Keynote about GPL3 at HOSC 2006

From Wikisource

Keynote about GPL3
by Eben Moglen

Keynote given at the Holland Open Software Conference 2006 in Amsterdam with questions from the audience, June 17, 2006

[Eben Moglen:]

Thank you very much, it’s a great pleasure to be here. You know when you’re invited to give a keynote speech on the last morning, that actually it’s less a keynote than a kind of cleaning up, you know.

I stand here in part, I think, to give a required class report. You are, after all, stakeholders in something I am doing on your behalf, and one part of my being here, I take it, is to offer the steward’s account of the state of the property.

That’s a little difficult for me to do, so I’m going to try to do it discreetly and without making too much news. I feel sure there will be more asked for in the questions than will be presented in my address.

But beyond summarizing where we are in the process of making GPL3, I think I can do a better job for you in my role as reporting to the citizens of the free republic if I offer a little bit of a sense about what I think follows the GPL3 process as we are now going through it. That is to say, at this stage, as we get towards half way, maybe it would be good to take a look at why we are doing this, again.

So, let me then begin with the status of matters as I understand them. On the 16th of January, the Foundation, following rules it had preannounced by a couple of months before that, released a first discussion draft of GPL in a global conference at MIT.

We had several hundred people in the audience from every part of the free republic; commercial, non‐commercial, and every major country, language, and participant group in our development community represented one way or another there.

We released a license draft which addressed all of the issues that we had said we thought a license updated for the current world needed to address, and we addressed those issues with some degree of tentativeness.

I noticed at the time that there was a little bit of a disagreement between the public statements of Mr. Stallman and me with respect to how tentative we were in some cases, but modulo a little bit of personal disagreement, which I think is thrilling and helpful, we were, I believe, all pretty clear about what we thought we were doing; we were releasing a first discussion draft.

We also published some technology for commenting on the draft, including some stuff we’d had to make for ourselves ’cause we couldn’t find any free software in the world that did exactly what we wanted. I had sworn up and down we wouldn’t do that, but you know how it is, you get an itch and you start scratching it, and next thing you know, you have more free software in the world, and I like stet, the tool that we made. I was unhappy to be spending a lot of money getting it written by a free software programmer, but he did wonderful work, Orion Montoya deserves a lot of credit for the look, feel, and operation of the GPL3 comment process through stet; those of you who haven’t been to gplv3.fsf.org might want to go.

The way in which we use highlighting in a document to represent the comments, the intensity of the commenting, to allow people to navigate the commenting on a single document I actually think is a useful collaboration tool for 21st century public deliberation. We noticed that some people are beginning to pick up the software and make other use of it.

Through that stet system, more than a thousand people around the world who are uninvolved in the formal comment process—just members of the citizenship of the free republic—came in and made comments about the license, some of them extremely articulate and thoughtful, some of them a little less so, but all of them carefully tied to a piece of the language by somebody who said “Here, this, I think you should do something about this.”, and it was in that sense a very well structured and helpful form of engagement. We read every single one of those comments, we thought about every one of them in the course of the process, through the documented process we said we were gonna have, as well as all of the participation by the members of discussion committees; people who came to Cambridge, and because they came to Cambridge were invited to participate in the conversation at a more intensive level if they cared to make the time, or people who were invited by those who were in Cambridge because they thought they particularly ought to be there.

We are still tinkering, of course, with membership in the discussion committees, adding people who want to be intensively committed to the process, and mixing up how those people engage with the comments of the public, for the advising of the Free Software Foundation.

We have, after many weeks of consultation, argument, writing of positions in regard to issues on all sides in all the committees, come to the point at which the second discussion draft of the license can be worked on and promulgated. Mr. Stallman and I had a lot of time in New York with other members of the Software Freedom Law Center and the Free Software Foundation to discuss proposals, issue resolutions, and pieces of the comment workflow, and we have largely completed a second discussion draft.

We are now considering the particularly difficult problems of how to position the draft with respect to ongoing discussions in the comment process around the world, so we’ll spend a few more weeks talking to people in and out of the discussion committees, making some decisions about what to do at this stage in the process, and on the 25th of July we’ll release the second discussion draft of the license and begin another round of accepting comment.

I regard the redrafting of the license at this point as probably about 90% complete; I think that the 10% of the license which is most complex to revise at the present moment in the process may change a little more again before release, but I think we know pretty much what we need to do in order to be fully responsive to the first round of issues and comments.

Let me say that again: fully responsive to the first round of issues and comments.

