Page History

Annual Technical Report 2004 on Patent Information Activities submitted by EPO (SCIT/ATR/PI/2004/EP)

Where URLs are requested below, it is preferred that either URLs which are likely to remain stable over time (three years or more) are provided, or home (main) page URLs are provided with a short explanation of how to access the corresponding information.

The term "patent" covers utility models and Supplementary Protection Certificates (SPCs). Offices which issue design patents should report their design patent information activities in their Annual Technical Reports on Industrial Design Information Activities.

I. Evolution of patent activities

...

In 2004, The EPO recorded 178 580 requests for European patent, 6.7% more than recorded in 2003 (167 350).

Of these, 120 100 (+7.1%) were made by designating the EPO in international applications, and 58 480 (+5.8%) applications were filed directly under the European Patent Convention. Out of the total number of applications filed, 67.3% were made under the PCT (67% in 2003).

As a consequence of the ongoing increase in the number of PCT international applications, the number of applications entering the European phase also increased by 6% to reach 65 230 in 2004. With the European direct applications, the total number of patent applications entering the European patent procedure rose by 6% to 123 700 (116 790 in 2003).

Trends or areas experiencing rapid changes with respect to the previous year

The growth in filings was nevertheless different among the technical units of the IPC. The number of European applications in the field of Electric communication technique (H03, H04) is the highest with 13 544 applications (+8.2%). The units Health (A61-A63 excl. A61K; 8 185; +16.4%), Agriculture (A01 excl. A01N; 1 120; +15%%), Lighting and Heating (F21-F28; 2 335; +14.8%), Foodstuffs and Tobacco (A21-A24; 1 410; +11.6%) and Instruments III( G09-G12; 3 942; +10.9%)) show the largest increases over the previous year, with a growth over 10%.

Seven units registered fewer applications in 2004 than in 2003. Among these were Macromolecular Compounds (C08; 3 113; -3.7%) and Fermentation, Sugar, Skins (C12-C14; 4 025; -4.2%).

In 2004, 55% of the applications in the European patent procedure, were applied for in the leading 10 IPC classes (A61, H04, G06, H01, G01, C07, C12, B60, F16, C08). In these classes taken together, the number of applications increased by 6.1% over the previous year.

II. Matters concerning the generation, reproduction, distribution and use of primary and secondary sources of patent information

...

A wide variety of announcements and notices are provided on a monthly basis in the Official Gazette of the EPO.

All such announcements (communiqués, press-releases, official notices) are also published on the EPO website under "Official communications" (visit http://www.european-patent-office.org/_new_off_comm/index.en.php

Mass storage media used (paper, microforms, optical storage, etc.)

BNS Image data: 5.8 million documents loaded in 2004

The EPO maintains a digital library of facsimile documents. In 2004, this collection, one of the world’s largest, increased by 5.8 million documents to 55.6 million complete patent and non-patent documents.

Quality control of the incoming facsimile data is systematically performed by specialised teams applying an ISO approach for the selection of the samples and making a decision on batch acceptance. Guidelines on backfile indexing and scanning of patent documentation from a vast number of national offices have been released and applied.

At the end of 2004, a new PDF Standard, was accepted after six months of comprehensive testing of various versions and will allow for the addition of new facsimile collections in 2005 and/or replace scanning of paper delivered so far. This concerns various publications from TW, RO, ES, KR.

Loading programmes were improved and completed. In a continuous effort to expand and improve the BNS collection, a total load volume of almost 5.8 million documents was achieved in 2004.

The following are some of a number of achievements regarding new acquisitions of patent documents (BNS):
JP: Addition of B and Y publications started in December 2004 with the loading of about 100 000 documents published in the reporting year.
DK: Image data loaded going back to No. 1, published in 1895
MA: New country added to EPO collection, image data from 1980-1999
NO: Addition of image data from 2000
SE: Addition of image data going back to 1895
CZ: Backfile scanning completed
BR: Image data loaded 1982-2002
FI: Image data loaded 1842-1919
GB: 350 000 pre-1920 documents loaded
BE: 42 000 documents loaded (published 1928-1986)
CS: 145 000 documents loaded
CA: 205 000 documents loaded (published 1999-2004)
HU: Complete collection 1980-1991 (23 784 documents) now available
RU : 31 092 documents loaded
FI, DE : Various loads to complete backfiles

Regular bulk extractions took place in 2004 at the request of other national patent offices or organisations.

Full text databases: more than 14 million patent documents in OCR format now

Much work went into the development of keyword-searchable full-text databases. A total of 1.3 million full-text patent specifications have been added to the EPO’s collection, extending it to over 14 million (patent) documents.

The most important additions to the full-text collection include French, UK and Belgian patent publications. For the FR full-text, more than 200 000 documents published in the period 1980 - 2004 were loaded. As to the GB full-text, the number is almost 200 000, ranging from 1979 - 2004. For both, the coverage changed from one member per patent family to all published GB and FR applications.

