Communications of the ACM
Viewpoint
Bounty Hunting in the Patent Base
Software patent issues have become a central concern in the software developer's world. We developers have lost a wide-open vista of technique and must now skip carefully over intellectual property landmines, as even college students writing simple homework programs can violate patent law.
Patents are costly to overturn, and are thus a strong statement of ownership. Such protection makes sense for worthwhile patents, but not for frivolous patents that attemptlike private companies erecting tollgates on public highwaysto monopolize commonly used techniques. Currently, the U.S. executive and legislative branches of government emphasize minimizing bureaucratic obstacles and government oversight in the patent application process, so as not to hinder commerce. The courts tend to accept issued patents at face value, which makes it difficult for an organization or individual to defend itself against even the most flagrant frivolous patent.
Although a patent can theoretically be rescinded if the technique is proven to be either prior work, or not the novel invention of the patent holder, many techniques that are public domain practices are currently being patented. This trend violates both conditionsthese techniques are prior work and are not novel. But often it is unclear to the patent reviewers that a patent is based on prior work. Patent applications arrive from a huge spectrum of human endeavor, and the Patent Office cannot be expert in all things. Take, for example, the infamous 1-Click patent [1], which restricts e-commerce sites from providing a single-button-based transaction completion facility. This idea of clicking a button to automatically finish a transaction does not strike this author as novel; it must have had prior use. Moreover, the technology at the heart of this patent is the Netscape browser cookie, a mechanism designed to retrieve previously saved personal settings. This is an example of patenting a trivial extension to prior art.
One possible remedy already established is Richard Stallman's famous common patent pool [2]. Stallman's concept is simple enough: a common pool of voluntarily submitted patents are shared among all software developers. Just as with the great open source software initiative Stallman inspired, this pool would start small, growing more useful over time with increased contributions.
It is uncertain whether the courts will uphold this mechanism, as it has not been challenged yet. Also unclear is who would carry the cost burden of obtaining patents, and whether the pooled patents would actually protect a technique from a frivolous patent. Since patents are costly to obtain, why would someone try to obtain a patent to countermand a frivolous patent?
Clearly an additional mechanism is needed to help the Stallman common patent pool. Since it is unlikely that every software patent issue may be avoided by drawing on a patent in the common pool, software developers as a group must contribute to the due diligence of software patent applications themselves. They also must work through the existing patent base to expose and help overturn frivolous patents. I propose to do this using modern information technology.
The patent bureaucracy is based on old, paper-intensive technology, supporting the idea that jurisprudence demands time, effort, and bureaucracy to keep everything square and in accordance with the law. But the life cycle for solely paper-based systems is nearly complete. At times the economic window for various software technologies slides by before the patent application is resolved. Just as the Internet needs ever greater bandwidth, so too do public agencies need more infrastructure to keep up with the exponential growth of technology patents under applicationthe Patent Office will eventually grind to a halt if it does not get an infusion of technology and capital investment.
Just as the Internet needs ever greater bandwidth, so too do public agencies need more infrastructure to keep up with the exponential growth of technology patents under applicationthe Patent Office will eventually grind to a halt if it does not get an infusion of technology and capital investment.
I propose the Patent Office use a public reward system similar to the one used by the criminal courts to police criminal activity. A patent applicant would be responsible for performing a thorough search before applying. Free agents, who I call patent bounty hunters, would be rewarded for finding existing patents not unearthed by the applicant. They would also be rewarded for discovering established common usage of techniques under patent application. Communities of interest would naturally police their own patents. Who but IT professionals could evaluate the quality of IT patents, just as who but medical practitioners could best evaluate a new medical device? The reward money for bounty hunters, awarded by a proper court of jurisdiction, would be balanced by fines levied on applicants who failed to do their prior art homework. As our judicial system sets bail in accordance with a defendant's risk of flight, patent fines would reflect the measure to which they violate notions of novelty and originality.
In software, the online community of developers has demonstrated its enormous intellectual muscle in the open source movement. Open source incubates more great code faster than any venture-capitalized incubator could dream of. Why not use this same power to patrol the intellectual space of programming techniques? As open source is housed on various repositories around the world, most notably SourceForge, why not establish open databases of prior art evidence? The developer community could add dated references to various techniques or inventions, which could defend against patent-based intellectual property claim jumps. In addition, a historical section could be evolved to provide evidence for the prior art bounty hunters going through the existing patent base.
Indeed, with the ubiquitous access the Internet provides, why not put all patents and filings online? Someday a person hoping to apply for a patent will simply submit keywords and descriptive text to the U.S. Patent Office Web site, and allow an automated system to examine the tentative submission. Such a system could perform a variety of searches and matches for the patent applicant. This alludes to the existence of software that can intelligently work with text. Such ontologically capable search engines are evolving quickly, and what better place to test them than on the database inventors will need to economically defend their inventions?
In addition to the idea of a centralized prior art database, the vast store of knowledge on the Internet needs to be searched before a software patent is issued. Several initiatives in development address the idea of tracking usage across the Internet. The code is out there. The source code of the famous Unix editor, Emacs, for example, has been publicly available for some time. It is unfortunate that source code for programs such as Emacs cannot be scanned intelligently and automatically every time someone applies for a general-purpose computing patent. It is common knowledge that Emacs contains everything but the kitchen sink in the way of software technique and capability!
Software Agents as Bounty Hunters
Mobile agent-based software can do much of the work of human patent bounty hunters. The idea is that a society of agents, acting at will over the patent base, can seek out evidence of prior work. Much of the significant amount of research going on in this area has been on applications of agents to financial risk systems, but why not apply this technology to something other than the ever-popular investment and e-commerce systems? The advantages of agents are well known: they reduce the need to create and organize a complex hierarchy, are free to jump from place to place or host to host, can react automatically to the changing dynamics of computer networks, and can make decisions without user intervention (alleviating the tedium of current patent search practices). Agent-based technology would also benefit those wishing to search patents to ensure they can use a software technique without violation.
Over time, human bounty hunters and the software agent communities would begin to specialize, say in compiler technology, or database technology, or wireless protocols. This band of highly skilled researchers would protect technological playing fields for the common use. In the end, this technology would actually encourage patent filing by making the patent search process robust and quicker, thus reducing costs. This, in turn, would allow the individual and the small organization to compete against the large companies that use frivolous patents as armor.
References
1. Amazon settles 1-Click patent dispute. Computer Wire (Apr. 2002).
2. Kelly, J. S. An interview with Richard Stallman. LinuxWorld (Mar. 2000).
©2003 ACM 0002-0782/03/0300 $5.00
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2003 ACM, Inc.
No entries found