Printer Friendly

The new cooperative patent classification system: improving patent searching.

A new era of cooperation between the United States and European patent offices began on Jan. 1, 2013. That's when the offices started using the new Cooperative Patent Classification (CPC) system. After a 2-year transition period, the CPC will replace the patent classification system now used by the U.S. Patent and Trademark Office (USPTO; By cooperating in the development and maintenance of the CPC system, the offices expect to speed up patent searching, eliminate redundant work, and shrink the backlog of patent applications awaiting examination.

Although the driving force for the new classification system is the enormous number of new patent applications filed every year and the inability of patent offices to keep up, the new classification system will also have ramifications for patent searchers outside the agencies themselves. Needs of internal patent examiner searchers, to some extent, dovetail with external patent information specialists. However, sometimes they diverge.

To put the volume of patent applications in perspective, consider this: Last year 503,582 new U.S. patent applications were filed with the USPTO. Its current backlog is 600,000-plus applications, and the average time to grant a U.S. patent is 4 years. The situation at the European Patent Office (EPO; is similar. Both struggle with increasing efficiency to keep up with new applications and to shrink their backlogs.


Patent searching and patent classification are two areas where the USPTO and EPO can gain efficiency by eliminating redundant work. Companies patenting an invention in the U.S. often patent the same invention in Europe, and vice versa. Although identical applications are submitted to the USPTO and the EPO, each office would independently classify, search, and examine the applications submitted to their respective offices. In addition, the EPO performs extra work by giving all U.S. patent documents a European Classification (ECLA) identification, whether they've been submitted to the EPO for examination or not.

By moving to a single classification system, redundant classifying will cease. Examiners in the U.S. can stop determining U.S. Patent Classification (USPC) codes for documents already classified in ECLA--such as applications initially filed in Europe--and the EPO can stop giving ECLA codes to U.S. patent documents.

Other efficiencies should ensue, such as eliminating duplicative searches. If both offices use the same classification system, then a search by one office should suffice for the other. Even if each office performs its own search, moving to a combined classification system will eliminate the need to perform two separate classification searches. Currently both ECLA and USPC must be searched to ensure complete results.

The CPC project is also the first step in the development of the Common Hybrid Classification (CHC), a single classification scheme for all the world's patent offices. The CHC is a project of the Five IP Offices (U.S., Europe, Japan, Korea, and China). Further information on the CHC is available on the Five IP Offices website (


Think of patent classification systems as taxonomies of inventions. The classification hierarchically categorizes inventions by area of technology and by specific features within each technology. Classification codes describe what an invention is and what it does. Inventions typically receive a primary classification describing the main inventive feature or use, along with one or more additional classifications describing secondary features of the invention.

Patent classification systems, and there are several of them, provide an alternative to text-based searching. While text searching is useful, its effectiveness is limited by the nature of language. A hierarchical index that classifies inventions according to features and functions avoids word-related problems. For example, since inventors act as their own lexicographers, there is no guarantee that your text-based search query will match the inventor's lexicon. In addition, words can have multiple meanings and alternate spellings, be misspelled, and appear in a different language. Facilitating multiple-language searching is a particular strength of patent classification systems.

There are, however, two problems with patent classification systems. First, they always lag behind the technology. If an invention is truly new, there isn't a classification code for it. Thus, classification systems must constantly be revised to handle new technology. Second, as the number of inventions grows, the classification system may be too coarse to distinguish between similar, but distinct, technologies. In other words, if there aren't enough subdivisions then each subdivision contains too many documents, and it becomes impractical to look at them all. To be effective, the system must scale.


The most widely used classification system today is the International Patent Classification (IPC), which was introduced in 1968 and is used by more than 100 patent offices throughout the world. Maintained by the World Intellectual Property Organization (WIPO), the system contains approximately 67,000 subdivisions and is updated annually. Codes are based on specifications, not claims. Because it is so widely used, you can perform a classification search in almost any patent database by using the IPC. The codes may be used alone, or in combination with text terms to limit the results to a manageable number of documents.

Though it's widely available, the IPC system isn't updated frequently enough, and doesn't have enough subdivisions, for the major patent offices. To address the shortcomings of the IPC, extended classification systems have been developed. In particular, the EPO developed the ECLA system, and the Japan Patent Office (JPO) developed the File Index (FI) system as well as a separate system, the F-term system. Both ECLA and FI use the IPC system as a foundation and then add additional subdivisions.

