GNU & The Free Software Foundation

Engineering Tech Talk at Google, June 11, 2004

1. Introduction

ED: Well, thank you everybody for making it. I'm Ed Falk and this man needs very little introduction; if you don't know what the letters RMS stand for, you probably don't belong in this room.

Richard was the founder of the Free Software Foundation, in 1984 I believe it was, and as such could be considered the father of free software and, of course, Google's infrastructure is based on free software. So we owe the free software movement quite a great deal of thanks. [And my mic is dying on this microphone so I won't talk too long.] This is Richard Stallman and we thank him for being here on short notice and we thank our mutual friend Lile Elam who arranged all of this and I think with no further ado, I give you Richard!

[Richard bows]

2. How it started

RICHARD: Please raise your hands if you cannot hear me. [Laughter] Yes, somebody raised his hand.

So, the topic of my speech is free software. I didn't begin free software; there was free software going back to the early days of computing. As soon as there were a couple of computers of the same model, people could try sharing software. And they did.

{This is not… This has a problem. How do we stop the feedback? Can someone do anything? I'm willing to get some feedback, but only from you, not from the PA system.

AUDIENCE: [unintelligible]

RICHARD: Well, that doesn't matter; I'm not an advocate of open source and never was and never will be.}

So free software existed before I started programming and I had the good fortune, in the 1970s, of being part of a community of programmers who shared software. So I learned about free software as a way of life, by living it. And I came to appreciate what it meant to be free to share with people, not divided from the rest of the world by attitudes of secrecy and hostility.

But that community died in the early '80s and I found myself confronted by the prospect of spending the rest of my life in a world of proprietary software. And, worst of all, confronted by the prospect of signing a non-disclosure agreement {which I}. And I had concluded that it is unethical to sign a non-disclosure agreement for generally useful technical information, such as software. To promise not to share with one's fellows is a violation of human solidarity. So when I saw that the machine downstairs was asking me to sign an NDA, I just said, “I can't sign an NDA.” Well, fortunately, there was an option; they let me come in here and speak without signing it, otherwise you would have had to go outside to listen. [Laughter]

(They asked a couple of other interesting questions; they asked about company, so I said I'm available tonight. [Looking at name tag][Laughter] And they asked for my host, so I put down But that's just the hacker spirit.)

So I found myself in a situation where the only way you could get a modern computer and start to use it was to sign a non-disclosure agreement for some proprietary operating system. Because all the operating systems for modern computers in 1983 were proprietary, and there was no lawful way to get a copy of those operating systems without signing a non-disclosure agreement, which was unethical. So I decided to try to do something about it, to try to change that situation. And the only way I could think of to change it was to write another operating system, and then say as the author “this system is free; you can have it without a non-disclosure agreement and you're welcome to redistribute it to other people. You're welcome to study how it works. You're welcome to change it.” So, instead of being divided and helpless, the users of this system would live in freedom. Ordinary proprietary software is part of a scheme where users are deliberately kept divided and helpless. The program comes with a license that says you're forbidden to share it, and in most cases you can't get the source code, so you can't study it or change it. It may even have malicious features and you can't tell. With free software, we respect the user's freedom, and that's the whole point. The reason for the free software movement is so that the people of cyberspace can have freedom, so that there is a way to live in freedom and still use a computer, to avoid being kept divided and helpless.

3. GNU operating system

You can't use a computer without an operating system, so a free software operating system was absolutely essential. And in 1983 I announced my plan to develop one: an operating system called GNU.

I had decided to make the system UNIX-like so that it would be portable. The operating system that we had used for many years at the Artificial Intelligence Lab was the Incompatible Timesharing System, or ITS. It had been written in assembler language for the PDP-10, so when Digital discontinued the PDP-10, our many years of work turned into dust and blew away. I didn't want to write another system and have the same thing happen, so I decided this system had better be portable. But there was only one successful portable operating system I knew of, and that was UNIX. So I decided to follow the design of UNIX, figuring that way I'd have a good chance of succeeding in making a system that was useful and portable. And then I decided to make the system upward-compatible with the interfaces of UNIX, and the reason for this was so that users could switch to it without an incompatible change.

I realized that I could take the best ideas from the various systems I had helped develop or use and add my pet ideas and make my dream operating system. But this would have been incompatible, and the users would mostly have rejected it, saying “it would be too much work to switch, so we're just not going to.” So, by making the system upward-compatible with UNIX, I could spare the users that obstacle and make more of a chance that users would actually use the system.

If the users had rejected it, I would have had a perfect excuse. I could have said “I offered them freedom and they rejected it; it's their fault.” But I wanted to make more than just an excuse. I wanted to build a community where people would actually live in freedom, which meant I had to develop a system people would actually use. So I decided to make the system upward-compatible with UNIX.

Now, UNIX consists of many components that communicate through interfaces that are more or less documented. And the users use those interfaces. So to be compatible with UNIX required using the same interfaces, which meant that the initial design decisions were already made, except one: what range of target machines to support. UNIX had been designed to support 16-bit machines, which was a lot of extra work, because programs had to be kept small; so I decided to save that extra work by not supporting anything less than a 32-bit machine. I figured it would take many years to get the system done and by then people would normally be using 32-bit machines anyway, and that turned out to be true.

So then the only thing that I needed before I could start work was a name. Now, to be a hacker means to enjoy playful cleverness—in programming, and in other areas of life, any area of life [where] you could be playfully clever. And there was a hacker tradition that when you were writing a program that was similar to some existing program, you could give your new program a name that's a recursive acronym, saying it is not the other program.

For instance, in the '60s and '70s there were many TECO text editors, more or less similar; typically each system would have a TECO and it would be called something-or-other-TECO. But one clever hacker called his program TINT, for “TINT Is Not TECO”—the first recursive acronym. And we thought that was very funny. So after I developed the first Emacs extensible text editor in 1975, there were many imitations, and some were called this-or-that-Emacs. But one was called FINE for “FINE Is Not Emacs” and there was SINE for “SINE Is Not Emacs,” and EINE for “EINE Is Not Emacs,” and MINCE for “MINCE Is Not Complete Emacs.” Then EINE was mostly rewritten, and version two was called ZWEI for “ZWEI Was EINE Initially.” [Laughter]

So I looked for a recursive acronym for “Something is not UNIX,” but the usual four-letter method was no good, because none of those was a word. And if it doesn't have some other meaning, it's not funny. So I thought, “what else can I do, hmm?” Nothing came to me, so I thought, “I'll make a contraction, then I could get a three-letter recursive acronym.” I started substituting all 26 letters: ANU, BNU, CNU, DNU, ENU, FNU, GNU! Well, “gnu” is the funniest word in the English language, so that had to be the choice. If you can call something “GNU,” it makes no sense to pick anything else.

So, of course, the reason why the word “gnu” is used for so much word-play is that, according to the dictionary, it's pronounced “new.” So people started asking each other, “hey, what's g-nu,” as a joke, long before you could answer “GNU's Not UNIX.” But now you can give that answer and the best part is, it sounds like you're obnoxiously telling the person what it isn't, instead of answering his question. But the fact is, you're giving the exact meaning of GNU; so you are, in fact, answering the question in the most exact possible way, but it gives the appearance that you're refusing to.

In any case, when it's the name of our operating system, please pronounce a hard G; don't follow the dictionary. If you talk about the “new” operating system, you'll get people very confused. We've been working on it for 20 years now, so it's not new anymore. But it still is, and always will be, GNU, no matter how many people call it Linux by mistake.

{[AUDIENCE: unintelligible] [RICHARD: Thank you!]}

So, having the name I could start work. I quit my job at MIT to begin writing pieces of the GNU operating system, in January 1984. I had to quit my job because, had I remained an MIT employee, that would have enabled MIT to claim to own all the code I was writing, and MIT could have turned it into proprietary software products. And since MIT had already done that kind of thing, I certainly couldn't trust them not to do so here. And I didn't want to have to argue with the MIT administration about all the details of the license I was going to use. So, by quitting my job, I took them out of the equation, and I have never had a job since then. However, the head of the AI Lab was nice enough to let me keep using the facilities, so I began using a UNIX machine at the AI Lab to start bootstrapping pieces of the GNU system.

