Richard Stallman interview

Ce document est une interview réalisée sur l'IRC par Robert Levin (lilo@varley.openprojects.net) le 13 juin 1998 sur le channel #gnu. Ce document est une retranscription non officielle, et peut être sujette à des erreurs.

Une traduction est proposée ici.


RMS : Hello

RL : Hi all On behalf of Some.Net and OpenProjects.Net, I'd like to welcome you all to the SomeNet Forum with Richard M. Stallman I'm Robert Levin, your moderator for this evening!
we've still got a few people taking their seats okay, I think we're ready to go the format for tonight's forum will be a moderated question and answer, and if you're interested in concurrent discussion, that can be found on channel #GNU-d later in the program, we'll take some questions....everyone be a bit patient with us though, and apologies if we don't get to everyone, we have a nice crowd this evening :)
okay, let me introduce our featured speaker Richard M. Stallman is possibly best known as the author of EMACS and the GPL, and the founder of the Free Software foundation his first direct contact with computers was in 1969, at the IBM New York Scientific Center he went to work at the MIT Artificial Intelligence lab in 1971....he walked in and was hired the same day (he was an undergraduate at Harvard at the time) Richard was the recipient of a $240,000 MacArthur Foundation prize fellowship, and has been instrumental in helping to open up and expand the Free Software community he's responsible as I noted for the FSF's General Public License, and Library GPL, which were instrumental in making Linux the success it's been in a moment we'll begin the program....but first, I want to take the opportunity to note some of the community's 'movers and shakers' in the audience....and say a quick hello to Alan Cox, longtime Linux hacker in addition, we have Mike McLagan (I always mess up the spelling of your name :), the maintainer of the linux.org web pages....welcome, Mike!
Arnt Gulbrandsen of Trolltech was here a few minutes ago, if you're still there, hey! we have a number of Debian developers over from #debian, Marcus Fleck, the moderator of comp.lang.python.announce, and I know I'm missing about ten other people I'd love to announce :) we just *had* (and hopefully will have again :) CmdrTaco, Rob Malda of Slashdot fame, who deserves a big applause for running Slashdot.org, and and Scoop of freshmeat.org! :)
I have to apologize to our speaker, by the way....not five minutes into our discussion I've referred to Debian, and I want to correct that reference to Debian GNU/Linux
Richard: maybe we could start our discussion by having you comment briefly about the difference in the way I put it, and my correction, and the reasons for that :)

RMS : Do you read me?

RL : yes, we can see your comments

RMS : GNU/Linux means a modified version of the GNU operating system, which uses Linux as the kernel.

RL : rms: a very large percentage of the utilities used with the Linux kernel are actually Free Software Foundation utilities, aren't they okay, maybe we can start by talking about your early days in computing what was the first really significant program you wrote, in your view, and what was the impetus for writing it?

RMS : The Free Software Foundation raises money to fund work on GNU. But most of the work on the GNU project is done by volunteers. the GNU project started with the goal of making a complete free Unix-like operating system. Doing this, we wrote many of the programs (utilities and others too) while we also found many programs that other people had written for other reasons but that would be useful to make the whole system. So the trees in this forest were planted by lots of people, but the reason there is a forest--a whole operating system--is because the GNU project decided to make a forest.

RL : you certainly did do that....it's hard to go over a list of any free software distribution without seeing a lot of FSF programs and GPL'ed code

RMS : Once again, you're counting the trees. At the beginning, people said that writing a whole operating system was such a big job that it was silly to try. So the most important thing that the GNU project has done wasn't any of the specific software, or the GNU GPL. It was to set the goal of making a whole free system and write software and licenses and manuals all with a view towards filling the gaps in the system-to-be.

RL : rms: in your view, what's the place of the Linux kernel in that tapestry....does it fit, or is it a placeholder?

RMS : The kernel is an essential major component of the operating system. So it's a very important piece of software.

RL : but is this a kernel you want to end up with? we've heard a lot about Hurd over the years....in your view, is it a logical successor?

RMS : I don't know whether the idea of 'logical successor' has a meaning here. Linux works fine as a kernel; we don't *need* another kernel. But since the Hurd has a more powerful and flexible architecture and since it is more-or-less working, I'd like to see it finished. Advancing technology is not as important as fighting for freedom, but it's still a good thing to do.

RL : that word 'freedom' seems very important to you....recently we've had a lot of discussion about terminology....'free software' versus 'open source' why is 'free software' the right term for what you and others in the community do?

