kurt@tc.fluke.COM (Kurt Guntheroth) (10/07/87)
The patent law says (approximately) "You cannot patent a law of nature." An interpretation of this rule says that an algorithm is a law of nature. That is, there is nothing substantial about an algorithm. The same thing goes for a chemical reaction. You cannot patent one because it is a law of nature that chemicals will react in ways determined by their environment and the laws of physics that control individual atoms of the reagents. What a patent protects is the application of an idea to a physical device or process. You can't patent a formula. You can patent the use of a chemical compound in the treatment of a disease. You can't patent a chemical reaction, but you can patent the manufacturing process including pressure and temperature and types of catalyst that permits the reaction to take place. This is often equivalent to patenting the reaction. It was formerly assumed that you couldn't patent a program, which was just the expression of an algorithm, after all. Now, patent law recognizes the patentability of programs when they are embedded into hardware for a specific purpose. For instance, you can't patent a complex heuristic search algorithm, but you can patent a "method for constructing a chess-playing machine employing a complex heuristic search." You can't patent an algorithm for manipulating numbers using large prime numbers and prime factoring, but you can patent a "method for encoding and decoding messages using techniques involving prime factoring and two keys, one of which is public." (or whatever, I am not too clear on RSA). Now, this covers more than just a specific machine for coding or chess playing. If you run a chess program on your computer you have at least temporarily constructed a chess palying machine. So software may after all be patentable, if it is employed for a specific purpose. I once attended a meeting for engineers here put on by the patent lawyers. They talked about all the things you need to do to get a patent. Then they talked about all the things that people did to break patents. It seemed that no matter how much protection you piled on, no matter how careful you were, that your patent was vulnerable. I kept asking "but if you add this protection, and that, and have your disclosure right, and the idea is really original and revolutionary, then your patent is strong, right?" What he said at the end was "Every patent suit goes to court and usually to a jury. Or at least to a judge who is not necessarily familiar with the esoteric aspects of your invention. Because people decide the law, you cannot know when you go in what the outcome will be, and there is no perfect protection. Your opponent may be a better public speaker than you, or his plight may be more sympathetic to the court, or the judge might have indigestion, or any number of other things. You just do what you can, you go to court, and hope for the best." Great world, huh. Disclosure: I am not a lawyer, but I know a lot of lawyer jokes. The preceding paragraphs do not constitute a legal opinion or advice of any kind. Some assembly required, void where prohibited.