A Technical Critique of Fifty Software Patents

The legal, economic, and technical literatures are replete with discussions on the desirability or otherwise of software patents. But each strand of this debate has some limitations. The legal literature is for the most part based on case law, and thus is difficult to generalize. The economic literature is typically based on the statistical analysis of large numbers of patents, but tends to ignore the technical arguments either for or against software patents. The technical literature - frequently hostile to patents - is most often based on an examination of a small number of pathologically bad patents, and thus like the law literature does not shed much light on software patents in general. This paper attempts to fill in these gaps in the analysis. We focus on the technical merits of what, on the surface at least, should be good software patents - a set of 50 most highly cited patents. Of course our approach cannot address all of the controversies surrounding software patents. Instead, we aim our inquiries on what insights technical specialists might provide lawyers, economists, and policymakers studying software patents. In particular, we ask, from a technician's viewpoint: Are software patents too obvious? Is the level of disclosure adequate? Are patents real - do they represent real innovations or are they strategic? Are software patents too broad in their technical coverage? Do software patents last too long? Do software inventions need additional copyright protection? We conclude that for the set of patents examined here: they are not too obvious in general; the level of disclosure is often less than optimal, indicating the need for reform; they are almost all for real innovations; some are too broad technically but most are not; they do last too long, but this has little to no apparent impact on the industry; and some software inventions need both patent and copyright protection. Taken altogether, software patent critics do have some valid points (especially regarding disclosure), but the accusations that they are generally obvious, trivial, and impede progress do not stand up to scrutiny