Patent classification
Patent classification schemes are used to organise and index the technical content of patent specifications so that specifications on a specific topic or in a given area of technology can be identified easily and accurately. This page explains in detail how to use it and should ideally be used together with our how to search for information page.
Patent classification schemes are constructed and maintained by and for patent examiners and their primary purpose is to help the examiners in their work. When examining a patent application, the examiner needs to search a collection of patent documents to identify relevant existing patent specifications and this task is facilitated by the use of a tailor-made classification scheme.
As part of the examination process an examiner will assign patent classification codes to the specification he is examining, so in its turn that specification becomes part of the classified collection of specifications available to examiners in the future. Therefore the classified collection of patent documents is growing constantly.
The usefulness of patent classification as a means of searching for patents information is a by-product of its primary purpose as a tool for patent examiners. Using patent classification as part of a search to identify patents in a particular field can help the non-expert searcher to focus and refine his search and produce a useful set of references.
The International Patent Classification
The International Patent Classification (IPC) is currently used by over 70
patent authorities to classify and index the subject matter of published patent
specifications. The IPC is maintained and is administered by the World Intellectual
Property Organisation and was first published in 1968. The eighth edition was
published in mid 2005 and came into force in January 2006.
The IPC is available on online as well as in hard copy. Note that the current
eighth edition's hard copy version is only an abridged "Core" version.
These and other editions of the classifications are listed elsewhere
on our site. It is a good idea to read these pages first before consulting those
sites.
Structure
The IPC divides patentable technology into 8 key areas:
A: Human Necessities
B: Performing Operations, Transporting
C: Chemistry, Metallurgy
D: Textiles, Paper
E: Fixed Constructions
F: Mechanical Engineering, Lighting, Heating, Weapons
G: Physics
H: Electricity
Within these areas technology is divided and subdivided to a detailed level,
which allows the subject matter of a patent specification to be very thoroughly
classified.
Each of the areas A to H is published as a separate volume and together they
are referred to as the schedules. The schedules are accompanied by the following
volumes:
- The Guide - which contains much useful information and advice on how the scheme
should be used
- Concordance - which identifies areas in the current edition which have been
revised since the publication of the preceding edition.
- The Catchword Index - which is a basic key word index identifying the area
of the IPC schedules where specifications on a given topic are likely to be classified
A sample page from Schedule A of the English language 7th edition of the IPC
can be seen below.
The IPC codes, which act as an index to the subject matter of the patent, are
printed on the front page of a patent specification and
are always identified by the INID code 51. There is usually
a superscript numeral indicating which edition of the IPC has been used to classify
the document. e.g.
The 8th edition
of the International Patent Classification has meant many changes. The IPC has
always been hindered by new and revised classes every 5 years since the 1st edition
in 1968, so that you often have to check back through the editions for the equivalent
(or as close as you can get) class in earlier editions.
A single classification code can now be used going back not just to 1968 but
often earlier. Much has already been revised and more will be coming on Esp@cenet®.
There were other changes. There is Core, a simple version which will be revised
every 3 years, and Advanced, which will be revised every 3 months. The classes
for the contents of Esp@cenet® will be reclassified as necessary. For those
familiar with the IPC, Advanced looks like the old IPC. Core cannot be easily
distinguished from it: it lacks many of the subclasses such as 25/02, 25/04 (but
does retain some). Those who wish to search for appropriate Core classes can select
the Core level on the Web classification, which otherwise defaults to Advanced.
It is to be hoped that future classification changes introduced to the Web version
will be well publicised as otherwise there will be constant checking to see if
a class long-used by the researcher has been altered.
Besides Core and Advanced, there will be a distinction in each area for Invention
and Non Invention (the latter were formerly called Indexing classes). Non Invention
is for interesting but not novel aspects of the invention. An invention involving
thermosetting materials may be meant for tights, in which case an appropriate
class can be added (tubular garments).
All this affects Esp@cenet®. When reviewing results in the hitlist and
in the Bibliographic format the IPC data is given in groups. An example is WO2005118664,
where classes in bold italics are for Advanced Invention which are followed by
italics, Advanced Non-Invention, and then by Core Invention, normal type. A further
group in normal type would have meant Core Non-Invention. Semicolons are used
to mark off the groups.
