Free-of-Charge Patent Information Sources:
Content and basic functionality
Stephen Adams Magister Ltd.
www.magister.eu
Magister ® is a registered trade mark in the United Kingdom
© Magister Ltd. 2003-2008, 2014 2
Topics
• Introductory remarks about the patent system
• What use is patent information?
• Information need – understanding a basic search classification
• “Information need” -v- “information source” – choosing a database
• Multiple authority collections
• Single authority collections
© Magister Ltd. 2003-2008, 2014 3
Topics
• Introductory remarks about the patent system
– how and why are patents published?
• What use is patent information?
• Information need – understanding a basic search classification
• “Information need” -v- “information source” – choosing a database
• Multiple authority collections
• Single authority collections
© Magister Ltd. 2003-2008, 2014 4
Basic patenting principles
• Patent offices publish information (gazette notices, complete specifications, amendments, current status etc.) as part of the patenting process:
The patent ‘bargain’
• What the patent holder gets:
– Acknowledgement as inventor
– A limited-life monopoly, in the form of…
– The right to exclude others from the
identical line of business
• What the general public gets:
– Information about technological
developments (who, what, when)
– Open access to the technology, once the patent has expired
© Magister Ltd. 2003-2008, 2014 5
“The internet changes everything…”
• The public web has become a medium of choice by which patent offices disseminate much of the statutory information which they publish…
– BUT only for those patent offices equipped to make their data available in electronic form
– AND there are many other sources available.
• Searchers in industry still mainly use subscription services for commercially important searches.
© Magister Ltd. 2003-2008, 2014 6
“The internet changes
almost/ everything…”
© Magister Ltd. 2003-2008, 2014 7
The official Nigerian patent register;
(original photograph courtesy of Spoor & Fisher, ZA)
Chemical fragment coding sheet;
(for use with Derwent World Patent Index database)
© Magister Ltd. 2003-2008, 2014 8
PATENT OFFICES generated ‘first-level’
raw data, e.g. paper specifications, official register copies, paper bulletins on subscription
DATABASE PRODUCERS acted as ‘information factories’, adding value to the raw material and generating electronic products e.g. magnetic tape
HOSTS were
‘information retailers’, who licensed the data and loaded it onto online systems, each with their own command
language USERS in industry
often needed multiple host accounts in order to access all the
databases that they needed
The old information supply chain;
almost all “for-fee”
© Magister Ltd. 2003-2008, 2014 9
PATENT OFFICES
DATABASE PRODUCERS
HOSTS USERS
The current information supply chain; a mixture of free and fee
‘First- level’, raw data, official registers
Additional editorial
‘value-add’
services
Multiple databases, common retrieval engines
€€€
€€
€ / 0
© Magister Ltd. 2003-2008, 2014 10
What sort of information is it?
• It’s big:
– approx. 1,127,000 patents granted by around 102 authorities worldwide in 2012
• up 13% from 2011
• equivalent to over 4.06 km. stack of paper…
– genetic sequence databases double in size within a year:
• single documents > 2,000,000 A4 pages
• It’s varied in format
– document driven: patents published covering either individual countries or regions (EP, EA, GC…)
– event driven: bulletin entries, register entries, complex legal status
• It’s in multiple languages unfamiliar to most searchers
– particularly from China, Japan, South Korea
– in non-Western languages including Arabic, Cyrillic and other scripts – China is now receiving and publishing the second-largest number of
application per year, behind the USA, and may hit more than 2m applications per year by 2015.
© Magister Ltd. 2003-2008, 2014 11
First-level data
• Most “free-of-charge” sources are either
– national patent-granting offices (e.g. USPTO), or – international patent-publishing bodies (e.g. WIPO).
• Some commercial organisations (e.g. Google) also obtain the data and re-publish for free on the public web.
• In both cases, the data supplied to the user is
– ‘raw data’, with little or no ‘value added’ content – generally limited to single country collections (no
family information)
– often limited to first publication stage only
© Magister Ltd. 2003-2008, 2014 12
PUBLIC SECRET
Time (in months) t = 0
First filing of invention
t = 12
Foreign filing
t = 18
First publication - laid “OPI”
t = 24
Decision to proceed Substantive examination
t = 36 +
Second publication - may be simultaneous with grant of patent
Outline of a typical patenting process
© Magister Ltd. 2003-2008, 2014 13
PUBLIC SECRET
Time (in months) t = 0
First filing of invention
t = 12
Foreign filing
t = 18
First publication - laid “OPI”
Many web sources ONLY cover first publications
This can have important implications for using the resulting patent information
Outline of a typical patenting process
© Magister Ltd. 2003-2008, 2014 14
Topics
• Introductory remarks about the patent system
• What use is patent information?
