Date: Mon, 31 Jan 94 11:50:57 -0800 From: Les Earnest To: cdh@gnu.ai.mit.edu In-Reply-To: cdh@gnu.ai.mit.edu's message of Sun, 30 Jan 94 16:58:33 est <9401302158.AA22721@goldman.gnu.ai.mit.edu> Subject: Patent Commission hearings Reply-To: les@cs.stanford.edu Here, in case you are interested, is the testimony I gave at the Patent Commission hearings in San Jose. Testimony on Software Patents Based on my 40 years of experience in computer system development, much of it before software patents were introduced, I believe that the alleged connection between such patents and the stimulation of innovation is tenuous at best and probably negative. Let me confess that even though I oppose the continuation of software patents, as a defensive measure I have filed for some that have been granted. When I entered the field as a programmer in 1954 there were only about a hundred of us in the whole world and each of us was turning out hundreds of inventions each year. Software was given the same kinds of protection as other documentation, namely copyright and trade secret. It was certainly a good thing that there were no software patents because my colleagues and I could have papered over the field and retired for 17 years or so to collect royalties. Since such patents didn't exist, we kept working and had quite a good time doing it, sharing ideas and standing on each others shoulders to see how high we could reach. In 1956 I went to MIT to help design the SAGE air defense system. It was a technological marvel, chock full of inventions, both hardware and software. It was the first real-time computer system and depended on the large software system that was cooperatively written by many people. This project helped transfer a lot of technology from MIT to IBM, but almost nothing was patented. Dozens of SAGE systems were deployed around the country, each with a vacuum tube computer that covered a floor area about the size of a football field. It is fortunate that the Soviet Union never attacked the U.S. in that era, because the marvelous technology in SAGE had several "Achilles heels" that would have caused it to fail catastrophically under attack. However, those shortcomings were kept well hidden from Congress and the public and, as a result, so-called "command-control-communications" technology became a major growth industry for the military-industrial complex. The most recent example of that line of development was the grossly defective "Star Wars" system. But that is another story. Beginning in 1959 I developed the first pen-based computer system that reliably recognized cursive writing. I believe that it was more reliable than the 1993 version of Newton, but the idea of getting a patent on such a thing never occurred to me or my colleagues. It wouldn't have done much good anyway, because the computer on which it ran filled a rather large room and the 17 year life of a patent would have expired before small, portable computers became available. In order to cope with a personal shortcoming, I developed the first spelling checker in 1966. I didn't think that was much of an invention and was rather surprised when many other organizations later took copies. Of course, nobody patented things like that. John McCarthy and I organized the Stanford Artificial Intelligence Laboratory and I served as its Executive Officer for 15 years. There was a great deal of innovation that came out of there, including the first interactive computer aided design system for computers and other electronic devices, early robotics and speech recognition systems, a software invention that became the heart of the Yamaha music synthesizers, document compilation and printing technologies that later came to be called "desktop publishing," the Sun workstation and public key cryptography. Few of these inventions were patented in the early period, but we later began to file for such coverage. The pace of innovation has necessarily slowed over time. Concurrently the amount of patent protection has increased. I suspect that these changes are connected. Yesterday in this forum, my friend Paul Heckel said that "Software patents stimulate new businesses." I'm afraid that Paul has that backwards -- in fact, new businesses stimulate software patents. Venture capitalists want the comfort of patents on the products that are being brought into the market, even though "know how" is far more important in most cases. In 1980 I co-founded IMAGEN Corporation, which developed and manufactured the first commercial desktop publishing systems based on laser printers. We filed for software patents to try to appease the venture capitalists, even though it was not actually important to our business. Of course, they didn't understand that and the lawyers were happy to take our money. Based on my experiences, I also joined the League for Programming Freedom to help resist the patent conspiracy and I later served for a time on its board of directors. In summary, for many years there was a great deal of innovation in the computer software field with no patents. Under the "stimulation" of software patents, the pace now seems to have slowed. I believe that there may be a connection, not only because of the time that must be spent in deciding what to cover and preparing the filing, but also because patents owned by other organizations, many of them actually representing prior art, constitute a mine field that must be carefully navigated. I recommend a return to the good old days, when success depended on moving faster than the other guys rather than trying to catch them in a trap.