Now, one of the things that I would say about the way in which the comment process has gone so far, and I won’t say much, but one of the things I would say about it is that it very rapidly demonstrated a surprising fact:

GPL version 2 was perfect.

I know that this is true because everybody told us that everything we wanted to do to make a change was a bad idea. It surprised me not all that much, because I know that human beings are reluctant to suffer change. And in that sense it was not surprising that any social process for accepting public comment began by locating the weak points in everything that was new. This is in fact good, this is the way enlightened rationality ought to work, but one of the byproducts was that it created an odd rhetorical situation in which we, who thought the license was pretty good, and who had been using it for fifteen years and believing it was pretty good and would stand another day or two of work, became the proponents of the idea that it really needed to be altered, whereas all the people who had complained for fifteen years about everything they thought was wrong with it had suddenly become the proponents of leaving it alone and why did we have to make a change?

As the person responsible for mediating this conversation I have occasionally derived a little bit of amusement about that. Sometimes I have even reminded people how interesting it is that they have suddenly become partisans of the strongest possible copyleft, “everything shall be done under this license, meaning this license only, never any changes again”, and how odd that is, in view of the license’s constant effort to be looking at the far horizon of the business of making and distributing software. but “the perfect GPL2 theory”, as I call it, has indeed been very strong all through the spring.

It is beginning to ease now, as people begin to realize that, one, It will be possible to live under the rules in GPL3 with the same degree of profitability and simplicity or non‐simplicity that it was to live under the rules of GPL2, and once stability in the new regime has been achieved, that regime will be just as stable and just as robust as the regime under GPL2, but, a lesson of great importance nonetheless is being learned by this:

The copyleft, far from being in any sense an anti‐commercial conception, has become so stable as a basis for business models worth so much money, that the agitation from those people motivated by commercial rationality is not in the direction of weakening the copyleft at all. It is very important to recognize that. If there is any element of the way GPL has worked over the last fifteen years which has proven to be of great value to the commercial parties, it is the license’s absolute, bright, simplicity as a license that requires copyleft.

Because everybody understands what the future of all investments made to code under the license will be. Nobody is required to conduct contingent analysis, “If we invest a million dollars in this program now, and we gain a certain result for a certain length of time, and somebody else comes along and makes a proprietary copy, what will happen to our business?” “If nobody makes a proprietary copy what will happen to our business?” And pretty soon you’ll have a sort of bushy analysis about what happens if you make an investment in a non‐copylefted program.

If you make an investment of time, effort, or the contribution of patent claims to a GPLed program, you know exactly what the future of your investment is going to be, it’s going to remain under GPL, there’s going to be a certain form of competitive market in the use of that software to enable the provision of services, but there will not be a proprietary alternative, of any kind, to complicate your analysis of your return on investment.

That so simplifies the return on investment calculation that businesses have discovered for themselves the enormous value in allocating their research and development expenditures under GPL. That’s a crucial fact in the transformation of the software business, whose crucial irreversibility was ratified by an event occuring since the last time I spoke to you all.

But we’ll return to the retirement of Mr. Gates in a moment. That is, after all, the small news. That’s about what’s over. We’re in the middle of talking about what’s not over, yet.

So, what I think is most interesting, then, about the second discussion draft is that it has to respond, in some deep way, to the pro‐copyleft position of business, while also responding to the equally pro‐copyleft but differently pro‐copyleft position of most of the non‐commercial hacker world.

Not surprisingly, though deeply committed to the copyleft on both sides—they are, after all, members of the free software movement if they are heavy users and reliers upon GPL whatever their politics may be in other respects—though deeply committed to importance of the copyleft, the commercial and non‐commercial components of our community still do fundamentally perceive different issues as crucial in the development of the license. It’s not surprising that the questions of patent rights are seen primarily defensively in one place, and with a certain degree of investment‐backed expectation in another place.

It’s not surprising that the question of how Trusted Computing might become Treacherous Computing, or how Digital Rights Management might become Digital Restrictions Management, are differently seen by people making devices and by people purchasing devices. Consumers, as they are known, are actually afraid that their ability to produce will be harmed. Producers, as they are known, are actually concerned that their ability to consume free software might be damaged by the development of rules favorable to their customers that are perceived as inimical to the interests of third parties important to their business.

The conversation that we are having, in other words, is a subtle and complicated one. It involves many thousands of people around the world, all of whom have equal stakes in the process, and need to be treated with equal respect and dignity. The difficulty in doing that is that some of them are individual developers in Kazakhstan and Bulgaria, and some of them are multinational corporations with hundreds of billions of dollars in assets and credit at their disposal, an enormous corps of trained people who are accustomed to participating in every public policy debate that affects the political or economic welfare of their business.

In order to create a process that holds anything like an equal playing field among those parties, requires a kind of social engineering which is infrequent, let me put it just that way, infrequent, in the world we now live in. It would not be right to think of our doing all that work trying to create such a process without thinking of that process itself as part of our objective. One does not do something like that only as a means to an end, one does something like that in part as an end in itself.

When we are done with this process we will have made fundamental rules for the allocation of billions of dollars in assets, the distribution of enormous gains from trade, we will have set rules for the conduct of a global industry, and we will have done so in an egalitarian, open, participatory fashion, in which the opinions of people were weighed and reasons were given for decisionmaking across a broader diversity of power and wealth than is characteristic of the way legislation is done by governments or states.

That’s a lesson to the 21st century and to the software industry. Who else offers software on terms that have been so thoroughly negotiated to consensus between all the parties whose interests are relevant to the exchange? Who else does business on those terms?

We will, when we are done, be able to ask that question of the world, with some justice, because we will be doing business on terms that have the legitimacy of enormous broad public consent, sought through egalitarian participatory processes. That’s a standard that other business addressing the same people for the same purposes ought to make.

So, we are, in my judgment, then, at the halfway point in a process which is about the process as well as the license. We are gaining an understanding of what the consensus for the copyleft is, and how the terms internal to the discussion need to be worked out.

There are, I think, when the second draft for discussion is released, going to be a much smaller number of issues than there were with the first draft; we will have closed a lot. That’s because a lot was listened to that people said.

I have heard much criticism, almost all of it from outside the process, and some from inside, based on the high proposition that the Free Software Foundation was not going to listen.

That the Free Software Foundation consisted entirely of some headstrong people, I assume they mean me rather than Richard, but at any rate, some very selfish and arrogant person, determined to run the whole situation solely for personal convenience in the expression of personal preferences, and so on and so forth. This mystifies me very much. If it were true it would be so much easier to have done on the 15th of January, by simply releasing the first discussion draft of the license as the license.

It’s perfectly obvious to those of us who sat through the two years of discussions about how to do it that preceded the beginning of our doing it that how to do it was regarded as a matter that ought to be governed by principle, just as what to do is a matter that ought to be governed by principle. Principle decision having been made about how to do it, you would think that observers of the Free Software Foundation would have come to the conclusion that having concluded how to do a thing by principle, that principle would be adhered to. Adherence to principle is, if anything, the sole distinguishing characteristic of the Free Software Foundation.

So my client has behaved as my client always behaves, by figuring out what the principle is, saying what the principle is, and living up the principle as best human nature and physical resistance to oppression can possibly accomplish.

And we are where we’d said we’d be. We have heard what everybody has to say. The second draft of the license responds to what people have said, we hope, in thoughtful ways. Where it is necessary to make a decision, of course, a decision gets made, something else gets put in the draft, and then we talk about it again.

I believe that the primary issues remaining for resolution over the course of the next several months will be in the area of patents and in the subject of the Digital Rights/Restrictions Management that we were briefly ventilating on Thursday.

I think, by and large, the internationalization of the license, which I regard as a very important part of the license’s work, has passed essentially unreviewed by people who should have looked, but is now in a state which is infinitely stronger than it was in GPL2. I have seen the last of the North American particularisms out of the license, i believe, in the work of all the lawyers around the world who have so graciously and energetically participated in helping us to think in new ways about how to express fundamental copyright principles in denationalized words.

Mind you, this is not the same thing as conforming to any individual country’s view of how copyright law formally works. This is a different enterprise, an attempt to generalize using language that is not owned by anybody, for the purpose of writing in a language which could be recognized by any legal system, not for domestic law talk, but for a reasonable expression of a licensor’s intent with respect to the legal principles as they are understood in that system. We should be able, in that way, to have a single license working around the world far better than the single license worked around the world in its last version.

I hope that there will be a little more attention to the internationalization aspects in the second round, because I think the more eyeballs on the work, the better the chances are that the work is well done.

The enhanced compatibility provision, as it was known in our first draft—the additional terms provisions for permissions and requirements in the license—will be, I think, extensively rewritten. Its complexity daunted the readers around the world, as it should have done; first drafts are first drafts, we said that.

It is clear now that there are other ways of organizing those ideas which will communicate better, and I believe that the next draft will see a much easier sledding time for all the people who want to help get it right. As I pointed out on Thursday, this question of compatibility among licenses is not merely a matter of adjusting GPL at the edges to get rid of particular problems with particular licenses, it is an attempt to address the problem of proliferation, which everybody agrees is a difficult and important problem, and which, like the weather, pretty much everybody has been doing nothing real about.

