|
!19990514 Here's a chance to critique Greg
Here's a chance to critique me. What follows is a rough draft of an article to be published in a IEEE publication this summer. Since it will be read by many, I want it to have some meaning beyond my usual rants and raves. I have to submit the final draft Monday, so if any of you have a bit of time today and tomorrow, I would appreciate any suggestions to make this more brilliant. I am constrained for the word counts. Thanks.
By the way, some of next week's PATNEWS will be all about PATNEWS and me :-) I figure after five years, I should use PATNEWS to promote my services and activities.
Greg Aharonian
====================
Does The Patent Office Respect The Software Community?
Gregory Aharonian
Source Translation & Optimization
srctran@world.std.com
How much blindingly brilliant innovation is there in the software world? I mean the type of development you might read about in a new IEEE/ACM article or conference paper that causes you to think to yourself, "That's cool." (I date myself). Probably not much, considering that software development is more evolutionary than revolutionary, with people constantly sharing ideas and techniques. The steady and collaborative development of Linux and the efforts of the Free Software Foundation typify such development in this Internet era. Is not the Internet an evolution of videotex (albeit with better interfaces and communications protocols)? Is not object oriented programming just a rediscovery of the Lisp world?
It is then with great concern, and some outrage, that programmers should view the tally of issued US software patents for the last few years.
1999 22500
1998 17500
1997 13000
1996 9000
1995 6100
1994 4500
1993 2400
1992 1600
1991 1500
1990 1300
That's right, the US Patent and Trademark Office (PTO) is issuing about 20,000 new software patents every year-a tenfold increase in a matter of six years. At this point, I would point you to an Internet FAQ for more information, but I can't because I might be inducing you to infringe a patent - U.S. Patent No. 5,842,221 entitled, "Dynamic Frequently Asked Questions (FAQ) System." Patents are supposed to be limited to inventions that are both novel (no one has ever made the invention the same way before) and unobvious (given the problem faced by the inventor, a skilled and knowledgeable programmer wouldn't have created the same invention). Considering these inexplicable numbers, organizations such as the IEEE and ACM should consider asking the PTO on what basis it thinks there are 20,000 novel and unobvious software inventions each year.
There are a few well known reasons why the PTO issues so many patents, including indifference to the prior art, the flood of patent applications, and assembly line working conditions for patent examiners.
Prior Art
One of the weaknesses of the current patent system is the absence of any searching requirements for patent applicants to prove their inventions are novel and unobvious. Prior to filing a patent application, the applicant is under no obligation to go to a library or access on-line databases to seek published prior efforts that relate to the patent application. The expectation is that the patent examiner will be doing so as part of his duty to issue the patent only if the invention is novel and unobvious. But the PTO and patent lawyers know that examiners do not have enough time, money, and resources to search thoroughly, especially for software.
In the last twenty years, more than 700,000 patents have issued for electronics inventions (hardware, software, circuits, optics, communications, etc.). Of these, approximately 80 percent have cited no IEEE/ACM/SPIE conference papers, journal articles, or books. In other words, the PTO has concluded that the inventions were novel and unobvious by examining prior issued patents but largely ignoring all other published literature. Tolerance of this evaluation system reflects contempt for both the technology and those who develop it (which the applicants are often just as guilty in ignoring their colleagues' work).
In part, patent examiners fail to cite non-patent prior art because they do not have access to an adequate library or database of non-patent prior art materials to work with. Regardless of the reasons for the shortcomings, the dearth of non-patent prior art cited calls the validity of many patents into question-particularly those that do not even cite prior patents (for example, recently-issued U.S. Patent No. 5,796,943 for a smart card cites no prior art at all). As late as May 1999, half of the issued software patents cited no non-patent prior art.
A 1988 COMPCON paper by a software patent lawyer outlined the same issue a decade ago. "The percentage of software written today that would satisfy the obviousness requirement is probably in the five to ten percent range. The question of obviousness is necessarily a case by case inquiry. Moreover, because the Patent Office has a limited collection of prior art, many of the software patents which are granted will ultimately be held to be invalid in litigation based on prior publications, foreign patents or commercial uses which were not available to the Patent Office." Apparently, little has changed in the last ten years.