– for industry, for academia, for lawyers
• Information need – understanding a basic search classification
• “Information need” -v- “information source” – choosing a database
• Multiple authority collections
• Single authority collections
The use of patent information
• Positive:
– It is a large, well-controlled body of literature covering all fields of technology
– It contains large amounts of experimental data or practical description which may never be duplicated in the journal literature
– It is accessible at much lower cost than most journals or conference proceedings
• Negative:
– It is unfamiliar to most academic researchers, and takes time and effort to understand properly
© Magister Ltd. 2003-2008, 2014 15
© Magister Ltd. 2003-2008, 2014 16
Obstacles to ‘easy’ use of patent information
• Two basic challenges:
1. Patent documents have a dual function
• as legal documents to define the scope of a monopoly and
• as technical documents to disclose the nature of an innovation
2. Patentable innovations vary enormously in scope
• “one small step for man” or
• “one giant leap for mankind”?
© Magister Ltd. 2003-2008, 2014 17
1. Dual function documents
• The language of a patent document is a compromise between
– legal definitions and terminology (needed if the end result is to be enforcable), and
– the “normal language” of scientific and technical disclosures.
• This does not encourage new users, even if they are skilled scientists or
technologists.
© Magister Ltd. 2003-2008, 2014 18
Legal terminology:
calling a spade a …?
• Cooperative Patent Classification
– A01B 1/02 (Spades and shovels):
– Sample titles of patents retrieved:
• “Digging implement”
• “Garden tool”
• “Digging tool”
• “Hand-operable excavation apparatus”
• “Material handling device”
• “Portable fulcrum”
• “Trencher”
• “Implement with a foot pedal”
© Magister Ltd. 2003-2008, 2014 19
• Most technical innovations are not “giant leaps”
but incremental developments of existing products or production methods.
– small technical developments pose great information retrieval challenges
• process chemistry ; a new solvent for a known reaction
• metallurgy ; an improved alloy with a slight change in composition
• electronics ; a smaller radio receiver due to better chip design
• mechanical ; a better mousetrap!
– words are often inadequate to describe the fine detail.
2. The nature of innovation
© Magister Ltd. 2003-2008, 2014 20
Search claim 1...
etc.
© Magister Ltd. 2003-2008, 2014 21
The implications for subject-based information retrieval
1. Document function and language
– search procedures which work for other literature need to be adapted to work with patents.
• may need to un-learn some of the techniques and tricks from other work
• word-based searches alone are rarely adequate
2. Innovation as incremental developments
– large volumes of closely related documents demand highly-discriminating retrieval tools
• a simple “Google-style” search simply doesn’t achieve the desired results in an effective manner
Information needs ? Answers from patents !
• Despite their problems, patent documents can be used to help in a great many technical problems.
• Some examples:
– workers at an SME want to find out how their competitors have tackled a similar problem
– academic researchers need to go beyond theory, towards working technical solutions
– economists want to assess national R&D spending and innovation ranking
– personnel managers want to hire new skilled workers with experience in a certain field
– licensing agents want to understand a fair price for purchase / royalty payments
– merger specialists need to consider the asset value of a company for acquisition
© Magister Ltd. 2003-2008, 2014 22
© Magister Ltd. 2003-2008, 2014 23
How the web has improved the access to patents…
• Enormous capacity for shared resources
– widespread access, large collections
• Convenient electronic publishing
– small organisations can become providers – users can integrate results within company
• Very fast access to documents
– typically within 1-2 days of publication
• Low or zero cost
© Magister Ltd. 2003-2008, 2014 24
…and where things have stayed the same, or got worse.
• Guaranteed fast connections
– or even getting connected at all!
• Consistent data quality
– the “same” files on different servers may be loaded in quite different ways
• (Non)-Standard searching methods
• Software or other technical support
© Magister Ltd. 2003-2008, 2014 25
Some concerns about patent searching on the (free) web
• Less information about actual content of database or website
• Relatively limited search functionality
• Content may not remain free
– particularly if hosted by non-governmental organisations
• How safe is my search ?
– encryption and confidentiality issues
© Magister Ltd. 2003-2008, 2014 26
Topics
• Introductory remarks about the patent system
• What use is patent information?
• Information need – understanding a basic search classification
• “Information need” -v- “information source” – choosing a database
• Multiple authority collections
• Single authority collections
Patentability;
Validity Commercial
intelligence;
marketing scope Trend analyses,
‘watch’ services
© Magister Ltd. 2003-2008, 2014 27
Different types of information need
Who’s this new kid on the block?
They never seek patents outside the US!
Wouldn’t any engineer do the same thing?
It’s simple, it works – but is it new?
Why is their MD visiting China?
Is there much demand for this outside Europe?
Who else is working on this?
Is there a
different solution to this problem?
What are our company assets?
Is there scope to work
together?
Should we try to buy, or license?
