pageok
pageok
pageok
Free Software and Copyright Law:

Yesterday I attended a talk given by Richard Stallman here at Temple, on copyright law's increasing dis-utility (and Stallman's proposals for reform of that law).

Stallman, needless to say, is a fascinating character. He's already a major figure in the history of computing and computers, and it may turn out that he's a major figure in the history of the production of creative works more generally -- time will tell about that. Twenty-five years ago, he had a ridiculous -- borderline insane, really -- idea: "free software" ("free," as he takes pains to remind us, in the sense "free speech," not "free beer"). Large numbers of people could collaborate to produce functioning and efficient software systems that would be outside of anyone's proprietary control? Why would anyone do that? Where's the incentive? Who's going to work "for free"? Who would be in charge? How could they all possibly make it work on the technical side?

Of course, he managed to pull it off -- not on his own, to be sure, but he surely deserves a great deal of the credit for the success of open source software, software which now dominates a number of important segments of the computer universe and which is becoming more and more central to the business models of even the giants in the industry (e.g., Sun Microsystems and IBM).

I have great admiration for men and women who manage to pull off things that are borderline insane when viewed against the conventional wisdom; it's why I like Jefferson so much. They've earned the right to have their crazy ideas taken seriously -- no small feat.

Stallman currently has copyright law directly in his sights. Like many people -- myself very much included -- he has concluded that copyright law is broken, in fundamental ways, that it no longer functions to encourage the production of creative works, but in fact has quite the opposite effect, serving primarily to stifle creative activity. I couldn't agree more with him on that. He's got a specific proposal for changes in the law. I'm going to wait until he commits those to writing before I comment specifically about them -- I'm not entirely comfortable relying on my memory of his talk as a basis on which to comment. But the basics were these: much shorter copyright term for all works (in the neighborhood of 10 years or so, rather than the insane life+70 we now have), and a division of works into various categories that would get different levels of protection, from low to high: functional works, works of opinion and reference, and works of entertainment.

The devil's always in the details, and I'll have more to say about this proposal when I see and understand the details a little better. But here's what's really interesting about all this. Open source software -- paradoxically, or perhaps only ironically -- is entirely dependent on copyright for its legal foundations. The entire open source system relies on a complex licensing scheme (of Stallman's invention), under which open source software is distributed under a special license that (a) gives all users certain rights (to use, copy, and modify the software) and (b) requires that any re-distribution include the same provision giving users those rights. It's a kind of recursive non-proprietary licensing algorithm -- quite ingenious.

What many people don't understand about open source licenses is that they can be (and are) enforced by asserting a claim of copyright infringement against violators, not merely a claim for breach of contract. That is, if you take open source code and copy it and modify it and then re-distribute it without the provisions providing your users with the rights set forth in the license that you received, you will be infringing the copyright in the underlying work (in addition to breaching your contract). That principle was reaffirmed recently by the Federal Circuit, in the Jacobsen v. Katzer case, and it is of fundamental importance to the whole open source movement. Why? Because a breach of contract action is virtually worthless as an enforcement device, while a copyright infringement action is a powerful weapon indeed. There are many, many reasons why this is so. In a breach of contract action, the plaintiff has the burden of establishing that there was a contract -- no small task, when the licensee here could've picked up this software from any of a million different places, all around the internet. Who's got a record of the "contract" that the defendant agreed to, and exactly where he/she agreed to it. And even if you establish a contract, your damages for breach of the contract are limited to the harm you suffered as a result of the breach -- demonstrating that that's more than negligible is going to be incredibly difficult. It makes a breach of contract action entirely ineffective as a means for enforcing the open source license. But copyright infringement's another matter entirely. Once you show that the defendant copied/modified/re-distributed your work, the burden is on the defendant to show that he/she was authorized to do that by the copyright-holder. And there are statutory damages independent of actual harm to which you are entitled for copyright actions -- now you're talking serious money.

Stallman understands this thoroughly -- though the vast majority of commentators on the open source movement have missed this point. I suspect that his ultimate aim is not merely to substantially weaken copyright (as in his proposal) but to eliminate it entirely, and I also suspect that he relishes the idea that he's been using copyright law as the main weapon in the battle to destroy copyright law -- another nice little recursive algorithm.

LTEC (mail) (www):
Do you favor eliminating copyright completely? What about patents?

If we really want to make a radical stand for freedom, I much more favor eliminating all libel/slander laws.
3.27.2009 10:56am
Soronel Haetir (mail):
I would actually prefer moving in the opposite direction and make IP much more akin to other forms of property. I'm not sure how to accomplish that however given the way new IP depends on existing IP.

Perhaps significantly higher fees or an IP tax to compensate for use of PD materials. If it's not worth paying the fees then the material reverts to public ownership.
3.27.2009 11:05am
PeterWimsey (mail):
Eliminating copyright is insane. Limiting copyright to 10 years may make sense in certain contexts involving technology perhaps, but is otherwise fundamentally theft.

If I write the next, say, Lord of the Rings, the law should not be that, 10 years and one day later, anyone can reprint the books or make a movie without having to obtain my permission or pay me anything. While I'm not unsympathetic to the idea that life plus 70 years is too long, it is much preferable to 10 years.

And, really, I don't think the fact that Disney still owns the copyright to Mickey Mouse is really harming creativity, nor do I think that the fact that a band has to pay some money to Jefferson Airplane if they do a cover of "White Rabbit" is particularly evil.

Again, while I'm sure that some tweaks are needed in the copyright area, throwing the baby out is really not necessary.
3.27.2009 11:16am
Ralphe (mail) (www):
I'll be the first to support Stallman and Post. Copywrite has become much removed from fostering creativity and instead stifles it for a profit. I think for most works twenty years or life of the author would be fair and am open to argument on the actual length. But what we have now is far too long and I suspect when that time runs out we will see a try for another extension.
3.27.2009 11:23am
Javert:

and I also suspect that he relishes the idea that he's been using copyright law as the main weapon in the battle to destroy copyright law -- another nice little recursive algorithm.
I think you mean: another vicious little attempt to cannibalize inventors and artists.
3.27.2009 11:25am
Some Random Bonehead (mail):
Consider that practically all of Disney's most famous animated pictures are reworkings of pre-existing public domain works. Would Ariel exist if Hans Christian Anderson (or his estate) still controlled rights to "The Littlest Mermaid?" How about the Brother's Grimm asserting ownership of Snow White, Beauty and the Beast, etc...? The existence of public, unencumbered works has resulted, in this case, directly in billions of dollars of economic activity. Would longer or shorter copyright periods encourage this?
3.27.2009 11:25am
what up doe? (mail) (www):
your first use of the word "recursive" is surely correct, but is your second? the second "algorithm" is self-referential, but is "recursive" the right word to describe it?
3.27.2009 11:25am
Douglas (mail):
One quick qualification. You note that statutory damages are available for copyright infringement actions. You can only get statutory damages if the copyrighted work had been registered within 3 months of its publication. Because much open source software does not have a valid registered copyright with it, getting statutory damages and attorneys fees for bringing suit against infringers is nearly impossible.

See 17 USC 412.
3.27.2009 11:32am
Oren:

I think you mean: another vicious little attempt to cannibalize inventors and artists.

Because those of us that willingly contribute to open-source projects are cannibalizing ourselves? Nobody requires you to lift a finger for the (L)GPL or BSD licenses.
3.27.2009 11:34am
statfan (mail):
Douglas, that's not quite true. First, FSF has registered the copyright in the software it holds copyright on. And second, it's possible to register after the problem is discovered. Since each act of distribution thereafter is a new infringement, there will likely be enough post-registration infringements to bring a case.
3.27.2009 11:40am
PeterWimsey (mail):
@ Bonehead - Assuming for the moment that Andersen's estate did still control rights to the littlest mermaid, why do you believe that they wouldn't sell the rights to Disney for use in a movie...or that Disney wouldn't pay for them. In fact, people make movies out of copyrighted materials every day of the year. The Tolkien estate having the rights to LOTR didn't prevent the LOTR movies from being made; the same is true with Harry Potter, or James Bond, or The Watchmen, just to name a few franchises off the top of my head.

As I mentioned above, I'm not averse, potentially, to a shorter copyright period, as long as it extends at least for the life of the author, and I could be convinced that copyrights on computer code should maybe be treated differently. But the argument that copyrighted material are unavailable for use during the copyright period is simply not true.
3.27.2009 11:41am
MichaelJ (mail):
Kudos on an excellent article that explains open source software well. However, I would like to point out one inaccuracy. It's not quite correct to claim that "the entire open source system" makes use of the "recursive" scheme (the jargon is "copyleft") that Stallman devised. Stallman's GNU General Public License is the most famous of this class of licenses, and a lot of famous open source software uses it. However, there also exists a class of licenses that do not require derivative works to use the same license. These include the BSD license, Apache License, and MIT license.

Personally, I've always wondered whether Stallman really wants to see the end or even a substantial weakening of copyright law if that would weaken his beloved recursive licenses. A lot of his rhetoric seems to focus on how restrictive licenses fundamentally infringe on the rights of computer users.
3.27.2009 11:47am
matt (mail):
I'll take this opportunity to reiterate my suggestion from an older copyright comment thread:

Ten years copyright, fee of $x.
Reregistration every ten years, fee of double (or 1.5x, or 1.1x, or 5x, ...) the previous fee.
Limited to, say, 300 years for constitutional reasons.

This would:
Guarantee fair compensation to authors.
Allow copyright owners to continue to exploit copyrighted material as long as it is commercially valuable.
Guarantee that the vast majority of works would find their way into the public domain within a reasonable period.
Promote progress in the useful arts :)
3.27.2009 11:49am
Some Random Bonehead (mail):
..why do you believe that [rights-holders] wouldn't sell the rights to Disney for use in a movie?

As you demonstrate, I don't doubt that they would, but that's not a particularly interesting question. The more important one is: would the studio have pursued the license in the first place? For Disney, the movie is only the first sale, merchandising is where you make the real money. Why would a company do all that work to benefit a passive partner? Pixar doesn't. They make original stories so that they don't have to pay royalties. It make a lot of financial sense not to do so.

Let's examine another of your examples: Spider-man. Are you aware that Marvel, the rights-holder, has leveraged the original movie creators and now has shouldered them aside? Iron Man, of last summer, was the first Marvel Studios independent film. So that's another risk for a potential licensor: the rights-holder, whom you have made wealthy, cuts you out of future development of the property.

So why would one want to remix an old character? Built-in audience. It's a lot easier to adapt and advertise Snow White than to create and market a new story. Pixar is incredibly good at it, but not every studio is Pixar.
3.27.2009 12:04pm
Joe Kowalski (mail):
My understanding of Stallman's position on copyright is that he is generally ok with copyright on works of art, literature etc. so long as the rights holders aren't using DRM to artificially limit the normal freedoms consumers of copyrighted works have. When it comes to software though, Stallman sees something different. Software, essentially is a tool. Historically, when people bought tools, they could study how they were put together, how they worked, make improvements to them, lend them to friends, and generally use them for any lawful purpose. With software however, the combination of copyright and the subsequent contracts software writers issue to make use of those copyrighted works, the general historical usages and freedoms with tools go out the window. Hence, Stallman's big push for Free Software.
3.27.2009 12:12pm
sonicfrog (mail) (www):
What I love most, is that the faulty IP / copyright lawsuit that was going to destroy the GPL and Linux, SCO vs. IBM,and Novell, and Autozone, and Chrysler, and, well.... Everybody, ended up proving the viability of GPL and actually, in the end, made Linux that much more viable in the marketplace.

