This Article is the first part of a wide study of the role of intellectual property in the software industry. Unlike previous papers that focus primarily on software patents-which generally are held by firms that are not software firms-this Article provides a thorough and contextually grounded description of the role that patents play in the software industry itself.
The bulk of the Article considers the pros and cons of patents in the software industry. The Article starts by emphasizing the difficulties that prerevenue startups face in obtaining any value from patents. Litigation to enforce patents is impractical for those firms. Efforts to obtain patents divert the firm's focus from the central task of designing and deploying a product, and the benefits of excluding competitors are limited for firms that cannot themselves exploit the relevant technology. Once the firm is larger, a number of potential benefits appear. First, despite concerns that patents are not effective to appropriate profits from innovation in the software industry, a substantial number of software startups do have patents of sufficient strength to exclude competitors. That important finding, taken with the fact that the principal targets of those patents are much larger firms, suggests patents are more beneficial to small firms than to large firms. The Article then considers indirect effects related to the use of patents in cross-licensing transactions and in providing information about the firm. The first benefit may be substantial to firms that obtain patents, but the Article rejects patent use in cross-licensing as a net benefit to the industry: absent some other benefit, all firms would be better off saving the costs of obtaining patents. The information benefits, in contrast, seem to be net improvements to the innovation system. The central question, which I do not attempt to answer here, is whether those benefits are sufficiently substantial to justify the costs of obtaining the patents.
The Article then turns to the prominent claims that the enforcement of software patents has hindered innovation in the software industry through creation of a patent "thicket. " The Article rejects those claims for two broad reasons. First, notwithstanding the empirical analysis of R&D spending in papers by Bessen, Maskin, and Hunt, direct evidence of high R&D spending in the software industry undermines claims that software patents cause firms to reduce R&D spending. Second, the actual structure and practices of the industry belie any claim of a patent thicket. Relying on interviews that I conducted and publicly available information, I show that the development of young firms in the software industry is not significantly constrained by large patent portfolios in the hands of incumbent firms.
The Article also contextualizes the role of patents by examining the relatively weak protections that copyright and trade secret afford. At bottom, neither of those systems can provide a useful mechanism that allows small firms to appropriate the values of their inventions. If such protection is a significant positive benefit of the patent system, it is equally true that neither copyrights nor trade secrets contribute (or can contribute) significantly in that respect, however useful they might be in other roles (such as preventing piracy).
The Article closes by considering critically the possibility of middle ground responses that would limit patent rights in the industry but not abolish them entirely. First, I criticize a possible registration system that might provide information benefits without the costs of excluding competitors. I argue that such an approach is impractical both because it would be difficult to disentangle the information benefits from the right to control technology and because of my sense that software firms would have an inadequate incentive to participate in such a system. Second, I consider the possibility of special limits on the rights of "trolls, " small nonoperating firms formed solely to litigate patents. …