It’s a shame to write about something with the sole intent of being critical, so let us first say that Benjamin Black’s commentary about the GPL in the context of cloud computing is of its time. That is, as the web starts to fulfill its promise of delivering more and more computing, the importance of Free software might appear diminished. Tim O’Reilly, for one, has been forecasting this for several years already, and the AGPL was written with this in mind. (Incidentally, fans of plain speaking might note that Greg Papadopulos has pioneered a new term for cloud computing: “the network”.)
Rather than make things even more complicated than they are, let’s be clear. Free software is about your rights in relation to a device that you own (iPhone owners: Apple controls your property). When we consider the increasing intimacy of technology in our everyday lives, work and social interactions, the importance of this becomes more and more apparent. You may want a vendor to control what your device does, but you should at least have the right to opt out. And this also explains why many people feel that the publc sector should have a proclivity mandate to use Free software. It starts to seem unreasonable that a government body, paid for by and representing the public, should allow its infrastructure to be controlled by a commercial interest.
But cloud computing is the use of someone else’s infrastructure. And so, Mr Black is entirely correct in saying that rules around, for example, data portability, are very important and not yet well-defined.
Now, none of this invalidates the importance of a strong copyleft like the GPL, and it certainly does not mean it acts unfairly. Mr Black talks about “those who control a software project” deciding who “gets paid”. In this model, the dual-licensing model (licensing software either as open source or with a commercial license), is the only way to “get paid”.
Leaving that dubious assumption aside, where does control come from? Control of a project is a function of its ownership of the copyright and the trademark of the software. Control over copyright is gained by creating the work, paying for its creation, or by paying for the rights or by receiving those rights as a contribution. By no means all GPL projects conform to this consolidated ownership, and those that do may be under the control of a commercial body or they may be held in trust elsewhere. I believe that the Free Software Foundation Europe offered this service a while ago. (And as an aside, I am delighted to read that some people welcome Sun’s policy of dual-copyright ownership. This was a labour of love for me and others to get the several hundred open source project at Sun to use the same contributor agreement).
So, when Mr Black is complaining that the GPL is “like DRM”, he is referring to projects which, in order for you to contribute back to the code base that will bear the trademark by which the project is widely known, require that you assign some measure of copyright ownership back to the original owner (or proxy owner) of that code base. It really, really, really is not like DRM.
Enough of this navel-gazing. Let’s simplify these problems:
The rights of the user in relation to their own property are reflected in Free software.
The rights of the developer in relation to a project they participate in are reflected in open source. (Some people consider that the use of trademarks are an important, under-investigated are of this – I am one of them).
The rights of the user in relation to using someone else’s infrastructure are not yet governed by any widely understood framework. The Open Cloud Manifesto is the first attempt (I am aware of) to address this. It will take time. That neither the Free software movement nor the open source movement address this is no criticism of either.
More to the point: exercising ownership rights which do not curtail anyone else’s freedom in relation to that work (c.f. DRM), but may impair their ability to make money on a piece of work you own is not, really, really not, wrong.