For those who find it useful to distinguish between them, the following codes
can be used on Esp@cenet® in the format ci:G11B5/62.
a, advanced
ai, advanced invention
an, advanced non-invention
c, core
ci, core invention
cn, core non-invention
Using the International Patent Classification
Since the IPC is used by virtually all of the active patenting authorities
in the world (and particularly by all the authorities of the major industrialised
nations) as a common means of classifying the patent specifications they publish,
it is possible to carry out an international search for patents on a specific
subject using the IPC as a key. However, it is a massive and complex tool designed
for an expert user group and when it is used by anyone outside that user group
it should be applied with care. The ECLA variant has made usage more complex.
Finding the "right" IPC classification code
Finding a classification code (or codes) upon which to base a search requires
a basic grasp of how the IPC works. Taking time to browse through the Guide may
help, but talking to someone who is familiar with using the IPC as a search tool
will be of most value.
The Catchword Index may be a useful starting point, but it would be most unwise
to search on a code identified from it without referring to the full IPC schedules
to check for the context in which the code is placed and for relevant notes. The
Catchword Index often proves to be inconsistent in the terms and concepts it includes.
One practical and efficient strategy for getting to the right area of the IPC
schedules is to identify one or two relevant patents (say by using the Esp@cenet®
database) by using a few keywords and, see how these have been classified and
then to consult that part of the IPC schedules for detailed guidance on which
codes to use. If using Esp@cenet® then the more detailed version of the IPC,
the ECLA classification, is often available as a hypertext
link indicating the definition of the ECLA class, and makes it easier to identify
and then request that class.
A surprising range (or scatter) of classification codes at different levels
of detail and even from different areas of the IPC can be assigned to specifications
forming part of a single patent family and describing the same invention, or to
specifications which are closely related in terms of their technical content.
There are a number of factors which contribute to this IPC scatter:
- The classification policy of individual examining authorities may vary. Local
practice may place emphasis on different features of the invention.
- The interpretation and subsequent classification of specifications on the
same topic may vary from examiner to examiner.
- In the various language editions of the IPC terms may not have exactly the
same significance and this can lead to some scatter in assigning IPC marks.
- Patent equivalents, though all describing the same invention, will not necessarily
be expressed in the same way or stress the same points: a particular application
of an invention may be stressed in a European specification while another application
may be stressed in its Canadian equivalent.
- Some patent offices only apply the IPC at a very general level.
- The USPTO assigns IPC codes to its specifications via an automated concordance
which is often "off target".
This scatter may not appear to be significant, but may have considerable impact
on the effectiveness of a search based on the IPC. Taking an example from the
extract from the IPC schedule for agriculture, forestry, etc. shown below,
it is possible that a British specification describing a tong-like hand tool for
uprooting weeds might be assigned the IPC code A01B 1/18, while a US specification
on the same subject might be assigned the slightly less specific IPC code A01B
1/16 (hand tools for uprooting weeds). If a search for other patents on this topic
was limited just to those classified at A01B 1/18, which appears to be the "right"
IPC code, the US specification would not be identified. Thus the effects of the
scatter become clear.
In short, the scatter means that it is important not to fix upon a single,
very specific IPC code and to base all subsequent subject searching on that single
code alone. Even if it is possible to identify an IPC classification code which
appears to express perfectly the subject to be searched, it is essential to consult
the full IPC schedules, to look at the hierachical context in which the code exists,
to read the notes, to consider searching using a less specific code and to consider
alternative codes. Note that the heirarchy is indicated on the schedules by the
number of dots and that for example a 4 dot code must be read in conjunction with
the 3, 2 and 1 dot group titles immediately above.
A sample page from Schedule A of the English language
7th edition of the IPC
IPC search tools
The IPC can be searched widely in databases on the Web, on other online databases
and on CD-ROM. In most databases the IPC codes are not updated when a new edition
is introduced so to search back in time the older codes, where different, must
be searched as well as the latest. We have a detailed
list of editions of patent classifications.
Using electronic search tools (CD-ROM or online databases) it is possible to
simplify and speed up the process by truncating IPC codes. In the hand tool example
given above it would be possible, for instance, to search electronically for any
specification assigned either the IPC A01B 1/16 or
A01B 1/18 in one go. Further refinement can be achieved by searching for any specifications
assigned an IPC with the stem A01B 1 and combining the results of this search
with a key word search based on the word stems Tong and Weed. With this approach
careful use of the IPC can greatly enhance the scope and effectiveness of a subject
search.
Other classification schemes
ECLA
This is the classification scheme applied by the European Patent Office to
its internal collection of search documentation and is based on the IPC, but is
often more detailed. ECLA classification codes can be used to carry out subject
searches on the Esp@cenet® database.
