why proprietary relicensing is unacceptable for a Free Software enthusiast

The past few months have witnessed a large debate about copyright assignment, which seems to has neared its peak around the time Project Harmony has released version 1.0 of its agreement templates. (Note: I use the expression "copyright assignment" in this post, but I'd happily include into it both actual Copyright Assignment Agreements — CCAs — and Contributor License Agreements — CLAs, as both could be used to permit various forms of proprietary relicensing.)

Mark Shuttleworth has been one of the most prominent voices campaigning in favor of copyright assignments (to for-profit companies). A couple of interesting blog posts of his where this is evident are "On balancing economic power in the FLOSS ecosystem" and "The responsibilities of ownership". Voices against the opportunity of increasing the popularity of copyright assignments are aplenty and it is pointless to repeat good arguments which I find fully convincing. If you are interested in the subject, here are a few good reads I recommend:

What I would like to add to the debate is the point of view of a Free Software enthusiast, not unlike yours truly. I contribute to Free Software not because it is superior to proprietary software in either its software products or in its development methodology. (It actually is superior in many cases, but that's just not my primary motivation.) I do contribute to Free Software to build a world in which users are in complete control of the software they run.

Let's assume we have a broken toaster at home. We are all free to (1) open it up and try to repair it by ourselves. If we don't have the needed knowledge, we are free to (2) hand it over to a person that we trust and who does have the needed knowledge to repair it. That failing, we can (3) ask the toaster producer to repair it for us. Many people like me simply don't accept to live in a world where, once all the toasters will be software-driven, the only repair option will be (3), just because the toaster software is proprietary. (Given various kind of [broken] metaphors have been used to explain copyright assignments, it seemed sensible to propose my broken toaster metaphor in exchange.)

When I contribute to some Free Software project, I not only care about being able myself to get a Free Software licensed copy of the whole product (i.e. the original software product + my own contribution). I care that all recipients of the product, present and future, will be able to get that contribution with enough freedoms attached to put them in control of the software that includes it. This desire can be fulfilled with a varying degree of success, depending on the license of the code base I contribute to. For instance, I find more pleasant to contribute to a copyleft licensed code base than to a weak-/non-copyleft licensed one, and more pleasant to contribute to an AGPL-ed web app than to a copyleft web app. It also means that, given the option, I will not allow the code base maintainer to re-license my contribution under terms that allow distributing it without suitable freedoms attached.

Different contributors will be ready to accept different re-licensing promises, but the mere fact we are debating all this — and that companies like Canonical feel the need of campaigning for copyright assignments — is evidence that many contributors like me exists and actually have a problem with allowing downstream relicensing of their contributions. What is feared the most is of course proprietary relicensing. According to my moral horizon this is the case not because the company might end up gaining money out of my code: I wouldn't have a problem with that. The reason is rather that the company will be able to distribute my own code in a way that grants less freedoms to users than those that were implicitly "agreed upon" when the contribution has been written.

In that respect, I can't help thinking at harmony agreements as a large scale smoke screen experiment meant to convince volunteers that copyright assignments are something Free Software needs. By exploiting the creative-commons-like "it's easy, just choose one" scheme, Harmony takes the chance of offering misleading options such as the one labeled "any OSI approved license". Via that option, which is seemingly fine, the copyright recipient will be allowed to relicense the contribution under some non-copyleft license (e.g. BSD) and from there do essentially what they want.

All in all, companies campaigning for copyright assignments seem to have failed to understand two main points. The first one is that contributors motivated by Free Software ideals want the code to remain free not only for themselves, but for all recipients of the code. I'm not claiming that all contributors to FOSS projects belong to this category, but the fact that there is a considerable amount of such people is undeniable. Hoping that they will be fooled by tricks like the "any OSI approved license" agreement is foolish.

The second is that while volunteer Free Software communities do respect companies that provide huge amounts of code contributions, they also expect companies to figure out by themselves which business models make their businesses sustainable. Telling volunteers that current business models are not sustainable without proprietary relicensing is worthless. Volunteers just don't care. They want software freedom guarantees and are not willing to renounce to them in exchange of the wealth of companies, not even under the (implicit) threat that otherwise those companies will be forced to reduce the amount of their code contributions.


Update: explain in the initial note why my reasoning here encompasses both CCAs and CLAs. Thanks Tollef for noticing that it wasn't clear.

Thanks for the blog post! It's good to see a calm, rational and informed response to the issue.

I would blog about it myself but would probably be too passionate about it and end up saying things that I shouldn't.

It would be awful if a client comes to me and say "Hey help me fix this Unity thing on my laptop!" and I'd have to tell them "Oh, I'm sorry, you're using a non-free OEM version that shipped with your laptop and there's nothing that I can do for you" just to have them tell me "Oh, but you said this is free software and you can change or fix anything!"

Thank goodness that there is Debian and the social contract!

Comment by jonathan Sun 24 Jul 2011 06:17:05 AM CEST

In a nutshell, I think you are saying: Harmony opens the door to permissive or proprietary licences, and people who care strongly about their work only ever going into freely licensed derivatives shouldn't accept that. I think that is just about self-evidently true: if you would hate above all else to see your own come back to you enclosed in a product you cannot modify or redistribute, you should never licence it such that this is possible. You should also never contribute to a LGPL, Apache, BSD or zlib-licenced project.

Where I disagree is the assumption implicit in the title, that only strict copylefters are free software enthusiasts, or motivated by free software ideals. You must know there are many people deeply involved in free software who either prefer permissive licences, or who prefer copyleft but consider other factors. (Personally my first changes were to Apache projects, and more recently gratuitous licence incompatibilities between Python libraries lead me to use Apache2 for my personal work.)