I had never used UNIX before that time. I was never a UNIX wizard and I chose to follow the design of UNIX for the exact reason that I've told you, not because UNIX was my favorite system or anything. Sometimes people write that it was changes in UNIX's licensing policy that inspired GNU. Well, this is not true; in fact, UNIX was never free software. They were more or less restrictive and more or less nasty about enforcing the requirements, but it was never free software, so those changes actually made no difference and, in any case, they took place long before I ever saw an actual UNIX machine.

4. GNU Emacs

So, at the time, I thought that I and the other people I was recruiting to try to help would develop all these pieces and make a complete system and then we'd say, “come and get it.” But that's not how it happened. In September '84, I started developing GNU Emacs, which was my second implementation of the extensible programmable text editor. And by early '85, it was suitable for me to do all my editing with it. Now, that was a big relief. You see, I had absolutely no intention of learning to use Vi. [Laughter, applause] So, until that point, I did my editing on other machines where there was an Emacs and copied the files through the net, in order to test them on the UNIX machine. Once GNU Emacs was running, I could do my editing on the UNIX machine.

But other people wanted to get copies of GNU Emacs to use it for their editing, to use it on their UNIX systems. There was no GNU system yet, there were just a few pieces. But this one piece turned out to be interesting by itself. People asked me for copies, so I had to work out the details of how to distribute it. Of course, I put a copy in the anonymous FTP server, and that was good for people on the net, but in 1985, most programmers were not on the Internet. So they asked me for copies; what was I going to say? I could have said, “I want to spend my time writing more pieces of the GNU system, not writing mag tapes, so please find a friend who can download it and put it on tape for you,” and they would have found people sooner or later, because programmers generally know other programmers.

5. Expensive habits

But I had no job, and I was looking for some way to make some money through my work on free software. So I announced, “send me $150 and I'll mail you a tape of GNU Emacs.” And the orders began dribbling in. By the middle of the year, they were trickling in, eight to ten orders a month, which, if necessary, I could have lived on.

That's because I make efforts to resist expensive habits. An expensive habit is like a trap; it's dangerous. Now most Americans have the exact opposite attitude: if they make this much money, they look for how to spend this much, [makes ample gesture] which is completely imprudent. So they start buying houses and cars and boats and planes and rare stamps and artwork and adventure travel and children, [laughter] all sorts of expensive luxuries that use up a lot of the world's resources, especially the children. And then, the next thing they know, they've got to desperately struggle all day long to get money to pay for these things, so they have no time even to enjoy them, which is especially sad when it's a matter of children. The other things, I guess, can get repossessed. So then they become puppets of money, unable to decide what they're going to do with their lives. If you don't want to be a puppet of money, then resist the expensive habits, so that the less you need to spend to live on, the more flexibility you've got and the less of your life you're forced to spend to make that money.

So I still live, basically, like a student, and I want it to be that way.

6. Definition of free software

But people sometimes used to say to me, “what do you mean, it's free software, if it costs $150?” Well, the English word “free” has multiple meanings and they were confused by that. It even took me a few years to realize that I needed to clarify this. One meaning, you see, refers to price, and another meaning refers to freedom. When we speak of free software, we're talking about freedom, not price. So think of “free speech,” not “free beer.”

Some users got their copies of GNU Emacs from me through the net, and did not pay. Some users got their copies from me on a tape, and did pay. And some got their copies from someone else, not from me, because everyone who had a copy was free to redistribute it. And did they pay that somebody else? Well, I don't know; that was between them. They didn't have to tell me. So GNU Emacs was gratis for some users and paid for for other users, but it was free software for all of them, because all of them had certain essential freedoms, which are the definition of free software.

So let me now give you the definition of free software. You see, it's very easy to say “I'm in favor of freedom.” I mean, even Bush can say that. [Laughter] I don't think he knows what it means. But the point is, unless you make a person get more specific, it's just cheap talk. So let me give you—let me get more specific now, and give you the definition of free software.

A program is free software for you, a particular user, if you have the following four freedoms:

Freedom 0 is the freedom to run the program however you like; Freedom 1 is the freedom to help yourself by studying the source code to see what the program really does and then changing it to do what you want; Freedom 2 is the freedom to help your neighbor by distributing copies to others; and Freedom 3 is the freedom to help build your community, that is the freedom to publish a modified version so others can benefit from your changes;

All four of these freedoms are essential. They are not levels of freedom, they are four freedoms, all of which you must have in order for the program to qualify as free software. All of these are freedoms that no computer user should ever be denied.


7. Freedom 2 moral dilemma

Why these particular freedoms? Why should we define it this way?

Freedom 2 is necessary so that you can live an upright life, so that you can be ethical, be a good member of society. If you use a program that does not give you Freedom 2, the freedom to help your neighbor, the freedom to distribute copies to others, then you are facing a potential moral dilemma that could happen at any moment, when somebody comes up and says, “could I have a copy of that program?” At that point, what are you going to do? You're forced to choose between two evils. One evil is to make a copy of the program for that person and violate the license. The other evil is to comply with the license, but be a bad neighbor. So you've got to choose the lesser evil, which is to make a copy for that person and violate the license. [Laughter, applause]

You see, in this case, this evil is lesser because it's directed at somebody who intentionally tried to divide you from the rest of society, and thus did something extremely wrong to you; and therefore deserves it. However, it's not good to live your life by lying to people. When somebody {asks you to promise that} says, “I'll let you have a copy of this, but you'll have to promise not to share it with anyone,” the right thing to do is say no. Once you have thought about this moral dilemma, you should anticipate that when you start using that program it's going to lead you to choose between two evils, and therefore you should refuse to use that program. You should just say “no, thanks” to it, and that's the principle that I believe in. If someone offers me a program that I'm not free to share with you, I'm going to say no, on principle.

In fact, I was once in the audience when John Perry Barlow was giving a speech and he said, “raise your hands if you have no unauthorized copies of software.” And he was surprised to see someone raise his hand, until he saw it was me. And then he said, “oh, of course, you,” because he knew why I have no unauthorized copies; that's because all my copies of software are free software, and everybody's authorized to make copies. That's the whole point.

8. Freedom 2 spirit of good will

The most essential resource of any society is the spirit of good will, the willingness to help your neighbor; not necessarily every time you're asked, but fairly often. This is what makes the difference between a livable society and a dog-eat-dog jungle. This spirit is not going to be 100% and it's not going to be zero, but it's going to be somewhere in between—and cultural actions can influence it, can raise it or lower it. And it's essential to work to raise it some, because that makes life easier for everyone. So it's no accident that the world's major religions have been encouraging this spirit of good will for thousands of years.

So what does it mean when powerful social institutions say that it's wrong to share? They're poisoning this vital resource, something no society can afford. Now what does it mean when they say that if you share with your neighbor, you're a pirate? They're saying that helping your neighbor is the moral equivalent of attacking a ship. Well, nothing could be more wrong than that. Attacking ships is very, very bad; helping your neighbor is good.

And what does it mean when they establish harsh punishments for anyone caught sharing? How much fear do you think it's going to take before everyone's too scared to help his neighbor? And do you want that terror campaign to go on in our society? I hope that the answer is no. We need to abolish the war on copying that is being imposed on our society. We need to say, loud and clear, “copying and sharing with your neighbor is good, it's legitimate, and laws that prohibit this are wrong.”

9. Freedom 0 to run a program, Freedom 1 to modify it

So that's the reason for Freedom 2; it's essentially an ethical reason. You can't live an ethical life if you don't have Freedom 2.

