Copyleft – the reversal of copyright’s abridgement of mankind’s cultural liberty, the restoration of freedoms suspended by copyright (and by patent to some extent) – is GOOD.
Without copyright we cannot have freedom?
Depressingly, there is an unfortunate tendency for copyleft supporters to conclude that because copyleft relies upon a copyright license to restore freedoms to the general public, that copyright is not only necessary and indispensable for ‘freedom’, but must also therefore be good – and so also to be supported.
Pointing out the irony of supporting a privilege that annuls people’s right to copy in order that a copyleft license can restore it doesn’t appear to ring ‘Internal contradiction detected!’ alarm bells in far too many copyleftists’ minds.
It is a ridiculous situation that people who would restore the freedoms suspended by copyright feel they must retain the very privilege that suspends those freedoms in the first place (in order to restore them). This is why copyright abolitionists receive very little sympathy from those in the free software movement. To be against copyright is to be against the enabler of freedom that is copyleft: the GPL is holy – without copyright there is no GPL – the abolitionist is a heretic.
As an instrument of injustice the privilege of copyright has no justification in the first place. While it is good that we use copyleft to restore freedoms suspended by copyright, it is ridiculous that champions of that freedom would attempt to justify copyright’s continued existence simply in order that the laudably libertarian licenses of copyleft can continue to undo it.
Freedom is the power to compel disclosure?
The best attempt at justification that copyleftists can come up with is the ethically worrying complaint that without copyright there would be no power to force or coerce people who publish binary derivatives to also publish their source code.
So I find it similarly ironic that instead of being at the forefront of a movement to abolish copyright that copyleft supporters assume they must support copyright in order to enjoy its power to coerce people to release source code to their published binary derivatives of other’s work.
Power corrupts. People enjoy too much the prospect of being able to control others, e.g. limit what others can or can’t do with what they publish, or coerce people to publish things they might prefer not to.
That this power to coerce disclosure of source wouldn’t exist without copyright should be a clue that it’s not a natural power and so not ethical.
Why is copyright/copyleft perceived as closed/open source?
We should examine why people think copyleft and the free software movement is more about coercing disclosure of source to binaries than it is about restoring freedom.
- Why do people want to coerce others to publish their source code?
- Who are these ‘others’ who will publish binary derivatives without source?
Given every developer espousing free software is evidently quite happy for source code to be publically visible to enable the considerable advantages of an open and collaborative software development process, there can’t be any ‘closed sourcers’ publishing GPL code. In other words, everyone developing free software wants their source code to be published. They don’t need to be coerced. So, who does?
It must be those ‘closed source’ publishing corporations that exploit copyright to sell binary copies at monopoly protected prices.
But why don’t they supply the source code too?
‘Closed source’ is an epiphenomenon of copyright
Copyright prohibits anyone from making unauthorised copies or derivatives, so the source would be of little use to anyone except those who’d like to help the developers pinpoint bugs – or competitors who’d like to reverse engineer the code and produce a cheaper equivalent. So, given how little benefit there is in providing the source (if not considerable disincentive), it shouldn’t be surprising that it isn’t provided. This is especially the case when you consider that copyright is actually pretty ineffective at preventing copies or improved derivatives.
Copyright doesn’t provide a reproduction monopoly to publishers of software binaries on condition the source remains unpublished. Keeping the source unpublished is entirely an option and decision of the binary publisher. But, it should be recognised that the decision not to publish the source is made soley because it is generally against the commercial interests of the vendor of copies. The more they can maximise the strength of their position as monopoly vendor of copies the better. Only they can produce improved binaries and sell copies thereof, because only they have the source. Certainly the source is provided in some cases, but this is usually reserved for wealthy clients who can afford the premium and legal/security overhead.
As an epiphenomenon, ‘closed source’ falls with copyright
It has to be recognised that people have a right to privacy and if there’s anything they don’t want to make public, they don’t have to make it public (unless a greater right, life is at stake). There’s nothing magical about software that makes it any different from a cookie recipe. People have just as much right to produce cookies or binaries and keep their recipe or source secret – even if their product results from their modifications to published recipes or source. ‘Freedom’ cannot be invoked to sequester the recipe or source code from someone who sells a cookie or binary. Privacy delimits such freedom, not copyright – and unlike copyright, privacy is a natural right.
Let’s remind ourselves as to the objective. It is to restore the freedoms suspended by copyright and undo its epiphenomenon of ‘closed source’, not to empower SWAT teams to break down coders’ doors to seize unpublished source code to the binary derivatives they’ve distributed.