The full-text for BE publications between 1984 and 2004 has been made available in three new full-text databases for the languages used in Belgium: French, Dutch and German. Around 15 000 documents were available at the end of the year.

The EPO’s full-text collection now also comprises the European unpublished patent documents through EPOQUE (TXTAEP*) with about 85 000 (unpublished) EP applications available to EPO examiners at the end of 2004.

With these permanent additions, the EPO has today far exceeded its initial aim of guaranteeing at least the same search quality with electronic means as can be achieved with paper documents.

Non-patent literature and library services: constantly improving

One of the areas of major activity in 2004 was once again scientific non-patent literature (NPL), a major source of information especially in the biotechnology, medical technology and computing fields. At the end of the year, 53 million articles in commercial NPL databases were accessible via the EPOQUE online search system. These publications are also accessed in their complete version via the EPO Virtual Library (EVL), a gateway to the Internet version of 3 900 journals (1 000 more than in 2003) and of the conference proceedings, books, encyclopaedias and standards of the major scientific publishers. The EPO has also developed an intranet repository of special documents including a full-text collection of 177 000 documents (100 000 more than in 2003).

In NPL, three new full-text databases were added (one containing American Institute of Physics Journals, and two for standards in telecommunications) and the existing full-text database of Elsevier journals was extended to contain the whole collection of their 1 800 journals, bringing the total number of NPL full-text records in EPOQUE to 1.2 million.
In the Library and Retrieval Sections, several developments that improved access to complete documents took place.

One of the main achievements in 2004 was the acquisition of the collection of the 730 Kluwer journals for EVL, bringing the number of journals available at the end of 2004 to 3 911. The number of freely accessible academic journals found on the Internet also increased by 12.5% to about 180 titles.

The number of orders for documents which could not be retrieved by the examiners themselves decreased by 6.8% in 2004 to 45 574, thanks to EVL and BNS. Indeed, the statistics of 9 important publishers in EVL show that 170 000 documents were consulted in 2004 by examiners, representing about 1.2 million Euros of savings in the OLIT budget. E-OLIT, put into production in 2002, was upgraded in 2004 to load automatically into BNS the articles retrieved internally, either in the paper library or in EVL.

A tender was launched in 2004 to select a single supplier of magazines for the EPO and therefore benefit from more competitive fees.

A large number of books were purchased, mainly because of the new version of the PCT Regulations, which had to be distributed in The Hague to 2 690 examiners and formalities officers.

New activities started or saw a strong increase in 2004: the central registration of incoming patent data, the acquisition of CD-ROMs for Network consultation and the retrieval of patents for the paperless project.

Word processing and office automation

New EPO Operating System (NEO)
The NEO server infrastructure has been migrated from Windows 2000 to Windows 2003. A new software for time and production management has implemented to support BEST procedure: MUSE (Management of United Search and Examination). Concerning the messaging infrastructure, the migration from OV to Lotus Notes was completed at the end of 2004.

III. Matters concerning abstracting, classifying, reclassifying and indexing of technical information contained in patent documents

...

The Munich team of PD Tools

In 2004, the Munich team was very busy with classification activities as, within the framework of BEST, the activities of Munich examiners in classification increase. Support was provided relating to DOCHELP queries, training of new classifiers and training in the use of ClassTool/DocTool within the framework of the NEO roll-out. Assistance was provided for ECLA training at national offices and classification training in the Asian area.

In the pre-classification area, a study was carried out and a quality control procedure developed. A pilot project is currently running in the Human Necessities Joint Cluster.

The Munich Intranet Team is in charge of all the technical issues, such as the implementation of the intranet, its technical platform, its structure and guidelines, as well as introducing new tools and features. This is done in close contact with all other EPO sites and DGs. The team publishes the local part of DG 1-2 Line sites, the President’s office site and DG 5 Line. The editing process has been simplified by separating content and presentation as well as using adapted technologies. A new tool (Contribute) enables content providers to edit their pages directly without having to contact the intranet team.

EPODOS (Electronically Prepared and Organised DOSsier), described elsewhere in this report, is managed by the Munich directorate. As a contrast to this most modern development, it is also in charge of a project requested by some national offices to preserve historic patent documents. Its object is to capture old patent applications and patent-related historic documents issued by the EPC member states before 1920 and to facilitate their access.

Following improved data coverage in BNS and EVL, the volume of traditional library services decreased. The pilot phase of a new library programme developed by the DOMOS team started in November 2004 to replace the current OASIS programme in Munich and BOCA in The Hague. Circulation of periodicals in paper form is still a major exercise with more than 22 500 new issues distributed over the year.