Flood of patent applications
A second reason for the large number of software patents being issued is that the PTO is simply overwhelmed with patent applications. The PTO issues patents for about half of the applications it receives, so that a ten fold increase in the output of software patents means that the number of patent applications has also increased about ten fold (in the last six years), a rate much higher than the increase in PTO personnel. Compounding this problem is high turnover among patent examiners and a favorable high-tech job market that makes it difficult for the PTO to hire and retain qualified examiners - approximately 66 percent of the current software patent examiners only have Bachelor's degrees and a few years of industry experience.
From where is this flood coming? Consider the top ten software patentees for 1998 (approximations based on a sample of 3300 software patents from 1998):
1200 IBM
360 Motorola
330 Fujitsu
330 Canon
310 Microsoft
300 Lucent / BellCore
280 NEC
260 Sun
260 HP
250 Sony
This might be surprising since many of these companies are not thought of as "software" companies in the Silicon Valley sense. They are large, well established companies with a long history of developing and patenting hardware and software, and using their patent portfolios to extract hundreds of millions of dollars of royalties (and control the entrance of newcomers to the marketplace), even though many of their patents are of questionable validity. Large companies with ample resources to seek out quality patents not doing so. In recent times, this has led to an interesting example of game theory and cooperation. Many other companies have decided that if they can't beat them by reforming the software patent examination process then they should join them by filing their own growing sets of patent applications. Much like these large companies, they hide behind the impotence of the current patent laws to ignore the prior art.
Assembly line patent examination
These problems are compounded by the harried conditions under which patent examiners must analyze patent applications, as explained in a private email message sent to me by a patent examiner:
One of the reasons so many "crappy" software patents issues is that examiners are forced to do so. The PTO's production system [that] discourages CONTINUED searches is a really big problem, for 2 reasons: (1) upon 1st amendment, PTO management's strong encouragement to either: issue or make final (do NOT continue searching) (2) in effect, actually taking AWAY counts for finding new art with which to make a rejection (for continuing to search). Why? We not only get no points for non-final-rejections, but a non-final rejection also takes much longer to finally go to a count (abandonment or issue). So if they rewrite their claims to get around our first action, they usually win, especially if they throw in additional claims to drag things out. Thus, we might throw a couple of good punches, but essentially we always are forced to quit the fight in the first round. And applicants know this.
These ingredients - large numbers of patent applications that overwhelm patent examiners who are not supplied with adequate prior art to evaluate novelty and unobviousness-combine in a recipe for explosive growth in software patents. The result is patents of dubious validity. It seems silly that the PTO actually issued some patents, such as U.S. Patent No. 5,491,779, issued in February 1996 for three dimensional pie charts, or U.S. Patent No. 5,764,906, issued in June 1998 for mapping phrases (such as "White House") to URL addresses.
At least for software, it is reasonable to conclude that the quality of issued patents is suffering greatly. Therein lies the dilemma for organizations such as the IEEE. Many IEEE members belong to large companies for which patent quality is less important than the size of the patent portfolio, while other IEEE members belong to smaller companies or are independent inventors, for which patent quality is very important in order to be taken seriously in the marketplace. Any reform that the IEEE might want to support (I encourage IEEE members to get involved with the IEEE's Intellectual Property Committee, http://www.ieeeusa.org/COMMITTEES/IPC/ )
will run into these conflicting interests.
But as a matter of professional pride, the IEEE should demand that the patent system better respect and pay more attention to non-patent prior art, including the many IEEE conferences, journals, books, and Web sites. Over the years, the members of organizations such as ACM, SPIE, ASME, APS, and others have worked together to provide forums to discuss, share, and advance software technology. Issuing patents without referring to the substantial body of non-patent prior art reflects a lack of respect that should not be tolerated.
|
