Testimony of James F. Cottone, President

National Intellectual Property Researchers Association - NIPRA

---

Oversight Hearing on the U.S. PTO of the

Subcommittee on Courts and Intellectual Property of the

House Judiciary Committee

----

Thursday, March 9, 2000

Good Afternoon Mr. Chairman.

My name is James Cottone, of Leesburg, Virginia. I’m here today testifying on behalf of the National Intellectual Property Researchers Association, NIPRA, which has been newly formed to represent the professional and business interests of nongovernmental professional people and firms who specialize in intellectual property information retrieval and dissemination.

Mr. Chairman, our members are intimately linked to the information dissemination functions of the PTO and typically have many years of experience - 10 to 20 or more - working with the PTO in various capacities, and in using all branches of the PTO’s facilities. NIPRA’s members serve as the world’s main concentration point for intellectual property (IP) information from the PTO. Every single day of the year, they receive thousands of requests from clients worldwide, and, using their knowledge of and access to the PTO’s facilities, locate and retrieve the information needed to respond to these requests.

The clients of NIPRA’s members include the full range of IP users: independent inventors, small start up companies, major multinational corporations, and patent law firms here in the Washington area, as well as national and international patent law firms. Functioning as this concentration point, it is absolutely crucial that our members have rapid and unhindered access to patent and trademark information, and that the information retrieved be of the highest possible accuracy. These three factors are the life’s blood of our members, and therefore of virtually the entire U.S. intellectual property community.

Over the last several years our members have seen a serious deterioration in the information dissemination functions of the PTO, and have especially noted in the past year or two, a near complete meltdown of several of the PTO’s most important facilities. Due to a rash of recent problems at the PTO, it has become exceedingly difficult - in some cases impossible - to access the information needed for our member’s clients. For reasons that defy rational explanation, the PTO is failing badly to fulfill its statutory requirement of "IN GENERAL... the Office...shall be responsible for dissemination to the public information with respect to patents and trademarks." This is clearly spelled out in the recently enacted Patent and Trademark Office Efficiency Act. [S 1948 and 35 U.S.C. 2(a)(2)].

Page 2

My testimony today will cover three main areas of special concern to NIPRA: [A] the PTO’s search system(s); [B] a brief overview of relevant testimony at the Commissioner’s Hearing on Issues Related to Prior Art during Examination, of July 1999; and [C] proposed amendments to Title 35 U.S.C.

**A**

With respect to the shortcomings of the PTO’s search system(s), consider a brief perspective on the nature and extent of automated search systems by a retired patent examiner with 25 years of experience who is now a member of NIPRA. As stated by Mr. Joseph E. Clawson, Jr. of McLean, Virginia -

"...from June of 1972 until my retirement in December of 1997, I worked as a patent examiner. I have done perhaps 10,000 to 12,000 searches of the Prior Art in the semiconductor and computer memory areas. In the early 1980s I had an opportunity to start using electronic database searching, using predominately "key word" search strategies. My initial enthusiasm was soon blunted as I found that I could not locate, electronically, the references that I personally knew were there from my manual searching. Others who were expert in the various technologies also came to much the same conclusion. This finding was further buttressed by scientific evaluations such as the March 1985 paper by David C. Blair et al., in the Communications of the ACM, "An Evaluation of Retrieval Effectiveness for a Full-Text Document-Retrieval System."

While this published paper reports the results done using a generalized database of only 350,000 pages, when applied to a Patent database, another unique problem arises: in many, if not most, cases the invention is never fully described "in the words." The Patent Law requires only that the Specification, including the Drawings, together be understandable and enabling to one of ordinary skill in the art to make and use the invention. "The words," in many if not most cases, merely "flesh out" what is shown in the Drawings and do not replicate "in words" what is in the Drawings, but are ancillary thereto. Thus, in a Patent database electronic search one is often presented the additional problem of "searching" for "words" which were never there to begin with.

Thus "automated electronic searching," while useful cannot fully replace an actual properly classified manual paper search file, in my experience. You cannot, quite literally, "find the references" searching electronically. A more correct way to view an electronic database is to consider that "ALL the information is lost," until you can discover some manner of getting a portion of it back. As above noted, in the Blair et al. paper, there was a staggering electronic loss of many tens of percent.