As to the usage of external databases monitored by the Munich team, the total expenditure, EUR 2 600 000, decreased by 8% compared to 2003. 97% of the expenditure was generated by the Joint Clusters P&A OC, Biotech, Industrial Chemistry and Polymers.

Data classification: improving document access

Documentation’s usefulness does however largely depend on how easy it is to access.

As in the years before, the Office has played a leading role in the IPC reform process, which aims to modernise the IPC in an electronic environment and make it more useful. Preparatory work is progressing in accordance with the new plan to meet the deadline of January 2006.

In 2004, the EPO received 782 000 newly published documents, about 51% of which were automatically classified by computer on the basis of a classified family member in DOCDB or because they correspond to patent applications previously classified by the examiners in DOCTOOL.

A total of 380 750 new patent documents were sent via Clipon to EPO classifiers at all EPO sites, resulting in the treatment of about 943 000 Clipon circulations (a document can be sent simultaneously to several classifiers) and the allocation of an estimated 993 000 symbols (ECLA, ICO, KW). Due to more classifiers being trained and batch transfer of documents to Munich, the number of treated circulations was 11.2% higher than in 2003.

The classification of 28 400 NPL documents resulted in 56 800 paper circulations with about 36 900 symbols allocated.

The reclassifications carried out via DOCTOOL resulted in about 2.3 million actions, eg allocation or modification, affecting the ECLA descriptors (EC, ICO, KW and TXT). This increase of 15% compared to 2003 was partly due to bulk transfers, for example resulting from the conversion of the CIS deep-indexing system in the field of C10M into ICO symbols, as well as a result of many reclassifications performed by examiner assistants in reorganisations. The fact that the number of examiners authorised to classify in Munich and Berlin increased also contributed to this increase in activity.
The project begun in 2001 and employing external staff (mainly students from technical universities) on a temporary basis to assist examiners with their reclassification work was continued and diversified. A new contractor was selected in 2004. Reorganisations took place in 29 directorates affecting 120 300 documents. This represents 13.1 examiner-years. A survey was conducted to evaluate the project and results will be published in 2005.

Time spent on documentation activities totalled 100 man-years in 2004, 5% more than in 2003.
Pre-classification of new incoming files is carried out by PD Tools pre-classifiers and examiners. In 2004, it was possible to increase the share of PD Tools pre-classifiers to over 50%. A total of 98 400 dossiers were pre-classified intellectually and another 2 600 automatically using software developed by Directorate Applied Research and Development. The pre-classification team also treated some patent documents for routing in Clipon and a small amount of NPL.
Deep-indexing for search in some selected fields was conducted for 4 600 documents.
One of the major activities of the Classification Board in 2004 was harmonisation of ECLA with the IPC, resulting in the creation of more than 400 new ECLA groups corresponding to the future IPC edition 8. A pilot study to test and improve a procedure for monitoring the quality of the classification process using a sample representing 250 000 documents was carried out by the Documentation Quality Board (DQB) . Results will be published in 2005. It continued its participation in the IPC Revision process.
A total of 897 requests for revision of the internal classification (DOC14) were received in 2004, 7% more than in 2003. For the correction of obvious errors, a simplified procedure was introduced using the newly created address eclafix@epo.org.

11 426 creations, additions or modifications of classification entries were performed by the ECLA team and 2 600 REPA requests were handled.
Turning to external classification systems, about 30 000 new Japanese documents receive FTerms and FIclasses in DOCDB every month. UC data (US DOCUS) are now fetched and loaded in DOC-DB2 on a bimonthly basis.
The Harmony project pursued by the Trilateral Offices aims to obtain similar classification schemes in ECLA, the US patent classification and the JPO’s FI classification by revising their respective systems in selected fields. It should lead to a reduction in the classification workload and easier access to each Office’s documentation. Seven new Harmony projects were started in 2004.
Training activities were organised in collaboration with L&D teams in Munich, The Hague and Berlin. Most of the effort was directed to the training of new classifiers in the use of DocTool/Clipon and ClassTool. Altogether 316 examiners participated in the “Classification for classifiers” course. Specific courses were offered to the 60 or so newly appointed documentalists in Munich to help them with their tasks. As several EPO member state offices had indicated an interest, the EPO now offers training on the use of ECLA. In 2004, participants from the German Patent Office followed such a training course and contacts were made with the Greek, Finnish and UK Patent Offices with a view to organising a similar course in 2005. A pilot remote training course was organised for examiners from the Swedish Office. An evaluation of this project will be available in 2005.

Coordinate indexing (domestic deep indexing systems, keyword indexing)

...

Data management: increasing the “intelligence” of our documents

The EPO is constantly acquiring and loading new data to update and complete existing databases. Many activities take place in the area of rebuilding and modernising the master databases to offer better quality and services to the users. This summary provides the highlights for 2004.

