Because many software patents fail to describe the necessary algorithms, the Aristocrat case will likely result in more software patent claims being invalidated by the courts (at least those written in means-plus-function format) and should encourage patent practitioners to include more details in their specifications. Though the algorithm need not be spelled out in detail, patent practitioners would do well to make sure that their software patents discloses algorithms.
The problem of unreadable software claims
In light of the "algorithm rule", one might ask why a patentee would ever write a software patent using "means-plus-function" language in the first place. One reason has to do with the difficulty patent drafters face writing claims that cover software.
Patent claims are a linguistic nightmare, a fact that tends to be all the more true for software inventions. Part of the reason is that no matter how complex the invention may be, every patent claim is written in one long run-on sentence. Though not required by Congress, Patent Office procedure insists that patent claims be written in single-sentence form. (See MPEP Sec. 608.01(m).) In the only case to challenge the one-sentence rule, the court sided with the Patent Office. See Fressola v. Manbeck, 36 USPQ2d 1211 (D.D.C. 1995). By its very nature, software tends to be a complex series of steps, often iterative and interrelated--which makes it hard to describe precisely in a single sentence.
For example, if a software invention included a mortgage calculator with a routine for calculating compound interest, one way to claim the subroutine would be to detail the steps of the algorithm: (1) receiving the principal, interest rate, compounding period, and term of the loan; (2) applying the compound interest formula to determine the amount to be paid; (3) determining the number of payments in the payment period; and (4) dividing the amount to be paid by the number of payments to determine the size of each payment. Writing this out, along with the other aspects of the program that presumably made it patentable, in a single sentence, all but guarantees unreadable prose.
The use of means-plus-function claims in software
One solution that patent lawyers sometimes use is to write claims using the means-plus-function format. Rather than detail all the steps performed by the computer program, the patent drafter recites a "means for calculating compound interest." This kind of patent shorthand makes the claim much easier to read. But, there is an important catch. Claims written in means-plus function language do not cover every means for performing the recited function. Instead, by statute, they are limited to the structure disclosed in the specification that is clearly linked to performing the function, and to equivalent structures. (See 35 USC § 112, ¶ 6) In the world of software, this means the claims are limited to the algorithm or algorithms disclosed in the patent, and equivalents thereto.
The problem arises if the patent never mentions the general steps that a computer would follow to calculate interest. There is no doubt that a skilled programmer could write a mortgage interest calculator, perhaps as a subroutine, perhaps as an object, perhaps as a lookup table. There could be many ways to write such a program, all within the expected ability of an undergraduate computer student. But, that's not the issue. If the patent claim recites a "means for calculating compound interest," then the patent must disclose an algorithm that is clearly linked to calculating compound interest.
One alternative is simply not to use means-plus-function claim language, and to find a different solution to the claim drafting challenge. But even so, it still makes sense to disclose some specific algorithms in the patent. Patents and the form of their claims often evolve during the years before they issue from the Patent Office. It will generally be good practice to err on the side of additional disclosure up front to provide support for the claims as they evolve.
For example, copyright is self-executing. That means that you get a copyright in your program as soon as you write it. In contrast, a patent is obtained only after filing and prosecuting - and in the process spending a lot of money on - a patent application.
Moreover, copyright is supposed to protect a specific expression - not broader ideas found in that expression. One major argument against software patents is, that in order to maintain the lightning pace of software innovation, developers must to be free to use general concepts pioneered by others so long as they don't copy the actual code.
... Read more
Maybe you are of a sufficient vintage to remember the game show Let's Make a Deal. But have you ever thought about the similarities between that show and the U.S. patent system?
In the game show, contestants would have to pay a price (a wallet containing $500) to see what was behind door No. 3 (maybe a live goat; maybe a brand new faux wood-paneled station wagon). Similarly, in the U.S. Patent and Trademark Office, the government pays a price (allowing a unique brand of monopoly) to see what is in envelope No. 3 (your invention). The analogy may seem far-fetched, but the basic premise is the same: that is, paying a price to see what is otherwise concealed. And even in the realm of patent law, sometimes the government ends up with...a goat.
Fortunately, unlike the game show, there are several ways the USPTO can get out of the deal even after the envelope is opened and the invention disclosed. To be worthy of a patent, the invention must be new, useful, and non-obvious. While the "new" and "non-obvious" requirements normally get most of the attention, the USPTO and the U.S. Court of Appeals for patent cases (the Federal Circuit) have taken a somewhat surprising approach in the past couple of months to back out of deals with potential patentees--rejecting patent applications on the basis of usefulness. In other words, the Federal Circuit has been deciding that certain classes of inventions just aren't patentable.
What is really creating a buzz in the patent world is that the USPTO and the Federal Circuit have recently addressed an almost decade-old class of patents that has developed a reputation as the runt of the litter as far as patents go--business method patents. Love them or hate them, the Federal Circuit's 1998 decision in the State Street Bank case has been widely interpreted to allow for the patenting of new and novel business methods. Since that case, the USPTO has been inundated with business method patent applications and, more specifically, software applications. The question is, will this trend continue?
... Read more
- prev
- 1
- next
