doubtful Apple would have made the same gamble (risking hundreds of millions of dollars in research and development) without some measure of protection. The proof: Apple filed more than 200 patent applications on its iPhone.

Considering that the cell-phone companies took their eye off the ball and missed that software is key in their industry, it's not too surprising the USPTO doesn't appreciate software's importance. But that lack of vision and foresight explains why an institution like the USPTO should not be the arbiter of whether the equivalent of the software wheel is patentable. Instead, the USPTO should concentrate on its job of discerning if patent applications are novel or non-obvious.

And, by the way, has anyone been paying attention to the economy of late? Perhaps this is not the best time to endanger new technologies, unless of course you make your living shorting Google and Apple stock.

http://www.businessweek.com/technolo...