About 98 600 English-language abstracts (+12%) were purchased or translated to complete the PCT minimum documentation and for other EPO member states. About 2.7 million new bibliographic data and 4.8 million backlog data from 73 countries were processed in the bibliographic master database (DOCDB), bringing the overall number of documents to 58 million, a 5.5% increase compared to 2003. Corrections were required to 983 000 records.

The bibliographic data of 34 300 patent documents received on paper from various countries were manually recorded and loaded into DOCDB. About 71 200 scientific and technical articles (NPL) selected and classified by the examiners were recorded in-house and loaded into the NPL master database. Bulk data capture in the Philippines continued in 2004 with the capture of bibliographic data of 929 300 patent documents mainly published prior to 1920 (GB and US).

In 2004, the citation unit treated 185 100 EP and International Search Reports, which resulted in about 330 000 corrections. About 0.5 million search reports containing 4.9 million citations were loaded into the master citation database, bringing the overall number of references up to 56.9 million patents and 7.6 million scientific articles. Reference to published documents is taking place throughout the EPO granting process and results in various types of citation which are all recorded in a central search process database. In 2004, about 1.2 million references were loaded into this database.

In order to reduce the volume of documentation to be maintained on paper, the "Paperless" project was further extended. By the end of 2004, only 35% of the documentation was still being updated on paper and 58% of paper documentation has been removed since the start of the project in 2001. About 11 million documents are still available in paper form in The Hague. As a result of the paperless office approach, a total of 22 379 documents have been scanned and loaded in BNS in 2004.

The DOCAREA Rebuild project

The EPO is in the process of reviewing and rebuilding the set of computer systems which support data exchange and data management activities and in particular the acquisition, management, classification and distribution of patent and non-patent documents throughout the Office.

The technical objectives are to produce a flexible architecture which is adaptable to future requirements and to make use of new technologies where appropriate, e.g. XML and JAVA. This major project enables PD Tools to shorten the processing time needed to load data, to cope with the increasing volume of incoming data and to secure better tools to manage the quality of the databases. It will also allow sufficient flexibility to incorporate new functionalities and to reduce maintenance activities.

The so-called “DOCAREA Rebuild” project has been phased in in a succession of sub- projects. A completely new master database (DOCDB-DB2) was loaded with production data in August 2004. In 2004, DOCDB2 started to be used for data exchanges and a major switch-over from the former DOCDB to the new DOCDB will take place by mid-2005.

New programmes for data exchanges have been validated and accepted by internal and external users. Procedures for English abstract and for citation data exchanges have been rebuilt and data is now extracted directly from the new DOCDB master database.

A new EPODOC exchange procedure (in XML format) has been put into production. A backfile delivery of 60 million documents has been used to load the new EPODOC database. In addition, the update procedure uses a so-called "Pull" mechanism which ensures full control by the EPOQUE Team of the extractions and is used to fetch the data, when needed, from the DOCDB master database.

The merging of data into one single master database has brought the benefit of data integration and has removed many data inconsistencies. This achievement represents an important milestone for the EPO and leads, through the new EPODOC, to many other major benefits for users, such as 34 million bulk corrections, 7 million English abstracts of Japanese applications, 1.5 million original abstracts, loading of EC classes allocated by national offices, synchronisation of classification updates from Trilateral offices, timely availability of new EP, DE, FR, GB publications in EPODOC and esp@cenet.

A new application was developed to allow a daily batch update procedure, online correction and ad hoc corrections. This new application will be put into production at the switch-over of master databases in 2005.

A new module has been finalised and integrated into CLASSTOOL to allow examiners to edit definitions in order to describe in detail the scope of ECLA and ICO symbols.

Other developments took place in the area of classification allocation and patent distribution which will allow better control and also diagnostics of the distribution process. The statistics will be more detailed than at present, and in the case of classification actions will distinguish between NPL and patents.

IV. Search file establishment and upkeep

...

Adding new countries and extending our patent data coverage

In 2004, the Front Office continued its tasks of updating and creating programmes to load the constantly changing delivery of patent data from over 70 patent granting authorities worldwide.

Some of the major achievements of the year are as follows:
MA - Morocco. Data for this new country was received covering the period 1980 to 2004. This has been loaded and work is ongoing to establish a frontfile delivery.
GC - Patent Office of the Cooperation Council for the Arab States of the Gulf. All data issued by this office since it started publishing in 2002 has been loaded.
AU - Australia. Over 260 CDs have been received containing missing data from 1970 to 1997. The programs to load this are in production and loading has commenced.
DE - Germany. New kind codes and number formats have been received and loaded.
MD - Moldova. New kind codes for plants (E, F & P) have been received and loaded.
CA - Canada. Extra data has been received enabling data from 1970 onwards to be corrected and extended.