Freedom 0 is needed for a completely different reason: so you can control your own computer. If you are restricted in when or how much or how you can run the program, clearly you're not using your computer in freedom. So Freedom 0 is obvious, but freedom 0 is not enough, because with Freedom 0 all you can do is use the program the way it was programmed by its developer. You're free to do this [makes hand sign] or nothing. To really be free, you've got to be in control of what the program does, so you need Freedom 1, which is the freedom to help yourself, the freedom to study the source code and then change it to do what you want.

If you don't have Freedom 1, you don't know what the program's doing. The developer is saying, “just trust me” and blind faith is the only way you can do it. And you have to be really blind, given that it's not unusual for proprietary programs to have malicious features, features that are put in not to serve the user, but rather to impose on, harm or restrict the user. For instance, spyware is quite common.

[51 seconds of missing audio were filled in by RMS in Aug 2010]

Microsoft Windows spies on the user; specific spy features have been found. Windows Media Player spies too; it reports to Microsoft whatever the user looks at.

[End replacement for 51 seconds of missing audio]

course do it. RealPlayer, for instance, spies on you. The TiVo spies on you. Some people were excited about the TiVo, enthusiastic about it, because it uses some free software inside. But it also has nonfree software in it and it spies on you. So this shows it's not enough. We shouldn't cheer when something uses some free software; we should cheer when it respects the user's freedom.

10. DRM, back doors, bugs

But spyware is not as bad as it gets. There are nonfree software packages that are deliberately designed to refuse to work. This is called DRM, Digital Restrictions Management, where the program says, “I won't let you look at that file; I won't let you copy this; I won't let you edit this.” Well, who the hell is this program to stop you? And sometimes nonfree programs will reconfigure your machine, for instance make it display advertisements, figuring that you won't know it's going to happen and you won't know how to undo it afterward.

And sometimes they have actual back doors. For instance, Windows XP has a back door: when it asks for an upgrade, it tells Microsoft who you are, so Microsoft can give you an upgrade designed just for you. And this upgrade could have secret accounts, it could have special spy features, it could just refuse to work. And there's essentially nothing you can do. So that's the back door that Microsoft knows about and we know about.

[Added in 2010: We later learned that Microsoft can force “upgrades”—a much nastier back door.]

There might be other back doors that we don't know about and maybe even Microsoft doesn't know about. When I was in India in January, I was told some programmers in India had been arrested and accused of working for Al-Qaeda, trying to introduce back doors into Windows XP. So, apparently, that effort failed. But did some others succeed? There's no way we can tell.

Now, I won't claim that all developers of nonfree software put in malicious features. There are some who try to put in features so that they will be convenient for the user and only for that. But they are humans, so they make mistakes. They can design features with all the best will that you don't like, or they can write bugs in their code. And when that happens, you're helpless too; you're the helpless prisoner of any decision that they make. Whether it's malicious or made with good will, if you don't like it, you're stuck.

Now, we, the developers of free software, are also human, we also make mistakes. I have designed features that users didn't like. I have written code that had bugs in it. The difference is, {with our} you're not a prisoner of our decisions, because we don't keep you helpless. If you don't like my decisions, you can change them, because you have the freedom to change them. I won't blame the developers of nonfree, user-subjugating software for being human and making mistakes; I will blame them for keeping you helpless prisoner of their mistakes by denying you the freedom to correct those mistakes yourself.

11. Freedom 3 having no master

But Freedom 1 is not enough. Freedom 1 is the freedom personally to study and change the source code. Freedom 1 is not enough because there are millions of users who use computers, but don't know how to program, so they can't take advantage of Freedom 1, not personally. And Freedom 1 is not enough even for us programmers, because there's just so much software, even so much free software, that nobody has the time to study it all and master it all and make all the changes that she wants.

So the only way we can really, fully have control over our own software is if we do so together. And that's what Freedom 3 is for. Freedom 3 is the freedom to publish a modified version, so others can use it too. And this is what enables us to work together, taking control of our software. Because I could make this change in a program and publish the modified version, and then you could make that change and publish the modified version, and someone else can make that change and publish the modified version. And now we've got a version with all three changes in it and everybody can switch to that if everybody likes it.

With this freedom, any collectivity of users can take control together and make the software do what they together want. Suppose there are 1,000,000 users who would like a certain change. Well, by luck, some of them will be programmers; let's say there are 10,000 of them who know how to program. Well, sooner or later, a few of them will make the change and publish the modified version and then all of those million users can switch to it. You know, most of them don't know how to program, but they can still switch to it. So they all get what they want.

Now let's suppose there are only 1,000 people who want some other change and none of them knows how to program. They can still make use of these freedoms. They can form an organization and each put in money, so if each puts in $100, that makes $100,000. And at that point they can go to a programming company and say, “will you make this change for $100,000 and when can you have it done?” And if they don't like the answer from there, they can go to another programming company and say, “will you make this change and when can you have it done?” Which shows us, first of all, that these 1,000 users who don't know how to program can, by using the four freedoms, get the change that they want. And second, it shows that free software means a free market for support.

Proprietary software typically means a monopoly for support. Only the developer has the source code in most cases, so only the developer can offer any support. If you want a change, you've got to go to the developer and beg. Now, if you're very big and important, maybe the developer will pay attention. If you're not, the developer will say, “go away, don't bother me.” Or maybe the developer will say, “pay us and we'll let you report a bug.” And if you do that, the developer will say, “thank you. In six months there will be an upgrade. Buy the upgrade and you'll see if this bug was fixed and you will see what new bugs we have for you.”

But with free software, you're dealing with a free market, so that those who really value support can, in general, get better support for their money by using free software. Now, one paradoxical consequence of this is, when you have a choice between several nonfree programs to do a job, this is actually a choice between monopolies. If you pick this program, the support for it afterwards will be a monopoly. If you pick this program, [points hand in different direction] the support for it will be a different monopoly, and if you pick this program, [points hand in different direction] the support for it will be yet another monopoly. So you're choosing one of these three monopolies.

Now, what this shows is that merely having a choice between a discrete set of options is not freedom. Freedom is something much deeper and much broader than having a few choices you can make. Many people try to equate freedom with having some choice and they're missing the point completely. Freedom means that you get to make the decisions about how to live your life. {It doesn't mean, you know} Having three choices about being able to choose this master or this master or this master is just a choice of masters, and a choice of masters is not freedom. Freedom is having no master.

12. Copyleft forbidding is forbidden

So I've explained the reasons for the four freedoms. And thus I've explained to you what free software means. A program is free software for you, a particular user, if you have all of these four freedoms. Why do I define it that way? The reason is that sometimes the same code can be free software for some users and nonfree for the rest. This might seem strange, so let me give you an example to show how it happens.

The biggest example I know of is the X Window System. It was developed at MIT in the late '80s and released under a license that gave the user all four freedoms, so if you got X in source code under that license, it was free software for you. Among those who got it were various computer manufacturers that distributed UNIX systems. They got the source code for X, they changed it as necessary to run on their platform, they compiled it and they put the binaries into their UNIX system, and they distributed only the binaries to all of their customers under the same license as the rest of UNIX—the same non-disclosure agreement. So, for those many users, the X Window System was no more free than the rest of UNIX. In this paradoxical situation, the answer to the question “is X free software or not?” depended on where you made the measurement. If you made the measurement coming out of the developer's group, you'd say, “I observe all four freedoms; it's free software.” If you made the measurement among the users, you'd say, “most of them don't have these freedoms; it's not free software.”

The developers of X did not consider this a problem, because their goal was not to give users freedom, it was to have a big success, and as far as they were concerned, those many users who were using the X Window System without freedom were just a part of their big success. But, in the GNU Project, our goal specifically was to give the users freedom. If what happened to X had happened to GNU, GNU would be a failure.

