Slashdot Mirror
Feature:On the Subject of RMS
The Advanced Learner's Dictionary of Current English defines ``eccentric'' to mean ``peculiar; not normal''. Is RMS eccentric? Yes, most definitly. He has got a meaning in his life and this alone would make him eccentric. True, he is also peculiar in many other ways. Who else would carry around his battered computer in a cloth bag? He does not care about what people think; he is on a mission. A mission to create a completely free operating system for everyone to use.
In his ideal world, all software is free software and sharing your code with your neighbor is standard practice. RMS lived in this world for some time when he was active around the AI lab on MIT. In some ways he lives in that world still. He is an eccentric idealist with a mission. It is not so surprising that normal people look at him strangely.
Someone once said to me that to be a hacker, you had to give up your normal life. Not true. There is a strong line between beeing a hacker and beeing a socially unfit hacker. Hackers like RMS, Alan Cox and Linus Torvalds are very clearly not socially unfit.
A few years ago, Linus got married with his wife Tova and they now have several kids. Alan Cox is married to Telsa (more commonly known as ``hobbit'' after her login name and email address). Even RMS had a sweetheat once named Alix. They eventually broke up, but that could have been for any reason. Neither I nor you have any business questioning what a hacker does on his private time. In many cases, I have a feeling we really do not want to know.
Hating, or atleast disliking, RMS has become a gimmick, much like hating Bill Gates and Windows. Linus Torvalds has managed to stay clear of that area by avoiding politics. When we judge Linus, we do it only by his coding. When we judge RMS, we do not do it based on his coding skills but rather on how his philosophical ideas differs from ours. I suppose that if Linus would get himself mixed up in these philosophical discussions, he too would be judged accordingly. It is fine that Linus does not do this as long as we all remember that we should not compare him with RMS the way most people do.
Free SoftwareWhen we speak of free software we speak of the right to study the program, distribute the program freely and improve the program with your own code. You do not need to do this, and many people probably do not even want to study how the program works or make their own improvements. But would it not be nice to know that you have that option?
Should you care? I think you should, but neither I, RMS or anyone else can force you to. We will try to get your attention to what we feel is important, but you don't have to follow us blindly. We want you to read our texts on philosophy and come to your own conclusion. If after you have read what we've written, you decide to walk some other path, we do not stop you. http://www.gnu.org/philosophy/philosophy.html is a good start. Read what is written there and then come back here.
I do not think the BSD license is a good idea, nor do I think that the artistic license is a good idea. However, I'm accepting that people want to use them because their values in life differs from mine. I dislike those licenses because I do not agree with the philosophical issues of them. The same way as when I vote, I do not vote for the socialistic party because I do not agree with their philosophical issues.
Thinking that everyone in the whole world could agree to one specific license would be naive. Unless we all suddenly start to think very much the same, that will not happen. RMS knows this. I know this. The BSD followers know this. Everyone should know this. Frankly, would you want to live in a world where everyone thought the same way?
When I was writing this, I got an email from someone who had read about me on my home page. The core of the email was "I find you interesting because you are not normal." For me, that is a great compliment.
I will not go any further on the philosophical issues. For me, it is enough that everyone is happy with whatever license they choose.
The Free Software FoundationThe Free Software Foundation has worked to support free software for 15 years now. The main project is the GNU project which aims at developing a completely free operating system. At least some of you should notice that 15 years is a very long time. Proprietary software companies get out new versions of their programs and operating systems almost every other year. What has taken the GNU system so long?
When RMS started out to create the GNU system, he began by publishing the GNU Emacs which is the standard text editor in the GNU system. He then needed a compiler, so he made the GNU C Compiler and the GNU Debugger. Pretty soon, there were utilities available for most standard tasks but you still needed a kernel to run them. So the GNU project started writing the GNU Hurd which is a kernel based on Mach.
While we were working on the GNU Hurd, Linus Torvalds released his Linux kernel and when people combined this kernel with the finished GNU utilities, they got a free operating system. For the first time, someone could run a completely free Linux-based GNU system. But it was still a system made by hackers, for hackers. To correct this, we sponsored the early development of the Debian GNU/Linux distribution which should enable everyone to use this Linux-based GNU system.
It might have taken some time to get this system done, but you should never rush software development. Doing so would result in bad coding and a bad design. We do what we think has to be done in order to create a completely free operating system; and we do it in our own time.
LinuxLinux is a great kernel and I use it every day. I will probably keep using it on some computers even after I switch to the Hurd. But it is important that people remember what Linux is; a kernel. It's not a complete system and it's not fair to the community to call it that. As a developer and user of both Linux, Hurd and the GNU utilities, I feel it is important that people understand the difference. Noone should come to me asking for help with the GNU Emacs just because I have said that I work with Linux; they should ask me for help because I have said that I am using several tools from the GNU system.
