I know many of you have read the article by Professor Richard Epstein, "Why Open Source is Unsustainable". I know because so many of you sent it to me. I was not originally inclined to write a rebuttal, because I felt he was a perfectly fine gentleman who just doesn't quite get the GPL. This isn't a sin. And I wasn't inspired to do it. So I set it aside.
Actually, there a teensy bit more to the story. I despaired of doing a good job. I sighed when I read "once someone incorporates open source software in his own programs, then any license that he issues cannot charge others for its use," because that isn't so, and my eyes got wide as I read on to learn he thinks open source development is like a workers' commune, but when I got to the question, "But how do the insiders, such as Linus Torvalds, cash out of the business that they built?" -- well, it was clear *I* couldn't write about it, because my pesky sense of humor started revving up, and it's like getting the giggles. Once you start to see a thing funny, you just make it can't stop, so obviously, this wasn't a job for me.
But you know how the community works. If one person doesn't feel like doing something, or doesn't know how, someone else will, and somehow it all gets done.
And two readers have stepped up to the plate. They would like to explain to Professor Epstein a few things about the GPL and why it is so valuable to those who use it. So here are their replies to Professor Epstein, one by Mark Levitt on how the GPL works in his experience, followed by a response by Tim Legge, "Is Knowledge Creation Unsustainable?"
Tim is a developer and speaks from that perspective on the real value proposition of the GPL, as the business types might say. He says this about himself: "I have been using GNU/Linux since 1996 when I downloaded Slackware (still my favourite distro) over a 14.4 modem. I have been a developer on the Etherboot project for two years. I live in Canada." And Mark is a technical author living in the UK, whose bio describes him like this: "He has been an enthusiastic Linux user ever since he carried a stack of thirteen floppies between his college computer lab and home to install Slackware." It sounds like I need to check out Slackware. I sincerely hope Mr. Epstein gets to read them both, because they are from the heart and respectfully present points of view he has not yet considered.
For some other responses, here's Larry Lessig's one sentence reply, and Joe Gratz, a law student, who explains why he believes the GPL isn't a "workers' commune".
To place this into perspective, and to explain why I finally decided to write about it, Mark brings this information to us:
"The Register has published a story that, I think, explains the behind-the-scenes reasons for the sudden anti-GPL in government article: http://www.theregister.co.uk/2004/10/28/ogc_oss_pilot_report/
"Basically, the UK Office of Government Commerce published a report yesterday that states open source is a 'viable and credible for infrastructure and for meeting the requirements of the majority of desktop users.' They apparently are reporting on a number of pilot projects in UK government. The actual report is published here: http://www.ogc.gov.uk/embedded_object.asp?docid=1002367 [PDF]
There is a bit more on an earlier version of the study here and if you like, you can read Microsoft's reaction to the study. I did some checking and Mr. Epstein is affiliated with the Hoover Institution, a public policy research organization, and here is their mission statement.
I am writing in response to your article "Why Open Source is Unsustainable" that was published in the London Financial Times. I hope to explain a number of areas where, I believe, you have made serious errors in your analysis. I hope you will take this in the spirit of discourse it is intended.
There are two serious flaws in your analysis.
First, you are confusing the General Public License with a contract. The GPL is a copyright license, nothing more, nothing less. And, in fact, you are advocating the exact disregard for the intellectual property right of copyrights that you accuse the free and open source movement of.
Second, you are incorrect in your assertion that a contributor to a free software project receives no value in return. In fact, the entire point of the GPL is to ensure exactly that. The participation in free software does provide direct and indirect rewards that are capable of sustaining the free software activities of developers. In fact, it is exactly these rewards that drive companies such as IBM, RedHat, and HP to invest heavily in free and open source software.
First, let's clear up a few things:
"In answering this question, first note that open source software relies on the very private property regime that its supporters, most noticeably Richard Stallman, disdain on moral grounds."
The GPL relies on copyright, actually. Richard Stallman is against software patents, not copyright. To lump two very different things (copyrights and patents) into a single umbrella term (intellectual property), and then imply that this is a hypocritical position of Stallman and the rest of the free and open source software movement is unfair.
Furthermore, your assertion that the "critical provision of the GPL has not been tested," is hardly worth the effort to debunk, except to say that the general principle that a creator of a work has exclusive right to dictate the terms of the copying and distribution of that work is well-established.
You continue by writing:
"It does not in so many words specify the appropriate remedy when some portion of the open source code is incorporated into an otherwise proprietary program."
As the GPL is a license, not a contract, it does not need to specify a remedy. The law of the land already specifies a remedy for the violation of copyright law. If you take a portion of my work, and that work is found to be protected by copyright, you will be liable for damages for the infringement of my copyright.
You assert the claim that if a portion of code licensed under the terms of the GPL is incorporated into a proprietary software project, that project magically "becomes open source software subject to the GPL." That is clearly false. If Microsoft, to use your example, copies and distributes code without a license, then Microsoft can be held liable for copyright infringement.
They have two options:
1) Comply with the terms of the license for the code they are infringing or,
2) Stop infringing the license by removing the code licensed under the GPL.
To suggest that the authors of the code licensed under the GPL can force Microsoft to open source their code even after Microsoft removed the code licensed under the GPL is not accurate. No copyright owner can force you to comply with the terms of a license to copy and distribute their work after you've stopped copying and distributing it. They can go to court and the court can force you to pay damages. They can ask the government to file criminal charges, conceivably. In neither case would the end result be the forced open-sourcing of Microsoft's code.
