05 Sep 2004

Patent Defense

As I wrote last time, this summer is seeing another round of concerns about the vulnerability of free software to patent infringement claims. But the free world is assembling the machinery with which to meet the risks posed by patents on software inventions. Rapid developments are occurring in the area, so it can be hard to interpret what the news means. In the last few weeks, for example, part of the solution has widely been reported on as though it were part of the problem.


One difficulty that patents create for everyone who makes software, whether proprietary or free, is that it’s impossible for a working engineer or programmer to have any idea whether her work infringes an existing patent. Patent searching is exceedingly expensive. Patent lawyers and agents have an incentive to make the claims in their patent applications as broad and as obscure as possible, right up to the limit the patent examiner will tolerate. Patent law does not require parties to conduct a search in order to avoid unintentional infringement. Moreover, because patent law in most countries provides heightened damages for “knowing” or “intentional” infringement, most large organizations would prefer to rely on patent-holders to inform them of any unintentional infringement, and then to negotiate over royalties or cross-licenses.

But the free software world cannot so easily ignore patent dangers. Upstream suppliers of software that can be freely copied, modified, and redistributed by downstream users are justifiably worried about their potential liability for passing along infringing software, as well as by their ethical obligation to offer people software that is as free as it looks. Large-volume enterprise users of free software—who are accustomed to bargaining for indemnity provisions in their software acquisition agreements with commercial software suppliers—would like to be able to insure against patent infringement risks. And everyone is aware of the monopoly, intensely concerned that it may try to rescue its increasingly untenable commercial position by using threats of patent infringement litigation to block the advance of free software.

So building a patent defense strategy requires some effort to understand the scope of the risks. Open Source Risk Management, a fledgling provider of insurance and indemnity for enterprise users of free software, recently announced the summary conclusions of a study conducted by my colleague Daniel Ravicher that identified more than two hundred US patents, including more than two dozen held by Microsoft, that might read on the Linux kernel.

This announcement was rapidly translated by the trade press into “Linux may infringe hundreds of patents.” But this fundamentally misunderstood what the value of the exercise, and the news, was. OSRM has not released the details of the study, but Ravicher’s own public statements make clear several very important points that were often missed in the coverage. First, not one of the patents the study identified was a patent that had been held valid in the US courts. At least one of the identified patents held by Microsoft is already under reexamination by the US Patent Office, as the result of a petition for reexamination filed by Ravicher himself through the Public Patent Foundation. Ravicher, who in conducting his study had only the assistance of one law student working over several months, was not attempting to determine whether Linux infringed any of the identified patents. That would be a complex and time-consuming inquiry for fifty patent lawyers. The goal of the study was to establish the scale of the risks against which an insurer, in the worst case, might be required to indemnify policyholders. It therefore identified patents one or more of whose claims, if valid, might be practiced by a Linux-like operating system kernel, solely by reference to the published claims themselves, without actually determining whether the existing kernel code actually practices those claims.

The OSRM study, which concerned only one free software program among thousands, does, however, establish an important baseline. Insurers like OSRM will undoubtedly refine their internal estimates of patent risk as to many free software programs in coming months, but already they can begin to determine the appropriate scale of premiums. As they announce affordable patent infringement coverage for business users, those users will be able to continue their ongoing large-scale adoption efforts with additional confidence.

The study is also a necessary part of our reconnaissance effort in building the free world’s patent defense strategy. As one stage in that process, approximate assessments of the possible patent interactions of a full range of free software programs are being conducted. Those efforts will identify some invalid patents dangerous to the freedom of free software, against which legal action can be taken by the Public Patent Foundation. They will help us to map the landscape faced by particular development teams in the near future, guiding free software in directions that avoid potential conflicts with patents held by hostile hands. Whether the bluster from the monopoly this summer has been an actual warning of impending hostilities, or just another round of expensive bluff, every day that passes finds us more equipped to defend the freedom of free software.

This column was first published in the UK in Linux User. It is also available in PostScript and PDF formats.

permalink | columns/lu | 2004.09.05-00:00.00

Comments are closed for this story.

Trackbacks are closed for this story.