If you are interested in how open source produces great software without the benefit of copyright/patent protection, Michele and my article on the Freeman is now available.

There is also voluntary renunciation of trade secrecy: the original creator publishes the source code the “blueprint” for producing the software along with the software itself. Some open-source software has the further requirement that as a condition of use, buyers make their modification available under the same terms.

The phrasing here could be misleading.

To clarify:

1) The GPL does not require any renunciation of trade secrecy, nor is anyone expected to renounce trade secrecy voluntarily.

2a) The GPL does not require that any modifications are made available.

2b) The GPL does not require that your modifications are made available under the same terms - you can make them available only under fully restrictive copyright if you wish.

3) The GPL has no conditions of use.

All the GPL says is that if you CHOOSE to publish the licensed work WITH your previously private/secret modifications, you have to publish the combined work on the same basis, i.e. according to the GPL.

The GPL assures the licensee that the covered work is available without obfuscation, i.e. if binary forms are published, the human readable forms that gave rise to them are also available.

If I understand correctly, you are talking about modifications that you make for your own use but do not redistribute? Our point was that if you choose to release something under the GPL then for that particular piece of software you have renounced trade-secrecy by releasing the software. Certainly you could keep other things secret, and an end user could modify their software for their own use without giving up any secrets.

Thanks for the clarification - the statement in the paper was broader than intended.

Yes, David, I'm not critiquing your understanding, just pointing out that unless you're very precise, it's very easy for people to misinterpret things (there are many myths thanks to such misinterpretation of the GPL).

The act of publishing something is obviously turning something that was private into something that is public. To describe this in terms of 'renouncing trade secrecy' is a bit clumsy and comes across as apostasy. :)

While we're being precise, we should also mention that "trade secrecy" is only "renounced" or abandoned for the specific source code published. In practical terms, there could well be trade secrets involved in how the source code was itself developed, or even in how the code may be put to useful application. Both of which could yield competitive advantage to a firm developing & publishing open source software.

True: anyone who has written software can tell you it is a lot easier to understand your own code. But never-the-less making the source code available is a significant reduction in secrecy.

Come on. It'd be like saying the following: "Publishing a book is a significant reduction in secrecy" Or "Getting into a bath is a wholesale abandonment of one's dryness"

The phraseology is exaggerating the significance of the act.

My point is that individuals (and corporations) will continue to prize their privacy and secrecy (and their dryness) - no matter how many things they publish (or baths they take). Moreover, free software licenses such as the GPL in no way suggest the renunciation of privacy or secrecy.

All the GPL says in this respect is "If you publish, you publish - you don't obscure your publication in Latin so only the clergy know what you're on about - or if you need to publish in Latin, you should also publish an English translation for the laity's benefit".

That's all. Nothing to do with renouncing secrecy.

Although there exist those that read and write assembly language code almost as fluenty as the average programmer high-level code, these people are very rare. They are called RCEs - reverse code engineers - and make a LOT of money.

In a "traditional" development model, the software is delivered as binaries. Nobody, except aforementioned group of hackers that are fluent in assembly language, knows what it is really doing. Specifically, nobody knows how it is doing what it does.

Compared to publishing just the binary version of the software, publishing under the GPL is a renunciation of trade secrecy in the sense that you don't just deliver a machine, but you deliver the blueprints on its inner workings as well.

A more efficient algorithm may have been discovered that gives a competitive edge. The results are the same, but it works faster. This algorithm could be a trade-secret. It goes without saying that the secret can be learned from the source code.

Therefore, I disagree, it IS a renunciation of trade secrecy, comparing binary form and source-code form.

The problem arises because the form for editing the work and the form for using the work are different, and the latter isn't easily converted into the former. That is what allows such secrecy to happen. I believe this to be a separate issue from "intellectual monopoly." To prevent this from happening, you would need to specifically outlaw it. The problem isn't caused by any existing law.

Even worse, the only way to prevent it nowadays depends on a very controversial existing law. I think that if intellecual monopolies were to be abolished, some legal framework for copyleft contracts would need to be created (or survive).

Well Kid, we can agree to differ.

When I publish my secret algorithms, I do not renounce trade secrecy, I stick to it like a clam to a rock.

As to copyleft depending upon copyright for its survival, this is like Noah's ark depending upon the flood for its buoyancy and the safety of the entire animal kingdom.

Without copyright copyleft is not required.

I'm disappointed that so many apparently intelligent people believe that suspension of the public's liberty is required in order to achieve its restoration.

You do not need to send armed storm troopers to break down a programmer's door in order to force them to release their source code.

Try paying them money (it works for most other professions).

Clue: It's a world in which copyright has been abolished.

I don't think anyone here is advocating forcing programmers to make source code available...Certainly I agree that would be a poor idea.