Surely, that’s not exclusive! - it’s common sense.
I’m being sued! What should I do?
Freedom To Operate
Portfolio review;
M&A, valuation
State-of-the-art review;
‘landscaping’
© Magister Ltd. 2003-2008, 2014 28
Patentability;
Validity
Freedom To Operate
Portfolio review;
M&A, valuation
State-of-the-art review;
‘landscaping’
Commercial intelligence;
marketing scope Trend analyses,
‘watch’ services
Different types of patent-related
question
© Magister Ltd. 2003-2008, 2014 29
Subject-matter
Patentability;
Validity; general
“learning”
Legal status
FTO
Portfolio review;
M&A, valuation
Assignee data
State-of-the-art
review; ‘landscaping’
Jurisdiction
Commercial intelligence;
marketing scope
Time Trend analyses,
‘watch’ services
Different types of searches using
patents to answer these questions
Commercial -v- free-of-charge
• Information may be free-of-charge to obtain, but gaining knowledge always costs
30
Retrieve Assess Understand Respond
$$
© Magister Ltd. 2003-2008, 2014
“free”
“fee”
© Magister Ltd 2002-2008 31
Key uses for free web services
• They are not comprehensive enough (data
sources) or flexible enough (search engines) to provide a good quality patentability search.
• Main uses by industry searchers are for:
– search strategy development
• low cost browsing, used to identify suitable search terms
– words/classifications/inventor names can be extracted to re- use in commercial search files
– preliminary (first-cut) searches
• some full-text (US, DE, some WO), mostly applicant’s abstracts
– legal status checks
• national patent registers increasingly available; not often suitable for bulk retrieval
© Magister Ltd. 2003-2008, 2014 32
Search types and choosing an information source
• A simple search classification
– Novelty (patentability)
– Infringement (“freedom of action”, “freedom to operate”)
– Validity & opposition – State-of-the-art
– Alerting (current awareness) – Family & equivalent
– Citation
© Magister Ltd. 2003-2008, 2014 33
1. Novelty
• Subject matter search focused upon the specific details of an invention.
• Designed to establish whether an
application for patent grant is likely to succeed.
• Ideally, results should be at hand for attorney as he/she is drafting the
application.
© Magister Ltd. 2003-2008, 2014 34
1. Novelty
• Requires :
– multi-country source (essential)
• not many web sources provide this
– multi-technology source (preferable)
– detailed subject indexing and search features
• web is normally raw data only
– patent and non-patent literature
– rapid updating (up-to-date content)
• web is very competitive here
© Magister Ltd. 2003-2008, 2014 35
2. Infringement
• Limited to specific geographic area(s) and to claims of valid, in-force patents only.
• Designed to establish the “freedom to
operate” i.e. to avoid the possibility of
infringing someone else’s patent.
© Magister Ltd. 2003-2008, 2014 36
2. Infringement
• Requires :
– single-country/regional sources – multi-technology sources
– detailed subject search facilities
– ready access to exact text of claims
• sometimes possible with the web, but can be laborious process
– links to legal status of retrieved records
• often quick and efficient on the web for small numbers of records
© Magister Ltd. 2003-2008, 2014 37
3. Validity & opposition
• Similar to novelty, but limited to matter published before the priority date of a known patent.
• Designed to gather material for legal
proceedings, questioning the validity of an opponent’s patent
– not new, no inventive step.
© Magister Ltd. 2003-2008, 2014 38
3. Validity & opposition
• Requires :
– multi-country, multi-technology sources – extensive subject indexing (full-text is
helpful)
• a few full-text sources on the web, but most are display only, not for search
– patent and non-patent literature
– search features allowing date limitation (useful)
• rarely available on the web
© Magister Ltd. 2003-2008, 2014 39
4. State-of-the-art
• Review of a defined subject area.
• Typically used to map out the major
technologies and/or companies in a field,
e.g. before starting a competing research
programme.
© Magister Ltd. 2003-2008, 2014 40
4. State-of-the-art
• Requires :
– broad-based subject indexing (not necessarily as deep as for novelty)
– multi-country coverage (optional) – good access to complete texts
• the preamble of patents can serve as mini-reviews
– statistical analysis / visualisation capability (optional)
• sometimes available on web resources
© Magister Ltd. 2003-2008, 2014 41
5. Alerting searches
• Ongoing awareness of new patents in specific industries or by specific
companies.
• Designed as an early-warning service to guide research programmes and avoid duplication of research
– also good for “me-too” work
© Magister Ltd. 2003-2008, 2014 42
5. Alerting searches
• Requires
– reliable high speed updating
• one of the great strengths of the web
– adequate broad-based subject capability (typically simple subject classifications) – standardised record formats for quick
browsing
• rarely available ; format almost always varies between countries/websites
© Magister Ltd. 2003-2008, 2014 43
6. Family & equivalent
• Identifies either
– the geographic extent of patent protection (family) and/or
– texts in preferred languages
• Questions of legal status under regional
patent systems may arise.