RMS : Eric Raymond told me he was shocked that I care more about freedom than about advancing technology, but I think I'm in good company. In the early 1980s I was faced with a choice: either use proprietary (non-free) software, and accept a way of life in which sharing with other people is called 'piracy', or stop using computers, or...make some other alternative To have another alternative, we needed a free operating system, one that we would not be forbidden to change. and forbidden to share. Hence the GNU project. The idea is that if people have a free operating system available, then they won't feel that the only course available is to use proprietary software.

RL : so, this is a political and philosophical point for you? freedom in the civic sense? does it have implications for society as a whole?

RMS : Absolutely. Along with freedom of speech, freedom of assembly, there is freedom to share generally useful information with other people. This should be an inalienable right. The GNU GPL is designed to make this an inalienable right, for the specific programs that are released under the GPL.

RL : how do you see this sort of freedom as it dovetails with the concept of property? are you a 'propertarian' philosophy, and is intellectual property the same sort of thing to you as physical property? s/philosophy/philosopher/

RMS : Property rights are systems that people set up for dealing with each other. These systems are complex and have many parameters that can be set in various ways. Property rights are just a part of the whole complex of systems that we use. We should judge them like all the other details of these systems: do they have good results for the people who work by them? I don't think that the idea of property, or the specific complex details of any particular system of property rights, is sacred. I think that property rights for material objects give reasonable results for the most part, here in the US, but some details of our system are not tuned very well. Markets are susceptible to instabilities--for example, thebusiness cycle and the rich-get-richer phenomenon. Various government policies have been designed to limit these instabilities. Unfortunately our government is now dominated by business and no longer tries to ensure that prosperity is spread throughout society. When Reagan advocated policies to redistribute wealth toward the rich, there was an opposition party to criticize it. But now that the Democrats have mostly jumped on board, Clinton does not have to say the words 'trickle down' or try to defend this policy.

RL : rms: is that why your approach to free software has been non-governmental, i.e., a private license agreement? to avoid trying to move mountains that are resistant to moving?

RMS : Anyway, I think that it is reasonable for people to own houses, cars and sandwiches. But if any particular kind of property rights causes general unhappiness and fear, then we should get rid of that particular kind, or modify it. Many people don't realize that copyright and patents are considered by the US legal system to be artificial government interference with the public's natural rights. The stated purpose of these systems is to promote progress: in other words, a public benefit. The concept of 'the copyright bargain' is used to describe this system: the public pays by losing certain freedoms, and in exchange gets more books published. I think this is a reasonable general idea, but I value the freedom to share with other people, and I'm not willing to give it up just to get more software written.

RL : so, you feel that patents for software, in particular, may not be a general good?

RMS : That's a bad bargain. Patents in software are completely a bad idea, because they obstruct progress instead of promoting it. But I was talking about copyright before. Don't ever try to talk about patents and copyrights together or generalize from one to the other. Almost every detail is different between them.

RL : but in practice, haven't both resulted in the collection of a lot of 'intellectual property' rights by fairly large, bureaucratic organizations? it seems that rarely do the productive rights to patents and copyrights remain with their authors

RMS : At that broad level of abstraction, yes they have similar results to a large extent. But when you want to ask what to do about the situation, the details start to matter.

RL : what do you see as the successes of the various GPL versions, and the failures? what has gone best for you, and worse, in the way your licenses have been used?

RMS : I don't know what to say, the question is so broad that it brings nothing to mind.

RL : is there any particular clause or wording that has given the community problems, that needs a serious tweak?

RMS : I'm thinking about changing the conditions for distributing binaries without sources, to say it is ok to distribute the sources on the internet. Right now the requirement is to make them available by mail order. Back in the 80s, mail order was more convenient for most people, but now the Internet is more convenient for most people. Maybe I will add wording to clarify the situation regarding dynamic linking.

RL : As I understand it, and let me know if I'm off, the GPL requires you to provide your patches if you publish a derived work the LGPL has the same requirement, but if you dynamically link (this would be good for libraries) your own linked source is not considered to be combined or derived?

RMS : I wouldn I wouldn't use those words to describe the situation, because they aren't precise enough. But it is more or less accurate, for both the GPL and the LGPL, as you've said it. The LGPL was designed to permit linking with non-free software so it makes sense to permit various kinds of linking. The GPL was designed to forbid linking with non-free software so it is essential that this not depend on details such as the method of linking.

RL : so, if you merely link dynamically with the GPL, a modified wording might serve to make it clear that you must still publish the source of what you link?

RMS : By the way, it is an important strategic question whether to use the GPL or the LGPL for a library. Sometimes it is better to use the GPL for a library so that the library is available only for free programs.

RL : what are the circumstances in which using the GPL, versus the LGPL, make the most sense?