Yes, I meant to outlaw leeching from an open-source project, adding a major feature, and releasing the resulting program in binary form (unless they can get permission from the original authors).

You are saying that this would not be a profitable practice in a world without copyright?

Yes Kid, in a world without copyright there is no commercial incentive to sell only the binary.

You can sell one copy of each.

The binary for $100, and the source code for $10,000.

You cannot sell a thousand copies of the binary (or at least, not for $100 each).

For one thing, you're competing with a load of others who ARE selling source code. Moreover, those others are giving the binary away free in order to promote and demonstrate the utility and value of the source code.

And those others are selling $10,000 of source code to 10,000 customers at $1 a piece (given the larger audience size created by giving the binaries away).

You tell me how post-copyright anyone can make more money selling binaries whilst keeping their source code unreleased, than selling the source code, and I'll fall off my chair.

So in the end, the problem IS created by an existing law.

Without the distortions of copyright, you get what you pay for and you get paid what it's worth.

There will hardly be a market for binary copies anymore, since the source code is far more valuable than the binary, but creating either requires the same amount of work.

Do I understand now?

Is that also where your trade secrecy argument comes from? Releasing the source code is obviously releasing (some of) your secrets, but releasing the binary is more of a demo than actually publishing your work.

Sounds like you understand, yes.

It's not where my point about secrecy comes from. That was mere pedantry.

"The market for open-source software uncopyrighted, freely reproducible computer programs"

GPL software IS copyrighted.

Even if the licence is used to preserve the freedom of users and "fair competition" between authors from "privatization" of the source code (exactly the reverse of "normal" licences), that's wrong to say uncopyrighted.

Now of course, without copyright at all there is no way for any licence to do the "preserve fair competition" between authors, if it's economically useful this right needs to be recreated by law.

Mea culpa. We certainly didn't mean to say that. GPL software is certainly under copyright. We meant that it renounces the monopoly power that comes with copyright.

You do like this word 'renounce' don't you? :)

BSD is the license that most closely 'renounces monopoly power' - waives the author's exclusive privileges (granted by copyright).

GPL is the license that actually disables monopoly power. It effectively nullifies copyright on the work and all its derivatives, consequently restoring freedoms to the public (including the original author) that copyright otherwise suspends.

Copyright can't simply be disabled by renunciation (if only).

It has to be retained and wielded in order to enforce the restoration of the freedoms it otherwise suspends.

This is why the GPL is so revolutionary, it enables the author to effectively abolish copyright from ever applying to their work.

David K. Levine, thanks for the clarification!

Crosbie Fitch, I'd argue that GPL maintains competition (or freedom to compete) more than standalone "freedom".

And more importantly, it does so where it matters, that is for support, custom development and incremental development, whereas traditional copyright is used only for initial development.

In France according to the government only 7% of the IT industry jobs are for software editors (including open source editors), 93% of jobs are for support and custom development.

Note: no economist interested in the software industry I met know about the numbers I just cited. No wonder existing literature is totally junk.

Laurent Guerby, well I'll be very glad to know of any sound argument that can demonstrate why the monopolistic privileges that copyright provides are required in order that the GPL can consequently nullify them all - and more importantly, why copyright+GPL is therefore superior to no copyright at all.

Isn't the point that the GPL uses copyright law to require people who release modifications to a GPLed program to make available their modifications under the same terms (that is release the source code)? Without copyright it isn't obvious that such a restriction could be enforced. That is if you released your source code without copyright I could modify it, sell my program, but keep my modified source code secret. I don't think there is anything very terrible about this - but it does mean there is case for a limited abolition of copyright - everything abolished except the GPL.

David, I strongly suggest you think long and deep about this (post copyright abolition is a whole new universe).

1) The GPL is not constructed from an ethical foundation, but as a reaction against copyright's suspension of the public's liberty and the iniquitous business practices it incentivises. Therefore, do not assume that you can infer from the GPL the ethical underpinnings of new law to replace copyright. The GPL undoes damage (it doesn't create new law). Its restraints are against those who suspend the public's freedom - not against the public and their right to liberty and privacy.

2) Without copyright there is no incentive to withhold source code and sell binaries. There is consequently no need for legal compulsion, nor prohibition against distributing binaries without source.

3) The point of the GPL requiring disclosure of source code is not because this is a right of the recipient of the binary derivative (it isn't), but that it is a weapon against the infringing (copyright supporting) publisher of GPL derivatives - to thwart any attempt on their part to illicitly exploit GPL software using the copyright based business model.

The idea is, the stronger a (copyright supporting) publisher demands that copyright law is enforced, the harder it is for them to infringe the copyright of GPL software. This is because if they take things to a court of law, then copyright will be upheld and they will be forced to publish the very source code that they want to keep secret (which also helps demonstrates infringement to all).