I've been using Linux since 2000. I was taking a Cisco IT class, which introduced me to the world of UNIX and Linux. I use Microsoft XP when I'm using Quickbooks for my business or preparing stuff for teaching. I want to use Linux when recording recording music, but my recording device is MS compatible only and won't work with Linux distributions such as 64 Studio, which is geared toward audio production.
3.27.2009 12:24pm
sonicfrog (mail) (www):
PS. For those who want more info on the fiaSCO, Groklaw has been following that lawsuit from the beginning.
3.27.2009 12:25pm
Aultimer:
Stallman is a great mind and certainly understands that 0 copyright is unworkable and counterproductive. I have no doubt he'll be happy with weakening copyright to a sane level while professing indignation that anything is left. I think any reform of copyright has to include firming up fair use in addition to more reasonable term (to me, something like minimum of 25 years, maximum of author's life and 15 or 20 years for corporate/psuedo authors and assignees).

matt - your proposal goes beyond "promting progress" with by focusing on the commercial life, that will inevitably lead to future Sonny Bonos.
3.27.2009 12:26pm
sonicfrog (mail) (www):
PS. I'm using Ubuntu Linux OS at this very moment.
3.27.2009 12:27pm
Aultimer:
Joe Kowalski - any links? Your descriptions reminds me of a paper on sui generis IP protection for software (maybe by Julie Cohen - Google didn't find it on first try). Seemed like a good idea when I read the paper.

You're probably thinking of the paper by Pam Samuelson et al in the Columbia L. Rev. (1994? or thereabouts) DavidP
3.27.2009 12:32pm
Javert:
I think you mean: another vicious little attempt to cannibalize inventors and artists.

Because those of us that willingly contribute to open-source projects are cannibalizing ourselves? Nobody requires you to lift a finger for the (L)GPL or BSD licenses.
The original quote referred to the destruction of copyright law, as such -- not just to your particular project.
3.27.2009 12:46pm
FWB (mail):

rather than the insane life+70 we now have


It is a bit difficult to understand how "to the author or inventor for a limited time" became life+70. Limited time to a person is easily understood to mean < life.

Typical of our government. Bought by the highest briber!
3.27.2009 12:55pm
Robert Shaw:
For most fiction, the royalties received peak soon after publication, then fade away fast, so a shorter copyright term makes no difference. In some cases, the royalties will come in at a steady level for many decades, but this is rarely more than pocket money. The authors who would be hurt by a short term are those whose popularity doesn't peak until several years after first publication, and those who remain best sellers for decades - two exceptional categories, which should be treated as such. Frame the law to deal with the usual case - books and software earning most of their royalties within a few years, for which a short copyright term is fine - and have special rules for the rare exceptions.

We should also consider whether all the rights which bundled together under the name of copyright should have the same term. In particular, I'd say that while the monopoly on printing actual copies should only be a decade or two, the monopoly on producing derivative works for fiction should last for the author's life, for trade mark like reasons. If Stephen King were to publish 'Harry Potter and the Spear of Destiny', he'd be trading on Rowlings' reputation, and affecting the market for any new derivative works she might decide to produce.

However, software isn't fiction. It shouldn't inherit the copyright rules that were suitable for books; rules should be found that fit its distinct nature.
3.27.2009 1:05pm
MCM (mail):
It is a bit difficult to understand how "to the author or inventor for a limited time" became life+70. Limited time to a person is easily understood to mean less than life.


What's worse than the actual number is that the Supreme Court allowed works that had entered the public domain TO BE RECLAIMED when copyright was RETROACTIVELY EXTENDED.

As Lessig says, this is an "easy case". If you can retroactively re-copyright things, how is that "limited times"?
3.27.2009 1:19pm
John D (mail):
Two thoughts about earlier comments:

Disney has been wrangling with the heirs of the man who bought the licensing rights to Winnie the Pooh from A.A. Milne. I think they had another legal tangle with the Great Ormond Street Hospital (which owns the rights to Peter Pan).

One member of the Tolkien Trust was against the LOTR movies. I've forgotten which of J.R.R. Tolkien's descendants was involved, but he did try to stop both LOTR and then any attempt at making a Hobbit movie.

Yes, people make movies from pre-existing works under copyright all the time. That doesn't mean it's free of legal questions.

Interesting further wrinkle: the opera The Fly, in addition to its crushing reviews, had copyright trouble. The makers of the opera bought the rights from the movie studio. Those rights were not actually the studio's to sell. The first movie was based on a short story for which film rights (not stage rights) were sold.
3.27.2009 1:24pm
Howard Gilbert (mail):
The larger and more sophisticated software projects understand copyright registration. Years ago when the copyright law had not been updated for binary computer files, IBM satisfied the requirements of the law by printing out in hex a copy of any new software product and shipped it to a previously notified cooperating customer. One day I received the first copy of a somewhat less popular software product, and because someone in shipping screwed up I also received without warning or explanation eight boxes of used computer paper containing a meaningless binary dump. Today the law has been updated, and I think the handoff can be done by two friends exchanging a DVD. I don't think the legal requirements of copyrighted software distribution have been forgotten by the major practitioners, but when the system works correctly they involve private exchanges that the rest of us do not know about.
3.27.2009 1:26pm
Sam H (mail):
The "free" software that Stallman is so proud of was created by theft. He is paid by the taxpayers to do a job, but he spends his time doing his "free" software. He is cheating the taxpayer just like Hanson at NASA does.
3.27.2009 1:32pm
Daryl Herbert (www):
Here's the Daryl Herbert Copyright Reform Act of 2010:

1 - Except as provided in this act, the term of all copyrights is the sooner of 40 years from the time of publication or 50 years from the time of creation.

2a - The term of Fictional Character Copyrights (identifiable fictional characters, individuals or a class of characters, including their name(s), appearance, and other distinctive traits, but not including their fictional backstory, and not including their relationship to any fictional plot elements) shall be protected for a term of 70 years from the time of creation.

2b - The terms of Fictional Character Copyrights may be extended up to 300 additional years, such extensions being contingent on payment of 50% of the estimated market value of the copyright during the extended term.

(This lets Disney get its way, but it also lets us stick our hand into Disney's pocket.)

2c - A "Fictional Character Copyright" is not violated by the unauthorized reproduction of a non-copyrighted work.

(comment: We would be free to copy, but not modify (except as permitted by Fair Use), all of the old Mickey movies)

3a - The term of Software Copyrights (source code and executable code) shall be the sooner of 10 years from the time of publication or 20 years from the time of creation.

4b - The term of a Software Copyright may be extended up to 50 years upon sufficient payment to the Copyright Office. Sufficient Payment shall mean 50% of the estimated market value of the term extension, but not less than $100,000. Monetary credits towards term extensions may be granted by the copyright office in exchange for the release of Source Code (the highest-level language under which a program was edited by humans, and all instructions necessary to compile it, if any) to the general public under the GNU General Purpose License v. 2.0. The number of credits shall be equal to 500% of the estimated market value of the Source Code.

(comment: If a software company wants to keep SOME program proprietary, it will have to release other source code to the general public, under GPL2. This will provide a net benefit to the authors of free software, and society in general, while allowing companies to continue to profit from their hard work/innovation--if you can make a program that still has market value 10 years later, you deserve the chance to keep profiting. I added a floor of $100,000 because I don't want people to seek extensions for software that just isn't profitable.)

4c - The term of Software Copyrights for software that is kept "in-house" (never distributed outside of an organization after its initial creation or purchase), regardless of non-disclosure agreements, and that constituted a trade secret prior to its improper disclosure, shall have no expiration.

(comment: this will allow companies to use copyright law to protect certain legitimate trade secrets even after improper disclosure of said trade secrets, namely software that they never intended to share with the world in any capacity. A single sale (beyond the initial purchase from the people who created it, if it was not created in-house) would remove these protections. If you want to sell your software, you should only get normal copyright protection and need to rely on contract law rather than Subsection 3c.)

Further comment: the biggest problem I can see with this is that software is often released in many, many versions (every time MSFT releases an update to Windows, it could be considered a new program). Term extensions would have to include different versions of the same software program. Companies should not be penalized for fixing bugs/security holes, or adding new features.

The second biggest problem is that we would need a small army of evaluators to determine the estimated market value of copyright term extensions. The fact that the copyright office would be raking in cash hand over fist should take care of this.
3.27.2009 1:42pm
Aaron Denney (mail):
If one feels that it is indeed truly necessary to keep Stephen King from selling "Harry Potter and the Spear of Destiny", than why isn't trademark indeed sufficient? Shouldn't he be able to create derivative works so long as they don't trade on the reputation of Rowling? How do you deal with a long running series, written by a corporation or long-running partnership that introduces new writers in the middle?

Generally, I think all the justifications for tieing the copyright term to life are flawed. Why should the expected time limits be lower for the aged and infirm? This directly lowers the value of the copyright they have, though admittedly this will usually be a very small effect.

What are the real problems with a fixed term? Are there any, or is life+n just a disgusied land grab trading on emotion? If there are indeed good reasons to let some works remain unavailable for longer than the average term, are scaling renewal fees as matt proposes not workable?
3.27.2009 1:44pm
Kirk:
I'd like to see someone defend the position that a copyright with a generous but strictly limited term of years--say 25 years--is not sufficient to completely fulfill any "promoting progress" requirement. Even just a single example of something in arts, literature, or software that genuinely has an actual payback period longer than that?

My strong assumption is that such a thing does not exist, but if you think you can provide an actual historical example, or a reasonable hypothetical (where "reasonable" == "anyone not so far out as Stallman would agree with it"), I would love to see it.
3.27.2009 1:46pm
Daryl Herbert (www):
To clarify, the 300 years is the max term extension; I expect Disney would buy extensions of 10-20 years at a time.

And to make a quick change: I would change the floor price to $10k, and add a floor price of $10k for each Fictional Character Copyright.
3.27.2009 1:46pm
Soronel Haetir (mail):
Why force use of the GPL? There are other licenses I have released code under, and others I would if appropriate, but GPL is not among them. I would sooner release all rights than accept GPL.
3.27.2009 1:57pm
Kirk:
Daryl,

You've got some interesting concepts there. Overall, however, your scheme is way too complicated, and in particular the provision for "payment of 50% of the estimated market value" for long terms of extension just sounds like you should title it the Endless Lawsuit Enabling Act of 2010 instead.
3.27.2009 1:59pm
einhverfr (mail) (www):
I am a software engineer and a self-published author. My software is all Free Software.

I do agree that copyright terms are way too long. There is no justification for keeping royalties flowing to my great-grandchildren until they are ready to retire. I personally would like to see copyright reduced to a fixed term of 30 years or so for literary works, and we need to rethink software intellectual propertly protections.

Patent and copyright law are both largely broken as they apply to computer software. Patents are problematic because software is sufficiently complex and the patents are sufficiently vague to provide no reasonable guarantee that distribution or use of any given piece of software is legal (since patents cover use too, this puts all software users at risk). Copyright is broken because software code reuse is done differently than it is with literary or scholarly works, and the expressive and functional elements are very closely tied together.

If someone wants to take ideas from my book, cite my book as the source, and then use those in their own studies, they don't need my permission to do this, provided they don't use my expressions of those ideas. So, for example, I can write a review of a Harry Potter book without getting the author's permission, and I can build my own book (nonfiction) on the works of many other scholars without running amok with their copyrights.

