Report:Thomson Innovation/Data Coverage/Patent Coverage/Bibliographic Coverage

Bibliographic Coverage

In addition to its full text resources, Innovation also offers a number of important bibliographic-and-abstract only collections that are essential to any worldwide search.

The Derwent World Patents Index (DWPI) and INPADOC worldwide family data collections are the backbone of the patent search industry, and DWPI is being offered here on a flat-fee basis with subscription, which is rare for the industry: generally it is only offered on a pay-per-use basis.

View the full Patent Collection Coverage Chart for Thomson Innovation here to see the most up-to-date bibliographic coverage information.

Salient Features of the Derwent World Patent Index (DWPI)

The Derwent World Patents Index, or DWPI, is a worldwide bibliographic family database produced by Thomson Reuters, the same company that owns Thomson Innovation. Thomson Reuters maintains a staff of 300 scientifically-trained editors who read over new patent applications and re-write their titles and abstracts, standardize their inventor and assignee names, apply other special indexing (such as chemical structure indexing and Derwent Manual Codes), and create patent families of technically equivalent documents, all with the end goal of producing a standardized, easily searchable database that promotes efficient search and retrieval of patent documents. The database contains data from over 44 patent-issuing authorities, representing over 43.5 million patent documents and 20 million unique inventions from all technical areas. It is widely used by patent search professionals.^[1]

The features of the Derwent World Patent Index are not covered in-depth here. To understand how these features enhance the user’s experience with Thomson Innovation, below is a summary of the DWPI features that can be searched exclusively, or accessed while viewing any patent document from Innovation’s multiple data collections. For further information, please see the DWPI article on Intellogist.

Value-added representative family titles and abstracts
To create a family record in the DWPI, human indexers translate new titles into English, and re-write vague, short and uninformative titles often found on patent documents. The new titles concisely describe the inventive content of the patent family. In addition, each DWPI family contains a value-added, re-written representative abstract in English, highlighting points of novelty, intended use, and advantage of the disclosed invention. These abstracts summarize the key content of the specification and claims in 250-500 words using common industry terms, so that the abstract can be easily skimmed to give the searcher an accurate idea of the patent family content in a short time. The DWPI file even includes “title terms,” a list of words including many synonyms to the actual terms used in the re-written title, to increase keyword retrieval.

A re-written DWPI abstract vs. an original abstract.

Standardized inventor and assignee names
The DWPI indexing team uses a patent assignee coding system to broaden searching. This means that multiple related assignee names are often assigned to a single patent record, to increase retrieval potential.

In addition, DWPI’s editorial policy on inventor names has long been to include only the surname and the first initial, with a separate term added including the surname and both the first and middle initials. For example, contrast the formal names of these inventors from US 7,313,752 B2 with their Derwent indexed names. A search on an inventor name where only the initials, “Park S E,” were known would pull up this related hit if the Derwent fields were searched. Using the Derwent fields can also be helpful to avoid missing references that may have incorrect first-name spellings or other name variations, although it may also broaden the result set to include false drops.

Derwent standardized inventor data vs. original data.

Chemical and Polymer Indexing
The Derwent World Patent Index originated as a pharmaceutical patent alerting file, and as such, the staff’s efforts as early as the 1960s were directed towards indexing chemical structures. Before complex computer search engines were available, indexing was performed using alphanumeric codes to represent chemical fragments, and these codes are still in use in the database today. The Thomson help file lists the following available DWPI Chemical Indexing systems and DWPI Polymer Indexing systems which are displayed in Record View for DWPI Entitled users, but are not currently searchable:^[2]

DWPI Chemical Indexing systems (for records classified in CPI Sections B (Pharmaceuticals), C (Agrochemicals) and/or E (General Chemicals))-

• DWPI Chemistry Resource Numbers (full coverage from 1999 to date)
• DWPI Registry Numbers (indexed from 1981 to date)
• Chemical Fragment Codes (indexed from 1963 to date)
• DWPI Compound Numbers (indexed from 1987 to date)
• Markush Compound Numbers (indexed from 1987 to date)
• Ring Index Numbers (indexed from 1972 to date)

DWPI Polymer Indexing systems (for records classified in CPI Section A):

• Plasdoc Punch Codes (indexed from 1966-1994)
• Plasdoc Key Serial Codes (indexed from 1978-1994)
• Enhanced Polymer Indexing (indexed from 1993 to date)

The Thomson Innovation help file states that search functionality for chemical and polymer indexing is planned as an enhancement for later releases of Thomson Innovation.^[2]

Editor's Note:

The DWPI file is a very well known and respected resource in the patent search community. The ability to search these records along with full text patent records in an online tool is a valuable one. Searching the DWPI text fields can increase the likelihood of retrieval for many vaguely-worded patent documents, and displaying these fields can help users understand the inventive concept of a patent without having to laboriously analyze the text. In addition, the standardized inventor and assignee names can increase accuracy over searching the original data. It is also unusual to come across this file open to subscription on a flat-fee basis to anyone, since it is normally offered only on a pay-per use basis.

The DWPI Chemical and Polymer Indexing are not currently searchable Thomson Innovation. The previous implementation of these codes in another end-user search engine, Thomson Reuters' Delphion (which no longer carries DWPI data), was deemed unsuccessful by some professional searchers, and this may be one reason that chemical and polymer indexing aren't searchable in Thomson.^[3]

Finally, although this article does not focus on patent informatics concerns, Thomson Innovation help materials state (and the editors tend to agree) that using DWPI text fields instead of original patent text can produce more useful results from the Text Clustering and ThemeScape tools.