Beware of Proprietary Drift

The Free Software Foundation (FSF) announced yesterday a campaign to collect a clear list of OpenOffice.Org extensions that are FaiF, to convince the OO.o Community Council to list only FaiF extensions, and to find those extensions that are proprietary software, so that OO.o extension developers can focus of their efforts on writing replacements under a software-freedom-respecting license.

I use OpenOffice.Org (OO.o) myself only when someone else sends me a document in that format; I’m a LaTeX, DocBook, MarkDown, or HTML user for documents I originate. Nevertheless, I’m obviously a rare sort of software user, and I understand that OO.o is a program many people use. Plus, a program like OO.o is extremely large, with a diverse user base, so extension-style improvement, from a technological perspective, makes sense to meet all the users’ requirements.

Unfortunately, the social impact of a program designed this way causes danger for software freedom. It sometimes causes a chain of events that I call “proprietary drift” — a social phenomena that leads otherwise FaiF codebases to slowly become, in their default use, mostly proprietary packages, at least with regard the features users find most important and necessary.

Copyleft itself was originally designed to address this problem: to make sure that improved versions of packages were available with as much software freedom as the original. Copyleft isn’t a perfect solution to reach this goal, and furthermore many essential software freedom codebases are under weak copyleft and/or permissive licenses. Such is the case with OO.o, and the proprietary drift of the codebase is thus of great concern here.

For those of us that have the goal of building a world where software freedom is given for all published and deployed software, this problem of proprietary drift is a terrible threat. In many ways, it’s even a worse threat than the marketing and production of fully proprietary software. This may seem a bit counter-intuitive on its surface; logic would seem to dictate that some software freedom is better than none, and therefore an OO.o user with a few proprietary extensions installed is better off than a Microsoft Word user. And, in fact, none of that is false.

However, the situation introduces a complexity. In short, it can inspire a “good enough” reaction among users. Particularly for users who have generally used only proprietary software, the experience of using a package that mostly respects software freedom can be incredibly liberating. When 98% of your software is FaiF-licensed, you sometimes don’t notice the 2% that isn’t. Over time, the 2% goes up to 3%, then 4%. This proprietary drift will often lead back to a system not that much different from (for example) Apple’s operating system, which has a permissively-licensed software freedom core, but most of the system is very much proprietary. In other words, in the long term, proprietary drift leads to mostly proprietary systems.

Sometimes, I and other software freedom advocates are criticized for giving such a hard time to those who are seemingly closest to our positions. Often, this is because the threat of proprietary drift is so great. Concern about proprietary drift is, at least in large part, the inspiration for positions opposing UbuntuOne, for the Linux Libre project, and for this this new initiative to catalog the FaiF OO.o extensions and rewrite the proprietary ones. We all agree that purely proprietary software programs like those from Apple, Microsoft, and Oracle are the greatest threat to software freedom in the short term. But, in the long term, proprietary drift has the potential to creep up on users who prefer software freedom. You may never see it coming if you aren’t constantly vigilant.

Originally published by Bradley M. Kuhn under the Creative Commons Attribution-Share Alike 3.0 United States License.

Hola! Looking for an answer to a question but did not find it? Then surf on over to the Questions & Answers section. It's a brand new addition to our site, and we are waiting just to answer your question(s).