Developed internally by the EPO for the use of its examiners, the ECLA system has 135,000 divisions and was updated monthly. In July 2012, EPO froze the system as part of the conversion to the CPC system. ECLA codes are not published on the cover sheets of European patents and published applications. Instead, they are only available through electronic databases. By confining ECLA information to electronic information sources, the EPO avoids having obsolete codes appear on older published patent documents. A detailed description of the ECLA system is available on the EPO website, and ECLA searches may be performed with the EPO Espacenet searching system (

The Japanese FI system is similar to the ECLA system. Like ECLA, FI is based on the IPC system, has more subdivisions (187,000), and is updated more frequently than the IPC. In addition to FI, the Japanese have also developed the F-term system, which contains 340,000 divisions, making it the largest classification system. Further information about the FI and F-term systems is available on the JPO website (


The U.S. Patent Classification system is the oldest patent classification system currently in use. It is also one of the few remaining national classification systems. The USPC was developed in 1836 and first revised in 1898. The system is applied to utility patents and has 167,000 divisions. Until revisions were halted in November 2010, the USPC was revised continuously, with updates published every 2 months. The codes published on the cover sheets of U.S. patents and published applications reflect the classification at the time of publication. As time passes, the original codes are increasingly likely to become out-of-date.

The USPC is applied to both published applications and granted patents. Published applications are classified by the contents of the specification, and granted patents are classified by the contents of the claims. Because USPC codes are based on patent claims, and because every patent has unique claims, every U.S. patent is separately classified. The USPTO also generates IPC codes for U.S. patents and publishes them on document cover sheets along with USPC codes. However, since the IPC codes are generated by a concordance, instead of through application of the IPC classification rules, the codes aren't reliable tools for searching U.S. patents.

Although utility patents make up the largest proportion of patents issued by the USPTO, the office also issues two other kinds of patents--plant patents and design patents. They have their own classification systems. The U.S. plant and design patent classification systems will be unaffected by the disappearance of the USPC system for utility patents.


The CPC project officially got its start on Oct. 25, 2010, when the agreement to develop a joint patent classification system was signed by David Kappos, director of the USPTO, and Benoit Battistelli, president of the EPO. Seven days later, on Nov. 1, 2010, the United States Patent Classification system (USPC) was frozen, and revisions were halted. The official CPC website, the clearinghouse for all CPC-related information, was launched 1 year after the signing of the agreement, on Oct. 25, 2011. On July 1, 2012, the European classification system (ECLA) was frozen. The first USPTO Cooperative Patent Classification User's Day was held on July 10, 2012. On Oct. 1, 2012, the CPC became available, for user testing and practice only, on the EPO website. The second USPTO CPC User's Day was held Dec. 14, 2012.

The official launch took place on Jan. 1, 2013, when both the USPTO and EPO began using the CPC system. Initially the USPTO will apply the codes to published applications, while patents will continue to be given USPC codes. By the end of the 2-year transition period, in January 2015, CPC codes will be applied to all U.S. patents and published applications, and the use of USPC codes will cease. A static database containing USPC data will remain available after January 2015, but it will quickly become obsolete.

The CPC system is based upon the ECLA system. ECLA contains codes for U.S. patent documents from 1920 forward, European patent documents, and a large collection of nonpatent literature. To build the CPC, additional subdivisions are being added in areas where the USPC system is stronger. The EPO is creating the initial version of the CPC by mapping ECLA codes to CPC codes. Once the conversion is complete, the CPC will be ready to use.

The initial CPC system is expected to have 260,000 subdivisions, making it second in size to the Japanese F-term system. The codes are specification-focused rather than claim-focused, and documents will be classified as families (groups of patents sharing a common parent application) rather than individually. Each patent family member will receive the classification codes applied to the first published application (the first A1 document). Additional codes, if needed, may also be given to later family members. After the CPC is fully adopted, the U.S. patent classification team will stop classifying U.S. patents individually.

To keep pace with rapidly changing technologies, CPC codes will be revised continuously. A joint task force has been created, and revisions will be performed by the USPTO-EPO team. The CPC system will have detailed definitions for each of the class and subclass codes, which will make searching the classification index much easier for casual users.


Search efficiency won't improve immediately for patent searchers, although it may for patent examiners. Since the USPTO will still be applying USPC codes to patents during the 2-year transition period, the USPC system will still need to be searched. Many searchers already use the ECLA system; after Jan. 1, they began using the CPC system. If in-depth searching is needed, then both systems must be searched. That's exactly the way it's done now.

