Richard Stallman (RMS) made a small (two hours) presentation here,
in Sophia Antipolis, on December 8th 1997, on the theme "The GNU Project".
Here is a brief summary of what he said, interspersed with my own comments.
The speech was made in French, which didn't contribute to making it dense.
RMS speaks a fine French, but I believe that this two-hour session could
have easily fit within a thirty-minute English speech. Actually, RMS got
tired by the end, and swiched to English for the rest of his lecture
(much to my relief).
When he started at the MIT, the Hacker mentality was at its apex. The information had to be free, and everyone shared it happily. Xerox had just donated one of the very first laser printers to the Institute. This printer used a driver for which source code was not available, which was a new notion at that time. Stallman once heard that someone at the CMU had these sources, so on his first trip to Pittsburgh, he visited the person and asked him the sources. To Stallman's great surprise, the person would not give them. RMS was shocked. Not only was this person concealing what he deemed a public information, but he was hurting the whole community by not helping him. This is what started his indisputable faith into "free software". From that day on, he set off on a quest to ban proprietary software and encourage the free sharing of source code by all means.
RMS was beginning to be successful with Emacs by that time, shipping more and more tapes. These tapes were sold $150 but, he insisted on that point, it was only the price of s&h. The software on it was both free from a pecuniary point of view, but more importantly, free of any intellectual rights. Fearing that these terms might change, RMS felt that he had to quit the MIT if he wanted to be sure that his subsequent works would belong to him completely. The Free Software Foundation was created and took over the distribution of tapes. RMS could now focus on his quest.
It was by that time ('84) that Non Disclosure Agreements were blossoming, and RMS certainly considered them as the vilest invention of all times. The Hacker community collapsed : some students wrote a word processor and instead of giving it away to the community, they decided to sell it to a company, which then proceeded to sell it
[already at that point, I felt an urge to utter a "So what ?..." but refrained myself. RMS has obviously never been concerned about money while he was a student, or he would have at least shown some understanding for this decision].
The PDP/10, which he was using, became obsolete and was replaced by a more efficient computer, whose OS was proprietary. NDA's became mandatory to be able to use the new material. RMS decided he couldn't stand this any longer and the idea of writing his own OS from scratch sparked in his mind. This OS would be free of charge and, more importantly, would make sure in its license that all users would all get an equal freedom to use and copy it.
Hackers love funny names. After Emacs, several clones appeared, which were called SINE (Sine Is Not Emacs), TINE, EINE, etc... Eine even got a successor called... ZWEI (Zwei Was Eine Initially). GNU was born.
In Stallman's eyes, his GPL gives uses three freedoms :
[At this point, we enter a much more disputable view of the world. One of the main reasons why X is not distributed under the GPL is that it was deemed as much too restrictive, imposing any supplied software to be GPL itself. This effect has been coined "the Midas syndrom", and has prevented a lot of code from from the FSF from being included in widespread programs. Instead, the X Consortium devised their own license, which relaxed the most stringent terms of the GPL and allowed for non-free parts of the software to be part of the whole distribution].
In 1985, the FSF started to ship tapes and began to receive donations. The GNU OS (to be known later as HURD) was progressing and most and more gaps were filled in its architecture. They chose X Window, TeX, etc... but there was still a major part missing : the kernel. They chose MACH as a base for it, but at that time, the government was imposing export restrictions, so they had to wait, and the project grinded to a halt. Meanwhile, a free Unix was slowly gaining momentum. Created by a unique person, Linus Torvalds, Linux was completely free and quickly gained a very wide acceptance. Stallman abandoned MACH and adopted Linux.
[Now we go one more step into computer history revisionism. RMS has been very silent on his initial "misunderstandings" with Linus. Actually, the initial exchanges with Linus were more than heated, and many people thought that RMS and Linus would be enemies forever. RMS just couldn't admit that someone had written an OS faster and more successful that his own, and more generally, despises everything that is not originating from the FSF. See below for more]
Stallman made an additional remark about Linux. Many different distributions are available, and one day, he tried to install one of them called "SUSE". He noticed that SUSE installed non-free (from a GPL point of view) software, but didn't tell you so. They were concealing the fact that non-GPL software was being installed on your computer. Asked about this, the SUSE people told RMS that it was intentional, that they didn't regard this detail as important, but that mentioning it might worry people and discourage them from using SUSE. Bottom line : RMS says "Don't Use SUSE" (for those interested, he recommends the Debian, which is one of the rare things him and I agree on :-)).
