There is an international crisis in software patents. The critical, unanswered question is this: When are software inventions eligible for patent?

Billions in corporate assets are at stake. Business is threatened when rules of law and intellectual property are in doubt.

Patent offices and courts around the world are paralyzed with backlogs of pending, rejected, challenged, and appealed software patents and applications.

Microsoft, IBM and SAP each make tens of billions in revenues from software. Each has thousands of patents and applications. The same is true of Lenovo, HP, Apple, and Samsung. Large, medium, and small enterprises, start-ups and IPOs, are spurred by software inventions and capitalized on the basis of software patents. Yet the validity of those patents—the very eligibility of software inventions—is hanging in the balance. The patent offices and courts of the U.S., Europe, China, Japan and Korea are upholding many such patents and applications even while striking down others, often with little rhyme or reason.

Wrestling with software patent eligibility

On Tuesday of this week, I started the day with a 6:00 a.m Central Standard Time international conference call with patent attorneys, lawyers and barristers from U.S., England, Germany, China, and Australia (where it was 11:00 p.m.). We are all struggling to make sense of the various, inconsistent, conflicting laws of patent-eligibility of software or, as much of the world says, computer-implemented inventions (CII).

In 2003, after thorough study and detailed reports by various national groups and much international debate, the International Association for the Protection of Intellectual Property (AIPPI) adopted a resolution as follows: “Computer implemented inventions should be eligible for patent protection and should not be treated more restrictively than other inventions.” Yet even today the question of patent-eligibility of software is very much alive and unresolved. On September 19, 2016, I presented, as a panel member, on the topic at the AIPPI World Congress in Milan, and the subject is designated for a full-blown study question for the upcoming 2017 Congress in Sydney.

The American Intellectual Property Law Association (AIPLA) also is struggling to make sense of the vicissitudes of software patent eligibility. The issue was a major subject of discussion at the AIPLA Annual Meeting in Washington, D.C., at the end of October. In a sign of the times, a large room devoted to a presentation on copyright law was filled beyond capacity with patent attorneys seeking alternative forms of intellectual property protection for computer software. See my article, 5 Ways to Protect Your Proprietary Software after Oracle v Google.

On the same subject (patent-eligibility of software), just Friday morning, I met with the American Bar Association, Section of Intellectual Property Law, Leadership Council, where we discussed the issue at some length, without coming to any consensus. We know it’s broken. But there seems to be no agreement on any easy (or not so easy) fix.

Here, as we say in AIPPI, I am “speaking for myself,” and not for the larger nongovernmental organizations (NGOs) of which I am a member. By the same token, these are my views and not necessarily those of my clients. Yet many share my concerns.

For software, the patent systems of the world are broken

Yes, the patent systems of the world are broken, for software.

For fifty years, the U.S. patent system has found itself unready to deal with software. A President’s Commission so found in 1966. A trilogy of U.S. Supreme Court cases in the ’70s tried to make sense of the issue. The U.S. Patent and Trademark Office and the U.S. Court of Appeals for the Federal Circuit struggled with the issue in the ’80s and early ’90s. Then it was decided that every invention under the sun—every technology created by humankind—is eligible for patent. A case known by the name of the party State Street Bank (Fed. Cir. 1998) stated that proposition. That case involved a financial or “business method,” then held to be patent-eligible, even while using a phrase (business method) that has since become the kiss of death for any patent or application so denominated.

The Supreme Court again wrestled with the subject in Bilski (2010) and most recently in Alice v. CLS Bank (2014). The current state of the law: Software is eligible for patent in the U.S., so long as the claim is directed to “something more” than an “abstract idea.” It does not suffice, however, to take a conventional method of doing business, such as currency hedging, and to say simply, “run it on a computer.”

Alice is not the answer

It is not so much that Alice is wrong as that it is often unhelpful in gatekeeping at the patent offices of the world.

All inventions, of course, are based at some level on “abstract ideas.” The laws of physics, and chemistry, and biology, can be reduced to abstract ideas. And there’s the rub. If every invention is based on an abstract idea, it is subject to rejection under Alice as patent-ineligible under 35 USC 101. (The Federal Circuit has held at least one software patent claim to be non-abstract, but the holding in that appeal does not necessarily ease the mind of the examiner.) And the holy grail of “something more” is elusive.

Software inventions are important

The thing about software is that it drives so much of modern technology. What used to be done with custom hardware is now done more elegantly and efficiently with software run on a general purpose computer. The former clearly is patent-eligible, the latter may not be.

In technical fields such as virtual reality and video games, inventions often are grounded primarily in software, yet they may or may not be patentable, and it can take many years and even more dollars to obtain final determinations from the various patent offices and courts of the world.