A similar error in using copyright as coercion occurs in the Creative Commons licenses that oblige attribution.
It is as much a mistake for the GPL to coerce disclosure as it is for Creative Commons licenses to coerce attribution. Like ‘closed source’, the lack of attribution is an epiphenomenon of copyright. This is because authors and artists who credit their (uncleared) sources open themselves up to copyright infringement lawsuits – by admitting they’ve copied. Without copyright there is no such disincentive. However, with or without copyright, it is ridiculous to threaten people with million dollar lawsuits if they fail to attribute their sources. There is no right to attribution. The right is to truth, to accuracy in attribution – in other words, a right against misattribution (qv plagiarism). Crediting one’s sources is a mark of respect that credits the creditor and credited alike. But you don’t undo copyright’s epiphenomenon of inhibiting attribution by coercing attribution – instead, you remove the threat from copyright that discourages it, by neutralising or abolishing copyright.
Similarly, you don’t undo copyright’s epiphenomenon of disincentivised source code publication by coercing the disclosure of source. What is ethical is to remove unethical constraints – not to contrive what would happen in the absence of privilege through unethical coercion (ends do not justify means).
Freedom needs no coercion, nor derogation of privacy
I contend that you can remove coerced disclosure from the GPL and still find that it undoes copyright’s ‘closed source’ epiphenomenon.
You undo or neutralise copyright and its abridgement of freedom by removing copyright’s constraints, by restoring the liberty they suspend in a license – and copyleft’s genius judo is in obliging the removal of those constraints in derivatives via the same license, thus liberating the public (including the licensor), not just the licensee.
However, you remove an epiphenomenon caused by copyright by removing its cause, the incentives for it, which is already achieved by the former: the removal of copyright’s constraints (ideally abolition). You need no additional coercion.
Compare coerced disclosure with liberty
Let us explore the difference between two copyleft licenses, one that coerces disclosure of source, the GPL, and one that doesn’t, the ncGPL (non-coercing – identical to the GPL except that distributors of binary/obfuscated derivatives don’t have to supply source on demand).
Those in the free software communities WANT to have the source open – or they wouldn’t participate, therefore the coerced disclosure clause of the GPL doesn’t come into play for this class of developer.
The critical example is of a corporation motivated by monopoly profits.
Why is such a corporation going to publish a binary under the ncGPL that it couldn’t under the GPL?
Remember, the ncGPL still breaks copyright’s ability to enjoy monopoly protected pricing.
Therefore the ncGPL holds no attraction to such a corporation – the largest publishers of binaries whose source remains unpublished/confidential.
Here are three of the obvious business models or propositions the ncGPL permits that the GPL does not:
- “As you can see from the attached binary, I’ve fixed the bug you asked me to. Once I’ve received payment I’ll provide the source.”
- “You’ll find via BitTorrent a gratis/freely copyable ncGPL binary we’ve developed as a promotional demo to help sell our secret source code (and thus cover our development costs).”
- “For a small fee we’ll develop a bespoke binary to your spec that you can freely copy or attempt to reverse engineer and modify as per the ncGPL, but because our source code reveals our exclusive technology we won’t let you have the source unless you pay us considerably more (to cover our R&D costs).”
These propositions could be made under the ncGPL or if copyright was abolished. They could not be made under the GPL, or if copyright was abolished and a new (unethical) law made it illegal to convey binaries without source.
There is nothing ethically wrong with them – being given a binary without source does not derogate from your liberty (it is copyright that is unethical, with its consequent epiphenomena antisocial).
These ncGPL business models do not engender copyright’s ‘closed source’ epiphenomenon that we suffer today because the ncGPL, like the GPL, defeats any market for binary copies. Either the binary is of interest to only a specialist client (bespoke), or it is given away as a free demo of the source to be sold. The likes of Microsoft and other sociopathic corporations exploiting the proprietary ‘closed source’ copyright enabled business model would adopt neither of these two ncGPL enabled models, and that’s because the potential revenue is equivalent to the labour expended – as opposed to the orders of magnitude greater revenue that a state granted monopoly can obtain. Moreover, the bespoke ncGPL binary developer must compete with ncGPL competitors who provide similar functionality with source included – a lack of coercion does not imply a lack of commercial incentive.