© Magister Ltd. 2003-2008, 2014 44
6. Family & equivalent
• Requires :
– accurate and timely collation of family data
• most web sources are single country
– intellectual added-value
• for legal variations such as US-Re or non- Convention cases
– variable quality links
– extensive time coverage
• publication of family members may spread over many years
© Magister Ltd. 2003-2008, 2014 45
7. Citations
• Analysing document clusters based on official patent search reports
• In support of other types of search
(particularly validity) or as an exercise in statistical evaluation of competitors
• Caution! requires :
– complete data sets (not yet routinely available)
– highly standardised data fields (esp. assignee)
© Magister Ltd. 2003-2008, 2014 46
Single-country citations
A B
A1 A2 A3 A4 B1 B2 B3
If A and B both cite A4, there is probably some link between A & B.
© Magister Ltd. 2003-2008, 2014 47
Single-country citations
A B
A1 A2 A3 B1 B2 B3
But if B2 is an equivalent of A3, the link would be missed in a simple citation database.
© Magister Ltd. 2003-2008, 2014 48
Family-family citations
fA fB
fA1 fA2 fA3 fB1 fB2 fB3
In a family citation database, if fA3 = fB2 the link will be found.
No web
source does this yet – commercial sector only
© Magister Ltd. 2003-2008, 2014 49
Topics
• Introductory remarks about the patent system
• What use is patent information?
• Information need – understanding a basic search classification
• “Information need” -v- “information source” – choosing a database
– directories and keeping up-to-date
• Multiple authority collections
• Single authority collections
© Magister Ltd. 2003-2008, 2014 50
Portals
• All portals / directories / link lists go out of data eventually !
– best advice is not to rely too heavily on any single source, but build up your own list of established sites
• Many patent-related portals are heavy on the legal side, and contain less
information about searching resources.
© Magister Ltd. 2003-2008, 2014 51
Manually-curated directories
• British Library Business & IP Centre
– www.bl.uk/eresources/business/cd-busin.html (click on
“Intellectual Property Resources”)
– www.bl.uk/reshelp/findhelpsubject/busmanlaw/ip/ipdoc/docum entation.html
• IP Menu
– www.iplawlink.com/ipmenu/
• Intellogist
– www.intellogist.com
• mixture of editorial and collaborative material
• WIPO directory of IP offices
– www.wipo.int/directory/en/urls.jsp
• link from homepage now missing; use Home>Reference>Country Profiles>Directory
© Magister Ltd. 2003-2008, 2014 52
Extracts from list of British
Library holdings of IP material
IPMenu entry for Italian data sources
© Magister Ltd. 2003-2008, 2014 53
Intellogist entry for Italy
© Magister Ltd. 2003-2008, 2014 54
WIPO directory including web links
© Magister Ltd. 2003-2008, 2014 55
Other sources of information
• Professional networking
– join AIDB !
– participate in an established “discussion”
with other professionals e.g. PIUG wiki
• Manual directory from the Dutch user group (WON) at www.won-nl.org
– websites sorted by information content (home page, bibliographic search, national register, official publications, machine translation, litigation, trademarks, other)
© Magister Ltd. 2003-2008, 2014 56
© Magister Ltd. 2003-2008, 2014 57
Topics
• Introductory remarks about the patent system
• What use is patent information?
• Information need – understanding a basic search classification
• “Information need” -v- “information source” – choosing a database
• Multiple authority collections
– IPOs; Espacenet, PatentScope, Depatisnet – Third party (non-IPO); Google, FPO
• Single authority collections
A word of caution
• Non-IPO suppliers of patent data are appearing and disappearing at a fast pace;
– the following survey includes only a few examples, and is limited to those which seem to have an established user base within industry.
– the survey also excludes websites which function primarily as document delivery (search by number only), rather than subject- searchable databases.
– be cautious about relying upon new resources until you understand what they can deliver.