In addition, the Front Office has been very proactive in meeting representatives of other Patent Offices when they visit the EPO, and promoting the benefits of supplying data to the EPO for loading into DOCDB so that the world can see their data via esp@cenet. In 2004, meetings with EPO member states took place, as well as with others including China, Russia, New Zealand, Moldova, Australia, Taiwan, Canada, Egypt, Morocco, Tunisia, Serbia and Montenegro, FYROM, Croatia and the LATIPAT countries. These meetings have improved the communication with many Offices and should result in higher data quality and enhancement of our coverage.

In the area of esp@cenet data delivery, two new local esp@cenet servers were put into production in 2004: LP (LATIPAT) and OA (OAPI). For the first time, a national esp@cenet server displays abstracts: at the end of 2004, the BE server offered 15 000 abstracts dating back to 1984. The DK server coverage was extended back to 1895.

Improving treatment of the data received: reducing manual corrections

Once again, over 200 changes to country programmes were made during the year, ranging from measures to improve the recording of priority data to complete format changes. Major changes included the preparation of the migration of US data to WIPO ST.36-compliant XML, the international XML standard for patent data. These changes improve automatic data handling and reduce the amount of costly manual corrections that have to be performed.

V. Activities in the field of computerized and other mechanized search systems

...

Improving examiners’ tools and documentation

Introduction

The EPO possesses one of the world’s most comprehensive collections of technical documentation. This huge amount of knowledge is indispensable for the Office’s search work, and is a key to the quality of European patents.

The main patent data collection managed by the EPO is also accessible to the public via commercial providers and the esp@cenet® patent information service. Databases managed by PD Tools are a main contribution to the European knowledge base society.

To maintain its lead in the field, the Office concentrates its documentation activities on two main aspects: improving the range and quality of its databases, and continuously improving its online search tools.

Implementing quality policy

In 2004, the EPO continued to rebuild its master database and procedures which support the acquisition, formatting, classification and distribution of data. By using modern technology and redesigning most of the processes, the EPO expects to improve further the quality, timeliness and completeness of the data.

In line with the EPO’s quality policy, a Quality Management System for the Documentation Acquisition Processes and the development of specific tools and processes to allow better monitoring of patent and non-patent data status are being introduced in 2005. Specific quality indicators have been set up and systematically monitored, allowing the process owners to trigger corrective/preventive actions whenever required.

Ideal and realistic targets have been set for the timeliness of frontfile data received from national offices as well as for the completeness and correctness of the data loaded into the EPO databases.

Computer-based search at the EPO: looking toward the future

The cornerstone of computer-based search at the EPO is EPOQUE, which is increasingly developing into a global tool covering all the applications used by examiners during the patent grant procedure.

The central backend part of this application, the search engine, is being rebuilt. This work, which started in 2003, should add new search capabilities to EPOQUE and improve its performance.

A contract concerning development work and provision of support and maintenance service with EMPOLIS-Bertelsmann, approved by the EPO Administrative Council in October 2002, was signed in March 2003 for the rebuild and development programme of the EPOQUE search engine. The programme is split into a series of increments, the first three of which were finalised according to plan:

Increment 1 – Analysis of the existing programme
Increment 2.1 – Bringing in the virtual database concept
Increment 2.2 – Technical concepts and global design

Increment 2.3, which includes the basic search engine, is the kernel of the query services and will be delivered in spring 2005. The acceptance of this search kernel is one of the major milestones and critical factors in the programme. Increment 2.3 - including the fusion - has therefore been carried out first during the development phase of the project. It allows for a point of evaluation and possible decision-making on the further steps and design philosophy of the programme.

Concerning databases and their management, the development of reformatting and loading programmes for new databases and the maintenance of existing ones are the main ongoing tasks, with EPODOC, WPI and EUREG remaining as the main databases in view of the workload.

At the end of 2004, the 90 EPOQUE databases contained more than 236 million records fully searchable in the INTERNAL retrieval session.

The electronic patent collection allows searchable access to 50 million patents or patent applications with more than 16.2 million abstracts in English (EPODOC and PAJ), more than 12.6 million WPI abstracts and more than 14.5 million full text descriptions and claims.

For scientific and technical non-patent literature (NPL), EPOQUE INTERNAL gives access to 54.5 million abstracts and bibliographic data with more than 2.9 million XP numbers and 1 450 000 full texts of scientific articles (+47.5%).

A total of eight new databases were added to the EPOQUE database portfolio. The main development activities have concentrated on the rebuild of the EPODOC database, following closely the DOCDB rebuild programme.

As to the client software, all 13 workstation applications have migrated to the NEO-Windows platform following their redevelopment in JAVA, offering many new functionalities. Two EPOQUE workstation clients were discontinued in 2004: EPOQUE O/S2, BDOOR Windows. All the users at the EPO and in all member states have been migrated to the full JAVATM SEA or EPOQUE NET.