[This was the second place where I kept expecting a good reason for not using SUSE, but it was never voiced. My guess is that you must refuse to use SUSE on the base of ethics. It Is Not Well... sigh]
Stallman spent a fair amount of time defending GNU's script language, Guile. Guile is a derived of Scheme, and should be used both becaust it is GPL, but also because it doesn't impose a choice of language : you can use any language you choose, and a translator will bridge to Guile. In brief, not using Guile prevents programmers from using "more decent" languages ("decent" probably means "issued by the FSF").
[I fail to see where Guile has an edge on other alternatives here. No language locks you in if you consider that a translator exists. You could just as well say "Tcl leaves you freedom to program in any language just as long as a translator to Tcl exists". As a matter of fact, very few Guile translators exist, and actually, very few Guile programs exist at all. Maybe it is fair to mention at this point that Stallman also had very difficult times with Ousterhout at the time when Tcl/Tk was becoming successful. Whatever evil Stallman thinks of Tcl (and he couldn't help slinging some mud at Tcl during his speech), Tcl still remains the scripting language of choice in the free community, while Guile has barely left the hard disks of the FSF. Note that I am not being sentimental here, but merely factual. I have always fancied Lisp-like languages to sh-likes].
The works in progress at the FSF include :
- a desktop, to replace the existing (non-free) ones
- translators for Guile
- end users applications
[Now something strikes me -- again. One of the main benefits of free coding, and often touted by Stallman, is that sharing source code allows reusing already existing works. It is crucial not to reinvent the wheel. And then, he boldly claims they are implementing a desktop, and along, all its cohort of most complex features like drag and drop and multi-visual displays. I really can't grasp the logic behind all this. Stallman despises Microsoft for their Not Invented Here syndrom, but he obviously suffers from the same disease]
Stallman concludes by encouraging everyone to write free code and contribute to the FSF. Don't write any proprietary code, or you will hurt the community.
[I must say that I do agree with the first sentence, but the second one left me speechless. This is kind of common with Stallman actually, he has repeatedly voiced definitive opinions on almost every topic in computer science, and he did start early by saying bitmap screens had no future, and later on, that mice were a useless peripheral to a computer, both positions explaining to a certain extent why Emacs19 lagged more and more behind Lucid Emacs (now XEmacs) despite Lucid's numerous attempts to collaborate with him and contribute to merging both codes. This debate eventually made it to Usenet circa 1993, and Stallman and Epoch/Lucid Emacs author(s) (among which Jamie Zawinski) exchanged very heated arguments. It eventually turned out that Stallman had not even launched Lucid Emacs once, and as things were turning out pretty bad for him, he retreated and eventually accepted the collaboration of the Lucid team. One year later, it turned out that all contributions sent to the FSF had been ignored. I must still have a transcript of these Usenet articles if anyone is interested].
I then made an attempt at bringing the debate back to solid grounds by mentioning that obviously, he had been successful at making a living out of free software, but that most of the students present in the amphitheater today, who would graduate within the next few months, couldn't afford to just write code for fun, giving it away, and hope that some day, maybe, they would get some money back.
Stallman's answer must have been one of the most vexing I have ever heard. Everybody was given a lesson on moral and ethics, and how money wasn't everything that mattered and that what had worked for him could work for others. Hell, he even knew people who also made a living with free software. He was so committed to his task that he had refused to receive a salary from the FSF.
His response actually reinforced my point. He didn't need money to live, and he was lucky enough to have developed Emacs in
- a place where his employer (the MIT) didn't care what the code he wrote would become
- a time where it was possible for a single man to write a software that would revolutionize the whole computer industry
However, I reckon that even the first point is not quite true. I suspect that the MIT didn't oppose to Emacs going public simply because they didn't perceive the potential. Stallman also felt this, since he decided to leave the MIT to focus on his work. He probably knew that now that he had raised a few eyebrows in his direction, the MIT would probably want to take advantage of his work. I don't know for him, but I would have been very proud if such a famous institute as the MIT would endorse and distribute my work. He obviously perceived things differently.
Now, as a conclusion, I must say that I am not that negative about Stallman. I respect him as a Great Scientist who left a very profound footstep in the history of computer science. What I contend is religion and bigotry. Nobody should allow themselves to lean toward sentiment and pretend that their choice is technologically sound and should be blindly adopted by everyone. Stallman has abided to a very precise ethics, but he shouldn't try to talk everyone into joining him, because he is someone whose concerns are now very far from the rest of us.]