So I looked for a way to stop this from happening. And the method I came up with is called copyleft. Copyleft is based legally on copyright law, and you can think of it as taking copyright and flipping it over to get copyleft.

Here's how it works: we start with a copyright notice which legally doesn't actually make a difference anymore, but it reminds people that the program is copyrighted, which means that, by default, it's prohibited to copy, distribute or modify this program. But then we say, “you are authorized to make copies, you are authorized to distribute them, you are authorized to modify this program and you are authorized to publish modified or extended versions.” But there is a condition, and the condition says that any program you distribute that contains any substantial part of this must, as a whole, be distributed under these conditions, no more and no less. Which means that, no matter how many people modify the program or how much, as long as any substantial amount of our code is in there, that program must be free software in the same way. In effect, we guarantee that nobody can put himself between you and me and strip off the freedom and pass the code on to you missing the freedom. In other words, forbidding is forbidden.

13. GNU General Public License

Copyleft makes the four freedoms into inalienable rights for all users, so that wherever the code goes, the freedom goes with it. The specific license that we use to implement the general concept of copyleft is called the GNU General Public License, or GNU GPL for short. This license is used for around two thirds or three quarters of all free software packages. But that still leaves a substantial number that have other licenses. Some of those licenses are copyleft licenses, some are not. So we have copylefted free software and we have non-copylefted free software. In both cases, the developers have respected your freedom; they have not tried to trample your freedom. The difference is, with copyleft we go further and we actively defend your freedom against anyone who would try to be a middleman and take it away from you, whereas the developers of non-copylefted free software don't do that. They have not tried to take away your freedom, but they don't actively protect your freedom from anyone else. So I think that they could do more for the sake of freedom. But they haven't done anything bad; insofar as they have done things, those things are good. So I won't say that they are wrong, I will just say that they could do more. I think that they're making a mistake.

But their work is free software, so it does contribute to our community and, in fact, that software can be part of a free operating system such as GNU.

13a. Developing GNU

During the 1980s, our work on the GNU Project was to develop or find all these pieces of GNU so that we could have a complete GNU system. In some cases, someone else wrote a program and made it free software and we were able to use it, and that was good because it shortened the work that we had to do. For instance, the X Window System is one of the programs that was developed by others for reasons of their own, but they did make it free software, so we could use it.

Now, people were saying the job was so big, we'd never finish it. Well, I thought we would eventually get a free operating system but I agreed the job was big; we had to look for shortcuts. So, for instance, I always wanted to have windowing facilities in GNU. I had written a couple of window systems at the AI LAB before even starting GNU, so of course I wanted that in the system. But we never developed a GNU window system because someone else developed X first. I looked at it and I said, “well, it's not copylefted, but it is free, it's popular, it's powerful, so let's just use it.” And so we saved one big chunk of work. So we took it, X, and we put it into the GNU system and we started making other pieces of GNU work with X. Because the goal was to have a free operating system, not to have a free operating system every piece of which had been written purposely by us just for that.

14. Making money off free software

However, it only happened occasionally that someone else released some free software that was useful in GNU and when it happened, it was a coincidence, because they were not writing this software in order to have a free operating system. So when it happened, that was great, but there were lots of other pieces we had to develop. Some were developed by staff of the Free Software Foundation. The Free Software Foundation is a tax-exempt charity to promote free software which we founded in October, '85, after GNU Emacs' popularity suggested that people might actually start donating money to the GNU project. So we founded the Free Software Foundation and it asked for donations, but also took over selling the tapes of GNU Emacs. And it turns out that most of the FSF's income for the first many years came from that, from selling things, from selling copies of software and manuals that everyone was free to copy. Now this is interesting, because this was supposedly impossible; but we did it anyway.

Now that meant I had to find some other way to make a living. As the president of the FSF, I did not want to compete with it; I thought that would be unfair and not correct behavior. So I started making my living by commissions to change the software I had written and teaching classes about it. So people would want some change to be made in Emacs or GCC, and they would think of hiring me, because they figured I was the author so I could do a better job faster. So I started charging as much as $250 an hour and I calculated I could make a living in 7 weeks of paid work per year—and that meant enough money to spend, an equal amount to save, and an equal amount for taxes. And [when I reached] that point I figured, “I won't take any more paid work this year, I've got other, better things to do.”

So I've actually had three different free software businesses during the period I've been working on GNU. I've described two of them; the third one is, I get paid for some of my speeches. Whether I get paid for this speech, I don't yet know. [Laughter] I said, “please pay me what you can.” Now, I think Google ought to be able to afford to pay me some handsome amount, but whether it will, I don't know. Anyway, I figured it's worth doing the speech just for the good it will do for the movement.

15. Why write free software

So this raises the question of why people develop free software. You see, there are people who believe that no one would ever write software except to get paid, that that's the only motive that anyone would ever have to write code. It's amazing, the kind of utterly stupid, simplistic theories that people will sometimes believe because that's part of a prevailing ideology.

Now, human nature is very complex. Whatever it is people are doing, they might do for various reasons. In fact, one person will often have multiple motives simultaneously for a single act. Nonetheless, there are people who say, “if the software is free, that means nobody's paid to write it, so no one will write it.” Now, obviously they were confusing the two meanings of the word “free,” so their theory was based on a confusion. In any case, we can compare their theory with empirical fact and we can see that at least hundreds, maybe thousands of people are paid to work on free software, including some people here, I believe, and there are about a million or so people developing free software at all for the many different reasons they have. {So to say that nobody} This simplistic theory about motivation is absurd.

So let's see what motivates people to write free software; what are the real motives? Well, I don't necessarily know about them. There could always be a person who has a motive that I don't know about or I've forgotten about. I can only tell you the motives that I recall encountering.

One motive is political idealism: making the world a better place where we can live together in freedom. Now, that's a very important motive for me, but it's not my only motive. And there are others who write free software and don't agree with that motive at all.

Another motive that's very important is fun. Programming is tremendous fun. Not for everybody, of course, but for a lot of the best programmers. And these are the people whose contributions we want most. In fact, it's so much fun, it's especially fun, when no one can tell you what to do, which is why so many people who have jobs programming like to write free software in their spare time.

But this is not the only motive; another motive is to be appreciated. If 1% of our community is using your program, that's hundreds of thousands of users. That's a lot of people admiring you.

Another related, but different, motive is professional reputation. If 1% of our community is using your program, you can put that on your resume and it proves you're a good programmer. You don't even have to go to school.

Another motivation is gratitude. If you've been using the community's free software for years and appreciating it, then when you write a program, that's your opportunity to pay something back to the community that has given you so much.

Another motivation is hatred for Microsoft. [Laughter] Now, this is a rather foolish motive, because Microsoft is really just one of many developers of nonfree software and they're all doing the same evil thing. It's a mistake to focus [solely] on Microsoft, and this mistake can have bad consequences. When people focus too much on Microsoft, they start forgetting that all the others are doing something just as bad. And they may end up thinking that anything that competes with Microsoft is good, even if it is also nonfree software and thus inherently just as evil. Now, it's true that these other companies have not subjugated as many users as Microsoft has, but that's not for want of trying; they just haven't succeeded in mistreating as many people as Microsoft has, which is hardly, ethically speaking, an excuse. Nonetheless, {when this particular motive motivates} this motive does motivate people to develop free software, so we have to count it as one of the motives that has this result.

And another motive is money. When people were being paid to develop free software, that's part of their motive for the work that they're doing. In fact, when I was paid to make improvements in various programs I had written, that money was part of my motive for doing those particular jobs, too.

[RMS, 2010: A motive I forgot to mention is improving a free program because you want to use the improvement yourself.]

So there are many possible motives to write free software. And, fortunately, there are many developers of free software and a lot of free software is being developed.