Number of users: the total increase in the number of users for EPOQUE Retrieval/search engine access in 2004 compared to 2003 is +5.8%, with +4.4% for The Hague, +4.5 for Munich, +5.9% for Berlin and +9.7% for the national offices. The monthly number of users was 4 800.

EPOQUE retrieval - Active search hours: a 2.0% increase in the connect hours was measured in 2004 compared to 2003. There was still a significant continuation of the effect of the deployment of BEST in The Hague, Berlin and Munich, although to a lesser extent than in 2003 compared to 2002. Whilst the first two sites are producing fewer searches than in earlier years, and consequently use less mainframe resources per user (decrease of 5.1% and 8.1% respectively), Munich has increased its search time (+6.4%). National office use of the EPO has continued to increase steadily (+11.3%).

EPOQUE and BNS Display: use of the EPOQUE Viewer Main display by examiners and EPO member state users increased by 23.8% in 2004. A total of 206 million pages were viewed.

SEARCH TOOLS - SUITE OF EXAMINER’S APPLICATION
In October 2004, the Office has released a new version of the EPO Suite of Examiner’s Application (SEA) with new functionalities relating to EPOQUE and the Viewer in particular. A corresponding version for national patent offices has been rolled out as well. In the EPOQUE VIEWER, the request for translation of Japanese patent documents can be issued by the examiner for translation of the claim and the description on the Japanese IPDL site. DOI's (Digital Object Identifiers) are introduced in all EPOQUE applications in order to allow access to the original article on internet by one click.

Specific tools

EPODOS

Over the past 20 years, automation of examiner tools has covered all major fields of the examiner’s tasks. With the launch of the EPODOS project (Electronically Prepared and Organised DOSsier) in 2004, the Office intends to develop an overall and comprehensive application for handling and processing patent applications together with the related documents and data. EPODOS will bring together all data and the toolbox needed by a BEST examiner to accomplish efficiently the tasks concerning either one or several related dossiers and not yet supported by current applications. EPODOS will integrate all phases of the procedure and apply to all types of application. Following project definition and selection of the team, actual developments will start in 2005.


Research and Development

In the framework of its Research & Development activities, the EPO experiments and develops new tools to further improve its efficiency, such as ranking search results according to the relevance for the search query or testing new products, which offer access in one click to multiple and highly relevant databases on the Internet. The fast localisation of text passages in patent images is another example that should be mentioned.


The project that was started in 2003 concerning automatic multilingual terminology extraction was resumed in 2004 to produce terms for translation purposes by automatic alignment.

The EPO Administrative Council, at its 100th meeting in December 2004, formally approved the European Machine Translation Programme (EMTP). A prototype was developed during 2004 and the basic services of the prototype system are fully operational and available to EPO staff. It was decided to perform some preliminary tests to assess the quality of translation before any customisation and development work, mainly involving lexicon enhancements, i.e. adding domain-specific and other technical and patent-related terms, was started. A general acceptance test was carried out among EPO examiners to serve as a basis for comparison with later surveys as the implementation of the system progresses. The Spanish patent office has already provided an initial set of documents for term extraction and the German and French offices will follow.

For automatic pre-classification, 2004 was a concept validation period for the automatic categoriser software used as a tool. A large-scale survey revealed that 85% of the 4 300 applications automatically classified were assigned to the correct directorate.

The query expansion project will continue to use the search keywords collected by examiners over many years in the EPOS database. Inconsistencies and redundancies in terms and expressions, which were extended to the same semantic family (related words and synonyms), were removed. This precious legacy is now generally available under the thesaurus in the full-text application.

Flowchart search was a further project in the reporting period. Its aim is to offer examiners the possibility of retrieving drawings in patent documents containing flowcharts.

A dedicated software, GetColls, was designed specifically for downloading standard documents.

A prototype has been developed to highlight passages in BNS pages. This will enable the user to select a passage of a document description in the text view and retrieve the highlighted passage in the facsimile view.

Within the framework of JAB Charter 806, aimed at providing examiners with means to collect and keep synonyms of chemical names efficiently, progress has been made by implementing in EPOQUE a generic query builder assistant and in-house access to the PubChem public chemical synonym database.

Biotechnology

The EPO is improving the specific search tools used by examiners in the biotechnology field by enhancing users' interfaces. The EPO performed an ergonomics analysis and is implementing the results of this study. In addition, integration of SNP (Single Nucleotide Polymorphism) related databases has started. The EPO also initiated the first move towards a non-redundant sequence database disclosed in published patent applications.

A total of 6 238 patent applications disclosing biological sequences were received in 2004, an increase of almost 50% compared to 2003. At the same time, output of the sequence listing unit doubled in 2004.