After the transition period, and especially after all U.S. documents have been reclassified in the CPC system, searchers will enjoy the benefit of searching a single classification system. Searching time should decrease, and answer quality should increase. The system will also provide an effective means for monitoring technologies for competitive intelligence.


At the EPO CPC User Day, held March 23, 2012, Fenny Versloot-Spoelstra described the EPO's plans for delivering CPC data. There will be little change for customers since existing systems will be used for data delivery. The CPC data will be published by replacing ECLA codes with CPC codes in the <patent-classification> field of the raw data XML files. Note that EPO customers are the producers of value-added databases sold through various search systems to patent searchers. The EPO expects to have the backfile, including U.S. patents, fully classified by the end of 2012. Likewise, during the first quarter of 2013, the USPTO is expected to deliver raw XML data containing the CPC codes.

What does this mean for the patent databases used by information professionals? The major companies are Questel, Thomson Reuters, and CAS. ProQuest Dialog also aggregates many patent databases.

"Questel," said Linda Williams, product manager at Questel, "will add CPC classifications in all our products where we get that content from the patent offices." Questel plans to make CPC data available during January 2013 in the PlusPat and FamPat databases. In addition, Questel plans to provide a CPC thesaurus.

"Thomson Reuters will make CPC data available during January 2013 in products and services that contain patent information," said Bob Stembridge, customer relations manager of scientific at Thomson Reuters. These products include Thomson Innovation, the Derwent World Patents Index, IP Data Feeds, WestLaw IP, IP Web Services, Patent Profiles, Custom IP Applications, Custom IP Solutions, and Thomson Data Analyzer. A CPC thesaurus is also planned.

"Features to support searching by CPC will be added to the entire CA/CAplus family of files," said Chris McCue of CAS. In addition, the USPAT and INPADOCDB files on STN will incorporate CPC data. Availability is expected during the first quarter of 2013. Like Questel and Thomson, CAS will produce a CPC thesaurus.

At the time this article was written, ProQuest Dialog did not have detailed information regarding its plans. To some extent, what happens there will be determined by the database producers who supply the databases to Dialog and, since Dialog currently has fields for IPC codes (IC=), ECLA codes (EC=), national patent classification codes (CL=), and Derwent class codes (DC=), it will probably create a similar field for CPC codes.


The CPC website ( is the source for information about the CPC project. The website contains the project history, current status, and latest news. There are slide presentations and videos of talks given at the EPO's CPC User Day. In addition, the slide presentations used at the first CPC User's Day at the USPTO are available.

Both the EPO and the USPTO are creating computer-based training modules that will be available on their respective websites. The first of the USPTO elearning modules is scheduled for delivery in the first quarter of 2013. The USPTO is also encouraging companies to contact the office ( to schedule individual briefings.

Patent Classification Internet Resources

Cooperative Patent Classification (CPC) website

CPC workshop videos, Vienna 2012

Espacenet--EPO Patent Searching system

European Classification System (ECLA)

Five IP Offices Common Hybrid Classification project

International Patent Classification System (IPC)

Japan Patent Office

United States Patent Classification (USPC)

United States Patent Classification Resources

David Gange ( is president of Altimatia, LLC and a professional patent researcher specializing in chemical and pharmaceutical patent research and education.

Comments? Email the editor-in-chief (

Anatomy of a CPC Code

Code Portion   Definition

A              Section
A23            Class
A23G           Subclass
A23G 9/00      Main Group
A23G 9/086     Subgroup

This table describes the component parts of a CPC
classification code. Patent documents are classified with
main group and/or subgroup symbols.

CPC Top-Level Section:

A   Human Necessities
B   Performing Operations: Transporting
C   Chemistry; Metallurgy
D   Textiles; Paper
E   Fixed Constructions
F   Mechanical Engineering; Lighting;
      Heating; Weapons; Blasting
G   Physics
H   Electricity
Y   General Tagging of New
      Technological Developments;
      General Tagging of Cross-Over
      Technologies Spanning Over
      Several Sections of the IPC

These are the top sections of the CPC hierarchy. Sections
A-H are derived from the IPC system and are also used in
the ECLA and FI patent classification systems.
COPYRIGHT 2013 Information Today, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2013 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Gange, David
Publication:Online Searcher
Geographic Code:1USA
Date:Jan 1, 2013
Previous Article:Virtual reference and in-depth assistance using shared workspaces.
Next Article:Finch and open access: debating the future of academic publishing.

Terms of use | Privacy policy | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters