1 2 3 4 5 6 7 8

Rivest, Shamir, and Adleman's work developing RSA at MIT was done with funding from the Navy and the National Science Foundation. As is the way with these things, although the government retained the rights to use the technology itself, the ideas were patented by MIT and the rights were handed over to a new company, Public Key Partners, for commercial exploitation. This outfit also had the rights to the patent for the Diffie-Hellman key exchange scheme and two other key patents in this area, giving it exclusive licensing rights to what seemed in 1992 to be all of public-key cryptography. The sole licensee for RSA at the time was the California- based company RSA Data Security. In a document dated December 4, 1992, covering legal issues with respect to PGP, Zimmermann stated that he had obtained the opinion of a patent lawyer before proceeding with PGP, and that he was therefore "convinced that publishing PGP the way I did does not violate patent law." He did not, he argued, steal any source code: "I wrote my PGP software from scratch, with my own implementation of the RSA algorithm."

Nonetheless, a complaint from Jim Bidzos, president of both Public Key Partners and RSA Data Security, speedily followed PGP onto the WELL. Within four days of its original June 7, 1991, posting, the original poster removed PGP from the WELL's libraries after a request from the system's management, who felt this was a legal dispute the WELL could afford to miss. The legality of using PGP within the United States was a hotly debated topic on the Net for the next couple of years. Endless megabytes were churned out in cryptography-related newsgroups like the scientifically oriented s c i . c r y p t, with one side arguing that it was important to respect intellectual property rights and the law, and the other side insisting that a mathematical algorithm should not be patentable.

This is not just sophistry and rationalization: the question of what constitutes intellectual property is constantly being revisited. In general, copyright law has covered the expression of ideas and facts, not the ideas or facts themselves. New designs for machines or drugs can be patented; but the sequence of words describing what those machines and drugs do and how they do it can only be copyrighted. Software, however, raised a new problem: was it a literary work or a new type of infinitely copyable machine? The European Community considered this question and decided to class it as a literary work. In the United States, however, the use of patents to protect software--even such basic routines as drawing a cursor on a computer screen--began proliferating in the 1980s. The difference is profound. Copyright protects the precise arrangement of words and letters on the page. Patents protect functionality, that is, what the program actually does. You can copyright a book's text; but if the book itself were invented today someone would patent it.

Because patenting was such a profound change from the early days of the industry, when software was generally free, in 1983 MIT researcher Richard Stallman set up the Free Software Foundation with the goal of creating a full suite of free software designed to run on the computer operating system UNIX.[10] UNIX itself was free, as it was written at AT&T at a time when the company was enjoined from selling such non-telecommunications products. At least one company, Cygnus Support, has been set up with the idea that software should be free and companies should derive their incomes from selling services and support for that free software, following the same kind of argument circulated by John Perry Barlow. Not surprisingly, one of Cygnus Support's founders was also one of Barlow's EFF co- founders and the man who invented the alt hierarchy, John Gilmore.

Stallman, Gilmore, and many other industry leaders argue that the widespread use of software patents can chill technological development.[11] They have a point: technology develops incrementally, as each new developer tries to improve on what others have done. The entire computer industry is built on the practice of