The EPO and EBI (European Bioinformatics Institute) continued their co-operation begun in 2004 to generate new specifications for the disclosure of nucleotide and protein sequences in patent applications. An XML International Nucleotide Sequences database format was presented by them in mid-2004 and during the trilateral Biotechnology working group meeting in Vienna in October 2004. The EPO is working towards adoption of this standard format by the trilateral patent offices in 2005.

In 2004, a total of 34 000 sequence identity searches were performed, of which 97% by EPO staff. About 4 200 patent applications were searched, representing an average of seven sequences per file. Sequence database entries were cited about 1 250 times in 2004. Further achievements in 2004 include the linking of full-page articles directly from sequence entries, access in SEA to annotations of the sequence database and direct import of sequence database entries in CAESAR.

Intranet moving toward a portal

As regards the intranet, the EPO is progressively moving existing content and applications under a Portal environment. In 2004, a Portal proof-of-concept was produced and some content conversion and infrastructure installation work was completed. This new tool will facilitate the internal communications and knowledge-management process within the EPO.

Improving our communication

PD Tools Documentation is continuously aiming to increase its user support and strengthen communications with its users.

The EPOQUE helpdesk is in permanent contact with the TECNO team (technical support to national offices) to ensure delivery of software releases and ad hoc second-level support, organise meetings and liaise with the EPOQUE and STC co-ordinators at the EPO sites. The EPOQUE helpdesk also gives training on request and provides internal and external users with valuable support for the use of EPOQUE, the VIEWER, EXTERNAL access and the EPOQUE databases, in close co-operation with DOCHELP for the data aspects.

DOCHELP (the documentation helpdesk) processed 5 128 requests, an increase of 13% compared to 2003. Most requests came from The Hague (59%), followed by Munich (28%) and Berlin (7%). The number of external requests continues to increase, with most received via the esp@cenet® helpdesk in Vienna. Enquiries received via IS Helpdesk are now processed directly at the Tivoli Service Desk.

The DOCHELP Intranet Homepage introduced a hotpage column containing the latest information on current, temporary issues and problems affecting a bigger group of users. The DOCHELP Frequently Asked Questions were updated and distributed to the examiners.

Three issues of DocToday, the paper bulletin of PD Tools Documentation, were released. DocAlert, the monthly electronic newsletter introduced in 2001, sent 12 issues to 3 200 examiners and other subscribers in 2004. The Gazette also published 11 articles on PD Tools activities and news.

Members of the communication unit gave numerous lectures on PD Tools during academies and other events. They also regularly contributed to the IPR helpdesk, an initiative promoting patent awareness and patent information among EU-funded research projects. Collaboration between the EPO, WIPO and the IPR helpdesk resulted in a common information tool: an IPE Newsfeed.

User Meetings held in spring 2004 at all EPO sites informed more than 1 300 participants about the latest and future developments in PD Tools. These events were greatly appreciated.

Training and more

In the framework of the training programme developed for A and B-grade staff concerning a proactive behaviour of staff towards objectives and priorities of PD Tools, about 60 staff members followed the seminar “Taking individual responsibility and initiative” which will continue in 2005. Other specific training was organized in collaboration with L&D, IS and the personnel department.

Newly recruited staff received information about the tools, products, and activities taking place in each section of PD Tools in a PD Tools Academy.

The yearly event for PD Tools' staff, the “PD Tools Day”, was organized in Brussels in 2004, including a visit to the European Commission.

External databases

There are no new developments to report for calendar year 2004.

...

Figures

Databases

• 90 different databases are available internally with a total of 236 million records.
• 56 million patent records are available in the EPO master patent database Doc-DB.
• 54 million records are available with searchable abstracts in non-patent literature databases.
• 16 million full-text searchable records: 14.5 million patent and 1.5 million non-patent literature articles can be searched and accessed by EPO examiners, a 9% increase compared to 2003.
• 55.6 million facsimile documents are available in the EPO digital library.

Usage

• In 2004, a total of 37 terabytes were downloaded from the EPO digital library BNS.
• an average 4 800 monthly users use EPOQUE databases, 6% more than in 2003.

VI. Administration of the industrial property office library and services available to the public (relating to facilities, e.g., for lodging applications, for assisting clients on searching procedures, for obtaining official publications and registry extracts)

...

The EPO Website
The EPO Website (www.european-patent-office.org) has been hosted since May 1995 and is managed by the Webteam in Vienna. The website content continues to grow in volume (more than 90 000 pages and some 4 GB of data), audience (some 4 000 distinct users per day) and popularity (some 4.5 million visits, 83 million hits and over 800 GB of data downloaded in 2004).

A number of “microsites“ with specific content and a new homepage were introduced in 2003 to demonstrate the “look-and-feel“ planned for the new website. This action was continued with the introduction of 14 new EPO microsites in 2004 including one for the Administrative Council of the European Patent Organisation (our governing body).

