Patents and Open Source Are Not Always Like Water and Oil

Patents have long served to enable innovation while assuring the complete disclosure of inventions and a reasonable return to inventors. Many open standards incorporate patented inventions. To ensure public access to IP in open standards, governments and industry have supported an equal rights provision: reasonable and non-discriminatory (RAND) licensing for patents included in standards. But for some, equality is insufficient. Some in the OSS community want to do away with these royalties (or any other RAND based licensing terms, such as scope of use or prohibitions on sublicensing) all together.

Many participants in the open-source community are implementing standards and licensing patents where necessary to create highly successful, innovative products, so why would others be calling for abolition of RAND licensing?

The answer lies in licenses: not the patent licenses, but certain open source software licenses.

Simply put, there are different flavors of OSS license. Some, such as the BSD and Apache licenses, allow OSS developers to design software that plays well with others. These licenses allow OSS developers to include software (any software, not just software that implements standards) that may require the developer to pay royalties. So if you are an OSS developer and you choose to use, for example, a BSD license, the question “can I use the standard in my OSS?” is invariably answered “yes.”

However, other open source licenses, specifically GPL licenses, are intended to ensure no IP royalties flow from not only the underlying OSS, but any add-ons to, or components of the underlying OSS code, including any standards implemented in the software.

So if you are an OSS developer who decides the GPL is the best license for your software, then yes, there is a problem. It’s a problem written by the authors of the GPL licenses who seek to exclude any patents from GPL licensed software – not a problem created by the patent license or the standards setting body that adopted RAND as the right approach to licensing of patents embodied in a standard. Patrick Durusau describes this well in “Self-Inflicted Discrimination and GPL.”

Many OSS developers, like commercial software developers, license these patents with royalties based on unit sales (so called “per unit” royalties). But this metric is proscribed by the GPL; so alternative valuations such as lump sum or royalties based on revenues can be substituted to enable OSS developers to use a standard in a way compatible with the GPL. In fact, according to an analysis of the Redhat/Firestar settlement not even OSS developers using GPL 3.0, the most stridently anti-patent OSS license, are entirely foreclosed from licensing such patents on terms that include some form of royalty.

So what is the best public policy?

Patents and RAND licensing stimulate innovation, remunerate inventors and enable interoperability, even interoperability for highly successful OSS. When the situation warrants, the market gets creative in working out reasonable royalties on GPL-licensed software. As Hellström, Kramler and Bulst of the European Commission Antitrust unit recently wrote in Holding Standardization to a Competition Law Standard (Concurrences Review, February 2010): “As competition law should be business model neutral, there should be no preference for open source based models over other models. From a competition law perspective royalty free access to IPR is not necessarily required.” It makes no sense to abandon patents and RAND licensing, fundamental aspects of standards-setting, to accommodate one type of open source software license.

Join the Forum discussion here: http://www.talkstandards.com/questions-for-standards-and-oss-open-forum/