We’re going to make a real proposal. A real proposal, that we think, after thinking about it very carefully for a very long time, works to achieve many of the purposes we have to have with a minimum of inconvenience and difficulty for everybody. We recognize that there are tradeoffs involved in making policy in the area, and this is a consensual activity. But we have a proposal to make, and I think it is a proposal much improved over the proposal in January. I do hope, therefore, that there will be more attention paid to that, as well.

Nonetheless, I have said what I think the central issues are, and we will see what progress we make over the next several months. There will be a third and last call draft in mid‐autumn, I’m not yet in a position to announce a date, and there will be a final adoption of the license, I believe, in sometime around mid‐January of 2007.

At that point, the “GPL version 2 only” projects will decide for themselves what licensing decisions to make, this includes, of course, the kernel Linux, and we will see, very rapidly, the automatic reversioning under GPL2’s automatic reversioning clause of pretty much everything else under GPL3.

So much for the status report, I am sure you will have questions.

Now, let me step back just a little bit to talk again about what I think this is for.

In the first place, it is for making sure that users’ rights are protected. If you think about it, what the issues are for users of computers, and their freedom to modify and share software, have changed a little bit since 1991. The technology is a little different; let me give you just one tiny example: In the world of 1991, you could pretty much assume That an information flow, even over a computer network, involving software, was unidirectional. A is giving binaries to B. And B, therefore, must have a right to get some source code from A.

Now consider the extraordinary importance of distribution of ISOs—of CDs—containing binaries, by BitTorrent. A practical mode of distribution which is terribly important to the world and which none of us would like to see disturbed. But if you and I join a torrent for the distribution of an ISO containing binaries, though I begin the day without any of the ISO, during the course of the day I not only receive binaries of software, I give away binaries of software. It is no longer correct to think of that as a unilateral transaction occuring in one direction: “A gave a binary to B, A must give source code to B.”

It is actually a world in which A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P gave binaries to one another, but most of them were only acquiring the binary, they didn’t really have the whole thing at any given moment, and they certainly didn’t have the source code, either, yet, or maybe ever.

Now, if you think about that for just a second, you’ll say “Well, OK, we’ll take care of it.” Well, yeah, “we’ll take care of it”, you understand, we have to think about what does that mean and what should we do about it, and how should we do something about it.

We actually didn’t do anything about it in discussion draft one. Thanks to somebody in the world, and I want to point out that this was just a citizen of the free republic, not a paid lobbyist or lawyer for a multibillion dollar organization, not the leader of a major free software hacking project, not a lawyer practicing in the free software world, not Eben, not Richard, not any expert at all in formal terms, just an expert, just one of those real experts, that is to say, a human being thinking clearly and carefully about something of importance. So we got a comment, so we said “You know, that’s right, we’d better think about this.”

So we spent a lot of time generalizing that comment and thinking about it. So the license is better.

What’s it for? It’s for protecting users’ rights by harnessing users’ thinking. In other words, it’s like free software; protect users’ rights by harnessing users’ thinking, by harnessing the innovation that is in all the minds that are connected by all the work we all do together.

Of course, there is another way to make software.

You just hire some people. If there are cheaper people available you fire the people you have and you hire the cheaper people. You make the software. And when they’re finished with it you take it away from them.

In fact, to be sure that you’ve taken it away from them, you take it away from them before they’re finished with it.

That is to say, you let them write it, but you don’t let them build it. Because if you let them build it, they would have the whole thing, and then they might take it away from you.

So you make a big company in which you pay people as little as you can get away with, and when they finish writing the software, before they actually build it you take it away from them and then you build it, and give it back to them in binary form, including the part they made, but not with any build authority, and you ask them to make it a little better, and you do that for ten years and then you release it.

And nobody has it. And nobody can fix it. And in order to reengineer it in the field you have to make binary patches to it.

And all of this is necessary because your software is a product. You have to get paid for it. And next thing you know, you start putting stuff in it to phone home from every computer to make sure that the people who have it actually paid for it.

And so you’ve got to start opening holes in the network and communicating in ways between machines that aren’t supposed to know that communication is going on, and pretty soon you’re ripping holes in the fabric of security again, in order to protect your investment in the software you took away from the people who you took it away from early ’cause you didn’t give them the right to build it.

I mean, this is madness, of course, you would say “This will never make good software.”. You know what? It never does.