Regarding the contract vs. copyright license nature of the GPL, you state that if a person doesn't know they are using code that is copyrighted, then they may not be bound by the terms of it's license. So, to carry on with your Microsoft example, suppose somebody gives me a copy of the recently leaked Windows 2000 source code. If I claim I "didn't know" what terms the code was licensed under, then I am legally allowed to use Windows 2000 source code all I want? Are the copyright laws somehow different for proprietary software than for free and open source software?
Regarding the argument that the GPL might not apply because of "restraint on alienation," then again, I ask you, I am free to ignore the license terms of Microsoft's Windows because they are "illegally" restricting what I can do with it? Is not the essence of copyright that the author can decide what happens to his or her work and under what terms?
Your statement that "Once the contract protection [of the GPL] lapses, then the open source movement is left only to its copyright remedies, which are likely to prove far weaker," seems to summarize the confusion. Of course, the GPL has no "remedy" beyond copyright law and nobody who understands the GPL thinks it does. It is a license to copy and distribute a work. If you don't have a license, copying that work is an infringement of the author's copyright. I'm not sure how those penalties are somehow "weak."
It seems that your attack on the GPL boils down to this: The GPL is likely to not be enforceable because copyright owners can't really force those who infringe their copyright to license their proprietary code under the GPL. This is, of course, a straw man argument. No copyright license has this power. It is not, as you imply, a fatal problem with the GPL. Nor does it change the fact that the penalties that do exist for infringing the copyright on GPL covered code are, nonetheless, fairly severe.
What all this has to do with whether the free and open source model is a viable economic model, I am not clear. You seem to be saying that since GPL-using copyright holders cannot force proprietary software companies to open source their code, there is a limit to the amount of open source code that can be created. You ignore completely the amount of free and open source code already running a large part of the IT infrastructure, including, I might add, the web server your publisher uses, created entirely without the use of proprietary code that was "forced open."
Your argument for why the free and open source movement cannot sustain itself likens open source to a workers' commune and asks the question, "Does that worker receive in cash or kind his share of the gain in value during the period of his employment?" The answer is, of course, yes. In fact, this is the primary purpose behind the GPL. The "worker" contributes a bit of his work. In return, all of the work that everyone else does on top of his work is available for his own use.
You are basically arguing that people who write free and open source software have no way of getting any value out of it at the end and, therefore, no lasting incentive. Indeed, you write: "But how do the insiders, such as Linus Torvalds, cash out of the business that they built?" Linus Torvalds made the initial effort to write a small UNIX-like operating system that ran on a single CPU model, only worked with a small amount of hardware, and didn't even do networking. Through the participation of many others, he now has the benefit of the use of one of the most stable, scalable, flexible operating systems on the planet, able to run on anything from tiny wristwatches to large mainframes, not to mention, of course, the fame and glory he has. You may not think these "non-monetary" awards are worth his investment but to say that he gets nothing out of it is not true.
Finally, in the last paragraph, you acknowledge that governments the world over are starting to realize that open source is a viable alternative. You attempt to characterize this as an attempt "to sway governments by a different" take, by which you seem to mean by something other than "cost." Furthermore, you seem to be arguing that cost should be the only consideration governments use to pick software. I wouldn't necessarily disagree with you. As I'm sure the governments who have adopted open source solutions can tell you, price is not the only measure of cost. There are many benefits of free and open source software beyond price, including the lack of vendor lock in, the existence of a free market for consulting, and the ability to upgrade software and systems when you want, rather than to meet the revenue needs of a foreign corporation.
So, in conclusion, I hope that you realize that the GPL is not, and not intended to be, a weapon to "trick" proprietary software companies into giving away code. It is intended to ensure that the work an author creates can be used by others but that the author gains from whatever work is added on to his own. Furthermore, it is exactly this characteristic of the GPL, along with the very human need for respect and recognition, that provides an incentive for individuals to contribute. And certainly, large corporations would not invest the money they do if they didn't think they get more out than they put in? I doubt that RedHat, HP, and IBM can be characterized as workers' communes.
Is knowledge creation unsustainable?
I don't often take exception to some of the opinions people have. Well, I guess I often take exception but I am rarely moved to write about it.
However, I read an article by on ft.com by Richard Epstein "Why open source is unsustainable" that essentially equates working on open source to a workers commune that produces some tangible goods.
I work on the free software Etherboot project and the code that I write or borrow from other GPL developers cannot be equated to a sack of potatoes.
Computer programming is essentially the creation of knowledge. Whether it is the knowledge of how to interact with a network card or how to download a file via tftp, it is documented in the source.
When I write an Etherboot driver for a network card I sometimes have access to the often flawed technical specifications from a vendor. However, I have written drivers with no technical reference other than the knowledge contained in a Linux driver. The difference is that while the knowledge contained in the driver C code may differ from the technical specifications, I know that the Linux driver works.
People outside the computer industry have been adding to the sum total of human knowledge for eons and I am doing my small part to help that. The fact that the knowledge can be "compiled" into an application that is useful to thousands of people is somewhat immaterial. The knowledge, unlike a sack of potatoes, lasts far past supper.
The creation of knowledge may be unsustainable but I suspect we have not yet reached that limit.