“Just as every writer dreams of writing the great American novel, every programmer back in the 1980s talked about writing the great American compiler”, Tiemman recalls. “Suddenly Stallman had done it. It was very humbling”.
“You talk about single points of failure, GCC was it”, echoes Bostic. “Nobody had a compiler back then, until GCC came along”.
Rather than compete with Stallman, Tiemann decided to build on top of his work. The original version of GCC weighed in at 110,000 lines of code, but Tiemann recalls the program as surprisingly easy to understand. So easy in fact that Tiemann says it took less than five days to master and another week to port the software to a new hardware platform, National Semiconductor’s 32032 microchip. Over the next year, Tiemann began playing around with the source code, creating a native compiler for the C+ programming language. One day, while delivering a lecture on the program at Bell Labs, Tiemann ran into some AT&T developers struggling to pull off the same thing.
“There were about 40 or 50 people in the room, and I asked how many people were working on the native code compiler”, Tiemann recalls. “My host said the information was confidential but added that if I took a look around the room I might get a good general idea”.
It wasn’t long after, Tiemann says, that the light bulb went off in his head. “I had been working on that project for six months”, Tiemann says. “I just thought to myself, whether it’s me or the code this is a level of efficiency that the free market should be ready to reward”.
Tiemann found added inspiration in the GNU Manifesto, which, while excoriating the greed of some software vendors, encourages other vendors to consider the advantages of free software from a consumer point of view. By removing the power of monopoly from the commerical software question, the GPL makes it possible for the smartest vendors to compete on the basis of service and consulting, the two most profit-rich corners of the software marketplace.
In a 1999 essay, Tiemann recalls the impact of Stallman’s Manifesto. “It read like a socialist polemic, but I saw something different. I saw a business plan in disguise”.[7]
Teaming up with John Gilmore, another GNU Project fan, Tiemann launched a software consulting service dedicated to customizing GNU programs. Dubbed Cygnus Support, the company signed its first development contract in February, 1990. By the end of the year, the company had $725,000 worth of support and development contracts.
GNU Emacs, GDB, and GCC were the “big three” of developer-oriented tools, but they weren’t the only ones developed by Stallman during the GNU Project’s first half decade. By 1990, Stallman had also generated GNU versions of the Bourne Shell (rechristened the Bourne Again Shell, or BASH), YACC (rechristened Bison), and awk (rechristened gawk). Like GCC, every GNU program had to be designed to run on multiple systems, not just a single vendor’s platform. In the process of making programs more flexible, Stallman and his collaborators often made them more useful as well.
Recalling the GNU universalist approach, Prime Time Freeware’s Morin points to a critical, albeit mundane, software package called hello. “It’s the hello world program which is five lines of C, packaged up as if it were a GNU distribution”, Morin says. “And so it’s got the Texinfo stuff and the configure stuff. It’s got all the other software engineering goo that the GNU Project has come up with to allow packages to port to all these different environments smoothly. That’s tremendously important work, and it affects not only all of [Stallman’s] software, but also all of the other GNU Project software”.
According to Stallman, improving software programs was secondary to building them in the first place. “With each piece I may or may not find a way to improve it”, said Stallman to Byte. “To some extent I am getting the benefit of reimplementation, which makes many systems much better. To some extent it’s because I have been in the field a long time and worked on many other systems. I therefore have many ideas to bring to bear”.[8]
Nevertheless, as GNU tools made their mark in the late 1980s, Stallman’s AI Lab-honed reputation for design fastidiousness soon became legendary throughout the entire software-development community.
Jeremy Allison, a Sun user during the late 1980s and programmer destined to run his own free software project, Samba, in the 1990s, recalls that reputation with a laugh. During the late 1980s, Allison began using Emacs. Inspired by the program’s community-development model, Allison says he sent in a snippet of source code only to have it rejected by Stallman.
“It was like the Onion headline”, Allison says. “`Child’s prayers to God answered: No.’”
Stallman’s growing stature as a software programmer, however, was balanced by his struggles as a project manager. Although the GNU Project moved from success to success in creation of developer-oriented tools, its inability to generate a working kernel-the central “traffic cop” program in all Unix systems that determines which devices and applications get access to the microprocessor and when-was starting to elicit grumbles as the 1980s came to a close. As with most GNU Project efforts, Stallman had started kernel development by looking for an existing program to modify. According to a January 1987 “Gnusletter”, Stallman was already working to overhaul TRIX, a Unix kernel developed at MIT.
A review of GNU Project “GNUsletters” of the late 1980s reflects the management tension. In January, 1987, Stallman announced to the world that the GNU Project was working to overhaul TRIX, a Unix kernel developed at MIT. A year later, in February of 1988, the GNU Project announced that it had shifted its attentions to Mach, a lightweight “micro-kernel” developed at Carnegie Mellon. All told, however, official GNU Project kernel development wouldn’t commence until 1990.[9]
The delays in kernel development were just one of many concerns weighing on Stallman during this period. In 1989, Lotus Development Corporation filed suit against rival software company, Paperback Software International, for copying menu commands in Lotus’ popular 1-2-3 Spreadsheet program. Lotus’ suit, coupled with the Apple - Microsoft “look and feel” battle, provided a troublesome backdrop for the GNU Project. Although both suits fell outside the scope of the GNU Project, both revolved around operating systems and software applications developed for the personal computer, not Unix-compatible hardware systems-they threatened to impose a chilling effect on the entire culture of software development. Determined to do something, Stallman recruited a few programmer friends and composed a magazine ad blasting the lawsuits. He then followed up the ad by helping to organize a group to protest the corporations filing the suit. Calling itself the League of Programming Freedom, the group held protests outside the offices of Lotus, Inc. and the Boston courtroom hosting the Lotus trial.
The protests were notable.[10] They document the evolving nature of software industry. Applications had quietly replaced operating systems as the primary corporate battleground. In its unfulfilled quest to build a free software operating system, the GNU Project seemed hopelessly behind the times. Indeed, the very fact that Stallman had felt it necessary to put together an entirely new group dedicated to battling the “look and feel” lawsuits reinforced that obsolescence in the eyes of some observers.
In 1990, the John D. and Catherine T. MacArthur Foundation cerified Stallman’s genius status when it granted Stallman a MacArthur fellowship, therefore making him a recipient for the organization’s so-called “genius grant”. The grant, a $240,000 reward for launching the GNU Project and giving voice to the free software philosophy, relieved a number of short-term concerns. First and foremost, it gave Stallman, a nonsalaried employee of the FSF who had been supporting himself through consulting contracts, the ability to devote more time to writing GNU code.[11]
7.
See Michael Tiemann, “Future of Cygnus Solutions: An Entrepreneur’s Account”,
10.
According to a League of Programming Freedom Press, the protests were notable for featuring the first hexadecimal protest chant:
11.
I use the term “writing” here loosely. About the time of the MacArthur award, Stallman began suffering chronic pain in his hands and was dictating his work to FSF-employed typists. Although some have speculated that the hand pain was the result of repetitive stress injury, or RSI, an injury common among software programmers, Stallman is not 100% sure. “It was NOT carpal tunnel syndrome”, he writes. “My hand problem was in the hands themselves, not in the wrists”. Stallman has since learned to work without typists after switching to a keyboard with a lighter touch.