If you ask a lot of people why Linus Torvalds and the Linux kernel that he wrote became one of the most prominent open source projects of all time, while Richard Stallman's GNU project has received much less attention beyond hacker circles, they'll tell you the difference has to do with Stallman's excessive commitment to an uncompromising ideology. Is that really accurate?

Below, I'd like to make the case for a more nuanced interpretation of Stallman and the Free Software Foundation, which were more pragmatic than many observers have appreciated.

First, let's take a look at what people have said about Stallman and the GNU project that he started in 1984. (He launched the Free Software Foundation a year later.) Here's what Torvalds wrote about Stallman in his 2001 autobiography:

The thing that drives me crazy about Richard is that he sees everything in black and white. And that creates unnecessary political divisions. He never understands the viewpoint of anybody else. If he were into religion, you would call him a religious fanatic.

At another point in the book, Torvalds added, in a discussion of Stallman's perspective on the license under which the Mozilla browser was released in 1998:

On the whole, the license was fairly mellow, but if you're someone like Richard Stallman you don't like mellow.

Similarly, Tim O'Reilly, the publisher who organized the meeting of free-software programmers in 1998 that led to the coining of the term "open source software," said that he excluded Stallman from the conference because he was "inflexible and unwilling to engage in dialogue." (That's a quotation from Sam Williams's 2002 biography of Stallman, "Free as in Freedom.")

And Bruce Perens, the sometime-Debian director who wrote the Open Source Definition (and who also apparently does not like pronouns), has said of Stallman, "I really like and admire Richard," but "I do think Richard would do his job better if Richard had more balance. That includes going away from free software for a couple of months." (That quotation also appears in Williams's book.)

St. Ignucius

Depictions like those above make Stallman out to be an overzealous ideologue, whose unwillingness to compromise on all things related to free/open source software has prevented the GNU project from having as great an impact on society as Stallman envisioned.

The implication of such portrayals is that Linux and the "open source" camp have enjoyed much more momentum because their leaders, most notably Torvalds, have been more willing than the Free Software Foundation to accommodate diverse viewpoints, and to see open code as a means to an end, rather than a sort of moral crusade.

To be sure, some of Stallman's actions have done nothing to help his image in this regard. He and the Free Software Foundation were indeed relatively inflexible regarding licensing issues and terminology when the open source initiative was launched in the late 1990s. Instead of seeking common ground with open source software projects, including Mozilla, that rose to prominence at the time but adopted licenses other than the GPL, the GNU crowd distanced itself from them.

Stallman has also made it hard for observers to view him as a totally rational human being when he does things like dress up as "St. Ignucius" and discuss "Emacs virgins."

(Stallman contends that he invented the Saint Ignucius persona in 1996 as a way of poking fun at himself and emphasizing that the free-software movement should not be taken overly seriously. It was only later, he said, that critics of GNU and the Free Software Foundation seized on activities such as the Saint Ignucius performances to identify Stallman as an overly zealous ideologue whose approach was counterproductive to the open source movement's goals.)

RMS the Pragmatist

In other respects, however, Stallman and the GNU project have proven much more willing to compromise in pragmatic ways, especially prior to the mid-1990s. Consider:

  • In its earliest iterations, the GNU GPL license was more restrictive than it is today. Between late 1986, when the first version of the GPL appeared, and the debut in 1989 of GPL 1.0, GNU's leaders watered down the licensing terms somewhat. They permitted developers to distribute GPL'd software alongside unfree code, allowed GPL'd programs to link to unfree libraries even if the libraries' source code could not be distributed and required that the source for GPL'd programs remain available only for three years, rather than in perpetuity. These compromises made the GPL more practical for many programmers, but they represented a retreat from the ideology of the Free Software Foundation.
  • Stallman and the Free Software Foundation were willing to make common cause with proprietary software companies when circumstances demanded it. That is why, for example, they protested Apple's "look and feel" lawsuit against Microsoft and HP that began in the late 1980s. They did this to protect the interests of free software, of course, not because they cared about the fate of the defendant companies. Still, it was a pragmatic move, which someone who was purely fanatical about free software, and viewed the software world in Manichean terms, would not have made.
  • GNU developers cooperated extensively with their counterparts at Berkeley who were working on BSD. At times, they even actively coordinated their work with BSD, and distributed BSD software. That was significant because the Free Software Foundation viewed the BSD licenses as too permissive, since they did not ensure that the source code in programs derived from BSD would always remain open. (If the BSD licenses had been more satisfying from Stallman's perspective, there may have been no need for the GNU project in the first place.) Here again, the Free Software Foundation and GNU pragmatically cooperated with another group whose ideology was different from their own.

It's also worth noting that, originally, Linux was distributed under a license that was even more restrictive than the GPL. Not only did it protect the openness of the kernel source code, but it also stipulated that distributors of the kernel may "not profit from the distribution. In fact even 'handling costs' are not acceptable."[1] In that sense, Torvalds was actually more ideologically strict when he introduced Linux to the world than was Stallman, who had no qualms about charging money for GNU software so long as the source code remained available. (In fact, GNU itself charged up to $5,000 for distributions of its software on tape or disk, although cost-free versions were always available online and it was perfectly legal for users to make and distribute their own copies of GNU physical media without paying).

If Stallman and GNU grew more zealous and uncompromising by the late 1990s, that was perhaps because they feared that the free-software movement was being co-opted by people who were quite uninterested in the philosophical and political implications of keeping source code open. It probably did not help that, by 1999, observers were calling Stallman a "forgotten man" and comparing him to a "revolutionary erased from a photograph" who was "written out of history," to quote from a March 1999 article in Wired.

Having watched others take credit for and control of the revolution he had started, Stallman perhaps was naturally inclined to radicalize his views, to distance himself from these people. It says a lot about just how deeply he thought the open source movement was undermining the Free Software Foundation's goals that he would not work with people who embraced the "open source" term, even though he had been willing to fight alongside the likes even of Microsoft in the 1980s.

In summary, at least until the mid-1990s, Stallman, GNU and the Free Software Foundation were more open to pragmatic compromise than popular memory today imagines them to have been. The St. Ignucius drill may have signaled that Stallman had come down the mountain, to borrow an analogy associated with another famous revolutionary, Maximilien Robespierre. But his views had not always been as provocative and obstinate.

[1]. This document, dated October 1991, appears to contain a copy of the original Linux copyright, although it does not explicitly say as much. If anyone has Linux source code with the original copyright notice intact (the license changed to the GPL beginning with the January 1992 release of Linux version 0.12), I'd love to hear from you as I continue working on a book about the history of free and open source software. The archived versions of early Linux kernels that I have been able to dig up do not contain copyright notices, and Torvalds only paraphrases the original copyright in his autobiography, rather than actually quoting it.