To help everyone understand this, I have made it a habit of calling my system a Linux-based GNU-system, in short a GNU/Linux system. This is not because I only use GNU-utilities, it is because I use programs that are part of the GNU system. Some people argue that XFree86 should be given as much credit as the GNU system. Those people do not understand that by calling my system a GNU/Linux system, I give credit not only to the GNU project but also to XFree86. Why? Because the XFree86 has been adopted by the GNU project for use in the GNU system, as has Lynx been which I use almost daily. So by calling it a GNU/Linux system I give credit to every program which are part of--or has been adopted by--the GNU project.
The specific name is not really important, though some people might want to think so. The importance is to give credit to the GNU project where credit is due. One way to do so is by calling the system a GNU/Linux system, but there are many more ways to do it and only your imagination will limit them.
XFree86 has been adopted by the GNU project for use in the GNU system, as has Lynx been which I use almost daily.
Ok, well then, I'll say that the Linux kernal community has adopted the GNU tools. Now it is completely correct for me to call the entire system Linux.
I personally don't use the term GNU/Linux much because it is such a mouthful. However, I do think it is important to understand why the GNU people feel so strongly about it.
In 1983 when Richard Stallman founded the GNU project, his goal was to produce a 100% free Unix-like operating system. This was a difficult goal because a Unix OS contains many components: the kernel, the shell, compiler tools, editors, windowing systems, etc. Because the task was so big, he looked around for components that were already free that he could use in the GNU system. He found, for example, TeX and the X-Window System, which then fulfilled the system requirements for a typesetter and a windowing system respectively.
For some components, there was no free implementation, so the GNU project set out to write them. GNU Emacs replaced the proprietary vi editor (though a free vi clone was eventually written by someone else). The GNU C Compiler replaced the proprietary pcc. Bash replaced the Bourne shell. Etc. These tools were all written because there was no other free program that did the job. Nobody at the GNU project wanted to re-write free tools that already existed. The goal was a 100% free operating system, not a 100% free operating sytsem written 100% by the GNU project.
Like many complex projects, the GNU system took a long time to develop. Rather than wait until the entire system was complete before releasing it, various components were released as they were developed. (Call it the Bazaar model if you will). They put their available code into a repository. This was the master GNU ftp server at prep.ai.mit.edu. This archive contained the GNU system as a work in progress, including many tools that were ready for production use.
When Linus developed his kernel, people obviously needed the rest of the operating system components to go along with it. So how did people get these? They simply ftp'd to the GNU archive and downloaded all of the GNU operating system components that were availble. They combined these with the kernel, to produce the system they called Linux.
While you might not agree with Richard Stallman, I think it is easy to see why this would upset him. If someone downloaded the CVS archive of the free software project I am working on, finished it off before I could finish my version, then released it under a name that gave me no credit, I would probably be angry too. I think the reaction of the community if someone did this would be very negative. The GNU ftp site was their code repository and the builders of the early Linux distros did exactly this. They built a system that was largely GNU code (especially in the early days before distros got loaded up with lots of user level applications) but did not give credit to the GNU project.
Again, you might not see it this way. But I think that if the GNU project was something that you had founded and invested years of effort into, you might be a little bit miffed someone did this to you. While Stallman is often accused of being a bit strange, I find his attitude on the "GNU/Linux" issue quite normal.
As usual I see many postings that show confusion over why the GNU project wants the "GNU/Linux" name to be used. These imply that the reason is because there is a large number of GNU components in the system. This is not the case. The reason is because the GNU Project had the vision and the goal of a completely free operating system, and systematically went about building it. Without the goal of a free operating system by the GNU project, Linux most certainly would not be what it is today.
The GNU project identified all of the major components of the operating system: the kernel, utilities, the shell, compiler tools, windowing systems, networking, and so on. It then found or built tools for each of these areas. When Linus wrote the Linux kernel and people looked around for the components necessary to build a complete operating system, they found them because that is what the GNU project set out to build.
It is tempting to say that the GNU Project did not write some of these tools - such as the X-Window system - and thus say that the GNU influence on the GNU/Linux operating system is exaggerated. However, the GNU project never set out to build every component themselves, nor did they ever claim that all components of the free operating system were there because of their efforts. Instead, they looked first for existing free tools to do the job, and wrote a replacement if one could not be found. X existed already, so a replacement was not needed. However, had the X-Window system not existed, or had it been proprietary, the GNU project would have developed a replacement for that too, just as they developed a replacement for the C library, and m4 macro processor, and many other things.
Call it GNU/Linux (if you choose) because the GNU project built ane operating system, not because they wrote any number of individual components. Without the GNU project, disjoint components is all that would exist.
Such non-sequitur's show the fallicy of the whole GNU/LINUX nonsense... I don't expect the XFree86
people share this opinion, or many others.
Here are several messages I've sent RMS on the general topic.