• Some companies not covered in detail (for reasons of space, not product quality) include:
– AztecIPSearch ® (AztecIP), Boliven Pro / Discover IP (CambridgeIP), IP.com Prior Art Database (IP.com), IPEXL (Intellectual Property
Exchange), Pantros IP (formerly Patent Café), Patent Lens ™ / The Lens (CAMBIA) and Prior Smart ® (Priorsmart.com)
© Magister Ltd. 2003-2008, 2014 58
© Magister Ltd. 2003-2008, 2014 59
The Espacenet family (xx.espacenet.com)
• National language interface for all databases
• Up to 5 databases available, depending upon the URL used for access
– local national file (e.g. GB)
– other European national files (e.g. DE, FR…) – EP applications full text
– PCT applications full text (Eng, Fre, Ger only) – worldwide file (90+ countries)
© Magister Ltd. 2003-2008, 2014 60
Note that available fields vary, according to database selection
© Magister Ltd. 2003-2008, 2014 61
© Magister Ltd. 2003-2008, 2014 62
Other members of the Espacenet family
© Magister Ltd. 2003-2008, 2014 63
© Magister Ltd. 2003-2008, 2014 64
Espacenet
• All national files remain in the national language
– searchable titles only, restricted file size
• EP and WO files
– searchable full text for EPO official languages only
– English abstracts and titles for all records
• Worldwide file (includes EP + WO)
– titles and abstracts, range of countries/years
© Magister Ltd. 2003-2008, 2014 65
Espacenet
• Worldwide database
– at least one member of each patent family is available as a scanned image, in PDF format – requires Adobe Acrobat to view
– some documents available with character- coded text (can cut/paste into other
documents)
• Other databases
– document availability/ form varies
Basic Espacenet coverage in the worldwide file
© Magister Ltd. 2003-2008, 2014 66
Additional country coverage is detailed at Help > Coverage of worldwide database (link off the Advanced Search page) or directly via www.epo.org/gpdc (Global Patent Database Coverage)
Example of detailed coverage for a single country (Canada)
© Magister Ltd. 2003-2008, 2014 67
Similar tables, including both overall coverage and timeliness of data delivery, are available for the other countries covered by the file.
© Magister Ltd. 2003-2008, 2014 68
Cooperative Patent Classification
• CPC marks are applied to a proportion of the documents within the worldwide file
– common classification across a range of countries, to assist worldwide searching
– caution : coverage is based upon the PCT Minimum Documentation collection
• Rule 34 PCT Regulations - EP, WO from 1978 + US, GB, FR, DE, CH, JP, SU/RU from 1920 +
• JP and SU/RU not given CPC codes normally, nor are KR or CN
– some other countries also classified, at examiner’s discretion
Family data in Espacenet
• The worldwide database is grouped according to patent families
– one document relating to each invention is chosen as the ‘representative’
– others are stored as ‘also published as’
– this makes the database useful for novelty searching, as the user does not have to de- duplicate equivalent patent applications
• The default family rule is ‘strict’, but the user can opt to view the ‘extended’ one
© Magister Ltd. 2003-2008, 2014 69
© Magister Ltd. 2003-2008, 2014 70
PatentScope
• ‘PatentScope’ is the name used for the
patent searching part of the WIPO website
– www.wipo.int/patentscope/en/
– initially only contained PCT applications, published by the International Bureau of the PCT, located at WIPO
• The WIPO website also contains the authoritative text of the International Patent Classification (IPC)
– www.wipo.int/classifications/ipc/en
PatentScope
• Complete bibliographic data, for PCT applications.
– Approximately 99.7% of the full texts – 100% with page images
• Additional country collections added
– often grew out of collaborative digitisation projects with smaller national offices
• e.g. Central and South American states, Morocco, Vietnam.
– now totals 38 collections + PCT ; 36m records,
approximately ⅔ in full text (9 different languages)
© Magister Ltd. 2003-2008, 2014 71
National collections
• WIPO is constantly adding new (and major) collections
– Canada (2.1m records, 50% full text) – China (3.4m records, 98% full text) – Japan (7.7m records, 94% full text) – Korea (1.9m records, no full text) – USA (10.4m records, 62% full text) – EPO (2.7m records, 67% full text)
– UK and Germany due to be added in 2014
© Magister Ltd. 2003-2008, 2014 72
PatentScope national collections in overview
© Magister Ltd. 2003-2008, 2014 73
Detailed PatentScope statistics
(full list at http://patentscope.wipo.int/search/en/help/data_coverage.jsf)
© Magister Ltd. 2003-2008, 2014 74
Features of PatentScope
• All full texts are loaded in the original language
– a Cross-Language Information Retrieval (CLIR) module facilitates synonym-finding in other
languages
• Proximity and range-searching features are available (not normally needed for Espacenet, due to lack of full text)
• Convenient download of results
• Graphical / tabular summary of results
© Magister Ltd. 2003-2008, 2014 75
Important note
• Although the PatentScope command language provides for full-text searching, almost all the collections apart from the PCT contain only a sub-set of records with full text
– most are mainly bibliographic data
• Do not confuse the presence of page images for full-text display with character-coded text which can be used for full-text search.
