The U.S. Supreme Court has done the right thing in agreeing to review the Court of Customs and Patent Appeals' decision on the patentability of Thomas R. Johnston's system of financial recordkeeping (CW, May 14 & 21).
The implied willingness of the high court to consider the broader question of software patentability in general is even more encouraging.
A patent on the particular system under review seems to be rather inappropriate. Don Leavitt, CW software editor, recalls writing a program that used an added digit in the Micr-encoded strip on checks and other docurments to separate the documents, and the values they represented, by application. That's apparently the essential logic of Johnston`s program.
But Leavitt's work was 10 years ago, in one of the first IBM 360/1419 Micr reader/sorter installations in the Boston area. Viewing himself as an average programmer, he argues that if he could develop such an approach, it couldn't really be unique or unobvious - basic prerequisites for a patent in any case.
But there are certainly some software systems that deserve special protection for the innovators that created them. Ideas as such are apparently not patentable, but the dodge of posing the programming logic in the guise of "apparatus" or a "machine", as Johnston did in his appeal, is clumsy at best.
In 1972, the high court rejected the pleas of inventors Benson and Tabbott to grant a patent on their program, finding the routine being presented was "essentially a series of mathematical calculations or mental steps" not suitable for protection.
But Justice William O. Douglas recognized the possibility of misinterpretation of that particular finding: "lt is said the decision precludes a patent for any program servicing a computer. We do not so hold."
The court seemed to look then to Congress for clarification on the rules that perhaps should be followed to allow patenting of software. Clearly Congress has had a multitude of other more pressing concerns since then, and the issue still hangs in abeyance. The software industry itself is split on the issue of patentability. Some feel strongly that conventional patent protection is fully justified. Others don't necessairily disagree, but note the mere act of applying for a paten puts all the innovator's code on the public record.
Use of even parts of the code thus exposed would be tempting to the unscrupulous and difficult for the patent holder to identify, gain compensation for or halt on the basis of patent infringement. Without - for now - any clear-cut pricture of what the solution should be, we nonetheless urge the Supreme Court to spend enough time on the question of patents for software so it will provide comprehensive and comprehensible guidelines for all parties concerned - developers, software vendors and users.
Editorial aus Computerworld