IBM - Software Patent Policy

Public Hearing on Use of the Patent System to Protect Software Related Inventions
Transcript of Proceedings Wednesday, January 26, 1994 San Jose Convention Center

Victor Siber, Senior Corporate Counsel, IBM

Mr. Siber: Good afternoon, and thank you for the opportunity of allowing me to deliver my comments this afternoon.

I am Victor Siber, Senior Corporate Counsel for the International Business Machines Corporation. My comments do represent the views of IBM.

The IBM Corporation spent 6.5 billion dollars on research and development in 1962. In that same year, over 22,000 software programmers were employed by the company, and we sold approximately 18.5 billion dollars of software products and services. So obviously, we are intensely interested in the subject matter of these hearings.

We protect the detailed expression in every one of our software products by copyright. And approximately 3 to 5 percent of these programs contain new and unobvious functions that are protected by patent. Patent coverage on these inventive functions protects our investment, gives us important business leverage, as well as access into foreign markets.

I understand that this hearing was scheduled, in part, to address questions raised by some broad software patents that have recently issued. These patents are alleged to cover old processes. The response by some is to call for changes in the law to prevent the issuance of patents in this technology.

In the short term that would hurt the U.S. computer industry. And in the long term all industries that use computers to gain competitive advantage will suffer.

The argument over these controversial patents is not that they cover unpatentable abstractions. These patents cover quite useful functions, which others in the industry want to use. The issue, and the key to the controversy is something quite different: whether these patents cover truly new and nonobvious functions and thereby add something to the useful arts. In the final analysis that is the job expected from the U.S. Patent and Trademark Office. And it is a job that is doable.

The United States was the first and continues to be the global leader in computer hardware and software technology. This technology is important not only in itself, but as a driver of innovation in other fields. And today, it is facing stronger and broader global competition. As the industry matures and competition from overseas increases, patents will be the key to protecting the most valuable US-originated innovations.

You have posed a number of questions at the today. Beginning with Question 3 relating to altering the standards for patent eligibility for software-related inventions, I want to make it clear that we completely oppose such a proposal.

Going into the next century, the key inventions will be in information processing. Altering the standards for patent eligibility for software-related inventions will shift investment away from this area.

The purpose of research and development in any technology is to gain an advantage over your competitor. But if your competitor can legitimately copy the fruits from your R&D and can create a product that can compete head-on with your product while you are still trying to build a market for the product, then you've lost.

The long term value of R&D in the marketplace is in the new functions implemented by software. If such new functions are protected, investment flows to the industry. If not, investment will dry up.

There are several other points I want to make on this issue. We can't divorce computer program-related inventions from computer hardware and other microprocessor inventions. The overlap between the two is so great that cutting back on one automatically cuts back on the other.

An alteration in the standards for patent eligibility will also put the courts and your office in a quandary. That is because computer program-related inventions can be implemented in either hardware or software. Applicants will simply cast their patent claims in terms of electrical circuitry. And if you limit claim coverage over computer program implementations of circuit inventions, you will turn electrical patents into nullities, because the circuit functions can be implemented by a programed computer.

Furthermore, if process patents cannot reach computer implementations of the process steps, there will be a negative impact on every industry that uses computer-controlled industrial processing or uses microprocessors in their products.

From an international standpoint, a cut-back on patent eligibility for computer program-related inventions sets a truly unfortunate precedent for the developing world. It would violate GATT and NAFTA provisions that prohibit discrimination of patent eligibility based on technology. The biotechnology area provides a graphic example of what happens when countries limit protection for a technology. Leading European companies working in this area of technology simply moved their biotech R&D operations to the United States.

Concerning the issue of sui generis protection, replacing patent protection for new and unobvious program and process functions, and replacing copyright protection for the original expression contained in computer programs with some form of sui generis right would be devastating to the industry.

In a single act, U.S. industry would unilaterally be disarmed relative to our competitors in Japan and Europe. We would lose patent priority rights in the 114 countries of the Paris Convention, and we would lose the copyright protection automatically afforded American program works in 102 countries as required by the Berne Copyright Convention.

As you know, the Trilateral Work Studies conducted jointly by the U.S. PTO and the Japanese Patent Office and the European Patent Office, concluded that the standards for patentability in the area of software-related inventions are generally the same except at the far margins.

