News 

In the GNU Manifesto, you state that you resigned from MIT because you 'cannot in good conscience sign a nondisclosure agreement or a software license agreement,' and that MIT 'had gone too far: I could not remain in an institution where such things are done for me against my will'. Do you think such institutions have changed 20 years on, and with the GNU project now in existence? Would you have made the same decision were you working at MIT today?

Things have changed, and today an MIT staff person could perhaps get MIT to agree to use the GNU GPL for releasing such software. It might be somewhat more difficult for me as the author of the GPL to be an MIT employee and use it, since MIT might have made me negotiate with MIT lawyers about what to put in the GNU GPL. Overall it is still better that I left.

On the other hand, MIT no longer employs operating system developers very much, and has not since the 80s. I cannot assume that MIT would have wanted to continue paying me all these years while letting me spend my time developing the GNU system and promoting free software ideals. I might have had to leave willing or not.

How does the GNU system, in its current state, compare with your original vision for the system?

The GNU/Linux system, which is GNU with Linux as the kernel, has gone far beyond the specific technical vision I had 20 years ago. At the same time, it falls short in one important respect: the widespread inclusion of non-free software in versions of GNU/Linux means that installing a GNU/Linux system does not reliably give you freedom.

The point of the GNU system was to be entirely free, so that the users would have freedom. If you include even one non-free program in the system, it is no longer a free system, and it does not achieve the goal.

A system including just one non-free program may come close to the goal, in a practical sense. How far it falls short depends on how painless it is, practically speaking, to remove that program. But inclusion of any non-free program always conveys the message that non-free software is acceptable, and that affects our community's further development. Whoever accepts one non-free program usually sees little reason to object to the second, the third, or the fifty-third. Thus, we face a tendency to add many non-free programs, often in important system roles that other programs depend on. In
particular, any time a law blocks free software from supporting a desired feature, someone is likely to offer a non-free implementation as a 'solution' to the problem.

People who are accustomed to seeing non-free software in the system tend to construct philosophies that legitimize it. We face the danger that the GNU/Linux system will degrade into a hybrid of free and non-free software, whose free portions will be of little use on their own. The goal of using a computer while living in freedom may be completely lost.

What has been the most significant application developed for GNU?

I cannot answer because that is not the way I think about the field. I focus on the obstacles we face, rather than on ranking the problems we have solved.

And the highs and lows of the last 20 years?

The adoption of the DMCA and the Mickey Mouse Copyright Act were some of the lows. The popularity of GCC in the 80s, the availability of completely free operating systems in the 90s, and increasing political support for free software in Brazil today, are some of the highs.

What are the most urgent development needs required to drive the GNU system further right now?

One vital need is for better free manuals (a free operating system should include complete free user documentation). Today users tend to use non-free manuals. A couple of publishers do publish and sell free manuals, but this practice needs to spread.

In software itself, we need a free complete Java platform, we need free software for speech recognition for dictation, and we need free 3D drivers for the common graphics chips (binary-only drivers are available gratis, but that is not free software and not adequate for living in freedom).

We need free software to play RealAudio and RealVideo and Quicktime and Windows Media Player files; these will be illegal in countries whose governments have caved in to the media companies, but they can be developed and published in countries that have not.

However, in the long term our main needs are not for specific technical developments. Those will come, sooner or later. What we need most is to block, and repeal, the laws and regulations and
measures that prohibit

ADVERTISEMENT

development of free software.

Right now the FCC in the US is considering a proposal to prohibit free software from receiving digital video transmissions. See www.publicknowledge.org. That would be the first ever specific prohibition on free software for a specific job. But there are already laws which have that effect.

The US already has the DMCA and software patents. The EU already has the EUCD, which has an effect similar to the DMCA. Additional laws, even worse, are now being proposed.

And how do you think the actions of SCO and the campaigns of Microsoft will impact on the adoption of GNU/Linux systems?

In the long term, I think SCO will have little effect. It seems unlikely they have any valid claims against Linux, the kernel of the GNU/Linux system, but even if they did have some, our community would persevere and solve the problem.

What's your vision for the GNU system 20 years from now?

My vision for 20 years from now is that all published software, aside from what is embedded in ROM in appliances that can't load new software or talk to any other devices, is free software. No one should ever be told, 'You can have this attractive software, but you cannot see what it does, change it, or redistribute copies to others.' The freedom to do these things is a human right that must never be denied.

How dangerous a threat is the abuse of notions of ownership, patents and copyrights to development for the software industry?

Copyright and patents are two very different laws; they have little in common except that both are conceived, in the US and UK, as artificial encouragements for conduct that serves the public interest. Properly and thoughfully, they ought to be judged in terms of their effectiveness at achieving the public benefits that are sought, and in terms of whether the burden they place on society is acceptable.

The grantees of these monopolies would like society to forget those goals and criteria, and treat both copyrights and patents as burdens that we must bear no matter how great the cost.

One of their methods for encouraging forgetfulness is to describe government-issue incentives as 'property' and describe themselves as 'owners'. We should firmly reject this usage even if others do not. The notion of physical property is not a problem for software development. There is no problem with the idea that you own a computer, or that you own a disk that holds a copy of a program. However, treating ideas as 'property' can turn software development into a quagmire.

Software patents monopolize software ideas, so they make software development legally dangerous. A nontrivial program combines many ideas, perhaps hundreds of them. If each of these ideas could perhaps be patented by someone else already, making a new and useful combination could get you sued.

The European Union is now considering a directive that can either allow or reject software patents. The free software community took the lead in lobbying against them and, amazingly, we convinced the European Parliament to reject them. Now the battle has moved to the individual states of the EU. See www.ffii.org for more explanation of the danger of software patents, and how you can fight back.

Since copyright law and patent law have essentially nothing in common, it should be no surprise that the situation for copyright is completely different. Copyright as traditionally understood does not impede the development of new software.

However, the recent EU Copyright Directive creates new restrictions on developing any means to bypass limitations intentionally imposed by a program in accessing data. That prohibits important free programs such as DeCSS. The proposed 'Intellectual property enforcement directive' would make things even worse. For more information about this threat, see http://www.cl.cam.ac.uk/~rja14/draftdir.html,
and contact your members of the European Parliament immediately.

www.gnu.org/philosophy/words-to-avoid.html explains why the term 'intellectual property' is biased and confusing, and should never be used.

Do you think the issues we've seen with patenting 'one-click shopping' and the Eolas '906' patent dispute, for example, will be an eternal presence in the industry, or can you forsee a time when laws will have to be changed to be more appropriate?

This is the wrong question to ask. If we believe in democracy, the right question is, how can we put an end to software patents? To treat this as a question to be forecast rather than a battle to be won
is to abandon democracy even in principle, and become loyal subjects of a corporate sovereign.

More information on GNU and the Free Software Foundation is at www.gnu.org. Richard Stallman's home page is at www.stallman.org.