Many current authors and commentators (though perhaps not as many patent attorneys) contend that the U.S. Patent and Trademark Office (PTO) is fine at examining patents for mechanical devices and pharmaceuticals, but argue that we need a whole new system for examining software inventions best understood by those under 40. They say that the PTO is not set up to examine software patent applications and, as a result, frequently issues patents on trivial improvements to existing practices. Several such authors and commentators have misinterpreted the outcomes of recent court decisions as confirming these notions, when in fact they debunk them. 

There is another point of view. As discussed below, although patent examiners are human, and determining whether an innovation is worthy of a patent is a very complex undertaking, when the facts and arguments are scrutinized, the PTO is likely, on balance, doing a pretty good job of examining software patents.

Software patent filings became very popular after the 1998 Federal Circuit decision in State Street v. Signature Financial Group, which upheld a patent on a data processing method used for managing groups of mutual funds. Faced with an explosion of software patent applications and popular commentary decrying what a supposedly bad job it did examining those applications, the PTO established a special unit that subjected many software patent applications to supplemental review. Furthermore, in the nearly 15 years since State Street, the agency has accumulated a voluminous record of prior art for examining new software patent applications. Anyone who practices in the software field can attest to how strict the office can be and that it has made great strides in improving the quality of the software patent examination process. 

The Inc. one-click patent (5,960,411) is frequently held up as an example that the agency will issue a patent on “anything” software-related. In fact, at least two different units within the PTO subjected this patent to extensive review. After it was issued, Barnes & Noble Inc. expended considerably more resources searching for grounds to invalidate this patent than the PTO could be expected to muster, but failed to convince the district court that the one-click patent was invalid. One can hardly fault the patent examiner for granting a patent that could not be discredited in court. Furthermore, procedures are in place whereby patents can be subjected to reexamination when new information comes to light, and the one-click patent largely survived reexamination. A recent PTO study confirms that courts uphold the validity of the vast majority of software patents when those patents are challenged.

An Apple Inc. patent (8,068,604) that played a prominent role in the company’s smart phone lawsuit with Samsung Electronics Co. has also been cited as an indicator that the PTO is too free in granting software patents. However, the agency tenaciously rejected Apple’s application nine times over almost 12 years, finally concluding that Apple had sufficiently narrowed and precisely defined the metes and bounds of its patent. When challenged with infringement, even Samsung’s monumental efforts could not prove that this patent was invalid. One might suppose that the office’s “decision on the field must stand” for lack of definitive evidence to the contrary.

Complaints about a lax PTO granting patents on old and/or obvious technology are not new. These same complaints have been raised since the 1800s. When John Glidden sought to enforce his patent on barbed wire, he encountered a parade of witnesses who claimed they had seen wire fences with barbs, that his fence was the obvious implementation of a thorn bush and that Glidden’s contribution of twisting the wire to hold the coiled barb in place was a trivial improvement. However, as the Supreme Court explained, it becomes very difficult to remember exactly what one has seen in the past. It becomes even more difficult when one has only a vague, oversimplified understanding of precisely what it is that a patent protects. Writers and commentators often share this shortcoming of not fully appreciating the precise complex interaction of different elements that must be present before the patent they are denigrating can be infringed. If an improvement really is trivial, then the market is free to ignore it. However, as in the barbed wire case, the almost universal adoption of what some call trivial software improvements often establishes just how non-trivial those improvements really are. As the Supreme Court noted: “Now that [the supposedly trivial invention] has succeeded, it may seem very plain to any one that he could have done it as well. This is often the case with inventions of the greatest merit.”

In fact, the PTO likely does about as good a job of examining software patents as it does in other technologies. What might appear to be trivial improvements are often very detailed, specific and critical refinements to existing technologies. Perhaps we ought to pause before we try to fix a system that likely does not need fixing.