The Japanese Patent Office has published for opposition in the last six years approximately (45) patent applications on software-related inventions, while the European patent office has issued approximately 2700 patents since 1980 which have resulted in multiple European member country patents.

The majority of these software-related patents in the European Patent Office and the great majority in Japan are issuing to non-U.S. companies. Without U.S. patents of a reasonable scope to bargain with, U.S. companies will potentially lose access to those markets.

With the introduction of a sui generis system, there would be --

Commissioner Lehman: Can I ask a question? Why do you think that is, that the majority of those patents -- we issued 8,600 patent this last year in the United States, so clearly we've got 8,600 people, and I think I cited those statistics before, two years ago it 6,500 or something like that. Clearly, we have the inventions, but people aren't seeking, American's aren't seeking patent protection in those foreign markets, why that is?

Mr. Siber: I think there's a questions of maturity of the industry, many smaller companies are not familiar with the process of filing overseas, possibly may not be able to afford it or feel they can't afford it.

Secondly, it is natural, particularly in Japan to have a large number of Japanese patent holders where the filings of large Japanese companies is very high.

Commissioner Lehman: So, but IBM, I assume does file in those.

Mr. Siber: Yes, we do.

Commissioner Lehman: And pretty extensive. I would assume when you file a patent application here, you file there too, since you're a worldwide company.

Mr. Siber: Correct.

Commissioner Lehman: Is that generally true? Are there situations where you don't file there and you would file here?

Mr. Siber: We do not file a counterpart patent in Europe and Japan for every patent that we file in the United States. It is selective, it is a selective process. We too have budget limitations, even though we just turned a substantial profit last quarter.


Commissioner Lehman: Sorry, to interrupt.

Mr. Siber: With the introduction of a sui generis system there would be an extended period of uncertainty, the bane of businessmen and investors, as the body of case law were developed interpreting the new sui generis law and determining its scope.

Additionally, there would be a complete absence of effective international protection for the new right. The internationalization of a sui generis law would require a sui generis treaty. Such a treaty would be negotiated without any of the international consensus on what sort of protection regime would be appropriate. As you know, the primary multilateral treaty to date for a sui generis right, The Treaty on International Property in Respect of Integrated Circuits, known as the Washington Treaty, was strongly influenced by developing countries that were hostile to IP protection generally. The result was a treaty so flawed that not one single major chip-producing nation supported it.

The only remaining way to internationalize that sui generis chip-protection right was through reciprocity. But reciprocity has it rewards and its vices. The European community now points to the reciprocity provisions in the U.S. Semiconductor Chip Law to justify its discrimination against America's authors through the reciprocity provisions in the E.C. Directive on Copyright Term Extension, and for denying U.S. authors the benefit of unfair extraction right in the draft E.C. Directive on Database Protection, and in refusing to grant movie and music producers and authors their fair share of the blank tape and movie levies.

Finally, regarding the series of claim format examples in Question 1, we believe that the format of the claim should be viewed from the perspective of patentable subject matter. If the claim as a whole is directed to a machine, article of manufacture, or a machine-operated process, then subject matter eligibility should not be an issue. The focus of the examiner should be on the normal statutory tests of novelty, nonobviousness, and utility of the claim as a whole. And to insure compliance with the formality requirements of 35 USC 112.

Specifically, a claim directed to an article of manufacture comprising computer-usable medium and a plurality of computer-readable program code means for causing a computer to effect a plurality of specific and interrelated functions should be eligible for patenting under the statute. Such an article of manufacture constitutes a machine part that is commercialized separately. And it should be protected separate, like other machines parts, if it meets the statutory test of novelty and nonobviousness.

In summary, the patent system is designed to instigate the invention of new nonobvious and useful functions that add to the arts. The key to seeing whether the system is working is to see if there is strong competition in the marketplace; to see if new products are introduced in the market on a regular basis; to see if employers and investors vote for the system with their pocketbooks by funding new development work.

As for my company, we rely heavily on the patent system to protect our investment in new products. And we are negatively impacted probably more than most by poorly-examined patent applications. Thus, we want the issuance of poorly-examined patents curtailed. And that is clearly doable if the PTO would invest in hiring and training the best possible examiners, as well as in the creation of an adequate database for software prior art.

Thank you for the opportunity. I would be delighted to answer any questions.