The 1996 World Intellectual Property Organization copyright treaty, the 1998 Digital Millennium Copyright Act (DMCA), the 2001 European Union Copyright Directive (EUCD), the 2006 French bill known as "DADVSI" all grant specific protection to so-called digital rights management techniques. We contend that this protection introduces a de facto new kind of intellectual property.

Authors (or publishers) of computer software are protected by copyright - unauthorized copying of software is illegal in most jurisdictions. This protection is very similar to that granted to books, photographs, songs etc. (though in droit d'auteur countries, software authors typically do not enjoy as many "moral rights" as authors of other kinds of work). Accordingly, this protection only covers specific programs, as opposed to the abstract ideas and techniques that software uses.

This is an important distinction to make. Computer software is generally written as source code (a human-readable, machine-processable text), which is then translated (or, in technical terms, compiled) into directly executable code. In most cases, the source code does not express novel techniques; yet, it is an original work, protected by copyright, in the same way that bad novels without novel ideas are protected by copyright. If the software presents interesting techniques, other software producers may decide to implement the same techniques in their own products, provided that they do not copy the software itself, but only the ideas that it implements.

In the industrial world, techniques are not protected by copyright, but by patents. Patents grant to the inventor of a novel technical technique an exclusive right over the exploitation of that idea, in exchange for the publication of the idea in the patent document. Whereas, previously, inventors had to keep their inventions as trade secrets if they did not wish competitors to copy them, patents allow inventors to publish their inventions and to grant licenses to other suppliers, thus encouraging the spreading of new ideas and new techniques. Also, the duration of patents is limited, typically to 20 years, which is supposed to be a good compromise between the general interest of having inventions free for everyone and the return on investment for the inventor, which is supposed to encourage innovation.

The application of patents to computer software techniques is highly controversial. While some forms of "software patents" are allowed by the US Patent and trademark office, a directive officially authorizing them was refused by the European Parliament. Reasons for opposing software patents are manifold:

  • Software "techniques" often consist in algorithms; that is, mathematical descriptions of computing processes. Traditionally, algorithms have been considered in the same way as mathematical formulas are - they are not patentable in themselves, though there are ways to work around this restriction.
  • The experience with the administration of "software patents" by the US Patent and Trademark office has been highly controversial. In many cases, patents were granted to techniques already known in the state of the art or immediately derivable by a technical person - both of which are normally precluded by patent rules.
  • The pace of the computing industry is very fast compared to other industries. If one for instance considers medical suppliers, medicines may be used for decades after they are invented, but patents only grant a monopoly over the first couple decades; in addition, the lengthy approval procedures subtract from the "exploitable" patent period. In comparison, 20 years may often go beyond the complete usefulness period of a software invention.
  • Finally, the economics of software are governed by powerful "network effects". This is important with respects to DRMs, and we shall thus explain these effects in more detail.

In many industries, the intrinsic qualities of a product are overshadowed by its compatibility with what other users have adopted. For instance, the Betamax format for video cassettes ultimately lost to the VHS format, whereas many have argued it was technically superior, because VHS was more widespread and the differences were not compelling enough to justify an alternate technology. In computing, there is the constant issue of "compatibility" - meaning the ability of several software or hardware components to work together.

Historically, it has been often the case that hardware manufacturers introduced gratuitous incompatibilities between their hardware and other manufacturers' (or even between different hardware lines in their own company) in order to lock customers in : for instance, they produced computers that would only work with their own terminals and printers. Patents over connectors, interfaces, and communication protocols may prevent competitors from manufacturing compatible systems.

The same applies with software, especially in today's wired world. We generally do not buy our software in isolation - because of the Web and emailing, we wish our word processor, our spreadsheet, etc. to be compatible with those that our colleagues and friends use, for instance. In the past, there existed a number of competing office suites with significant market share - but today there exists Microsoft Office and, far less used, its free competitor OpenOffice.org. With regards to operating systems - the essential software infrastructure on computers - there exists only two widespread desktop systems : Microsoft Windows and Apple's MacOS, though the free competitor Linux makes some inroads. One reason for this is that for many applications, one is more or less forced to use Windows or MacOS. For instance, there exist proprietary multimedia (audio and video) formats for which documentation is not freely available and whose designers only supply players for Windows and MacOS. If one wishes to play such content, the easy solution is to have Windows or MacOS. The same applies for, say, documents supplied in the Microsoft Word format - if one requests good compatibility, one has to have a copy of Word, which is available only for Windows and MacOS but not for Linux. Thus, there exist strong network effects that reinforce quasi-monopolies.

Despite these difficulties, some software makers manage to produce "compatible" software - software capable of reading proprietary formats. They often do so by "reverse engineering" software and file formats for compatibility purposes, which is allowed by current European law. Now, here's the catch with DRMs. DRMs "technical methods" are simply a new word for "copy protection" systems - systems that prevent unauthorized software from playing or copying content. If one wishes to produce software compatible with a DRM "secure" format, one will essentially have to understand how to work around the DRM system. For instance, if the DRM-encumbered file is encrypted for copy protection purposes, the "authorized" software will have to decipher it, and so will the "compatible" software.

The original language of the DADVSI bill criminalized working around DRM protections, without exceptions. Many people got concerned that this law would indirectly allow designers of DRM formats to claim exclusive rights of making software or hardware capable of reading such formats, since any maker of compatible software or hardware could be accused of circumvention of the DRM, a felony. Designers of DRM systems would thus enjoy rights equivalent to those of a patent (exclusivity of production of compatible systems) without the limitations (limited duration and obligation of publication), even though software patents had been refused by elected representatives. Because DRM formats are now used in video, audio, and even text, and because of the "network effect", this could have had severe consequences for competition. For instance, neither Apple nor Microsoft produce players for their popular DRM-encumbered formats for the competitor operating system Linux, and the law could have been used to outlaw any "compatible" player; thus, Linux would have been unable to play most videos downloadable from the Internet, making it unsuitable for many desktop deployments.

Because of this risk, the French parliament ended up including provisions that circumvention of DRMs was legal for compatibility and security reasons. We have already explained the compatibility issue; the security issue is justified by the actions of some manufacturers whose "copy protection" software was insecure and even tried to spy on the users (see for instance the Sony "rootkit" scandal). Should the original law have passed, producers of such systems could have prevented the revelation of their flaws by security researchers by threatening prosecution for circumvention.

There is certainly a serious debate to be had on the opportunity of allowing "software patents", and, if allowing them, under which conditions. There is also a debate to be had on business model of the entertainment industry in the age of the Internet. But, certainly, copyright for music and films should not be used as an excuse to introduce harmful de facto patents in other industries, especially when this "new kind of intellectual property" would clearly hamper competition.