"He responded derisively, stating that the university
was free but the compiler was not," recalls Stallman.
"I therefore decided that my first program for the GNU
Project would be a multi-language, multi-platform compiler."
Eventually Stallman found a Pastel language compiler written by programmers at Lawrence Livermore National Lab. According to Stallman's knowledge at the time, the compiler was free to copy and modify. Unfortunately, the program possessed a sizable design flaw: it saved each program into core memory, tying up precious space for other software activities. On mainframe systems this design flaw had been forgivable. On Unix systems it was a crippling barrier, since the machines that ran Unix were too small to handle the large files generated. Stallman made substantial progress at first, building a C-compatible frontend to the compiler. By summer, however, he had come to the conclusion that he would have to build a totally new compiler from scratch.
In September of 1984, Stallman shelved compiler development for the near term and began searching for lower-lying fruit. He began development of a GNU version of Emacs, the program he himself had been supervising for a decade. The decision was strategic. Within the Unix community, the two native editor programs were vi, written by Sun Microsystems cofounder Bill Joy, and ed, written by Bell Labs scientist (and Unix cocreator) Ken Thompson. Both were useful and popular, but neither offered the endlessly expandable nature of Emacs. In rewriting Emacs for the Unix audience, Stallman stood a better chance of showing off his skills. It also stood to reason that Emacs users might be more attuned to the Stallman mentality.
Looking back, Stallman says he didn't view the decision in strategic terms. "I wanted an Emacs, and I had a good opportunity to develop one."
Once again, the notion of reinventing the wheel grated on Stallman's efficient hacker sensibilities. In writing a Unix version of Emacs, Stallman was soon following the footsteps of Carnegie Mellon graduate student James Gosling, author of a C-based version dubbed Gosling Emacs or GOSMACS. Gosling's version of Emacs included an interpreter that exploited a simplified offshoot of the Lisp language called MOCKLISP. Determined to build GNU Emacs on a similar Lisp foundation, Stallman borrowed copiously from Gosling's innovations. Although Gosling had put GOSMACS under copyright and had sold the rights to UniPress, a privately held software company, Stallman cited the assurances of a fellow developer who had participated in the early MOCKLISP interpreter. According to the developer, Gosling, while a Ph.D. student at Carnegie Mellon, had assured early collaborators that their work would remain accessible. When UniPress caught wind of Stallman's project, however, the company threatened to enforce the copyright. Once again, Stallman faced the prospect of building from the ground up.
In the course of reverse-engineering Gosling's interpreter, Stallman would create a fully functional Lisp interpreter, rendering the need for Gosling's original interpreter moot. Nevertheless, the notion of developers selling off software rights-indeed, the very notion of developers having software rights to sell in the first place-rankled Stallman. In a 1986 speech at the Swedish Royal Technical Institute, Stallman cited the UniPress incident as yet another example of the dangers associated with proprietary software.
"Sometimes I think that perhaps one of the best things I could do with my life is find a gigantic pile of proprietary software that was a trade secret, and start handing out copies on a street corner so it wouldn't be a trade secret any more," said Stallman. "Perhaps that would be a much more efficient way for me to give people new free software than actually writing it myself; but everyone is too cowardly to even take it."
Despite the stress it generated, the dispute over Gosling's innovations would assist both Stallman and the free software movement in the long term. It would force Stallman to address the weaknesses of the Emacs Commune and the informal trust system that had allowed problematic offshoots to emerge. It would also force Stallman to sharpen the free software movement's political objectives. Following the release of GNU Emacs in 1985, Stallman issued " The GNU Manifesto," an expansion of the original announcement posted in September, 1983. Stallman included within the document a lengthy section devoted to the many arguments used by commercial and academic programmers to justify the proliferation of proprietary software programs. One argument, "Don't programmers deserve a reward for their creativity," earned a response encapsulating Stallman's anger over the recent Gosling Emacs episode:
"If anything deserves a reward, it is social contribution," Stallman wrote. "Creativity can be a social contribution, but only in so far [sic] as society is free to use the results. If programmers deserve to be rewarded for creating innovative programs, by the same token they deserve to be punished if they restrict the use of these programs."See Richard Stallman, "The GNU Manifesto" (1985). http://www.gnu.org/manifesto.html
With the release of GNU Emacs, the GNU Project finally had code to show. It also had the burdens of any software-based enterprise. As more and more Unix developers began playing with the software, money, gifts, and requests for tapes began to pour in. To address the business side of the GNU Project, Stallman drafted a few of his colleagues and formed the Free Software Foundation (FSF), a nonprofit organization dedicated to speeding the GNU Project towards its goal. With Stallman as president and various hacker allies as board members, the FSF helped provide a corporate face for the GNU Project.