16. The Kernel, Linux

So, during the 1980s we were filling in these missing pieces of the GNU operating system. By the early '90s we had almost everything necessary. Only one important piece was missing, one essential piece for an initial system, and that was the kernel. We started developing a kernel in 1990. {I was looking for some way to} I was looking for some shortcut, some way we could start from something existing. I thought that debugging a kernel would be painful, because you don't get to do it with your symbolic debugger, and when it crashes, it's sort of annoying.

So I was looking for a way to bypass that work, and I found one eventually, a microkernel called Mach that had been developed as a funded project at Carnegie Mellon. Now, Mach doesn't have all the features of UNIX; the idea is, it provides certain general low-level features and you implement the rest in user programs. Well, that, I thought, would be easy to debug, because they're user programs; when they crash, the system isn't dead. So people began working on those user programs, which we called the GNU Hurd, because it's a herd of GNU servers (you see, gnus live in herds).

Anyway, I thought that this design would enable us to get the job done faster, but it didn't work out that way; it actually took many years to get the Hurd to run, partly because Mach was unreliable, partly because the debugging environment wasn't very good, partly because it's hard to debug these multithreaded, asynchronous programs and partly because this was somewhat of a research project. At least that's as far as I can tell; I was never involved in the actual development of the Hurd.

Fortunately, we didn't have to wait for that, because in 1991, Linus Torvalds, a Finnish college student, developed his own kernel, using the traditional monolithic design, and he got it to barely run in less than a year. Initially, Linux—that's what this kernel's name was—was not free, but in 1992 he re-released it under the GNU General Public License and at that point it was free software. And so it was possible, by combining Linux and the GNU system, to make a complete free operating system. And thus, the goal we had set out for, that I had announced in 1983, had been reached: there was, for the first time, a complete modern operating system for modern computers, and it was possible to get a modern computer and run it without betraying the rest of humanity, without being subjugated. You could do this by installing the GNU + Linux operating system.

17. GNU vs. Linux confusion problem freedom

But the people who combined GNU and Linux got confused and they started naming the entire thing Linux, which was actually the name of one piece. And somehow that confusion spread faster than we have been able to correct it. So I'm sure you've heard many people speaking of Linux as an operating system, an operating system {most of which} which basically started in 1984 under the name of the GNU Project.

Now, this clearly isn't right. This system isn't Linux; it contains Linux, Linux is the kernel, but the system as a whole is basically GNU. So I ask you: please don't call it Linux. If you call it Linux, you're giving Linus Torvalds credit for our work. Now, he contributed one important piece of the system, but he didn't contribute the biggest part and the overall vision was there long before he got involved. We started developing the system when he was in junior high school. So please give us equal mention; surely we deserve at least that. You can do that by calling the system GNU/Linux, or GNU+Linux, or GNU&Linux, whichever punctuation mark you feel expresses it best.


Now, of course, part of the reason why I'm asking for this is that we deserve credit, but that's not really a very important thing. If it were just a matter of credit, it wouldn't be worth making a fuss about. But there more at stake here. You see, when people think that the system is Linux, they then assume incorrectly that it was mainly developed and started by Linus Torvalds and then they assume incorrectly that the overall vision came from him, so they look at his vision and follow that. Now, his vision is apolitical. He's not motivated to fight for freedom. He doesn't believe that computer users deserve the freedom to share and change software. He has never supported our philosophy. Well, he has a right to his views and the fact that he disagrees with us doesn't reduce the value of his contribution.

The reason we have the GNU+Linux system is because of a many-year campaign for freedom. We in the GNU Project didn't develop Linux, just as we didn't develop X, or TeX, or various other free programs that are now important parts of the system. But people who didn't share our values, who weren't motivated by the determination to live in freedom, would have seen no reason to aim for a complete system, and they would never have done so, and never have produced such a thing, if not for us.

But this tends to be forgotten nowadays. You will see, if you look around, most of the discussion of the GNU system calls it Linux, and tends to refer to it as “open source” rather than as “free software,” and doesn't mention freedom as an issue. This issue, which is the reason for the system's existence, is mostly forgotten. You see many techies who prefer to think of technical questions in a narrowly technical context, without looking beyond at social effects of their technical decisions. Whether the software tramples your freedom or respects your freedom, that's part of the social context. That's exactly what techies tend to forget or devalue. We have to work constantly to remind people to pay attention to freedom and, unfortunately, while we keep doing this, the users of our system often don't pay attention because they don't know it's our system. They don't know it's the GNU system, they think it's Linux. And that's why it makes a real difference if you remind people where the system came from.

People will say to me that it doesn't look good to ask for credit. Well, I'm not asking for credit for me personally; I'm asking for credit for the GNU Project, which includes thousands of developers. But they are right, it's true: people who are looking for some reason to see evil can see evil in that. So they go on and say, “you should let it drop, and when people call the system Linux, you can smile to yourself and take pride in a job well done.” That would be very wise advice if the assumption were correct: the assumption that the job is done.

We've made a great beginning, but that's all. We haven't finished the job. We will have finished the job when every computer is running a free operating system and free application programs exclusively. The job is to liberate the inhabitants of cyberspace. We've made a great beginning; we've developed free operating systems and free GUI desktops and free office suites and there are now tens of millions of users of these. But there are hundreds of millions of users of proprietary systems, so we have a long way to go. And, despite this wide range of free software, there are still a lot of application things that there is no free software to do; so we have a lot more work ahead of us.

We've come in view of finishing the job, you know. Maybe we're only one order of magnitude away, having come through many orders of magnitude. But that doesn't mean that what's left is easy. And today we have something that we didn't have before: we have enemies; powerful, rich enemies, powerful enough to buy governments.

18. Enemies of free software

At the beginning, GNU and the free software movement had no enemies. There were people who weren't interested, lots of them, but nobody was actively trying to stop us from developing and releasing a free operating system. Nowadays, they are trying to stop us and the main obstacle we face is this, rather than the work itself.

In the US, there are two different laws that prohibit various kinds of free software.

One of them is the DMCA, which has been used to prohibit the free software to play a DVD. If you buy a DVD, it's lawful for you to view it in your computer, but the free software that would enable you to do this on your GNU/Linux system has been censored in the US. Now, this affects a fairly narrow range of software: software to view encrypted media. But many users may want to do that, and if they can't do that with free software, they may take that as a reason to use nonfree software, if they don't value their freedom.

But the big danger comes from patent law, because the US allows software ideas to be patented. Now, writing a non-trivial program means combining hundreds of different ideas. It's very hard to do that if any one of those ideas might be someone's monopoly. It makes software development like crossing a mine field, because at each design decision, probably nothing happens to you, but there's a certain chance that you will step on a patent and it will blow up your project. And, considering how many steps you have to take, that adds up into a serious problem. We have a long list of features that free software packages don't have, because we're scared to implement them.


And now, the FCC is considering applying the broadcast flag regulation to software. The FCC adopted a regulation {prohibiting digital TV tuners unless} requiring digital TV tuners to have a mechanism to block copying and this has to be tamper-resistant, meaning it can't be implemented in free software. They haven't finished deciding whether this applies to software or not, but if they do, they will have prohibited GNU Radio, which is free software that can decode digital TV broadcasts.

Then, there's the threat from hardware that has secret specifications or is designed to interfere with the user's control. Nowadays there are many pieces of hardware you can get for your PC whose specifications are secret. They'll sell you the hardware, but they won't tell you how to run it. So how do we write free software to run it? Well, we either have to figure out the specs by reverse engineering or we have to put market pressure on those companies. And in both cases, we are weakened by the fact that so many of the users of GNU/Linux don't know why this system was developed and have never heard of these ideas that I'm telling you today. And the reason is that, when they hear about the system, they hear it called Linux and it's associated with the apolitical philosophy of Linus Torvalds. Linus Torvalds is still working on developing Linux. {which is, you know} Developing the kernel was an important contribution to our community. At the same time, he is setting a very public bad example by using a nonfree program to do the job. Now, if he were using a nonfree program privately, I would never even have heard about it and I wouldn't make a fuss about it. But by inviting the other people who work on Linux to use it with him, he's setting a very public example legitimizing the use of nonfree software. So when people see that, you know, if they think that's okay, they can't possibly believe that nonfree software is bad. So then, when these companies say, “yes, {we support} our hardware supports Linux, here is this binary-only driver you can install, and then it will work,” these people see nothing wrong in that, so they don't apply their market pressure and they don't feel motivated to help in reverse engineering.

