Few topics have roiled the normally staid field of patent law more than the expansion of patents into the software industry. As we've documented in some depth here at Ars, the last decade has seen the Patent Office grant an unprecedented number of software patents, which has resulted in a dramatic increase in software-related patent litigation.
Earlier this month, the Brookings Institution hosted a conference on abstract patents. Judging from the speakers at that conference, there's a big cultural gap between technologists and the patent lawyers over the issue of software patents. The lawyers recognized that the patent system isn't working well for the software industry, of course, but most favored incremental changes to patent doctrines rather than a fundamental reconsideration of the concept of allowing patents on software. Perusing the comments sections of Ars Technica or Slashdot suggests that many of the people on the business end of software patents wish the patent bar would simply leave the software industry alone.
Given the case law that now governs software patent disputes, it may surprise both groups to learn that the Supreme Court actually declined to extend patent protection to software algorithms, preferring to leave the matter to congress. In this feature, we'll take a close look at the Supreme Court's classic trio of software patent decisions. We'll explore what the Supreme Court originally said about the patentability of software, how the court distinguished between software and non-software patents, and what the consequences would be if lower courts more aggressively applied the limits on software patents that the Supreme Court articulated a generation ago.
Prologue: "Residual overhang of physicality"
One of the most basic disagreements in the software patent debate is over how to interpret the Supreme Court's rulings on the patentability of software. Software patent critics have long argued that the high court's precedents rule out software patents. Brookings Institution scholar Ben Klemens made that argument in a 2005 book (we covered Klemens' anti-software-patent organization last year), and his arguments were echoed by the Software Freedom Law Center's Eben Moglen in an amicus brief filed with the Supreme Court in 2006.
Software patent supporters have a very different perspective. For example, in a recent episode of the Intellectual Property Colloquium podcast, host Doug Lichtman talked with two leading patent scholars about the implications of last year's In Re Bilski decision, which placed new limits on patenting abstract concepts. Lichtman's guests, law professors John Duffy and Robert Merges, seemed to regard past Supreme Court limitations on software patents as confused and anachronistic. They lamented the "pointless metaphysical investigation" that the Supreme Court's precedents require of those seeking software patents. We are, they said, "stuck with this artifactual, residual overhang of physicality" that is "just the price we have to pay to get a software patent these days."
Interpreting legal decisions is never an exact science, so it's possible to find passages in the Supreme Court's decisions that support either side of the software patent debate. But an essential theme of all three decisions that we'll look at is that mathematical algorithms are outside of the scope of what may be patented, and that it's up to Congress, not the courts, to change that. Indeed, we think it may actually be an advantage that the Supreme Court decided these cases so many years ago. The fundamental principle that software is mathematics has not changed over the last 35 years, but the growing complexity of software systems may obscure that fact, especially for judges with minimal technical experience. As the problems with software patents become more obvious, judges and bureaucrats alike would do well to dust off their copies of the Supreme Court's classic software patent decisions. We think they'll find that the Supreme Court's decisions are more relevant—and more restrictive of software patents—than they might have thought.
reader comments
26