RMS : I don't think that we should be aiming to make any one program as popular or as useful as possible. 'Think globally, act locally.' I think the goal is to make the universe of free software as capable as possible, for the long term. For a given library, maybe the free software movement as a whole will benefit from the advantage of being able to use library L while proprietary software developers are not allowed to use it. If the library does something that lots of proprietary packages are available to do, then we won't get much advantage from limiting its use to free software, so we may as well use the LGPL.

RL : so, we might need to accomodate unfree programs in the short term, but the aim is to make the free world fill those needs?

RMS : Absolutely. The GNU project's goal is to replace proprietary software, not to enhance it.

RL : I have an audience question....one of our viewers wants to know, why does it seem so hard to get a truly free word processor?

RMS : Now that we have 100% free GNU/Linux systems on our machines, we're never going to install any non-free software on them, except perhaps if installing a certain program will directly help us to test a free replacement for that very program. As for word processors, we're working on extending Emacs into being able to do WYSIWYG word proc. The GNU system should provide a word processor, but it should not make users choose between that and all the power of Emacs. The only really acceptable solution is to provide them both together. If anyone wants to help work on this, please send me mail after.

RL : in the specific case, what is your feeling about the Corel announcement that they will be porting their proprietary products to GNU/Linux? is this a good thing, in terms of public relations? or does it encourage people to use proprietary software in a non-useful way?

RMS : A proprietary program is never part of our community. It might have indirect effects that are good for free software, but basically it is designed to separate people from their freedom.

RL : is it possible that organizations like Corel and Netscape will coopt and marginalize people via this process?

RMS : If these programs are useful, we need to write free replacements for them. I don't think that a proprietary application 'co-opts' people directly. But if people start considering a proprietary application a good thing because it might encourage more use of GNU/Linux systems, that is a problem.

RL : so, maybe we should just use these opportunities to take stock of what free programs the community needs?

RMS : It is important not to focus on the operating system and ignore the rest of what we want to do with software. If there's something we want to do with software, we shouldn't settle for anything less than a free prorgam to do it with, We needed a free operating system *first*, because you can't do anything without an operating system. But now that we have the free OS, we should be looking to make free software for other jobs too.

RL : on the subject of free software versus proprietary software, license problems,

RMS : To encourage use of a free OS by encouraging use of non-free other programs is a narrow vision. If some additional people use GNU/Linux because a Corel program is available on it, I won't say 'Go away' to them, but I also won't say that the Corel proprietary program is a good thing

RL : there's been a lot of recent discussion about the whole business with Qt and KDE....I had wanted to ask you, if you could, to talk about that whole business

RMS : just because it encouraged someone to use GNU/Linux. Qt is a library which appears to be technically useful, but is not free software. Like all the other non-free programs out there, it's outside of the free software community. KDE is a free program which was developed so that it needs Qt in order to run. This has a paradoxical result: although KDE is free software, it is useless as an addition to a free operating system, because there is no way to run it on a free operating system. In order to run KDE, you need to add Qt, which means that the operating system is no longer entirely free. The KDE developers thought that they would 'get the job done faster' if they used Qt--but the real result is that the wrong job got done. It's like saying, 'we can build this segment of track faster if we don't

RL : okay, for some of our participants who are maybe not clear on the problem with the Qt license....I don't want to take a lot of questions on this, but could you tell us what, in your opinion, 'breaks' the Qt license as free software?

RMS : make it connect with the other segment'. Maybe so, but if it doesn't connect, you don't have a railroad that works. I would have to look at the Qt license again; it has been several months since I looked, and I don't remember.

RL : okay, let's see....

RMS : I think it was either that it is limited to noncommercial distrbution only or that distribution of modified versions is not allowed, or both. Either kind of restriction makes a program non-free; I just don't remember for certain which of these restrictions Qt has. By the way, the Qt license says it gives permission to link Qt with GPL-covered programs; however, if you actually do that, you would violate the GPL. If the authors of a GPL-covered program want to give permission for linking it with Qt, they can do so.

RL : so regardless of the flaws you see in it, it just simply doesn't work connecting it to the GPL

RMS : It's clear that the authors of KDE mean to permit this, for example, so it is permitted *for KDE*. But if you wanted to link Emacs with Qt you would have to ask permission from the FSF, and I can tell you that the answer would be no. The purpose of releasing Emacs under the GPL is so that it can't be mixed with non-free software. Extended versions of Emacs must be free software.

RL : while we're on the subject of new and different licenses, what is your sense of what's going on with Netscape....the MPL and the NPL....are they truly 'free software' licenses? are they a good thing?