Practically, GPL code bases could still be in use in both examples anyway, possibly with the binaries supplied without declaring them to be illicit GPL derivatives. Why? Because the developer isn’t egregiously infringing copyright (though the public might). They are not engaged in the manufacture, distribution and sale of copies – the mode of an infringer that copyright law is optimised to prosecute. If the law cracked down on those who prepared and supplied unauthorised derivatives, the police would be waiting at the doors of publishing agencies and record label A&R departments ready to bust budding new authors and artists who’d failed to obtain licenses for their samples, quotes, and clips, etc. Much as publishers might be happy for people to assume, artists do not and cannot practically perform copyright clearance in advance of creating their art (unless extremely wealthy). The same copyright law applies to software developers as applies to any other artist – even if the process is different. And frankly, I do not see GPL copyright holders hiring private detectives to scour the land for binaries that have failed to declare themselves as GPL derivatives, and then hiring expensive lawyers to prosecute the developers for preparing them illicitly.
Finally, bear in mind that the ncGPL is still like the GPL in requiring that no constraints may be applied to preclude anyone privy to the source code from releasing it. The ncGPL simply omits to coerce those who publish binaries into providing the source on request. The ncGPL would instead recommend making source available on demand instead of making it a license condition. Being identical to the GPL in all other respects it maintains the GPL’s condition against preventing recipients of source from copying or otherwise conveying it. Thus any employees of a company developing an ncGPL derivative would remain just as free to release it unilaterally.
I appreciate that it is widely assumed that the GPL is so successful precisely because it coerces publishers of derivatives to publish their source, but I contend that it needs no such coercion, it needs only to remove copyright’s epiphenomenal incentive to keep source unpublished, and that is achieved by neutralising copyright, nothing more. No additional coercion is necessary. The unnatural power of privilege may be undone by using that power against itself, but freedom is not obtained through the use of that power beyond its undoing.
NB The ncGPL is quite different to the BSD in that the ncGPL requires derivatives to be similarly licensed.
If the proprietary ‘closed source’ business model is incompatible with the ncGPL then it is incompatible with copyright abolition. If you can see that the ncGPL would achieve the same source code transparency as the GPL then you need have no fear that copyright abolition would maintain a significant ‘closed source’ incentive.
The restoration of liberty suffices (copyright & patent abolition)
Copyleft is superior both ethically and economically because it restores the public’s liberty, not because it coerces the publishers of derivatives to disclose source code. It is the restoration of liberty, the dissolution of monopoly, that removes the incentive to keep source secret, not the threat or coercion of a license condition.
It is safe for supporters of free software to support copyright and patent abolition unconditioned upon a prerogative to compel source code disclosure.
Thanks again for such a wonderful stream of crystal-clear thoughts.
Would have never thought of the implications of coercion on the GPL.
I think this is the first time I read about the ncGPL, and there may be a good reason: I’ve googled a little about it and couldn’t find any other reference to ncGPL besides this very same article.
Not sure what the plans are, but… are you planning about formalizing this proposed ncGPL into a proper license to be used by developers?
It seems to me that it could be a great way to massively spread the way of thinking promoted by Cultural Liberty and help others to think in terms of copyleft.
I wonder what Richard Stallman would think of ncGPL, as it seems to take the fight for free sofware and liberties one step further.
The ncGPL is just a hypothetical variation of the GPL. Possibly it might influence GPLv4, or failing that GPLv5, but because this article is heresy to copyleft dogma (that ‘open source’ is only open because of a license obligation to publish source), I doubt this article is going to interest anyone except those who’ve already recognised that power to compel others to publish source code is not exactly natural.
I published this article primarily to prepare an argument on the matter with Mike Linksvayer. We’d exchanged words on Identica and it was agreed more explanation was required on my part.
One of the things that particularly bothers me about licensing, libertarian or otherwise, is that it seems designed to be continue forever. In the first few centuries of copyright, at least the terms were finite and so works would eventually be freed. What concerns me now is that we may be heading into future of perpetual licenses, with nothing ever again going into the public domain.
As a non-programmer, my concern with closed source code is that can turn my work into ransomeware.
If I purchase software to edit video, say, and edit all my home movies in it using the proprietary format, my work is now locked inside that proprietary format.
- I can’t take my partially edited file and work with it in different editing software,
- if the company goes out of business, or
- simply decides to stop supporting it
It’s all very well to say that no one should be compelled to publish sourcecode, but I disagree. Once the software is sold and out there, it must be open and accessible for consumer protection.
In the pre-digital world, I could cut my film on a Moviola or a Steenbeck, but was not locked into a relationship with a single manufacturer. Closed source puts consumers, artists and the culture at the mercy of manufacturers.
If you have a problem with the term of licenses, you have a problem with the term of copyright. And as only licenses can restore the liberties that copyright suspends, if you hope for liberty you’d better hope that such licenses last as long as the copyright (in some cases they don’t).