However, this is not the way it works with software. Typically software copyright protects derivative works which have been largely stripped of expressive (to humans) elements through the process of being run through a software compiler. The code is not archived anywhere, and so will never become a part of the public domain. Furthermore, software copyirights are effectively perpetual. For example, by the time Windows 95 gets out of copyright, there won't be a computer arount capable of running the software.

Finally, the limits of copyright protections relative to software have not been well discussed by the courts, leading folks like Stallman and Moglen to make up the law as they see it and then use the fact that people are litigation adverse to push their views on everyone else. For example, I don't see why dynamic linking of a program to a library covered by the GPL would even require the copyright holder of the library to give permission, and a lot of the arguments I hear from the Stallman camp seem to be very similar to those which the court shot down in SCO v. IBM.
3.27.2009 2:21pm
Andy Freeman (mail):
> open source software is distributed under a special license that (a) gives all users certain rights (to use, copy, and modify the software) and (b) requires that any re-distribution include the same provision giving users those rights.

As others have pointed out, there are other "open source" licenses with other provisions.

Note that GPL has another relevant/important clause - if you modify and distribute, you have to make the modified versions that you created available in source form.

One variant of the GPL takes this a step further. If you modify and use to produce services, you have to make the modified versions available in source form to the folks who use those services.

You might call that the "Google clause". Google is assumed to use modified versions of GPL programs to provide services to others. Stallman and others think that Google should be forced to release said modified versions.

Also, GPL comes in two big flavors, LGPL and GPL. Basically, if you use and modify an LGPL program, you can incorporate said LGPL program in other programs in certain ways without causing said other programs to be covered by the GPL or LGPL. ("L" doesn't mean "Library", it means "Lesser" although in practice it's mostly used for libraries.) However, if you derive something from a GPL program, that derivation is itself GPL.
3.27.2009 2:22pm
einhverfr (mail) (www):
Soronel Haetir:

Use or not of the GPL is strictly a business decision in my case. There are reasons to do so including:

1) You fork a GPL-licensed project.
2) You want to maintain control over how forks of your product can be licensed and so choose the GPL v2.

I personally find the GPL v3 scary because I had conversations about interop between the GPL v3 and BSD licenses with Even Moglen and with one other lawyer from the SFLC. Although both said the licenses were compatible, they gave reasons to my mind that seemed to be sufficiently contradictory to be worrisome.

I actually think the GPL protections are more limited than Stallman suggests because some things that Stallman seems to think are covered may be outside the bounds of copyright protections anyway.

However in the LSMB project, I pushed hard for (and got) a change in our documentation licensed from the GFDL to a close derivative of the BSD Documentation License. My reasoning was that copyright for software and books were sufficiently different that we were shooting ourselves in the foot by pushing it into a "copyleft" license.
3.27.2009 2:28pm
Andy Freeman (mail):
I think that the authors/owners of work subject to copyright should have to make a decisions, DRM and no copyright or no DRM and copyright, with no statutory protections for DRM methods. (In other words, if folks break your DRM it sucks to be you.)

Copyright theoretically expires and DRM doesn't. Furthermore, DRM often makes something useless even though the purchaser has done nothing wrong. (For example, some DRM systems "phone home" occasionally. When the vendor goes out of biz, said DRM keeps the folks who bought rights from exercising said rights.)
3.27.2009 2:28pm
Robert Shaw:
Stretching current trademark law to cover fictional characters would probably produce anomalies and undesired effects, as we've with patents, so it's better to devise an appropriate form a protection from scratch, and take care of long-running series into the bargain.

I'm not suggesting royalties for the production of derivative works, rather that the appropriate person has to give their stamp of approval. They can negotiate whatever fee they like for that, taking into account their expected lifespan, or do it for free, but whatever they do the public can be reasonably confident that what they're buying is more of what the original author produced, which is what they're after. Since this is for the public's benefit, not the authors, the varying term doesn't matter.

Of course, it'd be possible to define the right more narrowly. We could let anyone produce derivative works once the 10 year copyright term had expired, provided only that they clearly stated 'Not approved by the original author', so as not to mislead people.

However, that's a secondary consideration. The main thing is not to automatically assume that all the rights bundled together under the heading of copyright must have the same expiry date. There may well be advantages to splitting up that bundle, and treating its component parts differently.
3.27.2009 2:28pm
einhverfr (mail) (www):
Andy Freeman:

Exactly why I think the DMCA gives effectively perpetual copyright protections to digital works. One can't import tools to break DRM for copyright-expired works as long as other copyrighted works use the same DRM.
3.27.2009 2:31pm
einhverfr (mail) (www):
Robert Shaw:

Agreed We need more "Moral Rights" in copyright law in this country IMO.
3.27.2009 2:32pm
einhverfr (mail) (www):
Sam H:

I make money creating Free Software. I see no reason to call it theft. The idea of Free Software actually makes it possible for me to compete in some areas of software profitably where it wouldn't be possible otherwise (and as a producer, no less).

BTW, the way it works is I develop software for a fee. I release it free of charge so the cost of maintenance can be born by a larger number of people. Everyone (the original sponsor, me, the user community) wins. Not everything goes through this track as sometimes customizations are specific to how a customer does business and not generally applicable. In that case, the patch ends up maintained at their expense alone, but we try to minimize that because too much in that area ends up with long-term maintenance problems.

However, let me provide a very different example: Look at EnterpriseDB as a company. They contribute everything they can back to the parent project (PostgreSQL), but they keep the few things the community isn't interested in but their customers are, and they release a proprietary db based on PostgreSQL plus these extensions. Each contribution they make to the free version is a win because it simplifies their maintenance of their product. So once again, proprietary and free versions are not necessarily at odds.

I am sure Stallman hates the EnterpriseDB example because it goes against everything he stands for. However, it also shows how the economics of free software really does push for contributions even from commercial competitors.
3.27.2009 2:40pm
einhverfr (mail) (www):
SonicFrog:

What I love most, is that the faulty IP / copyright lawsuit that was going to destroy the GPL and Linux, SCO vs. IBM,and Novell, and Autozone, and Chrysler, and, well.... Everybody, ended up proving the viability of GPL and actually, in the end, made Linux that much more viable in the marketplace.


However it is interesting to compare the rhetoric of Stallman with the legal arguments of SCO. Personally I thank SCO because I think these suits clarify the scope of the GPL quite well and not in the direction Stallman would have liked.

It may take a few more lawsuits to drive the point home though. I think the GPL v2 in a post-SCO world is actually a very reasonable copyright license. I just don't like the GPL v3.
3.27.2009 2:46pm
ASlyJD (mail):
Einhverfr,

The problem with moral rights comes back to the nature of IP protection. Our Constitution enshrines a utilitarian view of inventions and writings, not a fruits of labor view. Moral rights in a work do not create incentives comparable to the actual copy and licensing rights.
3.27.2009 2:46pm
jfb2252:
As for Lord of the Rings, Tolkien let the US copyright lapse after 28 years. Lin Carter, the fantasy editor of Ace books, knew of LOTR and published the three volumes as soon as they were out of copyright. Ace sold about a quarter million copies before Tolkien renewed the copyright and sold US rights to Ballantine. I have both the Ace editions and the Houghton-Mifflin hardcover. I don't have the Ballantine paperback handy but I remember a box on the back containing something like: "Those who believe in courtesy at least to living authors will buy only this edition."
3.27.2009 2:53pm
einhverfr (mail) (www):
ASlyJD:

Moral rights could expire and could eliminate some disincentives for artistic works.
3.27.2009 3:03pm
Soronel Haetir (mail):
jfb2252,

That note is still printed in the BB editions afaik. It may even have been back ported to the HM editions.
3.27.2009 3:03pm
jfb2252:
Addendum to 2:53pm comment:

I suspect that the LOTR would have remained obscure for much longer, if not forever, had Ace not distributed so many copies with great cover art.

Ace (and Lin Carter) proved with LOTR that there was a market for adult fantasy.
3.27.2009 3:06pm
Soronel Haetir (mail):
Also, I would add that I'm not sure Tolkein needed any sort of incentive structure. He seemed to be very willing to create without any expectation of an eventual payout.

Perhaps the incentive was enough to make the push from rotted random ideas to a finished work, reading his WiP it is somewhat amazing that any of it satisfyed him enough to release to a publisher.
3.27.2009 3:14pm
Libertarian1 (mail):
Eliminating copyright is insane. Limiting copyright to 10 years may make sense in certain contexts involving technology perhaps, but is otherwise fundamentally theft.

If I write the next, say, Lord of the Rings, the law should not be that, 10 years and one day later, anyone can reprint the books or make a movie without having to obtain my permission or pay me anything. While I'm not unsympathetic to the idea that life plus 70 years is too long, it is much preferable to 10 years.



If I discover the cure for cancer, spend 10 years doing the required research, the law says, 10 years and one day later, anybody can manufacture and market my drug without getting my permission.

Why should Disney or an author have the rights I
am not allowed to keep?
3.27.2009 3:19pm
ASlyJD (mail):
Because nobody needs Disney. And FWIW, Disney will lose its copyright in Mickey Mouse in 2018.

Good luck doing anything with a famous trademark though.
3.27.2009 3:34pm
statfan (mail):
Sam H, what makes you think Stallman is paid by taxpayers?

He's not. He lives off of his MacArthur and Takeda grants and speaking fees. He has never been paid by the government.
3.27.2009 3:36pm
billb:
AslyJD: Re: Mickey copyright expiration. I'll believe it when I see it.

statfan: Stallman probably spent some time as a grad student funded under federal grant programs, but I don't know that for a fact.
3.27.2009 3:40pm
einhverfr (mail) (www):
Libertarian1:

Patents and copyrights are very different. And currently I think patents expire after 17 years. Originally, patents were extendable to 14 years and copyrights were extendable to 28 years. I think generally we should be moving back to that.

There are a few other elements I would like to see in copyright reform, besides the term moved back into what is reasonable:

1) Compulsatory licensing of out of print works. If the work is out of print, I think that compulsatory licensing needs to be set at a royalty fee of 3% (which is half of the industry-standard royalty paid to the author in book publishing). The incentive ought to be to continue to make works available. All that should be required in this case is notice delivered to the copyright holder.

2) Software ought to be covered by a separate copyright-like instrument that expires after 5 years and requires registration of the source code, not just the binary.

3) As an alternative to #1 and #2, copyright protections could be applied only to works which are actively commercially distributed, and when distribution ceases the author would AUTOMATICALLY regain all rights, and would have one year to bring the work back to market before losing them for keeping the book out of print.
3.27.2009 3:44pm
Oren:

The Tolkien estate having the rights to LOTR didn't prevent the LOTR movies from being made; the same is true with Harry Potter, or James Bond, or The Watchmen, just to name a few franchises off the top of my head.

You do realize that millions were spent litigating the Watchmen movie (which turned out a dub anyway).
3.27.2009 3:48pm
ASlyJD (mail):
ein,

I was referring more to the vignette of say, Heinlein writing "Battlefield Earth" then deciding "No, I would like the money, but I don't want to release this and have someone write another story about Johnny Goodboy Taylor." Or Rand and John Galt. Authors write stories because they want to write, they publish because they like the money, and the question of moral rights in their works doesn't generally factor into the calculation.

And I say this as one who writes and publishes online.
Shameless plug: Amy Sly
3.27.2009 3:49pm
Tony Tutins (mail):
Without copyright, there can be no licensing, so Stallman must support copyright.