RMS : The NPL is a free software license, but it has two drawbacks. One is that it looks like a copyleft, but the copyleft is so easy to bypass that for practical purposes it may as well not be there. The other is that the NPL and the GPL are incompatible; it's impossible to link NPL-covered code and GPL-covered code without violating one or the other of the two licenses. The reason that the NPL's copyleft is toothless is that it permits linking with proprietary object files. So if Greedy Bastards Inc. wants to add some proprietary extensions, all they have to do is put all the changes into subroutines, put those subroutines into a separate .o file, and puu calls to those subroutines into the existing Netscape files. Then they will have to release the subroutine calls as free software but that won't do much good without the functions they call.

RL : what about the MPL? and is it a good thing to have a license with a clause that includes a license for patents?

RMS : What I said applies to the MPL as well as the NPL. The differences are not very large. the patent clause in the NPL seems like a good idea, as far as it goes. Maybe I'll be able to learn something from that regarding GPL version 3.

RL : I have a question from an audience member....any words for the MacOS people out there writing GPL'd code?

RMS : There's no longer a boycott of Apple. But MacOS is a proprietary OS. If you are going to write programs for MacOS (or any proprietary OS), then it's good to make it free software, and good to use the GPL. But you can't have freedom using a proprietary OS. If you want to be free, you need to switch to a free system. GNU/Linux or ...BSD.

RL : rms: do you think of the free software movement as akin in a way to the civil rights movement in the US? is it important in the same way?

RMS : Segregation was a horrible infliction on millions of Americans. I can't say that proprietary software is as bad as that. But it may become something like the war on drugs: fear inflicted on everyone, and many people in jail, for sharing with their neighbors rather than bowing to the Owners. I hope not. But the War on Sharing is being launched now in Congress.

RL : how so?

RMS : Please look at www.dfc.org, and phone someone in the House Commerce Committee on Monday. there is just a little time to make the bill a little less harsh.

RL : I've been asked a question about the 'Digital Millenium Copyright Act'....is this related?

RMS : That is what the Senate called their version of the bill. It has already passed the Senate.

RL : what is the most damaging provision of the bill?

RMS : It lets the owners effectively write their own copyright law. Whatever restrictions they can implement through encryption or passwords have the force of law, in that bypassingthem is illegal. And products that are useful for bypassing them can be forbidden too. It is hard to be sure what would happen eventually,

RL : that sounds as if it pushes the concept of copyright way past the point of social utility....?

RMS : but even debuggers might be illegal. The legislators hardly even give lip service to the copyrightt bargain any more. I went with a group to plead with Barney Frank against new criminal penalties in this bill. (An alternative bill, Boucher-Campbell, does not have criminal penalties.) He said he was determined to have new criminal penalties, because 'The software industry is worried, the movie industry is worried, the music industry is worried... I asked, 'But is it in the public interest?' He said, 'Why are you talking about the public interest? These creative people don't have to give up their rights for the public interest!' HE is Barney Frank, Democrat from Massachusetts. Note that he has identified Gates and Eisner with 'creative people' (artists and musicians, I guess he means)

RL : but it sounds as if the creative people are really shut out anyway....I had always thought the first person to benefit from intellectual property was supposed to be its creator, but it seems

RMS : and he has turned the Constitution's idea of copyright to benefit the public on its head.

RL : the people who end up with the rights to most of this work (in the case of copyrights, well past the death of the artist) are bureacratic corps

RMS : Copyright-owning interests like to spread the idea that the purpose of copyright is to benefit the copyright owner. But that's backwards. According to the US Constitution, the purpose is to promote progress. From the public's point of view, the benefit that the copyright owner gets is a price to be paid for the increased progress.

RL : but to do that, don't you have to give people an incentive? does most of the progress happen after the copyright holder loses control?

RMS : Exactly. Copyright was supposed to promote progress by providing an incentive to writers. For the public, the incentive is a price to be paid. In the age of the printing press, the price was just money. And I think that the progress that resulted was worth that price. But now the price is our freedom. Also, the free software movement has proved that we don't 'have to' provide that kind of incentive to get lots of good software.

RL : at one point, didn't you write a fictionalization of

RMS : The software privateers always say that unless we give them power

RL : what might happen when copyright law got out of hand? is that still available on the net?

RMS : over us, we can't have software to use. They want us to feel that we can 't live without them and so we must give them whatever power they demand. But we don't need them. When they say, 'Give up your freedom or this company won't write any software for you,' my response is, 'Ok, don't.' The story is in www.gnu.org/philosophy/right-to-read.html.

RL : thanks very much, Richard Stallman

RMS : I have to go now to get on a train to go to a party. I don't want to work ALL the time!

RL : apologies to all who have asked questions we haven't gotten to, and thanks to everyone for coming by! and I've been getting a lot of messages thanking you for your work, and I think I should probably just pass them to you 'en masse' here :) oops, lost that :) thank you everyone....I'm going to throw the channel open