I see in the penultimate paragraph you somewhat back away from this, but I think it's equally foolish to pretend every free software enthusiast is a hard-copy-lefter.

companies campaigning for copyright assignments seem to have failed to understand two main points

I think, on the contrary, Canonical understands very well that some community participants will not sign an agreement that opens the door to permissive or proprietary dual licencing, and that some people don't care whether companies contribute to free software or not. Their judgement is rather that some, not all, people will agree.

Harmony takes the chance of offering misleading options such as the one labeled "any OSI approved license". Via that option, which is seemingly fine, the copyright recipient will be allowed to relicense the contribution under some non-copyleft license (e.g. BSD) and from there do essentially what they want.

Calling the option of "any OSI approved license" a "trick" is disappointingly FUDdish. Is there anyone who cares deeply about licensing and yet does not realize there are OSI-approved permissive licences, and doesn't infer that from the immediately adjacent option to allow only copyleft licences? Such a person really needs to do more homework. Would you say the DFSG tricks people by including permissive licences?

Integrated IT Lifecycle Management with Hitachi IT Operations Director

When replying to such an absolutist post I can't help noticing to the left of this form you are running nonfree Google Ads code in my browser, and using it to advertise nonfree software products. I realize I could block it, and I realize this is beside the point, but still, it seems you're prepared to compromise sometimes too, when there's a financial incentive.

Comment by mbp Tue 26 Jul 2011 03:09:20 AM CEST

Where I disagree is the assumption implicit in the title, that only strict copylefters are free software enthusiasts, or motivated by free software ideals. You must know there are many people deeply involved in free software who either prefer permissive licences, or who prefer copyleft but consider other factors

Quoting from my blog post: "I'm not claiming that all contributors to FOSS projects belong to this [people who want their code to remain free for everybody, not only for themselves] category".

Would you say the DFSG tricks people by including permissive licences?

No, not at all. Because DFSG is used to evaluate what gets delivered (or not) directly to users. Something which is DFSG-free and gets in the hands of users does guarantee basic software freedoms.

OTOH, what I referred to as the OSI "trick" is used decide the rights that an intermediary (e.g. a company) will have. I consider it a trick precisely because while it reminds of things like DFSG classification, it can be used to take away existing software freedoms to the final users.

When replying to such an absolutist post

In all honesty, I think of this post of quite pragmatic. It contains truisms for people for copyleft-ists. But that is not the target of current campaigns in favour of copyright assignments. The target is people who still needs to make up their mind about whether copyright assignments are "good" or not. I feel legitimate to target the same people giving an argument about why they are "bad". Fair is fair, no?

you are running nonfree Google Ads code in my browser

Oh, what a class!

But thanks for pointing it out. I've set up Google ads years ago and I've regretted it pretty soon. The reason why they are still there is that I'm waiting for them to reach the 70 EUR threshold for getting paid (I've never been). The day I'll get paid I'll remove them and donate the earned sum to Debian. (Oh, and of course: no one is perfect; but I refuse the assumption that since I'm not, I'm not entitled to argument on what I think is good and what is not.)

Comment by zack Tue 26 Jul 2011 10:42:42 AM CEST

The body of the article acknowledges that many free software enthusiasts are not strict copylefters and can potentially accept contributor agreements, but the title says its unacceptable for [all] free software enthusiasts. This irked since I am a free software enthusiast who releases non-copyleft code, and I am hardly alone in that. Perhaps this is just pedantic.

OTOH, what I referred to as the OSI "trick" is used decide the rights that an intermediary (e.g. a company) will have. I consider it a trick precisely because while it reminds of things like DFSG classification, it can be used to take away existing software freedoms to the final users.

If I understand correctly, you're drawing the distinction that the DFSG is only applied to projects that already exist to decide (inter alia) whether Debian will redistribute them, whereas a Harmony OSI agreement invites people to contribute new code under terms that may later turn out to be more permissive than they expected?

I don't think Harmony is trying to trick anyone or that they particularly get any benefit from tricking people, so if this can be changed to be more obvious I'd hope they would make those changes.

I'm not sure what I would suggest changing though; maybe you can think of something. The template agreement says: "any licenses which are approved by the Open Source Initiative on or after the Effective Date, including both permissive and copyleft licenses" and that to me is pretty plain language that should leave no doubt that non-copyleft licences are going to be allowed.

of course: no one is perfect; but I refuse the assumption that since I'm not, I'm not entitled to argument on what I think is good and what is not.

I am not perfect either, so I could hardly advance such a rule. I know it's irrelevant; I just couldn't avoid gazing at the ads as I composed my reply.

Comment by mbp Wed 27 Jul 2011 05:01:19 AM CEST

the title says its unacceptable for [all] free software enthusiasts. This irked since I am a free software enthusiast who releases non-copyleft code, and I am hardly alone in that. Perhaps this is just pedantic.

I think article titles might use hyperboles and I don't even think this one is particularly heavy. But you do have a point. Since I had another grudge with the chosen (sub-)title, I've just changed it. Thanks for your remark.

I don't think Harmony is trying to trick anyone or that they particularly get any benefit from tricking people, so if this can be changed to be more obvious I'd hope they would make those changes.

Don't get me wrong, I don't think it's actually malicious, that's why I've used quotes around "trick". I do think however that the options have been crafted to avoid that people think too much at proprietary relicensing. You can see it as some sort of rhetoric, if you please; after all that seem to perfectly fit with the clear desire of some companies which are campaigning for CLA/CAA. If they really wanted to make clear where proprietary relicensing is allowed and where is not, they could have embraced more the "creative commons model" they are using and offer checkboxes, having as one of those checkboxes "do you want to allow proprietary relicensing?". I think that would have been a more fair way of presenting the options on the ground, at least from my point of view of free software enthusiast.

Comment by zack Wed 27 Jul 2011 01:21:44 PM CEST