Report:Esp@cenet/Data Coverage/Patent Coverage/Bibliographic Coverage/Introduction

Bibliographic Coverage

The Worldwide Database

The worldwide bibliographic database collected and produced by the EPO, and made available on esp@cenet, is known as the Level II server to distinguish it from various national collections made available through the esp@cenet portal (see national collections, below, for more information). The bibliographic data coverage in this collection is sourced from the worldwide bibliographic and family data file that is produced by the EPO as the now-combined INPADOC/DOCDB file, which together contain records from over 80 patent authorities. (INPADOC and DOCDB used to be separate entities, but have recently merged, extending document coverage for some collections back to as early as the mid-1800s.) Esp@cenet contains bibliographic records, English abstracts, document facsimile images and sometimes even full text records for some collections. Unfortunately, it relies on the individual patent offices to produce and deliver this data; the result is that coverage, quality and timeliness varies widely between the different countries of publication covered by the database.

Keyword searching in the abstract field in esp@cenet is restricted to English-language searching. Therefore, wherever possible, English equivalent abstracts are taken from family members or produced by translators and included in the database. Abstracts are generally loaded into the database up to one month after publication. Some English abstracts are produced by the national patent offices themselves (such as the Patent Abstracts of Japan service), and are loaded into the database several months until after initial document publication. And some English abstracts are produced by human translators especially for the EPO; a general timeframe for how long after publication these are loaded is not known (with the exception of Japan: Japanese abstracts are usually available four months after publication).

The order of preference for which an English abstract is taken from an equivalent document and substituted for the original document abstract is as follows: EP English abstracts are given preference, followed by abstracts from US family members, and finally by abstracts from GB family members, if the first two do not exist.

Sources of Content Information for the Worldwide Database

A complete summary of all the data content in the esp@cenet database is difficult to produce, as there are three different (and often somewhat conflicting) sources of coverage information that could apply: one for esp@cenet itself, which does not contain much in-depth information, and two that deal directly with current INPADOC/DOCDB content, which contain more detailed information. However, these in-depth sources can be misleading, as INPADOC/DOCDB data coverage is actually more extensive than the esp@cenet's: a correspondence via e-mail with an esp@cenet help desk representative confirmed that esp@cenet does not include all data available in INPADOC/DOCDB.^[1].

The first source (and most reliable source for esp@cenet) is the coverage page available in the esp@cenet help guide. This table is reproduced in the Bibliographic Coverage Table in the Summary page of this article. It contains coverage date ranges, highest and lowest document numbers, and also a document count for each collection. However, this information is not broken down by document kind code for a number of countries in the table. The esp@cenet website does not have detailed coverage information regarding English abstracts, full text, or document images.

If more in-depth information is needed, users can consult the Global Patent Data Coverage (GPDC) report. An 80-page document produced by the European Patent Office (EPO), the GPDC provides more in-depth information about the national and regional patent office data delivered to the EPO, such as which collections contain titles, abstracts, facsimile images, etc. In this report, the date ranges for coverage are represented in blocks of five-year timespans based on the volume of documents received, and exact start and end dates for the collection are not provided. The information described in the report is not necessarily available via the esp@cenet portal: it more accurately describes the content of DOCDB, the BNS patent image database, and the EPODOC database. (BNS is a the EPO's database of patent facsimile images, and EPODOC is the EPO’s examiner database).

Finally, more information can often be found in the INPADOC/DOCDB PFS coverage spreadsheet. This spreadsheet of current coverage in the file is produced weekly, coinciding with the update schedule of the file itself. This document contains data coverage information that goes beyond what is available on the esp@cenet help guide table, such as breaking each country down more specifically by kind code (for example, making a distinction between "old law" and "new law" publications in cases where the country's patent law has changed), and providing information about any known gaps in coverage. Because it pertains to INPADOC, the coverage described in this file may be more extensive that that available in esp@cenet. Instructions for obtaining this spreadsheet can be found here.

As mentioned above, only the esp@cenet help guide data is presented in this article. However, the other two sources discussed here are freely available on the internet. A review of all three data sources discussed above was used to produce the INPADOC and esp@cenet coverage information in the Quick Table Comparison section of Intellogist.

Editor's Note:

The GPDC and the INPADOC/DOCDB coverage spreadsheet described above can add useful additional information to an understanding of esp@cenet coverage, especially on questions about whether certain bibliographic data (such as titles, abstracts, or images) could be available via esp@cenet. However, any information from either of these sources should be verified by testing esp@cenet, because these sources do not directly reflect the coverage of esp@cenet but rather its parent file, INPADOC/DOCDB.