As for obscure file-formats, these are a further epiphenomenon of copyright. If you can sell copies, then not only don’t you need to sell the source, but you don’t need to sell documentation of the file-format. Whereas, without copyright you have to sell the source (if you want to get paid for writing it), and if the source is published, the file-format is far easier to understand and document – moreover, the person paid for writing the source, is likely to get paid far more if they document the file-format.
Without copyright you might feel you still need the power to break people’s doors down, severely fine them or imprison them if they publish a binary without source, but such power is unnatural and unethical. This article attempts to explain why such power is not needed, indeed why such power could even be removed from the GPL without affecting people’s incentive to publish source code with their GPL derivatives. I suggest that, without copyright, the civilised approach and ethical incentive is to offer software engineers money in exchange for the publication of their work. Do not make it a crime for them to publish a free binary demo of their work in advance of being paid for their work. Similarly, don’t fine musicians for releasing a free MP3 demo file in advance of selling the score and FLAC versions of each component track of their digital master. Demand only your liberty back, don’t deny an artist the privacy that enables them to sell their unpublished work.
If you don’t like closed source then you don’t like copyright. Don’t make the error of inferring that an ephiphenomenon of copyright warrants a new unnatural, power to compel disclosure. Just abolish copyright. Don’t replace one instrument of injustice with another.
Actually, Crosbie, I’m not keen on either closed source or copyright.
But there is a world of difference difference between:
(1) a software engineer publishing “a free binary demo” (or a musician publishing “a free binary demo” of their work) prior to selling a copy of their work,
(2) a software engineer selling me a copy of their work with the source code locked up (or an RIAA corporation selling me anything locked behind DRM).
A publisher of anything generates income by selling copies. No copyright law anywhere in the world does anything to restrain any publisher from giving away free copies, in whole or in part. Nor would they. After all, that has long been one of the primary means a publisher has of promoting their wares.
Copyright uses the force of law to compel restrictions on people’s freedom to enjoy cultural works, just as closed source uses obscurity to compel restrictions on people’s freedom to enjoy software works. [The same software publishers who lock up their source code tend to avail themselves of copyright law and/or software patents to add the force of law to their assault on their customers’ liberty.]
I am in complete agreement with your statement that “It has to be recognised that people have a right to privacy and if there’s anything they don’t want to make public, they don’t have to make it public.”
Frankly, I would never advocate compelling software engineers to publish. Like novelists, software engineers are always free to write their code and hold it close, never publishing it at all. But, of course, like novelists, once software engineers elect to publish their software, they have themselves ventured outside of the private domain.
When I purchase a copy of a book that is “protected” by copyright, I at least have access to its entire contents. The effect of closed source software is akin to selling a copy of a novel, and then, after the customer has paid the agreed price for it, charging the customer a second time for the key to the final chapter. Effectively making closed source code ethically worse than copyright.
Most consumers are not aware we are being sold a pig in a poke. Disingenuous sales practices fool customers into believing we are purchasing software when in fact, thanks to the most outrageous legal legerdemain, we are merely licensing it. Even further, most users would never dream that it is not within our rights to use such software as we see fit, and are unaware they are legally restricted to using such software only according to the dictates of the publisher (who may, or may not, be the software engineer).
Consumer protection laws must at minimum compel a warning to inform consumers what they are getting. Software may be nearly universally adopted but neither software nor licensing is by any means universally understood by users. In the absence of such warnings, customers are essentially being defrauded. Society must be protected, particularly when the pace of innovation unfairly disadvantages human beings.
Still, society allows unfettered infringement of personal liberty only at its own peril, which is why I am disinclined to the idea of legal compulsion beyond consumer warnings. Education and freely available information will take up the slack. Even without such warnings and education, human beings are adaptable, which is why, as more people are becoming aware of the way closed source software infringes on their liberty, more and more are abandoning that Sisyphean treadmill.
If you give me something, like a “free demo,” it is conceivable that there might be strings attached. If you sell me something, even a copy, any imposition of restrictions, by either legal or physical means infringes on my liberty.
I am not sure you are taking into account the TIVO’s of the world.
There are certainly current developers of derivatives of GPL software that do not publish their sources and other who take other steps to prevent you from running changed sources even if you have them.
It is of course one thing to have certain freedoms restored in a generally non-free game and another to have a free game restored.
Good point Drew.
However, where the GPL fails, the ncGPL cannot do better.
