Why (not) to relicense VLC under LGPL?

On August 4th 2007, over 5 years ago, the conversion went like this (HTMLified for readability) on IRC:

littlejohn: About GPLv3, i've read the thread started by courmisch. I think in any case that LibVLC should be LGPL, since it can be treated like a library.
courmisch: littlejohn: LGPL? I take it you will handle the paperwork to obtain relicensing agreement from every contributor?
courmisch: littlejohn: I'd love to have LGPL for LibVLC, but that's not realistic.

After the developers meeting in 2009, Rémi Duraffort started mailing contributors for relicensing. Jean-Baptiste Kempf eventually did most of the work, and (kind of) completed the actual process of relicensing LibVLC to LGPL 3 years later in November 2012.

In retrospect, I was arguably proven wrong about relicensing being unrealistic. Technically, my claim still stands since streaming modules remain yet to be relicensed (only playback support is LGPL as of now). But I believe it is now only a matter of time before Jean-Baptiste disproves it.

He has documented the tedious and intricate process in details and even made a presentation on the topic at the FOSS.IN/2012 conference. However a number of curious have asked not how but why was that was done in the first place. The VideoLAN project has not released any clear statement. Before I go further, I must point out that what follows is solely my own personal views. (This article does not represent the views of the VideoLAN project or any other party, including but not limited to my employer.)

The most prevalent reason

...why VLC developers and former developers agreed to relicense is, I think, not so much as resolve to relicense as a lack of opinion. Most developers do not really care about what many in the open-source community refer to as the politics (often pejoratively). They wrote the code for fun and/or to address a need of theirs, i.e. to scratch their own itch. They did not really care about the license was. Or as BSD proponents would put it, they do not work to make software free, they work to make software better.

Politics

In fact, at least one, likely several, of the few developers that refused to relicense from GPL to LGPL, cited free software politics or ethos as the reason. To be fair, I think that is a valid point: The LGPL is more relaxed about mixing with proprietary software components, and so in a way it is less free than the GPL.

Relicensing as LGPL is a bet. I think there is a good chance that it benefits VLC and LibVLC. But there is also a risk that it ends up causing more harm than good. For instance, a company could develop and publish an useful but proprietary extension (for instance a plugin) and damage the thriving community. Users could end up considering VLC as inferior to the proprietary software, or become dependent on it. Whether you like it or not, most users care about software functionality much more so than software freedom (that is not a value judgement, just an observation).

Commercial interest

The main potential benefit, and also the main reason for relicensing was to attract more developers, and more potential business opportunities for existing contractors. Their are several reasons why I think this might happen:

Our most direct "competitors" within the open-source ecosystem,such as FFmpeg/libav and the gstreamer framework, are already licensed under LGPL. Less flexible means less competitive, and the GPL is (by design) less flexible than the LGPL.
Many companies insists on having proprietary codec backends, usually because of concerns involving intellectual property rights. Many well-known Linux distribution companies such as Canonical do use proprietary codecs.
The LGPL is intrinsically compatible with many more other open-source licenses than the GPLv2, or even the GPLv3. Relicensing enables the use of fine free software libraries that could not be used earlier.

VLC sorely needs more developer hands on it, as the number of active contributors has dwindled over the past years. That is why I personally think that the LGPL approach is at worth trying. Only time will tell how well that works out!

Mobile trend

Some people also hope that the LGPL will help release VLC on tightly locked mobile platforms. They are wrong though.

Firstly, VLC proper is set to remain distributed under the GPL, as there are absolutely no plans to relicense the existing user interfaces. Only the underlying library, LibVLC has been relicensed. In practice, mobile platforms need their own UI anyway. So VLC as we know it will probably never be published on the likes of Android or Windows Phone.

Secondly, the Android market terms are already compatible with the GPL: it has an explicit clause to allow open-source software.

The compatibility with the terms of the Windows Store is uncertain at this point, whether with LGPL or GPL. Additionally, the trademark conditions seem incompatible with the VLC trademark license (though VideoLAN could perhaps change it), so another application name would be required.

As for iOS, the audio and video output remain under the GPL because the developer of the short-lived MobileVLC explicitly refused to relicense. While I am not aware of an official reason, the free software politics from above would not seem to explain his decision. Otherwise he would not have published MobileVLC to begin with. I also do not believe that Apple would be to blame in this instance, since several former VLC developers-turned-Apple employees agreed to relicense (everyone of them stopped contributing immediately upon joining the fruit company, I did see a pattern there). By elimitation, I assume that he is spiteful for the removal of MobileVLC from the app store.

Reverse engineering

Last but not least, one of the original VLC developers agreed to relicense so that there would be more proprietary plugins to reverse-engineer VLC plugins that fully proprietary solutions. This might seem a bit perverse, but it could yield interesting results too.

Remlab