So when we face the various dangers that we must confront, we are weakened by the lack of resolve. Now, having strong motivation to fight for freedom won't guarantee that we win all of these fights, but it will sure help. It will make us try harder, and if we try harder, we'll win more of them.

19. Treacherous computing

We are going to have to politically organize to keep from being completely prohibited from writing free software.

Today, one of the most insidious threats to the future of free software comes from treacherous computing, which is a conspiracy of many large corporations. They call it “trusted computing,” but what do they mean by that? What they mean is that an application developer can trust your computer to obey him and disobey you. So, from your point of view, it's _treacherous computing_, because your computer won't obey you anymore. The purpose of this plan is that you won't control your computer.


And there are various different things that treacherous computing can be used to do, things like prohibit you from running any program that hasn't been authorized by the operating system developer. That's one thing they could do. But they may not feel they dare go that far. But another thing that they plan to do is to have data that's only available to a particular application. The idea is that an application will be able to write data in an encrypted form, such that it can only be decrypted by the same application, such that nobody else can independently write another program to access that data. And, of course, they would use that for limiting access to published works, you know, something to be a replacement for DVDs so that it would be not only illegal, but impossible to write the free software to play it.

But they don't have to stop at doing this to published data. They could do it to your data too. Imagine if treacherous computing is common in 10 years and Microsoft decides to come out with a new version of Word format that uses treacherous computing to encrypt your data. Then it would be impossible to write free software to read word files. Microsoft is trying every possible method to prevent us from having free software to read Word files. First, they switched to a secret Word format, so people had to try to figure out the format. Well, we more or less have figured it out. There are free programs that will read most Word files (not all). But then they came up with another idea. They said, “let's use XML.” Now here's what Microsoft means when they speak of using XML. The beginning of the file has a trivial thing that says “this is XML and here comes binary Word format data,” and then there's the binary Word format data and then there's something at the end that says, “that was binary Word format data.” And they patented this. {so that… I'm not sure} I don't know exactly what the patent does and doesn't cover, but, you know, there are things we could do, either reading or writing that file format, probably they could try suing us about. And I'm sure that, if treacherous computing is available for them to use, they'll use that too.

This is why we have a campaign to refuse to read Word files. Now there are many reasons you should refuse to read Word files. One is, they could have viruses in them. If someone sends you a Word file, you shouldn't look at it. But the point is, you shouldn't even try to look at it. Nowadays there are free programs that will read most Word files. But it's really better, better than trying to read the file is if you send a message back saying, “please send that to me in a format that isn't secret. It's not a good idea to send people Word files.” And the reason is, we have to overcome the tendency in society for people to use these secret formats for communication. We have to convince people to insist on publicly documented standard formats that everyone is free to implement. And Word format is the worst offender and so that's the best place to start. If somebody sends you a Word file, don't try to read it. Write back, saying “you really shouldn't do that.” And there's a page in which is good to reference. It gives an explanation of why this is an important issue.


20. Help GNU

Now, is the website of the GNU Project. So you can go there for more information. In the /gnu directory you'll find the history and in the /philosophy directory you'll find articles about the philosophy of free software and in the /directory you'll find the Free Software Directory, which now lists over 3,000 usable free software packages that will run the on GNU/Linux system.

[It is now over 6000, and located in]

Now, I'm about to close my speech, but before I do, I'd like to mention that I've got some stickers here to give away. These stickers show a flying gnu and a flying penguin, both rather unrealistic, but they're superheroes. And {I also have some things} if people don't mind, I've got some things I'm selling on behalf of the Free Software Foundation, so if you buy them, you're supporting us. I've got these buttons that say, “ask me about free software—it's all about freedom” and I've got some GNU keyrings and GNU pins that are sort of pretty. So you can buy those. You can also support us by becoming an associate member. Now, you can do that just through our website, but I also have some cards you can have if you would like to join [right now].

21. Saint Ignucius

So now I will close my speech by presenting my alter ego. See, people sometimes accuse me of having a “holier than thou” attitude. Now, I hope that's not true. I'm not going to condemn somebody just for not being as firmly committed as I am. I will try to encourage him to become more so, but that's different. So I don't think I really have a “holier than thou” attitude, but I have a holy attitude because I'm a saint; it's my job to be holy.

[Dons a black robe and a magnetic disk halo]
[Laughter, applause]
[Richard holds a laptop like a holy book and waves]

I am Saint Ignucius of the Church of Emacs. I bless your computer, my child.

Emacs started out as a text editor which became a way of life for many computer users and then a religion. Does anyone know what the alt.religion.emacs newsgroup was used for? I know it existed, but since I'd never read net news, I don't know what was said in it.

In any case, now we even have a great schism between two rival versions of Emacs, and we also have saints; no gods, though.

To be a member of the Church of Emacs, you must recite the Confession of the Faith: you must say, “There is no system but GNU, and Linux is one of its kernels.”

The Church of Emacs has advantages compared with other churches I might name. To be a saint in the Church of Emacs does not require celibacy. So if you're looking for a church in which to be holy, you might consider ours.

However, it does require making a commitment to live a life of moral purity. You must exorcise the evil proprietary operating systems that possess all the computers under either your practical control or your authority, and you must install a wholly [i.e., holy] free operating system, where “wholly” can be spelled in more than one way, and then only install free software on top of that. If you make this commitment and live by it, then you, too, will be a saint and you, too, may eventually have a halo—if you can find one, because they don't make them anymore.

Sometimes people ask me if, in the Church of Emacs, it is a sin to use Vi. Well, it's true that VI-VI-VI is the editor of the Beast, [laughter] but using a free version of Vi is not a sin, it's a penance.

And sometimes people ask me if my halo is really an old computer disk. [Points at halo] This is no computer disk, this is my halo. But it was a computer disk in a previous existence.

So, thank you everyone.


22. About anonymity, credit cards, cell phones

So I can answer questions for a while.

AUDIENCE: Yeah, do you know, or can you tell us why Linus Torvalds, who has very very different attitudes with yours, released Linux under your [unintelligible]? What motivated him?

RICHARD: I don't know why Linus Torvalds switched to the GNU GPL for Linux. You'd have to ask him that. I don't recall ever seeing the reason for that. I don't know.

AUDIENCE: Can you say something about the current effort to put security in the network itself?

RICHARD: I don't know… he said, “efforts to plug security into the network.” I don't know what that means.

AUDIENCE: [unintelligible] remove anonymity from the network itself.

RICHARD: Remove anonymity? Well, I don't know about those efforts, but I think it's horrible. I don't do e-commerce because I don't like to buy things with credit cards. I want to buy things anonymously and I do so by paying cash in a store. I don't like giving Big Brother any records about me. For the same reason, I do not have a cell phone. I don't want to carry a personal tracking device. We have to fight more to preserve our privacy from surveillance systems. So, although I'm not familiar with the specific efforts you're talking about, I find them dangerous, much more dangerous than computer insecurity. Now, perhaps that's because I'm not a Windows user; so I have less problem to deal with.

AUDIENCE: [unintelligible]

RICHARD: No, we can't. Basically he's asking if we can monopolize file formats. Well, the answer is, we can't do so using our copyright-based licenses, because copyright does not cover any idea, principle, method of operation or system; it only covers the details of expression of a work of authorship. So we can't, using our licenses like the GNU GPL, prohibit anyone from writing his own code to handle the same format.

