Eben Moglen's linux.conf.au keynote

The following is a poor attempt to summarize the talk.

The "legal state of the free world" is strong. In particular, attacks on the General Public License have abated. One year ago, the SCO group was claiming that the GPL was invalid and in violation of the U.S. constitution. That kind of talk is not happening any more. SCO "has not completely flatlined," but it is almost there.

What were the legal consequences of the SCO attack? Certainly the invalidation of the GPL was not one of them. There were two outcomes, one positive, and one less so.

On the positive side, the industry (as composed of large vendors who make money from free software) has decided that the community needs better lawyers. In particular, the industry has concluded that financing good legal advice for the community early in the game, before problems develop, is a good investment. The result was the creation of the Software Freedom Law Center, with almost $5 million in funding. That figure can be expected to triple in the near future. There should be, soon, abundant legal help available for nonprofit organizations and developers working in the free software area.

In this sense, the dotcom bust was a fortuitous event as well. As technology jobs went away, numerous technical people found their way into law school. Many of them were not too happy about it, but these were the students Eben had been waiting for the last fifteen years. Soon, there will be a new crop of lawyers who understand technology and who can read code - and they will be funded to work for the community. This is a very good outcome, and we owe thanks to Darl McBride for helping to bring it about.

The other outcome from the SCO attack is the general realization, in the boardrooms of companies threatened by free software, that copyright attacks are of limited value. SCO and its backers brought a heavily funded attack against a project set up fifteen years ago by a student in Helsinki who didn't think he had any need for lawyers - and that project sustained the attack easily. Copyright does not appear, any more, to be a legal tool which can be used to impede the spread of free software.

Patent attacks are a different matter, and "we are going to face serious challenges" in that area. There will probably not be much in the way of patent infringement suits against individual developers; those developers simply do not have the deep pockets which might attract such a suit. Instead, the attacks will come in the form of threats to users.

This is happening now: corporate officers will get a visit from "the monopoly" or others and be told about the sort of trouble waiting for it as a result of its use of patent-infringing free software. That trouble can be avoided by quietly paying royalties to the patent holder. This is happening "more than we would believe" currently - companies are paying royalties for their use of free software. It remains quiet because it is in nobody's interest to make this sort of shakedown public. The victims will not come forward; they will not even tell their suppliers.

Defending against patents is a complicated task. An important part is destroying patents - getting the (U.S, mainly) patent office to reevaluate and (hopefully) invalidate a threatening patent. This is what was done with Microsoft's FAT patent, for example. When it works, it is by far the most cost-effective way of dealing with patent problems; it is far cheaper than trying to litigate a patent case later on.

This process is tricky. Typically, a group wishing to invalidate a patent gets a single shot, in the form of its initial request to the patent office. After that, the process becomes confidential, and involves communications with the patent holder. So that first shot has to be a very good one. They are getting better at it.

Killing patents makes people in the industry nervous - they have their arsenal of patents too, after all. There is, however, an "agonizing reappraisal" of the patent system going on within the industry. Some companies in the technology industry are starting to get a sense that the patent system does not work in their favor. It will be interesting to see what happens within IBM, in particular. In general, patent reform is going to be a big issue over the next couple of years. Some parts of industry will favor reform, others (such as the pharmaceutical industry) are happy with the system as it stands now. There will be groups trying to redirect the reform process to favor their own interests, and many "false friends" appearing out of the woodwork. There will be opportunities for serious reform, but the community will have to step carefully.

Meanwhile, Samba 4, in particular, may not be safe; there are likely to be patents out there. "Expect trouble."

[In a separate session, Eben encouraged free software developers to record their novel inventions and to obtain patents on the best of them. Free legal help can be made available to obtain patents on the best ideas. Until the rules of the game can be changed, we must play the game, and having the right patents available may make all the difference in defending against an attack.]