Expensive devices (Tivos or games consoles as opposed to binary files) with which binaries are included should still be treated as expensive device plus free software demo (per the ncGPL the binary can still be freely copied).
The source is unpublished because the developer hasn’t yet been paid for it. Offer them an equitable amount of money for the source and per the ncGPL you are at liberty to share it.
If I sell you a black box that does something useful, then without copyright or patent, that box cannot constrain your liberty. Sure, reverse engineering is hard, especially without the source code and designs, but you remain free to do it. That I could make that task easier by giving you further information still doesn’t actually impinge upon your freedom.
Alternatively, if I’d bound you with manacles and omitted to provide you with the 4 digit code to the combination lock, then I’ve bound you, but that’s because the manacles constrain your freedom, not because I’ve omitted to give you the code. Not providing information to someone does not by itself impinge upon their liberty.
I’ve pointed out this error in the ‘four freedoms’ before. See The Seduction of Social Sanctity
Hey Crosbie, you drew me here from an article by Nina Paley where she wonders why some free software developers tend to make exceptions for opinion writings in what licenses they use and in rights they give (as if to preserve the integrity of opinions when they believe in allowing source code modifications). www.techdirt.com/art…
What I tried to explain in a comment there, as relates to this article, was that trade secrets in software are not easy to decipher and can lead to significant “harm”. There are security implications, and it is a path to significant lock in. A major reason to hide source is to facilitate a monopoly. It’s not that competitors want to know roughly how your software works in order to compete, is that potentially without exact duplication (bug for bug), they are at a significant disadvantage in terms of interfacing in important areas, including by understanding precisely the wide range of created user documents or data. A small misunderstanding (think of “Place paragraph in blue, 12 inches down” vs. “Place half of paragraph in red, 56 feet down”), can lead to horrendous results, even if the competing software is otherwise 10X better. We have network effects to deal with at a level that don’t exist for ordinary consumer goods. You can even create extremely complex protocols or security back doors where the details are changed in real-time over the Internet (eg, where the details aren’t known until run time when the missing key bits are downloaded, encrypted and time spread). The evidence can then be removed. It took a long time to decipher the Rosetta Stone. How much easier would it have been if we simply had a manual to the language? The issue with software is not identical to this, but, there are costs, like with any other reverse engineering, and one where the changes may occur very frequently (and compilers can take a tiny change in source code, in logic, and produce a very different iteration of the compilation because the mappings are many to many not one-to-one). Imagine the Rosetta language changing every hour as we are trying to “crack” the code (of the moment).
[some more…] As I mentioned on that thread, I think without copyright, we would come closer to cracking the codes or at least to better determine that the code is changing a lot. The real solution is to have people avoid the monopolies; however, software’s effect is not something we see, and lock in and network effects take hold and create unfair market places. Calling something unfair is a judgement call, but just as brand and past good experiences lead people to avoid many restaurants and go to particular ones, the problem with software is that it has network effects (“everyone” goes to the same ones) and also it really is easier to change food restaurants because you don’t have to invest a lot in learning the new foods or in interfacing with existing foods you’ve eaten in the past.
My decision on copyright: I would like no copyright (as you do) but not if it would only apply to software. My view is that society would gain so much overall from no copyright (or a weak version) that it would make up for risks in software.
Note, that the GPLv2 is “flawed” in at least 4 important ways. Patents are not resolved. We have “tivoization” (v3 issues). We have online hosting companies (AGPL issues). The copyleft creates inflexibility. These problems make it easier to agree that no copyright across the board (culture including) would be the best.
Great essay, apologies for taking nearly a year to read it in full. I’m mostly in agreement, but allow me to highlight two fragments that I disagree with.
“power to coerce disclosure of source wouldn’t exist without copyright”
Not necessarily. Regulation, including the variety backed by coercion, could force disclosure. Such could have nothing whatsoever to do with copyright. For example, a medical device regulator could penalize anyone installing medical devices without source disclosure.
“you remove an epiphenomenon caused by copyright by removing its cause”
Yes, but non-disclosure of source is not exclusively an epiphenomenon of copyright. Trade secret is a well-worn strategy that would make sense in some cases, not in others.
I agree that overall copyright and patent abolition would be a boon to software freedom. But it wouldn’t completely eliminate closed source, nor other avenues for requiring disclosure.
I have a post somewhat related to the points made above, see gondwanaland.com/mlo…
It also touches (see note 5) on possibility of a copyleft license that only neutralizes copyright, adding no regulatory conditions, as you’ve dubbed the “ncGPL”.
Congratulations on seeing copyleft more clearly than most who know of its existence. :)