If you don't believe in copyright, then you don't really have to worry about the GPL obliging you to supply the source code.

Go on, modify some GPL source code and sell the binary to someone.

Before you sell, you can even tell the recipient that it's GPL based. You can also tell them you don't believe in copyright nor the GPL. Then you can drop the bombshell and tell them that contrary to the GPL you're not giving them the source code.

For the recipient to prosecute you for non-supply: 1) They need to buy a copy of the binary from you 2) They need to convince themselves it is a GPL derivative 3) They need to find the copyright holder of the base work 4) They have to convince the copyright holder that it's worth prosecuting you for copyright infringement (or find someone else with big pockets who does). 5) The copyright holder has to get a warrant to search your premises 6) They have to find the GPL source code (ideally your modifications) and your build process such that they can demonstrate how the binary is a derivative of the base work - and that you produced it. 6b) NB The GPL doesn't prohibit modification, only publication of derivatives without source. 7a) They have to decide if you have anything to lose/suffer 7b) They have to decide if they would gain from this prosecution 8) Then they have to go for statutory damages 9) Ideally you'd have enough money for this or an out of court settlement.

This won't happen.

It only happens to commercial publishers of software who illicitly incorporate GPL code in their copyrighted binaries. And even then, they don't get prosecuted, they just get a pleasant letter from the FSF suggesting they think again.

The GPL is only a weapon against publishers who have copyright based business models.

More importantly, the GPL is only a defense against publishers who have copyright based business models.

The GPL obtains its strength (against copyright) from copyright, and it is only against those who believe in copyright that the GPL actually has any power.

Another way of looking at it entirely, if you really do think the holy GPL should be made law in place of copyright, is to consider just what law you'd create that would grant anyone in receipt of a binary the right to search anyone's premises in pursuit of the source code. Or would you simply make it illegal to distribute executables without source code?

The nonsense of such hypotheticals should persuade you that you're missing something...

Crosbie: the problem is that I don't disagree with this.

You may not disagree, but plenty do.

If you can help improve the arguments it would be helpful.

1. the principles of the GPL are at odds with those of copyright
2. that the GPL utilises copyright does not constitute support for its retention
3. if copyright were abolished, compulsory surrender of source code should not be enforced - nor does its non-disclosure even need to be frowned upon

So, when you suggest "it does mean there is case for a limited abolition of copyright - everything abolished except the GPL" then good arguments should demonstrate that this is a very poor case.

1 seems obvious to me, 2 is debatable (not all of copyright, but perhaps part of copyright), and I strongly disagree with 3, that is, I think publishing obfuscated source-code should be frowned upon.

People deserve the right to be in control over their own property, and they should be in control of what their computer is doing. Knowing what the software you run is doing is a prerequisite for that.

Tempting people to sell their bodies for money is frowned upon, and I think tempting people to sell essential freedoms for money is much worse.

Ah, it's misleading to call it an essential freedom here (I'm talking about the freedom to modify your property)

In fact, you do have that freedom without copyright. It's just made harder for you by methods of obfuscation. Nobody has an obligation to make your life easier. If you want that, pay for it.

Let me rephrase, then, tempting people to use drugs is frowned upon, and so should be tempting people to run your code if you're not going to tell them what exactly it does.

People are tempted into something that they probably don't understand the full implications of. If people really understood that they are giving (partial) control over their computer to the software developer, then they would probably charge for running that software instead of pay for it.

Kid,

If someone comes up to you and says "Psst! Mate. Do you want to buy some obfuscated code?"

Will you feel overwhelmed with the compulsive urge to purchase it?

And then no doubt you will reflect and realise you should have remained true to your principles and only purchased unobfuscated code?

Enumerating 'four freedoms' is not how to define an ethical foundation. We start off with freedom. Ethics is a matter of delimiting that freedom with human rights (ethical constraints upon our freedom). The GPL is a reaction against the unethical constraints of copyright (suspension of the public's liberty). Remember that copyright started off only constraining a few commercial printers (and hence remained underneath the 'ethical radar'). Copyright is now used to exert unethical constraints against the public's liberty to published art - and hence is now in the ethical radar as a big white blob of an elephant in each of our living rooms.

So, with or without copyright, if you buy binary code from a vendor, you do not have a right to the source code unless the vendor misrepresented the bargain. Caveat emptor applies here. Basically, if you want source code, then you ensure that that's what you buy.

The problem with copyright is that it incentivises the withholding of source code, i.e. so the market's flooded by binaries. Thankfully the GPL is able to demonstrate that copyright is unnecessary.

So, unnecessary, unethical, and ineffective, to me means that copyright is ripe for abolition.