One thing not mentioned: Stallman hates the Kindle, because it makes it impossible to share, lend, or sell the books one has bought.
3.27.2009 3:54pm
Guest12345:

Sam H, what makes you think Stallman is paid by taxpayers?

He's not. He lives off of his MacArthur and Takeda grants and speaking fees. He has never been paid by the government.


True. But there has been government funded software that ended up under the GPL license in violation of the law. For example SELinux was developed by the NSA. According to the law, such things are required to be public domain. But according the the Linux license, distributed changes must be covered by the v2 GPL license. The NSA distributed their changes.
3.27.2009 4:16pm
billb:
Guest12345: Can you cite chapter and verse on that? Plenty of software is developed by the government and government funded projects that isn't Public Domain.
3.27.2009 4:25pm
einhverfr (mail) (www):


Personally, I've always wondered whether Stallman really wants to see the end or even a substantial weakening of copyright law if that would weaken his beloved recursive licenses. A lot of his rhetoric seems to focus on how restrictive licenses fundamentally infringe on the rights of computer users.


I would add that a lot of the restrictions go well beyond copyright-related stuff too. For example, how many people are connected to the web server really is unaffected by copyright law. I doubt that if copyright law suddenly ceased to affect software that it would have much of an impact on these sorts of restrictions since they tend to rely on contracts instead.

Exempting software from copyright would also substantially weaken the GPL since suddenly it would become possible to add restrictions contractually to it without violating upstream copyright licenses.
3.27.2009 4:25pm
einhverfr (mail) (www):
Guest12345:

True. But there has been government funded software that ended up under the GPL license in violation of the law. For example SELinux was developed by the NSA. According to the law, such things are required to be public domain. But according the the Linux license, distributed changes must be covered by the v2 GPL license. The NSA distributed their changes.


Nothing prohibits the NSA's changes from being public domain, even if the work as a while would still be under the GPL.
3.27.2009 4:27pm
Guest12345:

Can you cite chapter and verse on that? Plenty of software is developed by the government and government funded projects that isn't Public Domain.


Section 105:

Copyright protection under this title is not available for any work of the United States Government,
3.27.2009 4:33pm
einhverfr (mail) (www):
Douglas Wrote:

You can only get statutory damages if the copyrighted work had been registered within 3 months of its publication.


Re-read the USC section you referenced. This means no statutory damages for infringement which occurs prior to the date pf registration unless it is registered within 3 months of its first publication.

What this means is that if you register within 3 months of first publication and someone else is illegally distributing our work, you can look back to the moment of publication for statutory damages. If you register after that point, you can only look back to the date of registration for statutory damages. Of course at $35 to register.....
3.27.2009 4:36pm
einhverfr (mail) (www):
Guest12345:

Ok, so suppose I take the NSA's original patches and port them to OpenBSD. That would strike me as perfectly legal.

However, if another contributor doesn't put their changes to the NSA patches in the public domain, that would strike me as being something that couldn't be immediately moved over and incorporated in the other project without specific permission.

Nothing prohibits the government from contributing to a copyrighted work. It just prevents the government from asserting copyright over their contributions.
3.27.2009 4:40pm
Guest12345:

Nothing prohibits the NSA's changes from being public domain, even if the work as a while would still be under the GPL.


The FSF disagrees. Your interpretation would make the GPL utterly toothless as it would be trivial to work around.

Additionally, they specifically released it under the GPL.
3.27.2009 4:42pm
einhverfr (mail) (www):
Guest12345:

The FSF disagrees.


No they don't. The FSF does NOT say you can't include public domain works inside a GPL program, but rather they say you can.

I am not the world's biggest fan of the FSF, but if I am going to criticize them I figure I should at least try to do so fairly :-)
3.27.2009 4:47pm
einhverfr (mail) (www):
Also if the statute bars the federal government from asserting copyright, wouldn't that make the work public domain regardless of whatever the NSA says?

I.e. the statute says copyright protections don't apply. It doesn't say that government shall make no representations to the contrary, does it?
3.27.2009 4:49pm
Guthram Est:
jfb2252:

As for Lord of the Rings, Tolkien let the US copyright lapse after 28 years. Lin Carter, the fantasy editor of Ace books, knew of LOTR and published the three volumes as soon as they were out of copyright.


That can't be right, as LOTR was published in the early 50s and I remember the Ace edition on bookstore shelves by the late 60s.
3.27.2009 4:56pm
Guest12345:

Ok, so suppose I take the NSA's original patches and port them to OpenBSD. That would strike me as perfectly legal.


Yes, as the NSA produced code is required by law to be without copyright protection.


However, if another contributor doesn't put their changes to the NSA patches in the public domain, that would strike me as being something that couldn't be immediately moved over and incorporated in the other project without specific permission.


Also true as changes to public domain works are not public domain, they belong to the creator. The creator can do anything they want with the resulting code. However the original public domain work continues to exist in the public domain.


Nothing prohibits the government from contributing to a copyrighted work. It just prevents the government from asserting copyright over their contributions.


Ah, but anything a government employee produces (excepting the statutory exceptions) is public domain/without copyright. So the Govt. Agency A could create from scratch some piece of software. That software is automatically not copyright protected inside the US. It cannot be released under license as it is not protected by copyright law.

On the other hand Agency B cannot derive and distribute a work based on a third party product that has a license requiring changes to be distributed under the third party license.

Agency B can only create non copyright protected works and the third party license requires copyright protection. There is an unresolvable conflict.

---

All this of course is re. the Federal Government. States are not covered by 105.
3.27.2009 4:56pm
Guest12345:
No they don't. The FSF does NOT say you can't include public domain works inside a GPL program, but rather they say you can.


That's not the question at hand. Ask them instead if you can take their GPLed product, make substantial changes and release your changes as public domain. The answer is no.
3.27.2009 5:00pm
einhverfr (mail) (www):

Ah, but anything a government employee produces (excepting the statutory exceptions) is public domain/without copyright. So the Govt. Agency A could create from scratch some piece of software. That software is automatically not copyright protected inside the US. It cannot be released under license as it is not protected by copyright law.


Hmmm... Not quite sure I follow this. If the NSA releases a patched Linux kernel, the Linux kernel is still under copyright, so the work as a whole, including the NSA patches must be released under the GPL v2. The NSA patches as separate matter would be public domain, however. So I think the difference becomes a matter of what is being licensed.

Now this sort of thing is possible under the GPL v2 because it is based solely on copyright law and has few restrictions other than as they relate to copyright. The GPL v3 is a fundamentally different matter, and I am not at all sure one can incorporate public domain code in GPL v3 software as the reason has to do with what is considered to be an "additional permission" under the license and who can remove such additional permissions. If I were to release changes to a GPL v3 program I would probably license my patches under the terms of "the GPL v2 or later"....
3.27.2009 5:02pm
Guest12345:

Also if the statute bars the federal government from asserting copyright, wouldn't that make the work public domain regardless of whatever the NSA says?


If they choose to distribute the GPL requires that they assert copyright to their derivative work. The law requires that the NSA not assert copyright.
3.27.2009 5:07pm
einhverfr (mail) (www):

That's not the question at hand. Ask them instead if you can take their GPLed product, make substantial changes and release your changes as public domain. The answer is no.


Can you provide a citation for this? And in particular a citation which states this as regards the GPL v2? The GPL v3 is a different matter entirely and outside the bounds of this discussion at least as far as SELinux goes since Linux does not have a license upgrade path.

I don't doubt that Moglen would say such a thing, but I think he is wrong. Even Moglen would admit you don't HAVE to assert copyright on your changes, and would probably find a loophole there. Now the context lines in the patch might still be copyrighted by the original author, but here we are talking about the change set, not the representation of the change set.
3.27.2009 5:08pm
einhverfr (mail) (www):
(i.e. nothing in any version of the GPL obligates you to sue for infringement, and nothing in any version of the GPL prevents you from giving additional permissions to your work.)
3.27.2009 5:08pm
einhverfr (mail) (www):
Guest12345:

If they choose to distribute the GPL requires that they assert copyright to their derivative work.


Where in the GPL is one required to assert/enforce copyright?

I didn't see a must-sue clause last time I read it.
3.27.2009 5:10pm
Guest12345:

The NSA patches as separate matter would be public domain, however.


No. The patches are a derivative work. They have intimate knowledge of, and made changes to, the process, device and security models of the Linux kernel. Read section 2 of GPLv2. In particular 2b.
3.27.2009 5:14pm
David Schwartz (mail):
I realize that courts have been holding that violating a license is copyright infringement, but this is a very wrong and very dangerous holding. It's precisely the same as the "exceeds authorized access" cases. Private individuals are not supposed to get to define what conduct is criminal. Criminal violations require precise laws approved by legislatures, principles of notice, and so on.

If I lend you my car and tell you to be back before seven, if you're back at seven-thirty, you have *not* committed car theft. The Courts have been losing this very vital principle.

Copyright infringement should exist only where the law says so. The law says you can't make a derivative work, so if you do so without permission, that's copyright violation. If you exceed the terms of your permission, that's breach of contract.

Private citizens should not be allowed to convert breaches of their contracts into crimes.
3.27.2009 5:16pm
David Schwartz (mail):
I realize I was a bit confusing up above. I don't mean literally criminal. I mean getting the special legal benefits that charges of copyright infringement get such as a presumption of irreparable harm, statutory damages, and so on.

While these are not as serious as criminal penalties, the same principle applies. Private citizens should not be entitled to claim irreparable harm, statutory damages, and son where the nexus of their claim is not that the person wasn't authorized to do the act but that he didn't comply with the terms of that authorization.
3.27.2009 5:19pm
einhverfr (mail) (www):
Guest12345:

One way to interpret this is that the federal government is essentially bound by statute to allow people to do ANYTHING with their works without suing for copyright infringement. The GPL (all versions) only require that one give all downstream users the permissions given in the GPL itself. No version of the GPL prohibits granting permission to do things not covered by the GPL to elements one owns the copyright to.

There are some concerns with the GPL v3 as to removal of these additional permissions, but this is somewhat outside the scope of this matter. For example, if I contribute a patch to Linux and release my changes under the BSD license I can do this by simply saying "This patch is licensed under the GPL v2 as a whole, but anyone who wishes to use my changes under the BSD license below may do so."

No version of the GPL I have seen makes one responsible to upstream copyright holders for failing to sue infringers where the GPL is violated regarding copyrighted elements one owns.

Now, I have argued with Eben Moglen about what the BSD license allows, and we disagree strongly on whether or not one can take a piece of BSD-licensed software and add licensing restrictions on that piece itself without making changes (he says one can, but they aren't effective, I say one can't because the license requires passing itself on but is not contagious, so one can still sue over licensing violations for one's own copyrighted elements). While these differences of opinion are relatively minor (and probably moot in most circumstances), they pose some rather technical issues in how the GPL v3 should be read when GPL v3 and BSD code is mixed together in a single program. FWIW, the general opinion of the Software Freedom Law Center seems to be closer to my view.
3.27.2009 5:22pm
Guest12345:

(i.e. nothing in any version of the GPL obligates you to sue for infringement, and nothing in any version of the GPL prevents you from giving additional permissions to your work.)



Section 2b. It requires you to distribute your derivative work under the terms of the GPL. For a totally original work you can do anything you want. For a derivative work you cannot.


Where in the GPL is one required to assert/enforce copyright?