This is done by either inserting an ECLA classification in the EC classification
field, if known, or by clicking on the highlighted ECLA field when a bibliographic
record of a patent specification known to be of interest is found. This provides
a "back door" way of exploring the classification with the ECLA code
for that record highlighted in yellow.
The advantages of using ECLA are that when the schedules are revised, which
happens quite frequently, the Esp@cenet®
database is revised so that only the latest codes need to be searched to cover
back in time. The codes are also applied consistently by one group of examiners
and are usually better than the IPC for the American patents. The data also goes
back much further than the IPC: to 1877 for Germany, 1909 for Britain, 1911 for
France and 1920 for the USA, for example. However the data is often only applied
several months after the publication of the specifications, so it is not suitable
for current awareness searching. There are also seem to be some gaps in its coverage
for the older non-German patents. The text of the classification schedules is
available on the web but appropriate
classes are best found on Esp@cenet® by clicking on the EC hypertext link
to see a definition of what sounds like a likely patent; then browsing to see
if nearby definitions are more useful; clicking the box next to the class; and
then clicking on "Copy" to insert that class into Esp@cenet®. Keywords
can of course be added as well to indicate a particular aspect of that class.
US Classification
This is the scheme used by the USPTO examiners as their primary classification
tool. The scheme can be used to subject search US patents as far back as 1790,
since all the affected documents are reclassified whenever the classification
schedules are revised. Uniquely, design "patents" are covered by the
classification in their own "D" classes, and are classified on the databases
along with patents. The US classification is only applied to US patent specifications
and cannot be used to conduct an international search. The scheme can be used
to carry out searches on a number of databases which are dedicated to US patents
and to search some CD-ROM products. A search based on US classification can also
be made on some free patent sites on the Internet. The text of the manual of classification
is published in loose-leaf format as is the alphabetical Index (both held in the
Library in Science 1 (North)) and is available on the web.
British Classification
This is the scheme applied in parallel to the IPC to all published British
patent applications by the examiners at the Patent Office. It has been used since
1962. The British classification schedules are revised and republished periodically.
Further information about the schedules is available on request from the Patent
Office:
The British patent abstracts published weekly are arranged in subject matter
groupings according to the British classification. File lists are available from
the Patent Office which give the patent numbers of all specifications classified
at a particular British classification code, or combination of codes. A charge
is made for the lists and again further information is available from the Patent
Office (tel. 0845 9 500 505). The British Library holds (at its Micawber Street
store) a massive file list in many volumes listing the British patents for each
class from 1911 to 1962.
Alternatively a search can be carried out in the Library reading room using
annual indexes to the British patent abridgments. The divisions involved need
to be identified first.
The British classification cannot be used to conduct an international patent
search, nor can it be used online. It can be used on the Espace UK compact discs
and also on the Europe-Access CD-ROM, which covers British patents back to 1980.
Further reading
The International Patent Classification as a search tool, W. Vijvers. World
Patent Information, 1990. 12 (1), 26-30.
The ECLA classification system, D. Dickens. World Patent Information, 1994.
16 (1), 28-32.
International Patent Classification, 7th ed., 1999. World Intellectual Property
Organisation. Volume 9, Guide, survey of classes and summary of main groups.
The sixth edition of the IPC, B Hansson & M Makarov. World Patent Information,
1995. 17 (1), 5-8.
Classification tools at the EPO, J. Rampelmann. World Patent Information, 1996.
18 (3), 149-153.
The function-application relation through a link between classification and
indexing, P. Faucrompre et al. World Patent Information, 1997. 19 (3), 167-174.
Online classification and indexation of documents at the European Patent Office,
J. van Thielen. World Patent Information, 1998. 20 (1), 17-20.
IPC revision: the future, J. Calvert. World Patent Information, 1999. 21 (2),
85-87.
Language related problems in the IPC and search systems using natural language,
M. Lyon. World Patent Information, 1999. 21 (2), 89-95.
Development of the IPC as a search tool, A Bruun. World Patent Information,
1999. 21 (2), 97-100.
IPC in the new millenium, S. Kunin. World Patent Information, 1999. 21 (2),
101-108. Addendum 21(3), 179-181.
IPC revision policy, S de Vries. World Patent Information, 1999. 21 (3), 165-168.
International Patent Classification in Derwent databases, B. Stembridge. World
Patent Information, 1999. 21 (3), 167-177.
