Many discussions around the interplay between Open Source Software (OSS) licensing and Reasonable and Non-Discriminatory (RAND) licensing of essential patent claims immediately jump to the question of patent royalties. To the lay person, it might appear a simple debate: patent royalties in RAND regimes are the basis of tension between RAND and “Open Source.” Peering deeper into the issues, however, reveals that neither RAND nor OSS are singular or uniform concepts capable of such a simple comparison as this. Moreover, lost in the complexity of the purely legal analysis are a range of important policy questions (some of which I have tried to set out below).

There are dozens of OSS licenses in use and each one is unique in the way it treats patents. But one thing is clear- it cannot be said as a general matter that open source licenses are incompatible with patents. For instance, a number of what some people might call “commercial style” open source licenses (such as the Apache 2.0 or Eclipse Public License) distinguish between those patents the distributor must license (roughly any of its patents that read on the part of the code it modified or contributed to) and those that it need not provide (patents that read on other unmodified portions of the code) in connection with distribution of the code. It’s clear that distributors of OSS licensed code aren’t thereby giving away all of their patents, even with respect to the broad functionality in that code. This analysis suggests that, at least among this category of OSS license, there is not an inherent tension between OSS and patents.

Aside from what a particular OSS license might require a distributor of the code to do with respect to patents it owns that read on the code, there is an even more important question to ask: How does a particular OSS license treat third party patents to which the distributor has taken a license? I’ve often heard the refrain “Open Source prohibits third party patent licenses with per unit royalties” as the underlying justification for the claim that RAND (which is after all a type of third party patent license) and OSS are in conflict. While it may be true that the very definition of “Open Source” precludes the distributor from collecting a royalty from recipients of the code, there is nothing in the Open Source Definition which precludes the distributor from taking a third party patent license (including one that has a per unit royalty).

The Open Source Definition is in essence a least common denominator for OSS, so there is nothing that would prevent the author of an OSS license from including a term that limited or outright prohibited the distributor from taking patent licenses from third parties. Few OSS licenses do this though, so for the vast majority of OSS licenses, there are no restrictions on the terms of patent licenses (including those negotiated in the patent context) that the distributor might take from a third party. I’d be remiss if I didn’t mention that the GPL family of licenses does include some restrictions on the types of third party patent licenses that a distributor of GPL code can take, but it’s not clear to me that the GPL would prohibit any kind of royalty associated with the economic activity of the distributor (e.g., fixed fee or based on product revenues as opposed to per unit).

This is a complex legal discussion and even experienced intellectual property and licensing lawyers can easily lose the thread. But there are important policy questions that we need to consider as we work through this set of issues. I’ve highlighted just a few of these that I hope might be the basis for continued discussion:

• Are RAND and OSS so in conflict that one really needs to yield to the other, or is this instead just a matter of finding solutions for a set of edge cases?
• Would one model support the full level of innovation that we currently have? Or are we better off in a world where we let multiple business models and innovation regimes battle it out?
• Are there issues beyond a per unit royalty term, such as restricting licenses to only “compliant” implementations or prohibiting sublicensing?
• Aside from royalties, don’t these other terms facilitate important objectives of the standards system (i.e., urging implementers to be compliant with the specification) and how do we balance those objectives into the debate?
• RAND typically covers only essential claims, but we know that there are a wide range of non-essential patents that implementers want to get access to as well (for so-called freedom to innovate). Can we contemplate changes to RAND (under which essential terms are licensed) without regard to the licensing practices for non-essential claims?
• Shouldn’t we distinguish between the implementer who can show that a particular RAND term is in tension with obligations imposed on him by a particular OSS license

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