Page 3

My own experience showed me that only an unacceptably small percentage of relevant Prior Art could be routinely retrieved electronically, and what was recovered was most often NOT the best, and NOT the most pertinent Prior Art. If one considers a "closed stack" paper search file where "nothing leaves," then one has absolute file integrity, and if properly Classified it then becomes easy for any person who can readily read English to find ALL the pertinent Prior Art in a particular subject area.

Thus in my opinion and experience, the use of electronic database searching is problematical, and particularly so when dealing with the unique nature of Patent specification disclosures. Paper search files, properly classified, are the only serious means by which Prior Art may be retrieved in an effective, efficient manner, and apparently the only search means fully in conformance with the requirements of 35 U.S.C. § 131. While electronic database searching gives many who use it a great illusion of "finding everything," this is true only to people unfamiliar with the particular technology. The reality is that this is indeed just an illusion, especially when dealing with the unique nature of Patents and Patent specifications.

Further, as was pointed out by Mr. James Cottone in his presentation at the July 14, 1999, PTO Public Hearing on Issues Related to the Identification of Prior Art During the Examination of a Patent Application, Figure 4 thereof shows that in a well maintained paper file Digest, one "...can reasonably expect to find all the best art in well under three hours of search." Costwise, it is estimated that every dollar spent by the PTO in establishing and maintaining these Digests can replace about $100 in attempting to provide equivalent results via an electronic database.

Thus, from both practical experience and scientific points of view, the use of large electronic database searching to providing a replacement for a properly classified paper search file has yet to be proven, and the present paper search files at the PTO - especially the Public search room files - need not only to be well maintained, but expanded to include all current publications, and foreign patents in order to rebuild the patent community’s eroding confidence in the U.S. Patent system."

The problems presented by automated searching are compounded significantly when the search system employed is not up to the task. The PTO's WEST Patent Search System (employing the BRS Database Search Engine) has repeatedly demonstrated itself to be an intrinsically flawed system, possibly fatally flawed. Evidencing both inconsistent and false results in numerous examples, NIPRA believes that a serious reconsideration of this system must be imposed on the PTO. Not only are public search room users fed up with WEST, but also

Page 4

examiners are complaining loudly about if not totally abandoning WEST, and are

using its cousin EAST unhappily by default. A large collection of instances of WEST system blatant failures (not merely those normally encountered when debugging a new system) are on file with the supervisor of the public search room. At the present time, it was not possible to gain access to that collection to share with the House Committee. However, a selected few well documented and easy to understand examples are at hand and will serve here as illustrative of the utter hopelessness of WEST, and to demonstrate why its confidence level is exceeding low. An additional thirteen severe problems may be found in an attachment to a November 19th, 1999 letter to Commissioner Dickinson submitted by a panel of professional public users of WEST. The few examples below in no way represents all of the kinds of problems associated with this particular search tool, and if space and time allow, hard copies proving these statements will be made part of this testimony. If not, please contact NIPRA by fax at (703) 415-0621 for complete documentation of these examples, or of other articles or published papers cited in this testimony.

CASE ONE: Patent Count for All Patents in Class 414 - (414/$). Results: 4754 Actual Count is 57128 (3/6/2000). The implications of this example are that a text search in the area of material handling where textual terms are juxtaposed to determine inventive features but limited to patents classified in Class 414 (material handling) would be limited to less than 10 per cent of the patents within this classification. This error would be occur to both public users AND examiners.

CASE TWO: Thesaurus Search for a Patent Assignee (Procter and Gamble). Results: 19 or 4551 or 4587 or 6845. A thesaurus search is used when a search may be directed to a plurality of names for the same entity. In this case, Procter and Gamble could be P&G, Procter and Gamble, Procter and Gamble Company, Procter and Gamble, Inc., etc. This is also known as an "EXPAND" function in other more reliable search tools. The implications here are that it is no longer possible to know accurately an entity’s complete patent portfolio.