I didn't see a must-sue clause last time I read it.


As someone who has claimed to distribute free software, you seem pretty vague on what the various GPL licenses actually say.

I'm not talking about general case, like where a private party develops and original work and releases it under the GPL. I'm talking about government employees making derivative works of third-party GPLed software or government employees creating original works and distributing them under the GPL.
3.27.2009 5:26pm
einhverfr (mail) (www):
David Schwartz:

Part of the problem tends to be the fact that website TOS's and open source copyright licenses both tend to be phrased as adherence contracts. I do agree that this sort of thing can give way too much weight to adherence contracts. At the same time, I can see plenty of cases where violating the license really should be seen as copyright infringement.

For example, if I say "You can reproduce my paper verbatem" and you modify it, that is copyright infringement. If I say "You can reproduce my paper and make changes as long as this permission notice remains intact" and you remove the permission notice, I think that should be copyright infringement too. (This gets to the core of my concerns over GPL v3/BSD license compatibility.)

When you have something more complex, like the GPL v3 (which I don't think anyone fully understands), I don't think it is reasonable to assume that every license violation is copyright infringement. I think that copyright infringement in this case ought to require a higher burden of proof than contract violation. I think copyright infringement would still be possible to find (for example, in closed-source distribution of GPL-licensed code), but anything short of wilful disregard for the license ought to be seen as a contract issue rather than a copyright issue.
3.27.2009 5:32pm
Guest12345:

One way to interpret this is that the federal government is essentially bound by statute to allow people to do ANYTHING with their works without suing for copyright infringement. The GPL (all versions) only require that one give all downstream users the permissions given in the GPL itself. No version of the GPL prohibits granting permission to do things not covered by the GPL to elements one owns the copyright to.


The GPL requires restrictions on derivative works. To enforce those restrictions you must own copyright. The government cannot create a work under copyright. I can't see what is so difficult to understand.

Try this:

If one wants to do A, one must do B.
One can't do B.
Therefore one can't do A.
3.27.2009 5:38pm
einhverfr (mail) (www):
Guest12345:

I am somewhat vague about various versions of the GPL because:
1) I don't think anyone understands the GPL v3
2) My view of the GPL v2 is fairly different from Stallman's. I don't see any reason why dynamic linking would be covered under the scope of the GPL unless the result was a change in other covered elements (for example, screen display in a game). However, I think it is anyone's guess how courts will interpret the GPL v2 in some of these distinctions.

First, copyright works on a "per-element basis" which is a fairly vague measure, but let me see if I can explain this (IANAL, etc).....

Suppose you take a scholarly paper. In it there are expressive structures and these range from the overall structure of the paper to individual sentences. If I am trying to see whether a second scholarly paper is derivative of this paper, I look at these expressive structures and see if I can detect sufficient copying of them from one paper to the other that it would be subject to the copyrights of the first paper.

In a computer program, we have some added problems, in that functional elements are not subject to copyright protection, so typically we have to abstract the program first so we can see those expressive elements. Then we have to filter out what is not protected by copyright law. Then we can do our comparison. This is the famous Gates Rubber test.

Copyright law only prohibits certain sorts of activities, such as distributing the software, or making derivative works. These terms are all defined in copyright law, and the definitions collide fairly heavily with the ideas in the GPL v2. Hence I am not sure what courts will say is permitted vs not.

Ok, so back to the question of what the GPL requires.

All versions of the GPL, whatever the other requirements, require passing on a set of rights to the work in question which must include every right the GPL provides. The intent here (and the text supports this interpretation) is that you can't encumber a GPL product with additional restrictions through your patches as long as your patches require permission from the copyright holder to distribute in the first place. Are you with me so far?

Nothing in any of these texts suggests that you can't additionally provide any other compacts, permissions, etc. to the copyrighted elements you created in your patches provided that you don't offer additional restrictions.

So if you release a change into the public domain, what is the effect? Are you allowed to do this with substantial changes? This seems to be what we are arguing.

If you release a change into the public domain, then any copyrighted elements YOU own in it are now free for anyone else to use. Any copyrighted elements which were taken from the original software are still owned by the original author who can assert and enforce that copyright however they see fit.

Now, I noticed a careful reading of the NSA site DOES NOT say that THEY are licensing their works under the GPL. Rather they are saying the works are licensed under the GPL because the patches etc. may still contain elements subject to the copyrights of others outside the federal government.

So the effect of releasing such a patch under the public domain is simply that more elements get discarded during the filtration portion of the Gates Rubber test. Thus it is allowed.

Nothing in the GPL prevents one from releasing one's modifications to a GPL program under a superset of rights to the GPL (i.e. a set of copyright permissions of which the GPL is a proper subset). The public domain is such a superset and so it does not conflict.

There is one problem spot however. Suppose you release your patch under the public domain. Suppose I take that patch and use it in my closed source app. Suppose someone discoveres that there were copyrighted, expressive elements in the original GPL program that got into your patch. I might be liable to the author of those elements (but not to you!) for my use of your patch.
3.27.2009 5:53pm
einhverfr (mail) (www):
Guest12345:

The GPL requires restrictions on derivative works.


Yes, but only on the copyrighted elements originally licensed.
3.27.2009 5:56pm
David Schwartz (mail):
For example, if I say "You can reproduce my paper verbatem" and you modify it, that is copyright infringement. If I say "You can reproduce my paper and make changes as long as this permission notice remains intact" and you remove the permission notice, I think that should be copyright infringement too. (This gets to the core of my concerns over GPL v3/BSD license compatibility.)
You have carefully picked cases that are distinguishable from the GPL case.

In the first case, assuming the modification amounts to the creation of a derivative work, I agree with you. But if it doesn't, then I disagree. Our private definition of what is verbatim and what constitutes a change can't bind copyright law. However, if I make a derivative work without permission, that's another story.

Copyright law says I can't make a derivative work. If I do so without permission, I've broken the law. If I do so but break the terms of the permission, I've broken the contract. It's really that simple.

In the second case, I agree if and only if I violated 17 506(d). If not, it's a normal contract dispute. 17 506 (d) states the terms under which it is copyright infringement to remove a copyright notice. You cannot add a new such law by contract.

It can't be an issue of copyright law whether or not I removed a permission notice against our agreement. That's a contract dispute.
3.27.2009 5:58pm
David Schwartz (mail):
The GPL requires restrictions on derivative works. To enforce those restrictions you must own copyright. The government cannot create a work under copyright. I can't see what is so difficult to understand.
The GPL does not "require restrictions on derivative works". It generally creates them as the automatic consequence of its operation. But if it fails to do so, then it simply fails to do so.

The GPL doesn't require anyone to enforce anything either. If you can enforce the GPL and want to, go ahead and do it. If you can't, again, oh well. The GPL fails.

The GPL doesn't say "you may only use works covered by the license if you can somehow ensure that your use furthers the intended goals and mechanisms of this license". It tries to do that by its own terms, but if it fails, that's not the licensee's problem.
3.27.2009 6:00pm
einhverfr (mail) (www):
Guest12345: Let me give you a tangible example of what I am talking about.

Suppose I have a product I built based on a GPL product. I release that under the GPL. Three years later, Acme Corp builds a product based on my product, and releases it closed source. I am under NO obligation to sue Acme Corp, and even if I do sue them, I can seek to provide them with a paid license allowing them to distribute a derivative work of my program in a closed source environment.

However, the question of whether Acme Corp's version is derived from the original GPL version that I did not write is an entirely different question and would only be settled by an analysis of those two versions. The GPL does not preclude me from offering other license terms to my own work. It only binds me to the extent that my program is derivative of another one, specifically.

Courts (most notably in SCO v. IBM) have not held the idea that software derivation is contageous.
3.27.2009 6:11pm
juris_imprudent (mail):
Copyright protection under this title is not available for any work of the United States Government

Does not apply to the work-for-hire products that the U.S. govt pays for, which is much more common than the product actually being a work of the U.S. govt. I'm not sure who actually did the work on SELinux, but just because the govt pays for it, does not automatically make it a public domain product.
3.27.2009 6:13pm
Soronel Haetir (mail):
Okay, so's here's a question I think might make this entire line far more concrete and interesting.

Larry Wall was working at JPL when he created early perl. He created it in order to make certain of his job functions easier. Did he actually have the right to release it under the GPL? Did he actually have the right to release it it with any sort of restrictions?

And if not, would newer versions be considered enough of a derivitive work that any modern versions would be bound by those facts?
3.27.2009 6:16pm
einhverfr (mail) (www):
David Schwartz:

First, let's come back to the original question which looks at web site terms of service in the Lori Drew case as parallel to this issue. I think that there are cases where contractual violations can rise to the level of unauthorized access and I think that there are cases where courts have rightly recognized this. The AOL case is a good example, where someone continued to relay spam through the web server after receiving specific notice that this behavior would stop and was rightly held to be in violation of the same statute that Drew was charged with (though in a civil court rather than a criminal one). Similarly, if you are at a party at my house and misbehave, if I ask you to leave and you refuse, at some point this might become criminal trespass.

The fact that I let you in my house once does not immunize you from future trespassing charges from me forever, relegating these to contract disputes. It allows contract disputes within a reasonable scope surrounding the contract, and there is another larger scope where trespassing charges might still be possible.

This is the same process I would argue with copyright licenses. Within a reasonable scope surrounding the license it ought to be a matter of contract. At some further point (well outside the scope of the contract), copyright infringement claims ought to be possible.

Let's look at a few other cases.

Suppose you write software and give me permission to install it on every machine in my organization, giving me copyright permission to make copies to further that task (for example, copying it onto a central file server first). Suppose I then take it and publish it, and offer it as a free download to everyone. Is that copyright infringement? Or is it a mere contract violation? I think in that case, it ought to be copyright infringement.
3.27.2009 6:28pm
einhverfr (mail) (www):
Soronel Haetir:

Can't derivative works of public domain works still be subject to copyright in their own right?
3.27.2009 6:29pm
Andy Freeman (mail):
> Stretching current trademark law to cover fictional characters would probably produce anomalies and undesired effects

Since Mickey Mouse appears to be covered by current trademark law....

Note that the "Mikey Mouse" copyright which may expire in 2018 is just the copyright of very old drawings and movies. The stuff produced last year will still be under copyright for a long time.
3.27.2009 6:34pm
Guest12345:

Now, I noticed a careful reading of the NSA site DOES NOT say that THEY are licensing their works under the GPL. Rather they are saying the works are licensed under the GPL because the patches etc. may still contain elements subject to the copyrights of others outside the federal government.


Huh? This is what is says:


All source code found on this site is released under the same terms and conditions as the original sources. For example, the patches to the Linux kernel, patches to many existing utilities, and some of the new programs available here are released under the terms and conditions of the GNU General Public License (GPL). Please refer to the source code for specific license information.


There is no qualification to even suggest what you say. It's very clear: All source code. It exactly says that the work they provide is covered by the GPL.

Yes, but only on the copyrighted elements originally licensed.


Section 2b of GPL v2:

You must cause any work that you distribute or publish, that in whole or in part contains or is derived from the Program or any part thereof, to be licensed as a whole at no charge to all third parties under the terms of this License.


The NSA cannot meet that requirement, therefore they cannot redistribute.

--

The GPL does not "require restrictions on derivative works". It generally creates them as the automatic consequence of its operation. But if it fails to do so, then it simply fails to do so.


No, section 2b requires that the derivative work, in its entirety, be covered under the GPL. Otherwise you have no license to distribute your derivative work.