The person who made the world believe that that’s a way of making good software—for the first time, because IBM didn’t think that made good software in the mainframe era when I worked there, nobody thought that made good software back then—the person who made the world believe that that’s a way of making good software just quit.

Not only that, but he just quit immediately before the shipment of the product that was supposed to be the greatest release in ten years, and is four years late.

[Unknown:]

You think he quit because [Unintelligible] ?

[Eben Moglen:]

No, I think that he quit because spending the largest ill‐gotten fortune on the planet is a better thing to do than taking responsibility for what’s about to happen to Microsoft. I think that's obvious to everybody.

The model that I described of making software, by taking it prematurely away from people who are paid to make it and then lose control over it, and who are aware that they are paid to do that only until somebody cheaper can be paid to do that, that model of making software has failed.

That’s not really a surprise to those of us who have been making software for 30 or 40 years. It was a brief experiment in an utter proprietization of software, a total productization of computer science knowledge.

It wasn’t a good way to go anymore than productizing mathematics would be a good way to go. Imagine that somebody had turned arithmetic into property, such that before you started doing any job in the world—engineering anything—you first had to go to him and buy the arithmetic you would need to finish the job.

The quality of arithmetic would rapidly decline, and people would start cheating, and use as little arithmetic as possible. And pretty soon stuff would be badly built because people were cheating on the arithmetic that they used. Rather than refining the arithmetic they would be reducing the quantity of required arithmetic.

Engineering is presently troubled by enormous difficulties that result from a distortion in the knowledge structure, which is now ending.

Mr. Gates’ quitting is the moment at which that becomes physically real to most people in the world who do not know what you and I know. For them, something has happened which now requires an explanation.

We know what the explanation is, we could even foresee the coming of that explanation. They now need to know: “What does this mean?” I would suggest that we go tell them what it means. I would suggest that as citizens of the free republic we go and tell people what has just happened. Mr. Stallman is 54. Mr. Gates is 50. Mr. Gates has retired. Mr. Stallman has not.

We’re going to make better, by an order of magnitude, the free software system this year.

The software makes itself better by an order of magnitude all the time, you know that ’cause you are the ones who make it better by an order of magnitude all the time. It never slides backward, it never gets regressed, it can never go anywhere but up, because what’s good is always available to be inherited by anybody who needs it; That’s what the copyleft accomplished, that’s why, in the end, everybody else retires. Because in the end, you’re always competing with the best there is, for free.

That’s the way it should be. There’s nothing wrong with that. “The competition is healthy!” say the pro‐competition people. “The freeness is healthy!” say the freedom people; they’re both right. They’re both right.

Competition doesn’t cease when everything is free; I offer you KDE and GNOME. A competition which is entirely for the good, I have to say. That doesn’t mean that I want to use either one of them necessarily, I will use Fluxbox whatever they do. But it is good that they are competing; it is making the world a better place.

So, we’re going to have inventiveness, we’re going to have fierce competition, we’re going to have extraordinary consensus and participation in the making of the terms of our activity in the world, we’re going to keep the computers free, in the sense of hardware that responds to the people who quaintly used to be thought of as its owners.

We’re gonna accomplish all of that, and the other side is actually going to quit. It’s gonna be a velvet revolution all the way to the end, guys.

What happens to the opponent is it vaporizes. And everybody’s going to say: “Where did the Soviet Empire go?” And we’re going to say: “It went into history as an experiment. It wasn’t very respectful of users’ rights, and users’ rights are really important, and after a while, it went away.”

Thank you very much.

[edit] Questions & Answers

[Unknown:]

About the first part of your talk, you said that there was a lot of, response from the corporate community about changes in the GPL.

Did that surprise you, that there were so many corporations who were following to the GPL version 2 something that they didn’t think of 15 years ago when the GPL2 was created?

[Eben Moglen:]

Well, first I want to put on record, yet again, the fact that I went to work for Mr. Stallman in 1993. I don’t want any credit for GPL version 2. Not because it’s not a thing that anybody in his right mind would want credit for, but simply because I deserve no credit whatever, OK?

I came to the license as a lawyer after it had already been long in use, I accepted the cards as they were dealt and I thought they were the most brilliant hand I’d ever seen. I thought anybody could play that hand. I thought “It’s a shame that nobody better than me has turned up to play this hand as a lawyer, because these are the nicest cards on earth, and they’re going begging.” Of course there were some people who thought they didn’t want Mr. Stallman as a client. That was a fundamental misjudgment, in my view, but nonetheless.