© Magister Ltd. 2003-2008, 2014 76
© Magister Ltd. 2003-2008, 2014 77
DEPATISnet
(http://depatisnet.dpma.de)
• Service provided by the German Patent and Trademark Office (DPMA)
• Wide country coverage of bibliographic data, plus some full text
• Provides a ‘beginner’ mode, plus a
sophisticated command language based upon the EPO MIMOSA software
• English and German search interfaces
Summary of principal Depatisnet coverage
© Magister Ltd. 2003-2008, 2014 78
Country Start year Documents Page images (%) Full text (%) Title (%) Abstract (%)
AT 1920- 1104707 39 0 100 3
CH 1888- 723789 98 0 89 5
DD 1978- 147164 99 98 96 81
DD 1946-1977 158076 98 98 8 0
DE 1978- 3050287 100 99 99 47
DE 1946-1977 2637734 96 97 89 1
DE 1877-1945 995750 100 100 44 0
EP 1978- 4965417 93 0 100 33
FR 1920- 3085073 76 0 76 22
GB 1920- 2703224 88 0 97 77
JP 1976- 22062182 47 0 50 47
KR 1970- 3621781 79 0 85 61
US 1978- 8981782 99 0 99 94
US 1946-1977 1819381 100 0 94 24
US 1790-1945 2535355 100 0 91 0
WO 1978- 3068502 98 0 100 99
+ basic bibliographic details from another 90 countries (based largely upon EPO data (= espacenet)
Features of DEPATISnet
• Addition of DEKLA (DPMA-enhanced IPC scheme for subject searching) to core records
• Search on single documents, but with
ability to create ‘on-the-fly’ family groups
• Substantial numbers of PDF copies per family, sometimes more than Espacenet
– good alternative source for document supply
© Magister Ltd. 2003-2008, 2014 79
Google Patents
(http://patents.google.com)
• Initially launched with US patents only
– contains OCR text from 1790-1975, not always high quality.
– some features still only available for US
• Progressively added additional collections
– EP, WO, CN, DE, CA
• Same basic search technology as main Google engine
– by default, results are “relevance ranked”
© Magister Ltd. 2003-2008, 2014 80
Issues with OCR text
© Magister Ltd. 2003-2008, 2014 81
Google search modes
© Magister Ltd. 2003-2008, 2014 82
Where am I ?
(www.google.com/patents or patents.google.com)
Where can I search?
(www.google.com/advanced_
patent_search)
Features of Google Patents
• Good display and download features
– but not for all records
• More hyperlinks than other sites ; facility to expand or modify search easily
• Still very biased towards US content (e.g.
patent type/status only applies to US cases)
• Unusual form of search engine
© Magister Ltd. 2003-2008, 2014 83
Free Patents Online (FPO)
(http://www.freepatentsonline.com)
• As with Google, based in the US and focussed largely on US documents and US practice
• Added EP-A and -B (1978+), JP-B
(1996+), JP-A (abstracts only, 1976+), WO (1978-?) and DE (in beta)
© Magister Ltd. 2003-2008, 2014 84
Free Patents Online
“advanced search” page
© Magister Ltd. 2003-2008, 2014 85
Features of FPO
• Full-text operators provided
– but no indication of handling non-English
• Unusual date range facility, may be useful in some circumstances e.g.
– [field]/[date1]->[date2]
– [field]/NOW-1YEAR->NOW
• Saved searches and alerts are possible for account holders (free to create)
© Magister Ltd. 2003-2008, 2014 86
© Magister Ltd. 2003-2008, 2014 87
Topics
• Introductory remarks about the patent system
• What use is patent information?
• Information need – understanding a basic search classification
• “Information need” -v- “information source” – choosing a database
• Multiple authority collections
• Single authority collections
– USPTO, JPO, SIPO, KIPO – Additional smaller offices
© Magister Ltd. 2003-2008, 2014 88
US Patent and Trademark Office (www.uspto.gov)
• Mounts 2 databases
– granted patents 1790-date
– pre-grant publications 2001-date
• Electronic Official Gazette (eOG:P) for browsing each week’s new grants
– no equivalent for published applications
• Applicant’s service site (pair.uspto.gov) for legal status (Public PAIR) and
electronic filing (encrypted, account-
based Private PAIR)
© Magister Ltd. 2003-2008, 2014 89
USPTO
• Accessing the search files:
– Use ‘Patent Search’ link from the ‘Popular Links’ listing on the home page
– or go via the ‘Patents’ button to a sub-menu and select either
• ‘search patents’ = granted documents, 18-month applications, re-assignments
• ‘Popular Links’ listing which provides additional links to ‘withdrawn patents’, ‘expired patents’
and ‘extended patent terms’
© Magister Ltd. 2003-2008, 2014 90
USPTO
• Browsing the Official Gazette
– Home > Patents > Tools > Official Gazette (Patents)
– or direct link to
www.uspto.gov/news/og/patent_og/index.jsp
• Contains most recent 52 weekly issues only; searchable by patent number,
assignee and (currently) US classification.
