Dr. Till Jaeger
Recent days have brought under the spotlights some interesting developments that invoke GPL licensing, at least partly. Unsurprisingly, Google’s purchase of Motorola Mobility has been receiving massive media attention. And as part of an ongoing debate, commentator Florian Müller argues that Linux developers can legally attack manufacturers of Android-based mobile phones for violating the GPL license in case those manufacturers have once breached the GPL conditions.
Fierce competition characterizes the mobile phones market, especially in the area of smartphones, as the combination of commonplace Internet technology and GPS-services opened a new battlefield. Growing centralization tendencies positioned in the forefront three main actors: Apple, with its pioneering hardware and software technology, the coalition between the old software giant Microsoft and former market leader Nokia, and Google. The latter, armed with its Linux-based operation system Android, has been supplying to mobile phones makers such as Samsung and HTC. Now, Google’s purchase of Motorola Mobility makes its aspiration to becoming an integrated software, hardware and Internet services group with its own smartphones warehouse impossible to overlook.
In this context, questions regarding the implications of GPL licensing come to mind. One of these questions have been receiving considerable attention after blogger Florian Müller argued that Android hardware makers that once have violated the GPL terms (e.g., by not offering the relevant source code) automatically lost their position as licensees and must acquire permission anew from Linux rights holders on an individual basis. The meager prospects of such re-licensing to occur in reality would place Linux developers in a convenient position to bring legal actions against once-GPL-violators quite easily.
To be sure, such interpretation of the GPL license is hardly true under German law. It is correct that the consequence of a GPL violation is the revocation of permissions stipulated therein (cf LG München I, Welte v. Sitecom). This does not mean, however, that acquiring GPL permissions anew requires the explicit authorization of every rights holder – a nearly impossible operation as a practical matter.
Also a licensee that once has breached GPL terms might acquire a new license and be granted the same permissions by bringing into circulation products that comply with the GPL conditions. After all, GPL licensors offer everyone the opportunity to use their software. A permanent termination of GPL permissions has no trace in the GPL license. Quite the opposite: Section 4 of the GPL, in the third sentence, stipulates that further use downstream in the chain of licenses is not affected by termination of the license upstream, inasmuch as users downstream themselves abide by the GPL terms. This implies that the “automatic termination” provision reflects the purpose of ensuring compliance with the GPL, as opposed to spelling out permanent termination and exclusion of violators. In short, there is no basis for differentiation between a user that once violated the GPL and a new user. In both cases, the license becomes effective as the licensee accepts the offer embodied in the GPL and implicitly attests to this acceptance by a corresponding conduct. Contacting each and every rights holder for obtaining permission is therefore not necessary (cf the ifrOSS-GPL Commentary, p. 6).
Müller’s reference to a blog post by Edward Naughton brought in order to support the permanent revocation theory is misguided. That post highlighted GPL enforcement by gpl-violations.org and Mr. Harald Welte. To be sure, all the relevant case law (see collection of court decisions available at the ifrOSS website) concerns instances, in which courts denied use of the software as a direct consequence of violation of the GPL terms and only as long as such violated continued. Rights holders have never asked the courts to recognize a permanent reversion of rights or any kind of long-term termination independent of an acute violation.
I do not consider here whether the result would be the same under U.S. law. The least that can be said more generally is that excluding users who are willing to comply with the GPL terms from the freedoms it offers appears difficult to reconcile with the very idea of Free Software to begin with. It is altogether a different question whether, under German law, willful and repeating GPL violations would trigger the general termination provision codified under Section 314 of the German Civil Code (BGB). The application of this provision must be examined on a case-by-case basis.
A question of greater practical implications is whether Android, in connection with its free access characteristics, can constitute a real alternative to Apple’s walled garden. The answer will greatly depend on the technical skills of users and their ability to develop their own Andriod versions for smartphones, thereby keeping a safe distance from the established Android market and from Google’s marketing strategies that rest on evaluation of personal data. From a technical standpoint, the opportunity to run Android operating system might be undermined by the application of digital signature, the so-called Tivoization. Whether the GPLv2 permits such practices is debatable. The FSF opines that in the lack of explicit arrangement in the license, preventing such technical restrictions is impossible. In the GPLv3, the FSF has already introduced an express regulation to reflect its aversion towards technological protection measures.
However, it remains questionable whether the above reading to the GPLv2 is the most plausible reading available. After all, its Section 3 makes plain that the obligation to surrender the “complete corresponding source code” includes technical means necessary for the installation and modification of the software version. According to this provision, one must provide users with the “scripts used to control compilation and installation of the executable” – apparently reflecting the wish to ensuring for licensees the ability to run the software on hardware. Applying German contract law, there is apparently an obligation to facilitate for licensees the ability to run on the original hardware more recently-developed software versions. Whether this might result in annulment of warranty is a different question altogether.
The attempt of hardware makers that integrate GPL software in their firmware to block access in nicely exemplified in the recent FritzBox litigation (see news report from 20.06.2011). In this case, it is not only technical but also legal measures that are been employed to build walls around firmware that include open software elements (disclosure – the author is counsel to one of the parties involved in this dispute). Time will tell whether Google would be following in the footsteps of such hardware makers, or rather, would prefer to promote free operating system in the real sense of the term. Whether injunctive claims made by Linux developers acquire a new dimension will significantly depend on Google’s choice at this juncture – and this, regardless of possible antitrust implications of the combination of market dominance and dubious commercial practices.
Thanks to Dr. Zohar Efroni for the tranlation of my news report.