- Jim Gettys
> Sender: owner-linux-kernel@vger.rutgers.edu
> From: Richard Stallman
> Date: Wed, 23 Dec 1998 18:17:02 -0700 (MST)
> To: fizban@tin.it
> Cc: linux-kernel@vger.rutgers.edu
> Subject: Re: Article: IBM wants to "clean up the license" of Linux
> -----
> Please show a
> bit more respect for Linus and all the other people and their efforts...:
> call the linux kernel "Linux" as Linus wanted to call it
>
> I always call the kernel Linux, for precisely that reason. Linus
> Torvalds started that program, and he says the name is Linux, so I
> call it Linux out of respect for him.
>
> I ask people to do the same thing for the operating system as a whole.
> It was started in 1984 by the GNU Project. For years, before Linux
> was written, we developed many components (not just "tools") of this
> system, and we did so as steps in the development of the system as a
> whole. (See the GNU Manifesto.)
>
> Linux (the kernel) doesn't come from the GNU project, and we never try
> to claim any credit for it. When people say that the GNU project
> "developed important parts of Linux", we explain that we don't deserve
> that honor, because none of Linux is our work. And we never call
> Linux a GNU program. (Some people have misinterpreted this as a
> gesture of rejection of Linux; actually, it is because we're not
> entitled to say so.)
>
> But while the GNU project played no role in the writing of Linux, it
> started the development of the operating system as a whole. That's
> what the GNU project was and is about. Writing dozens of programs
> such as GCC, Bash and libc--not only "tools"--was just a part this
> larger project.
>
> The system version most of us are using is the combination of Linux
> and the GNU system. "GNU/Linux" is a good way to describe that
> combination, and when I write that, it always means the whole
> combination. The kernel is simply Linux.
>
One might as well also say that the whole system should be called
"GNU/X/Linux"; the X Window system contribution, in terms of number of
lines of code of software, is very large. People should remember that
not only "hackers" contributed, but a number of major companies, including
my own, contributed large amounts to that code base, under fully free
terms (where the UNIX vendors went wrong was stuff built on top, and the
silly GUI wars of the beginning of the decade). In terms of total effort
and number of lines of code, both GNU and X represent much larger efforts
than the base operating system.
But the reality is that this is too cumbersome, whether you say "GNU/Linux"
or "GNU/X/Linux. The market and men on the street now associates "Linux"
with the whole combination, for better or for worse. I'm personally very
gratified that our (in this case, the X Window system community, GNU
community, and Linux community) are affecting a large and growing number
of people, rather than withering and dying from the effects of Redmond.
So long as Linus gives credit where credit is due to the various groups
that make up this community, there is little to be gained (and arguably,
much to be lost) by confusing people with a more complex nomencature.
Obviously, when writing for a technical audience, (rather than the mass
audience), being more clear what you mean may make sense and give credit
where credit is due. But lets not confuse the mass market, which
has enough trouble understanding Linux as it is.
- Jim Gettys
From: jg@pa.dec.com (Jim Gettys)
Date: Thu, 7 Jan 1999 11:36:03 -0800
To: rms@gnu.org
Cc: allbery@kf8nh.apk.net, fizban@tin.it, linux-kernel@vger.rutgers.edu
Subject: Re: Article: IBM wants to "clean up the license" of Linux
For better or worse (I believe for the better), the term Linux has grown
to cover the merged result of a large number of efforts of the last 15 years.
The components of the Linux system include (at least) in NO particular order:
o the Linux Kernel effort,
o BSD UNIX development,
o X Window System,
o Perl, Python, TK/TCL,
o and the large efforts that go under the GNU banner.
All of these are major efforts, by MANY talented individuals and corporations
(some of whom put many millions of dollars into the development of the
code, whether it be Digital/Compaq, HP, IBM, Sun, Red Hat, SuSE, Netscape,
Cygnus, and many others). They represent many man years of sweat, often under
very hard deadlines (at least in the X Window System and Netscape cases,
and probably others, at great personal cost). If you tried to remove any
one of them, you would end up without Linux (though some substitutes for
some pieces exist).
Calling out any particular one or subset of these efforts for particular
praise in a common term, when the current general term connotes all of the
contributors, slights the contributions of the others; this is why I believe
it is divisive. I have emotional scars left from the last round of
divisiveness (the UNIX GUI wars), and CANNOT condone any action that would
condone such divisions, and the results of such divisions, particularly
at a time when unity is needed.
I therefore believe that uses of the term GNU/Linux are divisive and wrong.
Therefore, I will not personally condone any such usage, and WILL NOT
make such distinctions, and strongly discourage others in doing so. Lets
spend our time giving credit to each other for what they have done for
us, rather than asking others to give credit to us for what we have done
for them. (to paraphrase Kennedy). One is inclusive, the other is divisive.
This is my final comment on this thread.
- Jim Gettys