• Links to the full text databases
© Magister Ltd. 2003-2008, 2014 91
USPTO - Main search files
• Full-text US patents from 1976
– includes page images in TIFF format
• additional viewer may be needed
– includes page images of certain amending documents such as Certificates of Correction
• Number and classification only from 1790-1975
• Content includes Design Patents, SIR’s,
Plant Patents and Re-issue cases
© Magister Ltd. 2003-2008, 2014 92
Main search files
• File typically updated within 1-2 days of publication
– bibliographic and text data
• Facsimile images of pages can be slightly delayed.
• Caution: amendment documents (e.g Re- issues, Certificates of Correction) are
sometimes missing from the facsimiles –
always check the PAIR record as well.
© Magister Ltd. 2003-2008, 2014 93
Main search files
• Pre-grant publications file
– First documents from Mar. 2001
– New legislation in 1999 (AIPA) required pre- grant publication for most US filings
(previously only once at grant) – New number format adopted:
• US YYYY/NNNNNNN A1
• Granted patents remain in the old continuous series US NNNNNNN (B).
© Magister Ltd. 2003-2008, 2014 94
Main search files
• Classification searching:
– use the associated files of USPC / CPC at www.uspto.gov/web/patents/classification/
(or link from ‘Search for Patents’ > ‘Popular Links’ > ‘Tools to help in searching by Patent Classification’)
– OR
– conduct a search, manually extract the class marks and then iterate via classification
listing.
Searching from within US classification
© Magister Ltd. 2003-2008, 2014 95
Hyperlink:
‘A’ to applications or
‘P’ to granted patents,
carrying this classification mark
© Magister Ltd. 2003-2008, 2014 96
Main search files
US 5,678,910
“Multiple angle projection for 3-D imagery”
US Cl. 353/7
Classification definitions cut-and-paste
436 additional patents in the same class.
Example titles:
“Stereo display apparatus”
“Three-dimensional display”
“Projector”
Class 353
Optics; image projectors Sub-class 7
Stereoscopic
Features of USPTO search
• All documents in a single language (English) and in full text from 1976+
• Basic proximity search available, with hit- term highlighting
• Citations are implemented as hyperlinks
– BUT only to/from other US patents, not all literature citations or other patents
• Provides for supply (at cost) of certified copies of patent specifications
© Magister Ltd. 2003-2008, 2014 97
© Magister Ltd. 2003-2008, 2014 98
Japan Patent Office (www.jpo.go.jp)
• The JPO has two websites, one in Japanese and one in English
– the two sites do not contain exactly equivalent resources for search
– however, more of the Japanese site is being translated and provided over time
• The main tool for English-language
searching is the Patent Abstracts of Japan
(PAJ) database.
© Magister Ltd. 2003-2008, 2014 99
JPO
• Access to PAJ: Home page > ‘Industrial
Property Digital Library’ > ‘Patent and utility model’ > ‘PAJ’
• Data content:
– unexamined applications (JP-A) from 1976-date
• does NOT include PCT national phase entry (500,000 series)
– human-prepared abstracts in English
• 3-month delay from publication date
• Better quality than applicant’s abstracts for searching
– basic legal status and link to machine translations of specification also provided from 1993 onwards
© Magister Ltd. 2003-2008, 2014 100
PAJ
• Searchable by
– Applicant (but not inventors) – Title / abstract
– Publication date – IPC
– Application / priority / publication number (separate search window)
• Initial display is hit-list (index), then link
to front-page style data
© Magister Ltd. 2003-2008, 2014 101
PAJ
• If English abstract is not yet available, display reverts to Japanese original
• Additional data
– limited legal status data via “Legal Status” button – on-the-fly machine translation via “Details” button
• translates the body of the specification and presents in segments, together with drawings where present.
– facsimile of document via “Japanese” button
© Magister Ltd. 2003-2008, 2014 102
Additional resources
• Patent and Utility Model Gazette database
– simple number search, access to facsimiles
• Patent and Utility Model concordance
– links together application number,
unexamined publication number, examined publication number (old law) and grant
number
• FI/F-term search
– Japanese classification systems for non-word search of more extensive databases
© Magister Ltd. 2003-2008, 2014 103
© Magister Ltd. 2003-2008, 2014 104
FI/F-term search
• Classification searching by two methods
– FI system (JPO enhanced version of the IPC)
• typically 1-5 per document
– F-terms (deep-indexing terms)
• typically 15-50 per document
• In addition to an alternative search key, use of classification allows access to a wider collection of documents
– Patents back to 1885
– Utility models back to 1912
State Intellectual Property Office (of the People’s Republic of China)
• Free databases of Chinese patents and utility models are hosted by SIPO (www.sipo.gov.cn)
– two separate sites:
• China Patent Inquiry System (www.cpquery.gov.cn)
– includes account-only e-filing module – public file inspection from 2010 onwards
• Patent Search and Service System (www.pss- system.gov.cn)
• More detailed databases (C-Pat Search) are
provided by the Intellectual Property Publishing House (IPPH), a separate division of SIPO
– registration required, partly fee-based
© Magister Ltd. 2003-2008, 2014 105
English version of SIPO home page
© Magister Ltd. 2003-2008, 2014 106
China Patent Inquiry System
• Simple search interface offering search by
application number, title or applicant name(s)
– no classification, abstracts or other numeric data (e.g.