We could conceivably get patents; however, it turns out patents are very, very different from copyright; they have almost nothing in common, and it turns out it costs a lot of money to get a patent and even more money to keep the patent going. And the other thing is, {Microsoft doesn't need to get} you shouldn't assume that what Microsoft is getting a patent on is important because it's a big improvement. It just has to be different. Microsoft can get a patent on something about a file format that's different and then they can force most users to switch over to a new format that uses that idea. And Microsoft can do this because of its market power, its control.

We can't do that. The whole thing about the free software is, the developers don't have any power; the users are in control. We can't force users to switch over to anything, not even for their own safety.

You know, we've been trying since around 1992 or so to convince users to stop using GIF format, because that format is patented and some users will get sued. So we said, “everybody please stop using GIF format for the sake of those who get sued if the public uses this format.” And people haven't listened. So the thing is, we can't do what Microsoft does, because that's based on using the power that they have, and since we have chosen to respect people's freedom, we don't have power over the public.

24. Dangers of webmail loss of freedom

AUDIENCE: So, when somebody's using Google, they don't have access to the source code that we use, so they have no way of [unintelligible] what we do, so using that violates their freedom.

RICHARD: When a person is accessing the Google server, they don't have either the binaries or the source code of the program that Google is using, because it's Google that's using the program; that person is not using the program. So I wouldn't expect to have the authority to change the software that's running on your computer. You should have the freedom to change the software that's running on your computer, but I would never expect that I would have the freedom to go into your computer and change the software there. Why should you let me do that? So that's the way I see it when a person is using Google server to do a search.

Now, there is a possible danger there. The danger doesn't come from things like Google. The danger comes from things like Hotmail. When people start using a server on the net to store their data and to do the jobs that they really could be doing on their own computer, that introduces a danger. I've never understood the people who said that thin clients were the future, because I can't imagine why I would ever do things that way. I've got a PC and it's capable of doing things like running a mail reader; I'm going to have the mail on my own computer, I'm not going to leave it on anybody's server. Especially not a server I have no reason to trust. And these days, of course, if you allow your personal data to be on somebody's server, you might as well be handing it straight to Ashcroft and his gestapo.

[RMS, 2010: Gmail is comparable to Hotmail in this regard. See also for another issue that applies to some, but not all, network services.]

AUDIENCE: unintelligible

RICHARD: He's asking, “if people were using a thin client and all the computation were done on a remote server.” Yes, it does mean that people lose freedom, because, clearly, you can't change the software that's set up on somebody else's server, so if you're using the software on somebody else's server, instead of running it on your own computer, you lose control. Now, I don't think that's a good thing, and therefore I'm going to encourage people not to go along with it. People will keep on developing the software to do these jobs on your own machine.

{Leaving so soon? [Laughter] I hope it wasn't something I said. And gee, now I won't get to meet her. Anyway.}

AUDIENCE: Are the Creative Commons a different denomination of the same religion or a different religion?

RICHARD: {Creative Commons} Well, first of all, this isn't a religion, except as a joke. The Church of Emacs is a joke. Please keep in mind, taking any church too seriously can be hazardous to your health, even the Church of Emacs. So this has nothing to do with religion.

This is a matter of ethics. It's a matter of what makes for a good society and what kind of society we want to live in. These are not questions of dogma, these are questions of philosophy and politics.

The Creative Commons licenses are designed for artistic works, and I think that they are good for artistic works. The issue for artistic works is not exactly the same as for software.

Software is an example of a practical, functional work. You use it do to a job. The main purpose of a program is not that people will read the code and think, “boy, how fascinating, what a great job they did.” The main purpose of software is, you run it and it does something. And yes, those people who are interested in software will also read it and learn, but that's not the main purpose. It's interesting because of the job it will do, not just because of how nice it is to read. Whereas with art, the main use of art is the sensation that you get when you look at it or listen to it. So these are very different ways of being used and, as a result, the ethical issues about copying and modification are different.

For practical, functional works, people have to be free with the four freedoms, including free to publish a modified version. But for art I wouldn't say that. I think that there's a certain minimum freedom that we must always have for using any published work, and that is the freedom to non-commercially distribute verbatim, exact copies. But I wouldn't say that it has to go further than that necessarily. So I think the Creative Commons licenses are a very useful and good thing to use for art.

26. Malicious free software

AUDIENCE: Since everybody has the freedom to modify the code and republish it, how do you keep out saboteurs?

RICHARD: Well, you don't. The point is, you can't ever. So you just look at these different versions and you see which one you actually like. You can't keep the saboteurs out of nonfree software either; in fact, the developer could be the saboteur. The developers often put in, as I said, malicious features. And then you're completely helpless. At least with free software, you can read the source code, you can compare the two versions. If you're thinking of switching from this version to that version, you can compare them and see what's different and look for some malicious code.

27. Patented file formats

AUDIENCE: Do you happen to know which popular file formats are secret and which ones are public?

RICHARD: Well, of the popular file formats, the only ones that I know of that are secret are some Microsoft ones. But, on the other hands, there are others that have patent problems. For instance, there's still a patent covering LZW compression, which is used in GIF format. And someone has a patent he claims covers JPEG format and is actually suing a bunch of companies. And then there's a patent on MP3 audio, so that the free software MP3 encoders have been driven underground in the US [1]. That's why people should switch to Ogg Vorbis format. And then, if you look at, say, MPEG-2 video, there are 39 different US patents said to cover aspects of MPEG-2. So there are a lot of such problems.

28. Games as free software

AUDIENCE: Is there any software that sort of mixes between the Creative Commons and functional software, such as games or…?

RICHARD: Well, {you can say that a game} in many cases you can look at a game as the combination of a program and a scenario. And then it would make sense to treat the program like a program and the scenario like a work of fiction. On the other hand, what you see is that it's quite useful for the users to edit and republish modified versions of these scenarios. So, although those are like fiction and art, not like software, it really seems to be useful for users to be free to change them.

29. GPL freedoms for cars, saving seeds

AUDIENCE: Do you envision this free software philosophy to go across, off the boundary to products, commodities…

RICHARD: When you say, “products, commodities,” could you be concrete?

AUDIENCE: [unintelligible] cars

RICHARD: So should the free software philosophy apply to cars? Okay, well the free software philosophy is, you should be free to copy and modify them. So, if you have a car copier, I think you should be free to copy any car. But there are no car copiers, so that really is a meaningless question. And then, second, modifying. Well, yeah, I think if you've got a car, you should be free to modify it and, in fact, lots of people do modify their cars. So, there may be some restrictions on that, but to a large extent that freedom exists. So what you see is that this isn't really a meaningful question when you're talking about physical objects. There are, in general, no copiers for physical objects.

If we imagine, someday in the future, that such copiers exist, well that will be a different situation and yeah, that change would have consequences for ethics and politics. If we had food copiers, I'm sure that agribusiness would be trying to forbid people from having and using food copiers. And that would be a tremendous political issue, just as today there's a tremendous political issue about whether farmers ought to be allowed to save seeds. Now, I believe that they have a fundamental right to save seeds and that it's tyranny to stop them. A democratic government would never do that.

30. No software is better than nonfree software

AUDIENCE: [roughly] Do you see a problem with free software being under-produced because nobody wants to invest money [unintelligible]?

RICHARD: I don't know what you mean by “under-produced.” We see that some people develop free software and some don't. So we could imagine more people developing free software and, if so, we'd have more of it. But, you see, the tragedy of the commons really is a matter of overuse. And that's something that can happen maybe with a field, but it doesn't happen with software; you can't overuse a program, you don't wear it out. So, really, there's no analogy there.

AUDIENCE: Well, the example you gave is, let's say there's a useful program and a thousand people want a change to it. You said they could get their money together and go hire a programmer to make the change. But each individual in that group can say, “well, I'll just let the 999 pay for the change.”

