2
I have aimed to make this edition combine the advantages of my knowledge and Williams’ interviews and outside viewpoint. The reader can judge to what extent I have achieved this.
I read the published text of the English edition for the first time in 2009 when I was asked to assist in making a French translation of Free as in Freedom. It called for more than small changes.
Many facts needed correction, but deeper changes were also needed. Williams, a non-programmer, blurred fundamental technical and legal distinctions, such as that between modifying an existing program’s code, on the one hand, and implementing some of its ideas in a new program, on the other. Thus, the first edition said that both Gosmacs and GNU Emacs were developed by modifying the original PDP-10 Emacs, which in fact neither one was. Likewise, it mistakenly described Linux as a “version of Minix.” SCO later made the same false claim in its infamous lawsuit against IBM, and both Torvalds and Tanenbaum rebutted it.
The first edition overdramatized many events by projecting spurious emotions into them. For instance, it said that I “all but shunned” Linux in 1992, and then made a “a dramatic about-face” by deciding in 1993 to sponsor Debian GNU/Linux. Both my interest in 1993 and my lack of interest in 1992 were pragmatic means to pursue the same end: to complete the GNU system. The launch of the GNU Hurd kernel in 1990 was also a pragmatic move directed at that same end.
For all these reasons, many statements in the original edition were mistaken or incoherent. It was necessary to correct them, but not straightforward to do so with integrity short of a total rewrite, which was undesirable for other reasons. Using explicit notes for the corrections was suggested, but in most chapters the amount of change made explicit notes prohibitive. Some errors were too pervasive or too ingrained to be corrected by notes. Inline or footnotes for the rest would have overwhelmed the text in some places and made the text hard to read; footnotes would have been skipped by readers tired of looking down for them. I have therefore made corrections directly in the text.
However, I have not tried to check all the facts and quotations that are outside my knowledge; most of those I have simply carried forward on Williams’ authority.
Williams’ version contained many quotations that are critical of me. I have preserved all these, adding rebuttals when appropriate. I have not deleted any quotation, except in chapter 11 where I have deleted some that were about open source and did not pertain to my life or work. Likewise I have preserved (and sometimes commented on) most of Williams’ own interpretations that criticized me, when they did not represent misunderstanding of facts or technology, but I have freely corrected inaccurate assertions about my work and my thoughts and feelings. I have preserved his personal impressions when presented as such, and “I” in the text of this edition always refers to Williams except in notes labeled “RMS:”.
In this edition, the complete system that combines GNU and Linux is always “GNU/Linux,” and “Linux” by itself always refers to Torvalds’ kernel, except in quotations where the other usage is marked with “[sic]”. See http://www.gnu.org/gnu/gnu-linux-faq.html for more explanation of why it is erroneous and unfair to call the whole system “Linux.”
I would like to thank John Sullivan for his many useful criticisms and suggestions.