priorities)
• Provides basic bibliographic information and some legal status in English (e.g. request for examination filed)
– display text is mostly in Chinese
• Link to ‘view’ full documents connects through to same system as PSS System
© Magister Ltd. 2003-2008, 2014 107
Patent Search and Service System
• More sophisticated command language than CPIS
• Provides bibliographic details in both
Chinese and English versions, including abstracts and drawing, plus facsimile of original documents
– link to Chinese → English machine
translation available once facsimile is on screen
© Magister Ltd. 2003-2008, 2014 108
© Magister Ltd. 2003-2008, 2014 109
Korea Intellectual Property Office (KIPO, www.kipo.go.kr)
• Simple search in Korean Patent Abstracts (KPA, manual translations into English) via KIPO home page.
– 1973-date (grants), 1999-date (unexamined)
• Main searching resources via KIPRIS (Korea Intellectual Property Rights
Information Service) at www.kipris.or.kr
– additionally provides access to KPA and K2E-PAT machine translation
© Magister Ltd. 2003-2008, 2014 110
Additional services
• Search databases in addition to KPA
– Patents ;
• grants from 1948, unexamined from 1983
– Utility models ;
• grants from 1948, unexamined from 1983
– Designs ;
• grants from 1948, unexamined from 1996
– Trademarks ;
• applications and grants from 1950
• K2E machine translation
– available on a pay-per-view basis or flat fee for unlimited translations during a given period e.g. daily, monthly
subscription
© Magister Ltd. 2003-2008, 2014 111
© Magister Ltd. 2003-2008, 2014 112 Colour-coded hit list:
Light blue = registered (granted)
Pink = rejected
Dark blue = unexamined
Green = withdrawn
© Magister Ltd. 2003-2008, 2014 113
Minor country sites
• Smaller patent offices with searchable files
– Australia, Brazil, Bulgaria, Canada, China, Czech Republic, Finland, France, Germany, Hungary, India, Italy, Lithuania, Mexico,
Moldova, Netherlands, New Zealand, Poland, Romania, Russia, Singapore, Spain,
Switzerland, Thailand, United Kingdom….
• Some have sophisticated search files,
others are simple browsable PDF gazettes
© Magister Ltd. 2003-2008, 2014 114
UK Intellectual Property Office (www.ipo.gov.uk)
• Local link to Espacenet server
– All 1977 Act GB cases (first published 1979)
• National Patent Status service
– direct link to OPTICS register
– UK Supplementary Protection Certificates
• UK Design and trade mark search files
• Other Patent Office publications and
services, including Patents & Designs
Journal
© Magister Ltd. 2003-2008, 2014 115
Canadian Intellectual Property Office (CIPO) (http://cipo.ic.gc.ca)
• Patent Office Record and corresponding Trade Marks Journal available
• Select ‘Patents database’
– searchable back to 1869 (some text from 1920)
– title, abstract and claims (from 1978) – inventor and applicant details
– full facsimiles in PDF format
© Magister Ltd. 2003-2008, 2014 116
IP Australia
(www.ipaustralia.gov.au)
• Main database
– Integrated AusPat service launched – Full text back to 1904
– Many specifications as PDF
• Special features
– Proximity operators for full text search
– Name Selector allows viewing of inventor and applicant names from consolidated listing, before adding to search form.
© Magister Ltd. 2003-2008, 2014 117
INPI, France (www.inpi.fr)
• Similar arrangement to Espacenet
– title, abstract, applicant, inventor, IPC, publication number
– provides simultaneous FR/EP/WO search in default file
• Complete texts available, not searchable
– FR national collection from 1989
– Translation of EP-A claims into French from 2004.
© Magister Ltd. 2003-2008, 2014 118
Some others...
• Taiwan
(www.tipo.gov.tw/en/index.aspx
) – comparatively crude, site not alwaysavailable… but at least it’s there!
• Mexico
(http://www.impi.gob.mx) – Patents from 1980 to date– Full abstracts and bibliographic details, including front page drawings
Summary (1)
• There is a great deal of free of charge information available.
• Most comes from national IPOs with minimal value add.
• More and more full text is available, but usually only in the original language
• The commercial sector still provides, and charges for, added value in patent
databases.
© Magister Ltd. 2003-2008, 2014 119