Defining Software Patents: A Research Field Guide

Abstract

What, exactly, distinguishes a patent on a software innovation from patents on other kinds of innovations? While the question may appear esoteric at first glance, it is the crucial first step in any empirical study of software patents. The policy debate over software patents - whether they should be allowed at all, what should qualify, and whether they are used to block or encourage innovation - has been raging for quite some time on both sides of the Atlantic. Meaningful contributions to this debate should be rooted in firm empirical analysis, as opposed to individual anecdotes. But supplying firm empirical analysis requires data on software patents, and thus the question posed above. In this paper, I examine the definitional choices made thus far by the emerging empirical literature on software patenting and explore the implications of those choices. The emphasis here is on how the definition selected can affect the conclusions made. I find that different definitions do lead to datasets with distinct characteristics. While no one definition emerges as clearly right, some methods appear better than others. Much more empirical research on software patenting is needed to improve patent reform efforts, and that research should begin by justifying the definition of software patent chosen.