The GPL doesn't say "you may only use works covered by the license if you can somehow ensure that your use furthers the intended goals and mechanisms of this license". It tries to do that by its own terms, but if it fails, that's not the licensee's problem.


The GPLv2 is a distribution license. It has nothing to do with usage. That's why it's a question of copyright and not a more general case of contracts.
3.27.2009 6:38pm
Andy Freeman (mail):
> [Stallman] lives off of his MacArthur and Takeda grants and speaking fees. He has never been paid by the government.

He's been at this long before he received those grants or was popular enough to get speaking fees.

According to http://en.wikipedia.org/wiki/Richard_Stallman , he has worked as a programmer at MIT AI Lab and was an MIT grad student in Physics. It's possible that the latter wasn't supported via a govt grant but the vast majority of MIT's AI Lab funding in those days came from the govt.
3.27.2009 6:42pm
Guest12345:
einhverfr:

Guest12345: Let me give you a tangible example of what I am talking about.


Your example has nothing to do with anything I'm talking about. I don't dispute anything you relate re. private parties.

juris_imprudent:

Does not apply to the work-for-hire products that the U.S. govt pays for, which is much more common than the product actually being a work of the U.S. govt. I'm not sure who actually did the work on SELinux, but just because the govt pays for it, does not automatically make it a public domain product.


That is understood. However this:

The National Security Agency (NSA)
Researchers in NSA's National Information Assurance Research Laboratory (NIARL) designed and implemented flexible mandatory access controls in the major subsystems of the Linux kernel and implemented the new operating system components provided by the Flask architecture, namely the security server and the access vector cache. The NSA researchers reworked the LSM-based SELinux for inclusion in Linux 2.6. NSA has also led the development of similar controls for the X Window System (XACE/XSELinux) and for Xen (XSM/Flask).


Sure sounds like NSA employees.

Soronel Haetir:

Larry Wall was working at JPL when he created early perl. He created it in order to make certain of his job functions easier. Did he actually have the right to release it under the GPL? Did he actually have the right to release it it with any sort of restrictions?


If he was an employee of JPL then no that early version of perl had no copyright protection. That assumes that JPL employees are government employees and not Cal Tech employees.
3.27.2009 6:54pm
einhverfr (mail) (www):
Guest12345:

I think you fundamentally misunderstand the way copyright works, including where the government is involved.

Suppose I publish "The 20 Greatest Supreme Court Cases of All Time" and do nothing other than select and order them. The book, consisting entirely of content in the public domain, would still be subject to my copyright. If you copy my book I can sue you. Now, if you copy one case out of my book, that might not be actionable (de minimis considerations). But the book itself would be copyright-worthy and subject to whatever terms I put on it.


You must cause any work that you distribute or publish, that in whole or in part contains or is derived from the Program or any part thereof, to be licensed as a whole at no charge to all third parties under the terms of this License.


I think you are misreading this. If I add public domain elements (by my authorship or not), this does not prevent the work which is derived from the program from being licensed under the terms of the terms of that license. I could even add elements I might license under the BSD license or assign my changes to the public domain.


No, section 2b requires that the derivative work, in its entirety, be covered under the GPL.


Sure, but this does NOT prohibit you from adding additional permissions to elements you created up to and including the public domain. It just prevents you from adding those permissions to elements included in the process of derivation.

Let's take a simple example. Suppose I port a driver from OpenBSD to Linux for a wireless card. There is no reason I can't license my changes under the BSD license even if it is now distributed as part of a work under the GPL because the licenses are compatible (i.e. the BSD license grant is a proper superset of the GPL license grant). In fact, the SFLC official recommendations for maintaining such code in an GPL appication are that one should NOT change the licensed before significant changes are made, and even after that, one is not REQUIRED to change the license.
3.27.2009 7:03pm
einhverfr (mail) (www):
Guest12345:

You seem to be arguing that the federal government is under special restriction not to release contributions to open source projects. I think that this is untenable and a simpler reading is that any elements created by federal government employees are simply not covered by copyright. This doesn't mean that the patches created by government employees are public domain, but rather that copyright elements created by such employees over the course of engineering the change set are public domain.

The basic thing is that the government can't sue you for copyright infringement in this matter, and so it is solely a question of whether there are private parties who hold copyright to elements in the change set.
3.27.2009 7:08pm
einhverfr (mail) (www):

Sure sounds like NSA employees.


It still doesn't mean that the NSA is licensing their work specifically under the license. It could be just that they are not representing that the work is sufficiently free from outside encumbrances (other elements from in the Linux kernel) to be considered fully public domain.
3.27.2009 7:13pm
Splunge:
he surely deserves a great deal of the credit for the success of open source software

That is such ahistorical bullshit. Every single bit of the GNU project with the exception of Emacs (which almost defines the concept of "feature bloat" and has had just about zero substantial influence on modern computing) consisted of porting brilliant ideas from somewhere else.

The success of the free software movement, at least insofar as Stallman is concerned, is his ability to parasitize the superb work of the people who designed and implemented PL/I, Multics, C and Unix and pass this off as some kind of revolutionary creativity.

I was never surprised that after the magic GNU project failed to deliver -- was essentially hijacked by Torvalds -- the original FSM showboater "set his sights" on copyright law. It's pretty much always the case that technical people who prove to have no lasting ability to be usefully creative go into management, and, if they can't even do that very well, become self-righteous paranoid crusaders against The Man, or against whatever tool the jihadi thinks The Man is using to keep him down.

Blech.
3.27.2009 7:41pm
leogex05 (mail):
. PS I'm using Ubuntu Linux OS at this very moment.


Shh.... I won't tell anyone.

A happy Sidux user.
3.27.2009 7:43pm
einhverfr (mail) (www):
Splunge:

The success of the free software movement, at least insofar as Stallman is concerned, is his ability to parasitize the superb work of the people who designed and implemented PL/I, Multics, C and Unix and pass this off as some kind of revolutionary creativity.

You do realize that that pattern is not at all contained to the open source world. In fact it also describes pretty much all major software on the market.

Where did DOS come from? How about Windows?

At the same time, there have been revolutionary advances in software which have come out of the Free Software world. These include:
1) W (the ancestor of X, and also of other modern GUI's)
2) X (you can do REALLY cool stuff with X)
3) Kerberos
4) Zephyr (the prototype of pretty much all IM stuff)
5) Mozaic
6) CERN-HTTPD
7) HESIOD (which IMO was quite a bit better than LDAP which has largely replaced it)
8) TeX
9) LaTeX

I am sure I am missing a bunch. However, the fact is that most successful open or closed software is basically a clone of something that came earlier. Innovation is rare on both sides and IMO overrated.
3.27.2009 7:55pm
Guest12345:

I think you fundamentally misunderstand the way copyright works, including where the government is involved.


I think it's you. I've already provided an exact quote of the relevant statute.

I think you are misreading this. If I add public domain elements (by my authorship or not), this does not prevent the work which is derived from the program from being licensed under the terms of the terms of that license. I could even add elements I might license under the BSD license or assign my changes to the public domain.


You seem to misunderstand what I'm talking about. I'm not saying add some public domain code to a copyright protected program. I'm talking about taking a copyright protected program, creating a derivative work and releasing it into the public domain. As in taking GPLed code, having government employees make a derivative program and then, by law, being required to provide it to me (assuming they distribute) without the restrictions required by the GPL.

In the SELinux case, there are numerous sections of code that only make sense in the context of the Linux kernel. They use internal linux kernel data structures. They require linux specific code flow. They cannot be built independent of linux. There is no case for arguing that they are not a derivative work.

Sure, but this does NOT prohibit you from adding additional permissions to elements you created up to and including the public domain. It just prevents you from adding those permissions to elements included in the process of derivation.


Sure. If you have an independent piece of code and you just use a piece of glue code to stick the two together, then your independent code won't be covered. But I'm not talking about independent code here. SELinux isn't independent of Linux.


You seem to be arguing that the federal government is under special restriction not to release contributions to open source projects. I think that this is untenable and a simpler reading is that any elements created by federal government employees are simply not covered by copyright. This doesn't mean that the patches created by government employees are public domain, but rather that copyright elements created by such employees over the course of engineering the change set are public domain.


No I'm pointing out that the government cannot release anything under a license. It doesn't matter if that doesn't please you or seems untenable to you. It's just what the law is. If what the government employee wants to distribute requires a license, because it's derived from a GPLed code base say, then they cannot distribute it.
3.27.2009 7:58pm
einhverfr (mail) (www):
One more thing. There are three showcases for perpetual vaporware:

1) Duke Nuke'm Forever
2) GNU HURD
3) Perl 6.

I generally joke that Duke Nukem Forever is being rewritten in Perl 6 with HURD as the platform of choice, and we should see released together....
3.27.2009 7:58pm
Guest12345:

It still doesn't mean that the NSA is licensing their work specifically under the license.


No. That would be the page where they explicitly say: THIS IS LICENSED UNDER THE TERMS OF THE GPL!!!! Their licensing page is very clear on this. If I say something is red, I'm not suggesting, hinting or insinuating that it is blue with some possible red contamination. I'm saying it's red.

The line you quote was identifying that NSA employees were involved in the development. Which brings the conflict.

Christ. I'm turning into jbg.
3.27.2009 8:08pm
David Schwartz (mail):
Suppose you write software and give me permission to install it on every machine in my organization, giving me copyright permission to make copies to further that task (for example, copying it onto a central file server first). Suppose I then take it and publish it, and offer it as a free download to everyone. Is that copyright infringement? Or is it a mere contract violation? I think in that case, it ought to be copyright infringement.
Of course this is copyright infringement. 17 USC 106(3) specifically makes distribution to the public a reserved right, one not granted under the contract in your hypothetical.

But my distributing the work to the public is copyright infringement only because 17 USC 106(3) says you need permission to make a work available to the public. The copies were lawfully made under copyright law.

A better example would be if the contract says I can only sell the work to the public on weekdays and I sell it to the public on weekends. Copyright law says nothing about the days you can sell something.
3.27.2009 8:17pm
einhverfr (mail) (www):
Guest12345:

A couple points....

First, you misquoted the statute. The full sentence was:

Copyright protection under this title is not available for any work of the United States Government, but the United States Government is not precluded from receiving and holding copyrights transferred to it by assignment, bequest, or otherwise.


This tells me:
1) The Federal Government can in fact hold copyrights from others and re-license derivative works in accordance with those agreements.
2) The Federal Government can hold copyrights for their own work through other laws. For example contractors can assign rights, or the federal government can register their copyrights in foreign countries but forego protection in the US.

This means that, if the work was prepared by NSA employees, then we can assume that the work is public domain only within the scope of the US. It is quite proper for the NSA to license their work globally under the GPL in this case. This means that the work could be properly licensed indernationally under the Bern Convention.
3.27.2009 8:20pm
David Schwartz (mail):
The fact that I let you in my house once does not immunize you from future trespassing charges from me forever, relegating these to contract disputes. It allows contract disputes within a reasonable scope surrounding the contract, and there is another larger scope where trespassing charges might still be possible.
This seems like an odd straw man. Nobody is saying that if you once authorize someone to come on your property you can never revoke that authorization. But *you* have to unequivocally revoke the authorization for their to be a trespass.

What you cannot do is say "you may only come on my property wearing a green shirt" and then use trespass law to frame our dispute over whether my yellow-green shirt was "really green". This is actually what GPL enforcement as copyright violation does, taking an argument over whether the correct source was made available or appropriately available, and giving it presumptions of irreparable harm, statutory damages, and so on.