So, with that in mind, I would say that I was not surprised by any of the businesses participating in the process, because I deal with those people all the time in my ordinary life as the lawyer responsible for enforcing GPL on behalf of the Free Software Foundation, and as the founder of the Software Freedom Law Center.

In fact, I should point out that those businesses are themselves major contributors to the Software Freedom Law Center; they are the reason it exists. It’s all the money that they are willing to put up so that hackers can have lawyers for free that made it possible to be doing this in the first place. So in fact, if there is anything unusual in this situation it is that I find myself negotiating with those organizations on behalf of my client the Free Software Foundation, while those organizations are themselves being asked to, and in many cases still routinely paying the bills for what they might, at this moment, call the other side’s lawyer.

We have had conversations over years together, those lawyers and I, about what might be in GPL3, we’ve talked about it formally and informally as we’ve talked about it with lots of people around the world. I know those lawyers very well, I deal with them all the time, I have a very good idea about how they think and about how their organizations think. That’s part of my job for the Free Software Foundation, is knowing how those organizations and lawyers think.

So, if anything, I would say that we have been very comfortable together, though the talk has sometimes been rather tense. We have heard things from one another that we don’t always hear from one another, and I have taken threats from old friends, and maybe even made one or two. But that’s just business. That’s just lawyers negotiating over what is, after all, a deal that those people consider just as important as you and I do. They may have different reasons for thinking it’s important, but the importance is as real to them and deserves to be as real to us.

So, it’s like all the other conversations that we have with one another on this subject; They should go in a pretty straightforward way: “What do you need? What do you want? What are the constraints acting on you? What do you think we don’t understand about your situation that we would need to understand in order to see matters your way? How can we come to agreement here?” This is what we are doing in a whole variety of relationships, of which those relationships are a few. Does that help?

[Unknown:]

Yes, thank you.

[Eben Moglen:]

Yes.

[Unknown:]

Why are you claiming that Mr. Gates has quit? Our news say different things. Aren’t they true news?

[Eben Moglen:]

Sure, but when a CEO of a publicly traded company in the United States resigns or retires, there’s a drill. You go through it exactly this way. You tell the markets two years in advance, you announce the transition, you do exactly what Microsoft did: First, pure press release, day one. Second, sit‐down with New York Times to give account of “How I see the world”. Day three… This is a ritual formality.

If that happened in another way, as when Jeffrey Skilling quit Enron—without any of that previous drill—in the fall of 2001, that would mean the company was going bankrupt immediately, and of course that’s not happening.

Microsoft’s a company with tens of billions of dollars in cash; even when it is having a bloody disaster, which it is having, it doesn’t have to do anything sudden. It can announce everything with the greatest suavity and decor. You will see the greatest suavity and decor here, but the underlying lesson is very simple, it has two parts: One, Mr. Gates has quit, two, Mr. Ballmer has not quit. Those are the key lessons, and put together they are devastatingly bad for Microsoft.

[Unknown:]

You said that version 2 was perfect, I think for all practical intents and purposes it was very good. However,

[Eben Moglen:]

I didn’t say version 2 was perfect, well, I said they thought it was, I was joking, as you know.

[Unknown:]

However, there’s been an ongoing dispute about version 2 whether it should be treated as a grant, a contract, and I’ve not seen anything new in the initial draft of version 3 which addresses the dispute.

[Eben Moglen:]

It’s a dispute that I think is not necessary to address directly because it’s based on a misimpression. We said: “This is not a contract, you don’t have to accept it in order to get a copy of the program, or run it.” We said: “Copyright law alone is enough to provide for all the legal technology that the licensor of this code needs in order to achieve her or his intentions.”

That means, if this is not a contract in some place at some time, that won’t matter at all; all the necessary legal consequences will follow from its being simply a unilateral copyright permission, but, if it is treated by local law as a contract, fine, it’s also a contract.

We merely call attention to the fact that in our home legal system, and in many legal systems, a contract without an acceptance doesn’t bind anybody, and we wish to point out that that matters not at all to the workability of GPL.

[Unknown:]

I agree with you that it doesn’t really matter for the workability, generally speaking, but due to a historical mistake in the European Union, I don’t think it can be construed as a grant, as, like you, by the fact that that we have limited warranties in GPL, and whether you also take away a few use rights which I think is reasonable given that contract. However, you need acceptance for that, and I would like to see the new version, does the new version 3 together…

[Eben Moglen:]

That will not happen, but there are alternate ways of achieving the same outcome that not only will happen, but are being provided for. I don’t think this is going to be a problem at the end, as I don’t think it is a problem now.

Of course, however, your view, as a lawyer working in a place, is absolutely crucial, and we will listen, once again, to everybody who has anything to say on these subjects.