At least 500,000 new businesses are started every year in the U.S. alone. See WSJ (Nov. 16, 2016). Many of them are based on software development, which requires great creativity and inventiveness, and relatively little capital. How many of those new businesses are launched to make computer chips in the U.S.? You probably won’t need all your fingers to count them.

To develop and make a new computer chip (hardware) requires millions if not billions in capital equipment; that development, no matter how trivial, will be eligible for patent. Yet to accomplish the same functionality with software, no matter how inventive, will be slammed with rejection at the threshold of eligibility.

On a related theme, in the U.S., it is held that one cannot patent a “law of nature.” This rule is used to strike down software and medical patents, among others.

Japan has picked up on this principle, but stating it in exactly the opposite way, holding that one cannot patent an invention unless it based on a law of nature. This might make more sense than the U.S. rule, because, after all, applied technology should be based on laws of nature.

Europeans have their own challenges

In Europe the situation is not much clearer. There, computer-implemented inventions (CII) are not eligible for patent “as such.” (Software is the word that should not be named.) The European Patent Office (EPO) goes through an elaborate process of determining whether an invention comes from the field of technology, as compared, for example, to the field of design. The former is eligible for patent; the latter is not. There must be a technical problem, with a technical solution, and a technical effect. The German practice of subtracting everything that is not technical seems to be applied by the EPO. Everything nontechnical is ignored. What is left? Not much, it seems, in software cases. Whatever is left of a technical nature often is deemed to be eligible for patent, however, that little pile of scrap, sad to say, may be found to lack “inventive step.” Sorry, no patent for you. One thing missing from the European approach is the need to look at the invention “as a whole,” a principle embedded in U.S. patent law.

Patience and persistence are needed in US patent prosecution

The U.S. Patent and Trademark Office has struggled, and continues to struggle, with patent-eligibility of software. First one and then another set of “guidance” is issued. Example after example is given of claims found eligible versus claims found ineligible. It’s a mess. Examiners don’t know what to do. What’s safe for them to do is to reject, and to reject again. But that is costly to applicants and to the world of business, which runs best on established rules and predictability.

The victims of the software patent crisis are patent applicants, patent owners, and accused infringers. A Silicon Valley-backed foundation rails against software patents even while Silicon Valley companies file and acquire wholesale numbers of patents and applications. Silicon Valley hedges its bets.

Sit this one out?

One might be tempted to sit out this dance. The day is short, but patent term is 20 years from filing. Smart companies can’t afford to let their guard down—to cease filings or abandon portfolios—to let competitors run ahead and stockpile patents and applications. A head-in-the-sand approach may save money in the short term but it is far from certain to yield profits over a score of years.

Legislation on patent-eligibility of software will be introduced in the U.S. Congress. AIPPI, AIPLA, ABA, and other NGOs, and, indeed, governments, will tackle the question again.

It may not be pretty: Efforts to define “technology” and to provide definitive lists of eligible versus ineligible subject matter are unlikely to be helpful. More law—more lists—will spawn more questions and lead to more maneuvering by applicants to present inventions on the affirmative side of patent eligibility.

The existing U.S. statute is workable

Section 101 of 35 U.S.C. has stood the test of time since at least as early as 1952. It states as follows:

“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.”

“Useful,” here, should be interpreted as “applied,” in the same sense as “useful Arts” in the U.S. Constitution, Art. I. Section I, clause 8. There was no word “technology” in 1787. That’s what “useful Arts” meant then: applied technology. That, dear readers, should be the threshold for patent-eligibility of software. As AIPPI previously resolved, “[c]omputer implemented inventions should be eligible for patent protection and should not be treated more restrictively than other inventions.”

The other conditions and requirements, per 35 USC 101, et seq., include novelty and non-obviousness (inventive step) and a fully detailed written description sufficient to enable one of ordinary skill to practice the claimed invention.

When a software invention is disclosed and claimed, objectively and without semantic gamesmanship, as a new and technically-useful-and-applied process, it should be allowed to pass the broad gate of patent eligibility and to be searched and examined for the other conditions and requirements of the patent laws.


New, technically useful and applied software programs should be eligible for patent. If it takes another act of Congress, spurred by resolutions of NGOs, to make that clear, so be it.

What to do now? Software developers may wish to take a belt-and-suspenders approach and protect themselves with patent applications, trade secrets, copyright registrations, and written agreements. See my article, 5 Ways to Protect Your Proprietary Software after Oracle v Google. If you’re wondering whether patents are right for your business, see my series on the 4 W’s of Patent.

Some of the most successful software companies in the world—Microsoft, Google and Facebook—blazed their trails by launching new technologies that took the world by storm. Creativity, drive, capital, market leadership, and belt-and-suspenders IP protection: These remain the key ingredients of a recipe for success.