Arguments over what source code was released and how are as alien to copyright law as arguments over how green or not green a shirt is. They should not be entitled to the special processes that actual copyright violation claims get.
3.27.2009 8:29pm
David Schwartz (mail):
Oh, one more thing: Am I the only one who finds it more than a little ironic that that "information wants to be free" folks are the ones arguing for these astonishing increases in the powers of copyright holders and their ability to control the use of their works?

Is winning this battle really worth losing the war?
3.27.2009 8:33pm
einhverfr (mail) (www):

A better example would be if the contract says I can only sell the work to the public on weekdays and I sell it to the public on weekends. Copyright law says nothing about the days you can sell something.


In that specific case, I would agree that it should be a contract issue, assuming no aggrevating factors.

However, I think that one could have malicious behavior sufficiently far outside the contract as to be a separate matter. For example, I say "You may print and sell 1 copy of my book to the general public" and you print and sell a million copies while I have, in the mean time, given an exclusive license to someone else to do that. I think that at some point, the behavior is so far outside what should be expected within the bounds of the contract, that one should not consider its existence.

If I say "You can enter my house tomorrow" this doesn't preclude me from pressing trespassing charges at some later point and you can't just hide behind my license and say it is a civil matter and a contract dispute.
3.27.2009 8:48pm
einhverfr (mail) (www):

This is actually what GPL enforcement as copyright violation does, taking an argument over whether the correct source was made available or appropriately available, and giving it presumptions of irreparable harm, statutory damages, and so on.


Arguably not. Arguably, the GPL terminates on the first violation. So arguably the first act of distribution in violation of the license might be a contract matter. Arguably after that point, you no longer have any right to distribute the software, so if you continue to do so, you are infringing on the copyrights.

Personally I think the courts need to provide more room around such licenses in terms of what is a contract matter or what is a copyright matter. Differences of interpretation of an open source license should never subject one to copyright infringement matters. This means a good faith error or dispute ought to be settled through contract means.

So I agree with you to a strong point. I just think there are limits to how far a having any copyright license agreement immunizes one against copyright infringement.
3.27.2009 8:52pm
einhverfr (mail) (www):
David Schwartz:

One possible bright line would be that behavior up through the first court decision could be deemed to be contract violation if it is not fundamentally abusive of the license, and then if the behavior continues after the initial court decision, the matter could be an issue of copyright infringement. I would support this line.

However, as an author, I would not want someone to say "You told me I could pass this copy off to a mutual friend of ours. I published it and sold 100000 copies, and that is thus just a contract dispute because your initial permission in some way involved making it available to a member of the public!"
3.27.2009 8:55pm
MLS:
If one deems it appropriate to talk about 17 USC 105, it is useful to quote this remarkably short statute:

Copyright protection under this title is not available for any work of the United States Government, but the United States Government is not precluded from receiving and holding copyrights transferred to it by assignment, bequest, or otherwise.

Note key phrases such as "under this title" and "not precluded". These have not gone unnoticed by federal agencies. Some try and mandate assignments of copyright from its private sector contractors. Some take works of the USG and assert them internationally. Some have even taken the very questionable step of claiming that while government works are not protected under copyright because of Section 105, they are fully enabled to claim rights under trade secret law.
3.27.2009 9:02pm
einhverfr (mail) (www):
David Schwartz:

Am I one of the "Information Wants to be Free" folks?

Actually, I just want to live an ok life, put food on my table, make a difference in the world, etc. I built my business on the prospect of helping folks use open source software.

Seriously, copyright doesn't protect information in US law anyway. It only protects expression, so I don't see what copyright has to do with freedom of information except as the law has become extremely bloated.
3.27.2009 9:02pm
einhverfr (mail) (www):
Guest12345:

You say the federal government cannot release works under a license (like the GPL). What exactly is such a license in your view?
3.27.2009 9:35pm
Guest12345:

First, you misquoted the statute. The full sentence was:



Copyright protection under this title is not available for any work of the United States Government, but the United States Government is not precluded from receiving and holding copyrights transferred to it by assignment, bequest, or otherwise.


This tells me:
1) The Federal Government can in fact hold copyrights from others and re-license derivative works in accordance with those agreements.
2) The Federal Government can hold copyrights for their own work through other laws. For example contractors can assign rights, or the federal government can register their copyrights in foreign countries but forego protection in the US.


I disagree that I misquoted, I quoted the relevant section. I even included the comma so that people with sharp eyes would realize that it wasn't a complete sentence. I never disputed that the US Government can own copyright. I never disputed that the government can be assigned copyright. My only point is that in the US, the federal government cannot exert claims of ownership to the works of federal employees employees, except where allowed by law.

That fact means that SELinux, derived from Linux by US government employees, should not have been distributed since they could not satisfy the requirements of both the GPL and the law.
3.27.2009 9:37pm
Guest12345:

You say the federal government cannot release works under a license (like the GPL). What exactly is such a license in your view?


If you want to continue this you have to understand some things:

1) I'm only talking about works produced by US government employees.
2) I'm only talking about inside the US.
3) I'm not talking about the US Dept. of Commerce or any other agency that has statutory exemptions.

To answer your question, that would be any license. Qualified by my three points, the US Government has no copyright protections on works produced by government employees.
3.27.2009 9:50pm
David Schwartz (mail):
I'm not sure a bright line rule is really possible. But I think there are clearly cases on one side and cases on the other side. The basic question is this -- if this is copyright infringement, what is the least that would have had to change to make it not copyright infringement?

IMO, a sensible rule is that it has to be obvious at the time that the act occurred, wasn't authorized. Any plausible contract argument that the act was authorized should be sufficient to defeat an infringement claim.

Certainly one strong enough to survive summary judgment should be sufficient.
3.27.2009 10:14pm
einhverfr (mail) (www):
David Schwartz:

Then we substantively agree. I would add that good-faith, inadvertent errors, and minor contract violations should be contract matters.

For example, the contract says in big bold letters, "You may only display my content on your web site on weekdays" and you display it on week ends by accident. That should be a contract matter even though it was obvious at the time when it happened that such was not permitted. Similarly you overlook a clause in a reasonable way and rectify the problem when it was pointed out, that, again, should be a contract matter.

Where I would draw the line is where there is reckless disregard for the contract, bad faith, and/or particularly egregious violations.

So after this, it seems that I might hope courts would put even more under the contract umbrella than you.....
3.28.2009 12:26am
einhverfr (mail) (www):
But Guest12345:

SELinux is distributed internationally. Also if it is only public domain inside the US, it is perfectly reasonable for the government to attach an open source license to it. They can't enforce that license in the US, but they can internationally.
3.28.2009 12:28am
Guest12345:

But Guest12345:

SELinux is distributed internationally. Also if it is only public domain inside the US, it is perfectly reasonable for the government to attach an open source license to it. They can't enforce that license in the US, but they can internationally.


It's not that complicated. The GPL is a distribution license. They cannot meet the terms of the license in the United States. Therefore they cannot distribute in the United States.
3.28.2009 1:21am
J.T. Wenting (mail):
If copyright is abandoned, no creative person will be compensated for his or her work.
After all, why would a publisher publish a book if after a single copy is printed a million copies are created using office photocopiers or by scanning it into a computer and placing a PDF on some website?
Same with software, same with music.
The only money to be made in any creative work would be life performances and the merchandise sold there (which would not be protected either, anyone could create picture perfect copies and sell them or give them away, so don't dream of selling at a profit).

At the moment "free" software, the copyright on which Stallman guards viciously while claiming to want copyright laws scrapped (he only wants them scrapped for other people so he can steal their work and call it his own), survives solely because companies pump money into it in order to use it as components in their own commercial systems. If selling those systems no longer is a viable business because any pundit can legally create copies of it and sell them at a lower price (or give them away) that flow of money will dry up very quickly as those companies either go out of business or have to restructure their efforts.

When there's no more money to be made writing software, people won't choose to do it as a profession and won't spend the time and money to get trained as professionals (the same would be true in other creative fields as well, think architecture and industrial design).

While Stallman claims that a lot of very good people will pour over source code to create great works for free as a gift to society, that's a foolish assertion.
It may hold true for a generation at most while people with prior training maintain what's there, but with no or very few new and trained people entering the market from the bottom that workforce will quickly dry up as people leave it for lack of time (they'll have to retrain and take up other jobs) or because they die.

The same will hold true for other creative arts and crafts.
While there will still be work in some fields for creative craftmen (think woodworkers and stonemasons working on commission to create single pieces for restoring buildings) software engineering and creative writing will most likely die, as will recorded music (life music will continue as it did in the middle ages).

The perfect socialist world, where everything is owned collectively and noone gets compensated for his or her work because there's no need for compensation when everything is free.
Except of course that the rent still needs to be paid, the children still need to be fed, and the landlord and supermarket won't take kindly to your claims that everything is owned by the Collective and therefore you have a right to take it without payment.
3.28.2009 3:56am
Anonymous!:
Partial disclosure - I'm somewhat affiliated with the NSA, so I'm posting this anonymously.

Guest12345, IANAL but I think einhverfr's reasoning is sound, and if you guys are quoting the only relevant copyright statute, it sounds like he's right about the government being allowed to redistribute a work under their original licenses, even if it has been modified.

All this may be moot because I don't think we can be certain whether government employees or contractors wrote the actual code in question. For-profit contractors are fairly common in the NSA community.

In any case, I'm not sure whether the NSA has special standing in regards to intellectual property rights, but I know that they can and do file many patents:

http://www.google.com/search?q=nsa+patents

And yes, they are sometimes licensed, with fees, to the private sector.

Now, I don't know enough about IP law to know whether other federal agencies can do this, or where the NSA gets this authority from, but they have the authority to file patents, and possibly even secret patents whose term of protection begins upon someone else filing a similar patent with the US PTO.

And to J.T. Wenting, who claims that abolition of IP rights would lead to the end of creative works - I'm pretty sure creative works like music, poetry, literature, and art have existed long before copyright protections have. Things would change, sure, but to claim that people can't create quality works without strong IP rights is ridiculous. Take a look at the works of top military photographers like Andy Dunaway or Adam Mancini. I enjoy distributing my works, whether it's software code or political commentary or landscape photography, for free, because I enjoy creating it in the first place.
3.28.2009 8:02am
Soronel Haetir (mail):
I would actually think the problem is even worse, that the people in question didn't even think all that much about the legal aspects of what they were doing.

Somewhat like my question about perl above. Any private employer I've ever been associated with would have claimed ownership over a tool someone created to make their job easier. So, with the caveat that Wall may have actually been employeed by CalTech I see two possibilities: 1 the US government allows people to retain ownership over work private employers wouldn't or (2) no one realized that the general copyright release applied to them.

Even more problematic, even if he were working for CalTech, universities generally claim ownership over the work of their employees. Look at the general U of California licence for that case. (I realize CalTech is private, not part of the UofC but would expect them to have a similar requirement.)

I use perl as my example here because the early versions at least are clearly the work of a known individual without the derivitive work issues seen with selinux.