RICHARD: Well, they can do that, but that would be pretty stupid, because if they saw that the result was, it wasn't getting done, then if it's of some importance to them, then they're much better off joining and contributing their money so that the change gets made. And whether they do this or not, either way I won't agree that anything tragic has happened. If they join and they pay for their change and they get it, that's good, and if they don't join and they don't pay for that change, that's good too; I guess they didn't want it enough. Either one's okay.

Nonfree software is evil and we're better off with nothing than with nonfree software. The tragedy of the commons can happen either through overuse or under-contribution, but overuse is impossible in software. Under-contribution happens when a program is proprietary. Then it's a failure to contribute to the commons. And so I would like that proprietary software to stop being developed. A nonfree program is worse than no program, because neither one allows you to get a job done in freedom, but the nonfree program might tempt people to give up their freedom and that's really bad.

31. Portability of free software

AUDIENCE: Is their a potential conflict between the free software philosophy and the portability of [unintelligible]?

RICHARD: No, {I don't see} this makes no sense to me at all. I see no conflict between the philosophy of free software and portability. And in the free software world we've worked very hard to achieve portability from all sides. We make our software very portable and we make our software standardized so that other people can easily have portability, so we are aiding portability from every possible direction. Meanwhile, you see Microsoft deliberately introducing incompatibilities and deliberately blocking interoperability. Microsoft can do that because it has power. We can't do that. If we make a program incompatible and the users don't like it, they can change it. They can change it to be compatible. So we are not in a position where we could impose incompatibility on anybody, because we have chosen not to try to have power over other people.

32. Is some free software obfuscated on purpose?

AUDIENCE: Something [unintelligible] obfuscated [unintelligible] understand it.

RICHARD: Well, I disagree with you. Please, this is silly. If you're saying a program is hard to understand, that's not the same as the people are restricting it. It's not the same as saying, “you're forbidden to see it.” Now, if you find it unclear, you can work on making it clearer. The fact is, the developers probably are trying to keep it clear, but it's a hard job and, unless you want to compare our software with proprietary software and see which one is clearer, you have no basis to make the claim that you're making. From what I hear, nonfree software is typically much worse and the reason is that the developers figure no one will ever see it, so they'll never be embarrassed by how bad it is.

33. Proprietary keeping an edge

AUDIENCE: You hear the argument a lot from people who manufacture devices or [unintelligible] hardware that they need to have proprietary software in order to give them an edge, because, if they gave away the software for free, then a competitor could manufacture the device [unintelligible].

RICHARD: I don't believe this. I think it's all bullshit, because there they are competing with each other and each one's saying, “we need to make the software proprietary to have an edge over the others.” Well, if none of them did it, they might all lose their edge? I mean, so what? We shouldn't buy this. And I mean, we shouldn't buy what they're saying and we shouldn't buy their products either.

34. Forbidding is forbidden how is this freedom?

AUDIENCE: I might be saying [unintelligible]

RICHARD: Please don't. The issue that you want to raise may be a good issue, but please try to raise it in a neutral way, rather than raising it with an attack.

AUDIENCE: There's something in my mind, so I'll just speak up. The thing is, by actually registering [unintelligible] thing and saying that “you can redistribute this software but you have to comply with these four freedoms,” is that not restricting my freedom too?

RICHARD: No, it's restricting you from having power. To stop A from subjugating B is not a denial of freedom to A, because to subjugate others is not freedom. That's power.

Now, there may be people who would like to exercise power and we're stopping them, but that's good and that's not denying anyone freedom.

I mean, you could just as well say if you're overthrowing a dictator, the dictator's saying, “you're taking away my freedom to dictate to everyone!” But that's not freedom, that's power.

So I'm making the distinction between freedom, which is having control over your own life, and power, which is having control over other people's lives. We've got to make this distinction; if we ignore the difference between freedom and power, then we lose the ability to judge whether a society is free or not. You know, if you lose this distinction, then you look at Stalinist Russia and you say, “well, there was just as much freedom there, it's just that Stalin had it all.” No! In Stalinist Russia, Stalin had power and people did not have freedom; the freedom wasn't there, because it's only freedom when it's a matter of controlling your own life. Controlling other people's lives is not freedom at all, not for either of the people involved.

35. Can Google help free software

AUDIENCE: In your opinion, is there anything that Google as a company could do better in the spirit of free software?

RICHARD: I actually don't know enough about what Google is doing to have any opinion. But if Google would like to donate some money to the Free Software Foundation, we would gladly accept it. {I gather that, I mean} I met some people here who are working on a particular free program, namely Linux, the kernel. And I didn't ask actually if they publish their improvements. [AUDIENCE: They do] Oh good, so that's contributing. I mean, if you want to contribute to other pieces of free software, that would be nice too, but I don't know if you have a need to do that. And, of course, if you ever have a chance to release some other generally useful new piece of free software, that would be good too.

[RMS, 2010: Google now distributes some large nonfree programs. Some are written in Javascript, and servers install them without your noticing.]

36. Free software on windows, good or bad

I'll take three more questions.

AUDIENCE: So, if I develop free software for a proprietary system such as Windows, essentially I'm supporting the proprietary system. Am I doing a good or a bad thing here?

RICHARD: Well, there's a good aspect and a bad aspect. In regard to the use of your code, you're respecting other people's freedom, so that's good, but the fact that it only runs on Windows is bad. So, really, you shouldn't develop it on Windows. You shouldn't use Windows. Using Windows is bad. {That is, in itself} It's not as bad as being the developer of Windows, but it's still bad and you shouldn't do that.

AUDIENCE: So you're saying, just don't do it at all.

RICHARD: Yeah, don't use Windows. Use GNU/Linux and develop your free program for GNU/Linux instead. And then it will be good in both ways.

AUDIENCE: But couldn't it open Windows users to this ideology?

RICHARD: It could, but there's enough free software available for use on Windows to have that effect. And the thing is, developing software for Windows is going to create a practical incentive for people to use Windows, rather than use GNU/Linux. So, please don't.

[RMS, 2010: to put it more clearly, making free programs run also on Windows can be useful as he said; however, writing a free program only for Windows is a waste.]

37. SCO's suit

AUDIENCE: What would be the impact of SCO winning their argument against Linux? So what would be the impact on…

RICHARD: I don't know, it depends. It would have no effect on the GPL. But {it might have some effect} some code might have to be removed from Linux. And whether that would be a big problem or a tiny problem depends on what code, so there's no way of saying. But I don't think SCO is a real problem. I think software patents and treacherous computing and hardware with secret specs, those are the real problems. That's what we've got to be fighting against.

38. Stallman's problem typing

AUDIENCE: I have a non-ideology question. I'm personally very interested in your battle with repetitive stress injuries and the impact that it had on the development of GNU Hurd.

RICHARD: None, because I was never working on the GNU Hurd. {I've never} We hired a person to write the GNU Hurd. I had nothing to do with writing it. And there were a few years when I couldn't type much and then we hired people to type for me. And then I found, by using keyboards with a light touch, I could type again.

39. Open source, good or bad Pat-riot Act.

AUDIENCE: Can you give us your opinion of open source?

RICHARD: Well, the open source movement is sort of like the free software movement, except with the philosophical foundation discarded. So they don't talk about right and wrong, or freedom, or inalienable rights, they just don't present it in ethical terms. They say that they have a development methodology that they say typically results in technically superior software. So they only appeal to practical, technical values.

And what they're saying may be right and if this convinces some people to write free software, that's a useful contribution. But I think they're missing the point when they don't talk about freedom, because that's what makes our community weak, that we don't talk about and think about freedom enough. People who don't think about freedom won't value their freedom and they won't defend their freedom and they'll lose it. Look at the USA Pat-riot Act. You know, people who don't value their freedom will lose it.

40. The end

So thank you, and if anyone wants to buy any of these FSF things or…



  1. All the patents on MP3 will have expired by 2018.