Back to the GPL: the work done by Harald Welte getting the German courts to recognize and enforce the GPL has been a very good thing. Eben, however, is also pleased by the fact that, over the last decade or so, he has not had to take the GPL to court. Threats to enforce the GPL are entirely credible - there are few volunteers to be the first defendant in a GPL infringement suit in the U.S. It also helps that the Free Software Foundation, in enforcing the GPL, seeks neither money nor publicity. Instead, what they want is compliance with the license. "I get compliance every single time."

Enforcement against embedded manufacturers ("appliances") has been problematic in the past. These manufacturers have less motivation to comply with the GPL, and the costs of compliance (especially after a product has been released) are higher. The working strategy in this case recognizes that the company actually guilty of the infringement (usually a relatively anonymous manufacturer in the far east) is highly receptive to pressure from its real customers: the companies who put their nameplates on the hardware and sell it to the end users. If you go to a company with a big brand and get that company to pressure the initial supplier, that supplier will listen.

Meanwhile, the appliance manufacturers have started to figure out that posting their source is not just something they have to do to comply with the GPL - it can be good business in its own right. When the source is out there, their customers will do some of their quality assurance and product improvement work for them - and remain happier customers.

In summary, the problems with GPL compliance by appliance manufacturers will go away in the near future.

There is not much to be said, at this point, about what will be in version 3 of the GPL. Much, however, can be said about the process. The GPL currently serves four different, and sometimes conflicting goals. Any attempt to update the GPL must preserve its ability to serve all of those goals. The components of the GPL are:

• A worldwide copyright license. Worldwide licenses are exceedingly rare; they are typically tuned to each legal system in which they operate. The GPL cannot be issued in various national versions, however; it must work everywhere.

• A code of industry conduct - how players in the free software world will interact with each other. Any new code of conduct must be negotiated with the industry; it cannot just be imposed by fiat.

• The GPL is a political document; it forms, in a sense, the constitution of the free software movement.

• It is the codification of the thought of Richard Stallman, and must continue to adhere to his beliefs.

Updating the GPL will be a long process. Eben will be putting together an international gathering of copyright lawyers to help with the crafting of the copyright license portion of the GPL. A separate gathering of industry representatives will be needed to hammer out the necessary compromises on the code of conduct; this is a part of the process which may not sit well with Richard Stallman, but it must happen anyway. The constitutional part of the GPL, instead, should see minimal changes - there has been no fundamental change in the wider world to motivate the creation of a new constitution. On the last point, there will be no revision of the GPL which does not meet with the approval of Richard Stallman and the Free Software Foundation.

When a new license nears readiness, it will be posted with a long explanation of why each decision was made. Then will come the comment period, as the FSF tries to build a consensus around the new license. The revision of the GPL is, perhaps, the most difficult task Eben has ever taken on, and he is not sure that he is up to it. The job must be done, however.

As for when: "soon." He did not want to undertake revisions of the GPL while it was under attack - updating the GPL should not be seen as a defensive maneuver. Now, however, the GPL is not under attack, and "the monopoly" is distracted for the next couple of years trying to get its next big software release out. This is the time to get the work done, so something is going to happen.

In response to a question about software-controlled radios: that is a global problem, not just limited to the United States. Japan, it seems, is the worst jurisdiction in this regard; there have been threats to arrest foreign software radio developers should they set foot there. Fixing the software radio problem is a key part of ensuring freedom of communication in the future, and it is currently Eben's most pressing problem. There has been little progress so far, however, and new strategies will be required.

In general, freedom is under threat worldwide. The events since 9/11, in particular, have accelerated trends toward a repressive, surveillance-oriented world. If we want to ensure our political freedoms in this environment, we must work for technological freedom. Without the ability to control our own systems, to communicate freely in privacy, and to interact with others, we will not have the wider freedoms we hope for. The free software movement is the heir to the free-speech movements which started in Europe centuries ago; we are at the forefront of what has been a very long and difficult fight for freedom. The difference is that "this time we win."

Standing ovations for speakers at Linux conferences are a rare thing; Eben Moglen received two of them.