CASE THREE: Search for a Patent bearing text of interest in combination with Patent Assignee's Country. " sheltered device mount" and Kuwait). Results: US 5950372 assigned to International Design Systems Company of Kuwait is NOT FOUND by Assignee Country but IS FOUND by the cited textin its abstract). The implications of this search is that either all the data on this patent has not been properly indexed in the USPT database or that the BRS search engine itself cannot keep track of data collected from polling the index when it must search for the confluence of two (or more) sub-searches.

These and other examples clearly support the contention of most WEST system users that this search tool is not adequate or accurate enough for the job it is intended to do. Both public users and examiners report a LOW-to-NO

Page 5

confidence level for this engine (WEST/BRS). We at NIPRA agree.

**B**

On July 14th, 1999, the PTO sponsored a valuable Public Hearing on Issues Related to the Identification of Prior Art During the Examination of a Patent Application in Crystal City. At that time, a paper authored by myself was presented providing the first publicly available quantitative analysis of patent searching comparing the efficacy of manual searching of prior art vs. automated searching using the then available PTO’s APS/Messenger system. As a result of several hundred search efforts, a statistical analysis was presented in a ground breaking series of probability of detection graphs, culminating in a Figure 4 which curve fitted manual vs. online searching results. Remarkably, it was shown that given the existence of suitably established and maintained paper file subclasses - or alternately a "motherlode" paper file Digest - a typical search that produced a

§ 102 reference (effectively blocking the issuance of a duplicative, invalid patent) could be completed in just under three hours of search effort with just as high a probability of detection as the best automated searching combined with manual searching of standard (regularly classified and maintained) paper files. Reference to Figure 4 of that paper shows a cumulative probability of detection of a § 102 reference, given that one exists somewhere in the PTO, to be 0.92. This was a startling find that readily supported a conclusion stated then as:

"Presuming that online searching will assume a significantly larger role in prior art searching as time goes on, it is critically important that the theoretical benefits of online searching - of which there are many - are actually realized. At the very minimum, much more training time must be provided to the examining corps to make sure that the online potential is obtained in a professional and usable manner for patent applications being examined right now. Equally critical is the necessity for the PTO automation developers to slow down their rate of change so as to get better control of their work product, and to take the uncertainty and confusion out their systems. These three factors - training, uncertainty and confusion operate to greatly reduce the potential of the online systems to contribute to better prior art searching."

Since that time, sadly, the PTO has retired the APS/Messenger system and substituted for it the BRS EAST/WEST systems that are considered to have about 40% of the capability of its predecessor search system.

**C**

I turn now to NIPRA’s proposed three amendments to Title 35 to address in part the several problems noted above.

Page 6

Consider first a proposed amendment to 35 U.S.C. § 1

Section 1(a) of title 35, United States Code, is amended by adding the following sentence:

" The United States Patent and Trademark Office shall continue as an agency where all records, books, drawings, specifications, and other papers and things pertaining to patents and to trademark registrations shall be kept and preserved."

Consider second a proposed amendment to 35 U.S.C. §4

Section 41(I)(1) of title 35, United States code, is amended by adding the following sentence:

"The Director shall not cease to keep and maintain, for use by the public, paper or microform collections of United States patents, foreign patent documents, and United States trademark registrations."

Finally, consider:

41(I)(4) The Director shall submit to the Congress an annual report on the automated search systems of the United States Patent and Trademark Office and the access by the public to such systems. The Director shall also publish such report in the Federal Register. Prior to the submission and publication of such report, [The] the [Commissioner] Director shall provide an opportunity for the submission of comments via public hearings by interested persons [on each such report] on the automated search systems and shall include these comments in their entirety in each such report.

Mr. Chairman, I want to thank you and the other officials who have provided us this much needed opportunity to discuss the many issues facing the PTO and its many constituencies. I especially want to thank those on the Subcommittee on Courts and Intellectual Property for the invitation to testify today, and for the first time to present the views and concerns of our new National Association, NIPRA. I pray that the views expresed herein find symphathetic ears and that the three proposed amendments to Title 35 are given swift consideration and approval for the benefit of the entire patent community.

That concludes my prepared remarks, and if my oral presentation today has

Page 7

departed from this written testimony, please consider this written version as the official position of NIPRA on all matters covered. I’ll stand by for a few questions if our chairman will so allow.

James F. Cottone

President

Washington. D.C.