I am both retaining legal advice around the world, under privilege, for the Free Software Foundation, and also accepting a lot of public commentary from lawyers around the world looking at these and other questions within the parochial lenses of every legal system, in order, instead, to build up a transnational view of what the institutions are which achieve the maximum degree of coverage with the minimum degree of parochialism. I am satisfied that there are legal technologies for reaching those goals. I am addressing that language when the Software Freedom Law Center does technical drafting for the Free Software Foundation, I am addressing that language to you.

If you find that there are problems then we need to talk about them, as I said to you on Thursday, I take seriously your point concerning whether the anti‐DRM provisions of the license are expressed as use restrictions. The fact that the first draft expressed some things in language which might have amounted to use restrictions is not a highly recommended part of the first draft, but first drafts are first drafts.

We will talk about each one of these issues; what to do about acceptance, what to do about making rules against DRM that do not make use restrictions that bind users in unacceptable ways, we will talk about all those questions with everybody who wants to engage. I know that this is an issue that needs addressing.

I disagree that the way to address it is to require acceptance in GPL, that has the wrong consequences in too many places for too many things. But, nonetheless, something needs to be done about the other side, too.

[Unknown:]

Will GPL be translated?

[Eben Moglen:]

Will it be translated? The existing GPL is widely translated. Will it be officially available in more than one language? I don’t know the answer to that question, but I suspect the answer is “No”.

I do not think that formal translation of the license, risky as it is with respect to legal infrastructure, is something that the Free Software Foundation wishes to do. The fundamental question is this one: When you make an official translation of the license into another language, what you’re actually doing is making an official translation into a legal system.

That is to say, the official translation into French for Quebec and the official translation into French for Sierra Leone, and the official translation into French for France are very likely to be different in themselves. Because what it being translated are not words, but legal technologies.

So when you translate the license officially, into new languages, you are effectively producing specific legal system copies of the license, and you need two things: One, you need legal infrastructure consisting of lawyers, and the support structure to defend and protect that license in that legal system as a thing in itself. You have to be prepared to be in the courts of Quebec, with Québécois lawyers to discuss the Québécois consequences of the Québécois license. And, two, you must have a way of delineating that it is the Québécois license, not the French license or the Sierra Leone license that you are talking about.

Those are solvable problems, they are by no means insuperable. Over time, they will be solved. The question is: “What is the state of the infrastructure now?”

The Free Software Foundation, believes, I think—though I am not here speaking for the Free Software Foundation—the Free Software Foundation believes, I think, that infrastructure has not advanced almost everywhere to the point at which that would be a safe thing to do. And there are other, safer, things to do. There are ways of employing wrappers, there are ways of employing unofficial translations, with various ancillary adjective mechanisms attached to them; there are strategies short of complete official translation and the hiving off of particular subdomains within the GPL.

Moreover, the GPL3 process is designed to produce a universalizing license, actually designed to produce a universalizing license. I want the lawyers who are working on it to be thinking about the process of universalizing, not the process of adapting to local circumstances; those are two completely different jobs.

If you ask a lawyer to do something which is hard, and something which is easier, he, like all other human beings will do the easier thing first. The problem is that universalizing is really hard. So, it’s easy to get lawyers to spend time parochializing the license. It’s harder for all of us—and I am the worst offender in this—it is hard to universalize, it is hard to let go of your categories and try and think in universal categories. And the more expertise you have in a particular locality, the harder it is, that’s why I’m the worst offender.

I think that our primary task right now is to universalize the license. And then our task is to grow the infrastructure. That’s what SFLC is for, to make more lawyers. When we have a lot more lawyers and a really strong universal license, then the time comes to begin talking about how to make domestic national implementations, where that is inescapably necessary.

Erik… Yeah, look, go ahead.

[Unknown:]

In connection to that, I wondered what part does the European Commission take in the discussion in GPL3?

[Eben Moglen:]

The European Commission has taken no official role in the discussions. I invited Jesús Villasante to come to Cambridge and he was present. Subsequently the European Commission, which I think would ordinarily participate in a process of this kind through a contractor or some other independent party, has found no independent party to participate, and has not been participating eo nomine.

But, I have every intention of doing everything I can to facilitate easy participation in any way that the commission or any directorate within the commission finds useful and productive.

Paul.

[Unknown:]

What is the status of LGPL?

[Eben Moglen:]

LGPL will be announced in a first discussion draft version with GPL3 discussion draft two on July 25th.

