Here Come the Suits: The Dramatic Rise in Software Patents Issues Promises an Equally Dramatic Rise in Cases to be Filed
While much of the computer industry has been focused on the Y2K problem and many software companies are worried about what their applications will do on Jan. 1, 2000, another significant situation has been quietly brewing for the past few years that will also present a major challenge to the industry in the new millennium. Over the past decade, software companies have increasingly sought patent protection for their products, as shown by the dramatic increase in the number of patent applications filed and patents issued for software- related inventions. When these figures are combined with the rise in patent litigation generally, the new millennium "doomsday" for the software industry may not be Y2K, but an onslaught of lawsuits to enforce those newly minted patents.
The reason why the millennium bug is attacking now is all too obvious, but what dark fate set off this patent explosion? The rapid growth in the number of patents sought and obtained may be attributable to a settling of the case law in recent years on whether software and computer-related inventions are patentable in the first instance.
Back in the Iron Age of computer technology, it was thought that a computer program should be considered a "mathematical algorithm," and therefore a "law of nature" not subject to patent protection. Then, three Supreme Court decisions, handed down between 1972 and 1981, established that computer programs are not unpatentable per se, and that the mere fact that an application claims a mathematical algorithm does not automatically make it non- patentable subject matter. See Gottschalk v. Benson, 409 U.S. 63 (1972); Parker v. Flook, 437 U.S. 584 (1978); and Diamond v. Diehr, 450 U.S. 175 (1981).
However, it was not until the mid-1990s that the U.S. Court of Appeals for the Federal Circuit rendered its crucial en banc decision in In re Alappat, 33 F.3d 1526 (Fed Cir. 1994). Alappat articulated a test for evaluating algorithm-based software and cleared the way for the patenting of software inventions. And in early 1996, the U.S. Patent and Trademark Office (PTO) issued examiner guidelines for software patent applications.
In general, under Alappat, claims containing algorithms are patentable if they (1) produce a "useful, concrete, tangible result," and (2) are narrowly drawn so as not to preclude use in unrelated applications. See also State Street Bank & Trust Co. v. Signature Financial Group Inc., 149 F.3d 1368 (Fed. Cir. 1998). Of course, to be patented, the software-related invention must also meet the threshold requirements of "novelty" and "nonobviousness."
The statistics available on the number of software patents issuing on a year- to-year basis are imprecise, at best, due to the fact that software appears in patents falling within many different fields of art. Some rudimentary statistics from the PTO, however, confirm what virtually every law firm with a patent prosecution practice, and many without such a practice, know to be true:
The software patent business is booming.
Software Patents Issued
Total 6135 6417 6986 8158 9271 10605 13498 12768 73838
To perform a rough evaluation of the rates at which software patents issued during the 1990s, the authors considered the total number of patents issuing from PTO art groups highly likely to consist predominantly of software patents: 341, 345, 364, 382, 395, 701, 704, 705, 707, and 711. Although the numbers within each individual group are not particularly telling because of reorganizations at the PTO, the totals clearly reveal a trend.
While the number of all patents issuing from the PTO each year has increased by 30 to 40 percent since 1990, the number of patents issuing in the software art groups in the same time period has increased by more than 200 percent--from a total of 6,135 in 1990 to a total of 12,768 in 1997. (Actually, 1996 was the busiest year, with a total of 13,498 patents issuing in the software art groups.)
Furthermore, searches for patents issuing to companies that one might consider to be today's software giants reveal that most of them had few, if any, patents issuing in 1994 and then saw exponential increases during the next three years. For example, just five companies--the Microsoft Corp., the Oracle Corp., Sun Microsystems Inc., Apple Computer Inc., and Novell Inc.--saw a nearly 600 percent increase in patents issued in art groups 364 and 395 (the most frequently assigned software related groups) from 1994 to 1998. Interestingly, despite these general increases, many well known software companies have limited or no patent portfolios and will be vulnerable to patent infringement suits in the coming years.
This trend in software patents is likely to continue into the new millennium. The Federal Circuit's 1998 decision in State Street Bank appears to have opened the door further for software patents by allowing protection of business methods. That decision may also draw nontechnology industries into the patent arena. In fact, since State Street, many businesses that have not traditionally been active in applying for patents, such as banks, have been making inquiries regarding protection for their electronic commerce efforts.
The greatest limitations on the rate at which the number of software patents increase may actually be the numbers of patent practitioners qualified to write the applications and the number of patent examiners qualified to examine those applications. The PTO has highlighted computer software as one of three fields that have had "significantly higher growth rates" in applications, and its current budget calls for additional training and increases in staff to address this rapid growth. (The other two fields are also computer-related: communications and semiconductors.)
In light of these statistics, companies that write any kind of software would be wise to prepare for the coming onslaught of software patent litigation.
First, if it has not already done so, the company should file defensive patent applications for its own software products. These patents are invaluable in the event that the company is sued for infringement by a competitor, because they can be asserted in counterclaims against the competitor. The presence of applicable defensive patents might also dissuade potential plaintiffs from bringing suit in the first place for fear of opening themselves up to infringement claims. (Even if no litigation is brought against the patentee, the patents obtained as part of this strategy might provide some licensing revenue.)
Second, the company should begin building a library of prior art relevant to its software products. The purpose of this prior art will (hopefully) be to prove the invalidity of any patents asserted against the company in litigation. The fact that most software patents have been recently issued means that, unlike in most other fields, one will have difficulty finding useful prior art in previously issued patents. The early versions of software that contain the basic elements that might be improperly claimed in a more recent patent often were not patented themselves, so no easily accessible public record of their attributes exists.
In addition, because software develops at what is now known as "Internet speed" and obsolete versions are quickly discarded, finding working versions of early software that serve as prior art can be very, very difficult. For example, if a software patent claims some feature that was actually present in MS-DOS Version 2.0, one might have to find a computer that is still running that 15-year-old version of DOS in order to prove the patent invalid.
Moreover, software patent claims are often directed to minute details that do not appear in trade publications or instruction manuals. While some expert witnesses have begun to build their own prior art libraries for use in software patent litigation, a company should collect samples relevant to its own particular products.
Nobody has a crystal ball in which she can see the software industry mired in patent litigation for the first decade of the new millennium. But with the limited pool of prior art and the rapidly rising numbers of patents issuing, all the elements of a free-for-all exist, and computer-related industries should be prepared.
Timothy R. DeWitt is of counsel and Jule L. Sigall is an associate with D.C.'s Arnold & Porter. They are members of the firm's intellectual property and technology practice group.