It would be interesting to see such a case litigated where the defendant claims that the plaintiff simply didn't have the right to attach a particular license. Also interesting would be whether that killed the licence on patches or not.
3.28.2009 9:08am
Roger_Z (mail):
I fully support Stallman's right to:

a) create software for free,
b) license such software any way he wants,
c) encourage others to license their software (whether they got paid for producing it or not) in precisely the same way that he does,
d) imagine and advocate for other ways for programmers to earn a living such as writing documentation and giving technical support

What I don't accept is his moral approbation for those of us (myself included) that choose to create proprietary software and then do (b) above.

And, I disagree with him about the nature of at least some software. In my case, I invent novel algorithms for natural language processing which are used to improve the medical transcription process. The way these algorithms are used is to encode them in software which is distributed in binary form to run on computers. Stallman thinks this is evil, since I am not making the source code available to the users. I say: my users could never have invented the algorithms which I implemented in code, and I am not forcing them to buy my code, so they must consider this to be a fair trade (their money for my binary code).

Would the users be better off if I included the source code in the purchase price, so that they could then modify the program to do something else that they wanted done? Possibly, but so what? If Stallman's argument holds, then all sales of goods and services which do not include full access to the underlying means of production in the purchase price are immoral. I conclude that Stallman supports unadulterated communism.
3.28.2009 9:31am
Ryan Waxx (mail):

@ Bonehead - Assuming for the moment that Andersen's estate did still control rights to the littlest mermaid, why do you believe that they wouldn't sell the rights to Disney for use in a movie...or that Disney wouldn't pay for them.


Why would you believe Disney would have bought the rights? Look at their list of movies... all or nearly all are of uncopyrighted stories. Snow white? The Hunchback of Notre Dame? The list goes on.

Don't simply assume that modern copyright straitjackets aren't causing problems based only on your wishful thinking... look at the EVIDENCE.

The evidence is that they in general don't use copyrighted works as the basis for their works, which suggests that Disney's acts of creation were facilitated by the absence of copyright.

The evidence is also that Disney is not providing the same useful source material function that the original story did, because no one in turn is creating (legal) works based off of Disney's works.

So essentially we are dealing with a group of companies who freely take others' ideas and produce new works from them, but when it comes time for them to be the sources, they are a creative desert judging from how many branches they sprout.

Modern copyright law and practice is the intellectual equivalent of deforestation.
3.28.2009 10:02am
Daniel J. Wojcik (mail) (www):
ASlyJD said:
I was referring more to the vignette of say, Heinlein writing "Battlefield Earth" then deciding "No, I would like the money, but I don't want to release this and have someone write another story about Johnny Goodboy Taylor."


Eeeeee! Hubbard, not Heinlein. 8^0
3.28.2009 11:20am
einhverfr (mail) (www):
Soronel Haetir:

The problem is not as you characterize. The question might be whether early versions of Perl were public domain, but it certainly doesn't affect whether current versions of Perl are public domain. The US Government doesn't claim ownership (and hence public domain) on work that occurs after employees leave, and since reasonably modest changes would be sufficient to create a copyright interest on the software.....

Unless you are saying you want to go back, take the Perl 1 source code and create some entirely different proprietary application from it, that issue is moot.
3.28.2009 12:04pm
einhverfr (mail) (www):
Roger_Z:

I like Tim O'Reilly's position, that the most fundamental freedom for a software developer is the choice of license.

BTW, I don't see anything immoral about closed-source software. However, it is a business model that I think is fundamentally more risky than an open one. It also scales better inside an organization, but there are other considerations too.

Once again, I see NOTHING wrong with, for example, EnterpriseDB's business. What they do is take the BSD-licensed software called PostgreSQL and release a closed-source version which has better compatibility with Oracle. The idea is to help people move away from Oracle at a lower cost, and their work is subsidized by the work of the PostgreSQL community.

However, this path of contribution is a two-way street. The pace of development with PostgreSQL is very high, so to simplify EnterpriseDB's code maintenance, they give back pretty much everything the PostgreSQL developers want. This works out great because there is a general consensus that Oracle does some things incorrectly and hence Enterprise gets to develop certain things that the PostgreSQL community doesn't want and they can sell that for a license fee. Stallman seems to consider that to be "theft" but really it is an arrangement that works out well for everyone.
3.28.2009 12:11pm
MCM (mail):
Would the users be better off if I included the source code in the purchase price, so that they could then modify the program to do something else that they wanted done? Possibly, but so what? If Stallman's argument holds, then all sales of goods and services which do not include full access to the underlying means of production in the purchase price are immoral. I conclude that Stallman supports unadulterated communism.


Unless information is different from traditional property because it's non-rivalrous. Then your analogy is stupid.
3.28.2009 12:49pm
einhverfr (mail) (www):

Would the users be better off if I included the source code in the purchase price, so that they could then modify the program to do something else that they wanted done? Possibly, but so what? If Stallman's argument holds, then all sales of goods and services which do not include full access to the underlying means of production in the purchase price are immoral. I conclude that Stallman supports unadulterated communism.


I join MCM in saying this analogy is stupid too :-). I am going to flesh out my arguments a little more though.

The first important question is we have to decide what we mean by moral and immoral. I think Stallman argues that closed source software is harmful to the public good, at least compared to open source software. I think there is an argument to this effect, and if I didn't I would be building closed source software. However, this isn't uniform. My view is that if there is no business model to support an open source application for one reason or another, then one can't provide the benefits in an open source application that one can in a closed source app, so this breaks down.

Now, it is true that in open source, typically the developers own the means of production. However, this is built on a market economy at its best, and thus represents a move not from the free market capitalism to the neofeudalism of the USSR but rather a move from big business capitalism to something very different: self-organizing communal networks across national boarders. The communal networks though depend on and feed the system of free market capitalism. So while it can be seen as the fulfilment of Marx's predictions regarding the fall of big business capitalism, it does not evolve in a way we would recognize as communist (which generally involves nationalization of means of production, not private, shared ownership of these means).

Open Source Software is fundamentally dependent on the free market in most areas. The big differences really have to do with how the "intellectual property" is managed and how it is monetized. I would like to eventually see all software be open source, but I see a lot of closed source being helpful in the mean time.
3.28.2009 1:16pm
Tony Tutins (mail):

Where did DOS come from? How about Windows?

1. DOS was an improved version of CP/M.
2. Windows was a copy of XWindows, for the crippled x86 series.
3.28.2009 8:00pm
Kirk:
Tony Tuttins (3.28.2009 8:00pm):

Oh good grief.
3.28.2009 10:16pm
EW1(SG) (mail) (www):
Guest12345:
That fact means that SELinux, derived from Linux by US government employees, should not have been distributed since they could not satisfy the requirements of both the GPL and the law.
Hogwash. The government never held copyright to Linux in the first place, and 17US105 is inapplicable~meaning the government also has no standing to change the existing copyright~which covers SELinux as a derivative work. The FARS cover this in much greater detail.

einhverfr:

I am sure Stallman hates the EnterpriseDB example because it goes against everything he stands for.
I'm not sure where this idea springs from. The principals of Ada Core Technologies have built their business model on the GPL principles (after long and close association with Stallman). Your description of EnterpriseDB sounds very like their business model.
3.29.2009 12:35am
einhverfr (mail) (www):
EW1(SG):

The difference is that EnterpriseDB sells software licenses to closed-source versions of open source projects. Stallman has, in the past, made harsh comments about similar players in the Apache market.
3.29.2009 8:40pm
einhverfr (mail) (www):
Tony Tutins:

I might disagree with a few details (Windows was inspired by an off-shoot of W iirc, not X), but the basic message holds:

Nearly every major piece of software any of us uses is build on that which has come before, and this is a good thing. It is not different in open source than it is in closed source software development.
3.29.2009 8:42pm
Paul Lockett (www):
J.T. Wenting: "When there's no more money to be made writing software, people won't choose to do it as a profession and won't spend the time and money to get trained as professionals"

This is a fallacious argument when discussing copyright on software, as the vast majority of programming is carried out within businesses for in-house use only. It is never released outside the business and therefore copyright is irrelevant.

Arguably, if a greater proportion of software is free/open source, it becomes easier for end users to adapt the software to suit their own needs and in the process, creates more opportunities for software writers, in much the same way that we are all writing profusely on here because nobody holds control over the English language.
3.30.2009 9:52am
einhverfr (mail) (www):
Paul Lockett:

Additionally, many of us make money writing open source software. The "No money" argument probably holds in some niches (consumer software, games, etc), but doesn't hold in most areas of software development. In fact, there is a LOT of money to be made doing open source development but it is harder to monopolize the market.
3.30.2009 11:44am
themighthypuck (mail):
Roger_Z's argument is the clever "camel's nose" argument, although it has recently been referred to as a Godwin. I should add that I think Roger_Z was kidding.
3.30.2009 6:07pm
David Schwartz (mail):
PL: Nothing anybody holds copyright to is in any way analogous to the English language. It is only specific expressions, where large numbers of other expressions not protected would be equally functional, that you can hold copyright to.

If copyright ever creates functional restrictions, that's a bug. Fortunately, courts have been pretty good about fixing those. (See for example, Lexmark v. Static Controls, merger doctrine, Apple v. Franklin, and so on.)
3.30.2009 7:16pm
XX (mail):
@PeterWimsey:

> but is otherwise fundamentally theft

And you find it appropriate to comment on copyright, not understanding that infringement of copyright (either by law or in theory of property) cannot be theft, because no property has actually been removed (i.e. nothing is missing).
3.30.2009 10:22pm
David Schwartz (mail):
XX: If I charge $15 for people to tour my garden, and you sneak in without paying, tour my garden, and leave, is that theft?

"Theft is the criminal act of dishonest assumption of the rights of the true owner of a tangible or intangible property by treating it as one's own, whether or not taking it away with the intent of depriving the true owner of it."
http://www.businessdictionary.com/definition/theft.html
3.31.2009 3:19am
einhverfr (mail) (www):
David Schwartz:

Copyright infringement has been held by the courts over and over to be a fundamentally separate issue from theft. This isn't analogous to theft of services either since services consume your time. The theft analogy only works if you treat copyright as some sort of "services in advance" but that doesn't work quite well either.

A better analogy would be if you charge people $15 to photograph your garden and I take high resolution photos from a low-flying aircraft (say, a helicopter) without paying you if that is theft. I would argue that it is not.

Copyright infringement is a matter entirely separable from theft. Basic point: If you sell 1000 service items and people steal 1000 more service items from you, you are still worse off than if you only sold the 1000 service items in a similar way than you would if people stole other merchandise. However, if you sell 1000 copies of a book and people photocopy another 1000 copies, you are in exactly the place you would have been if you only sold 1000 copies of the book.
4.1.2009 7:15pm

Post as: [Register] [Log In]

Account:
Password:
Remember info?

If you have a comment about spelling, typos, or format errors, please e-mail the poster directly rather than posting a comment.

Comment Policy: We reserve the right to edit or delete comments, and in extreme cases to ban commenters, at our discretion. Comments must be relevant and civil (and, especially, free of name-calling). We think of comment threads like dinner parties at our homes. If you make the party unpleasant for us or for others, we'd rather you went elsewhere. We're happy to see a wide range of viewpoints, but we want all of them to be expressed as politely as possible.

We realize that such a comment policy can never be evenly enforced, because we can't possibly monitor every comment equally well. Hundreds of comments are posted every day here, and we don't read them all. Those we read, we read with different degrees of attention, and in different moods. We try to be fair, but we make no promises.

And remember, it's a big Internet. If you think we were mistaken in removing your post (or, in extreme cases, in removing you) -- or if you prefer a more free-for-all approach -- there are surely plenty of ways you can still get your views out.