As a consequence of the way the additional permissions section of GPL has worked—one of the reasons we made the additional terms provision the way we did—LGPL can now be quite elegantly reexpressed as GPL plus a single additional permission. So a very much smaller, more compact, more operable version of LGPL as an additional permission on top of GPL will be revealed when the next drafts come out.

I think this is a very great conceptual simplification of the world of the copyleft licenses; I’m very happy with the way that that has gone, I look forward to a conversation about it with everybody else in the world who cares about this, rather recondite, subject.

[Unknown:]

Can you explain to me a little more about the the dynamics of automatic reversioning? Suppose I don’t like version 3, what happens, can I go back from version 3 to version 2?

[Eben Moglen:]

Remember that these terms are set by the license we have all been living with, for 16 years now. This is about what the language of GPL version 2 says. So, it is familiar language; let me just go through how it works.

The license says: “We may publish revised versions of this license from time to time. If you receive a program under this license, and it doesn’t say what version applies to it, or it says ‘version X plus any later version’, then when you pass the program along, you may make a choice what license to use.” Right?

So, this is a power addressed to each person who receives a copy of a program under the license. You may make a decision about which version of the license you distribute on when you distribute. If the license says “version X plus any later version”, then that is the scope of your choice. If the license says “Under GPL”, and doesn’t state a version, then you may use any version ever published by the Free Software Foundation.

All right? So, the choice of version in license if the choice of the party who receives code, it’s part of your right as a user to pick the version of the license. All right? Under GPL version 2, if a program is licensed just “GPL”, you as an individual recipient could put GPL version 1 on it. GPL version 1 is a little hard to find, but you could find it, and apply it to the code. If the code says “GPL version 2 or any later version”, then you could only apply, for the last 15 years, version 2, ’cause there was no later version. When there is a version 3, you’ll be able to apply version 3, as an individual. If the license says “Version 2 only”, however, then that’s the version you must apply when you redistribute. All right?

So, the existing terms of GPL2 plus the existing terms of code actually in distribution determines what will happen in the future, when there is GPL3; That has nothing to do with what is written in GPL3 now. The transition provisions from this license were written a long time ago, and everybody has had a long time to come to think about them. Many people haven’t thought about them yet, they didn’t need to. They’re about to see how it works; that’s how it works.

Now, what the foundation will do when the new license is announced, is that it will relicense its works “Version 3 or any later version”. And therefore, downstream regression will not be possible with respect to FSF works.

Some people out there in the world will simply continue to license “GPL”, some will license “Version 3 or any later version”, some will license “Version 2 or any later version”, and some may choose to license “Version 2 only”.

All of those choices will be available to licensors who hold the copyright, and some of those choices will be available to downstream redistributors.

Yes.

[Unknown:]

You told that competition between software seems good, do you think that the same might apply to licenses?

[Eben Moglen:]

That competition among licenses is good?

[Unknown:]

Yes.

[Eben Moglen:]

Sure! It’s not a question, in my judgment, mostly, of competition among licenses, it’s mostly a question of the variablility of clients. The Software Freedom Law Center is not a place where you come in the door and we clamp the GPL on your wrists like handcuffs. Right?

The Software Freedom Law Center is a law firm which exists to help hackers make and distribute free software in a non‐profit way. When our clients come to us and need to talk about licensing, we don’t say “Take the GPL or God damn you!”. We say “Let’s talk about what you need. Let’s talk about how you want your code to flow through the world, let’s talk about what you want to have happen, let’s talk about what you don’t want to have happen.”

And then out of that conversation a license can be chosen or made. Maybe you even need to make a license; I don’t say “Don’t make licenses anymore”. What we think is that good legal advice early results in intelligent license choices for projects. That’s all.

Not that there is a single intelligent choice of licenses, or that there will come a day when license development has ended and there is no more history. I said Thursday that if you are concerned about proliferation, copyleft is the answer to the problem. That wasn’t the same thing as “Copyleft is the answer to all problems.”, or “Proliferation is the one thing you ought to be worried about.”. I did not even say “If you are interested in copyleft, there is only one license, use GPL.”, none of those things is true.

I am here to talk about GPL and the work I do for the Free Software Foundation; GPL is a wonderful license, it will get still better. But that’s not to say “GPL is a wonderful license, there shouldn’t be any others”, not by any means. Remember all the effort which is going into compatibility with other licenses here. Our goal is not by any means to prevent licenses. Our goal is to make software freedom as full and as available as possible.

Keynote about GPL3 at HOSC 2006

From Wikisource

[edit] Questions & Answers

Views

Personal tools

Search

Navigation

Toolbox

Create a book