The Internet is a great medium for promoting patent information products and services, enhancing public patent awareness and introducing visitors to the concept of patentability. This should lead to an increase in the number of users of patent information and subsequently the patent system. The range of EPO services accessible via the website also includes esp@cenet® and epoline®, part of the website's evolution from a conventional website to a portal for patent information products and services.

The new website project was slowed while the EPO looked into a consolidated approach to Internet and Intranet activities, likely to be based on a new common platform. This will require convergence in the Office‘s various web-based services. The project includes the procurement of a Content Management System (CMS) for the EPO website in order to facilitate management and maintenance of this increasingly complex site. The new CMS (to be implemented during 2005) will be used for both the external website and the new, office-wide Intranet project which will supersede the current constellation of DGxlines.

In the framework of the Viennese training activities on patent information products, we will continue to develop distance learning activities taking into account new technologies. By creating a so-called virtual classroom, registered participants can participate interactively via the internet in instructor-lead "live events". Such new services will be accessed via the EPO website and will fully conform with the EPO website design guidelines.

At the same time, we will progressively introduce the use of data in XML format and collaborate with other intellectual property players such as WIPO and the IPR HelpDesk.

PATENT INFORMATION SERVICES

esp@cenet®
esp@cenet® provides access to patent information to the broad general public via the internet at no charge.

Usage of Version 3 over the last year has continued to grow, with up to
1.3 million searches, 4 million views, 11 million page downloads, and 65 000 different IP accesses per week. The majority of users (61%) have migrated to the latest version of esp@cenet®.

In addition to the installations in each of the member states, there continues to be increasing demand for the creation of servers for non-member states. Long term test installations still exist for the Brazilian patent office, additional installations for Rospatent, the Eurasian patent office and “Latipat” (a single server hosting data for up to eighteen Latin American countries) are now in place. An esp@cenet® installation will be implemented in Japan during the latter half of 2005 and work is in progress on a Chinese version.

The new Version 3 interface for esp@cenet® has now been adopted by the majority of offices, with ongoing plans in place to ensure that in the second quarter of 2005, all offices will be running the latest version. The "old" version will be closed at the end of the second quarter of 2005.

Work has continued on extension of esp@cenet® to allow local collections to include multiple document types (applications, granted patents, utility models etc.), as well as locally supplied data in addition to the standard supply that comes from the Office. This possibility will be used primarily for loading backfile data on to the Level I servers, but will also enable the rapid loading of new documents, such as translations arising from a European grant where these are not currently available from the Office’s systems.

Major efforts have been focussed on the rebuild of the level 1 application.
The Verity Information Search engine has been replaced by the K2 search engine. In line with the overall development strategy, the application is designed to run on the common EPTOS Linux platform, although it should also be possible to run it, if necessary, on the existing esp@cenet® Windows based hardware. Other aims of the rebuild were the harmonisation of the code base between the level I and level II applications, for cost and ease of maintenance reasons. Support for UNICODE has been incorporated to resolve a number of existing codepage issues

An extended test phase of full document downloading will be launched in second quarter of 2005 prior to making the facility freely available to the public in the latter half of 2005.

ESPACE® Series

The ESPACE® series offer high density data storage for publications. They enable simple and cost-effective distribution of a large number of documents and, in combination with the MIMOSA Software, full-featured search access to full documents, indexes and other type of data.
A wide range of ESPACE® series has already been in production for a long time (the first ESPACE® series was started in 1989). Many MIMOSA-compatible intellectual property information products are also in production, some within the framework of EPO co-operation activities with member states. Third-party products are also in production by other patent offices and commercial data providers.
The following products are currently in production:
EPO Series
ESPACE®-EP (previously ESPACE® EPA and EPB), ESPACE® ACCESS EP-A, ESPACE® ACCESS EP-B, ESPACE® FIRST, ESPACE® WORLD, ESPACE® BULLETIN, ESPACE® LEGAL
National series member states
ESPACE® AT, ESPACE® BNL, ESPACE® CH, ESPACE® DK, ESPACE® ES , ESPACE® Fl / Backfile, ESPACE® GR, ESPACE® IE / Backfile, ESPACE® IT, ESPACE® MC, ESPACE® PT, ESPACE® SI, ESPACE® UK, , ESPACE® Access-EPC, ESPACE® PRECES Access, ESPACE® PRECES
National series non member states
ESPACE® ASEANPAT, ESPACE® ID, ESPACE® MY, ESPACE® PH, ESPACE® TH, ESPACE® OAPI, ESPACE® DOPALES, and other series from Australia, Canada, Commonwealth of Independent States, Japan, Korea, Mexico, Taiwan, Russia
Commercial providers
PCT Gazette, US Patent Images
These are stand-alone products which are used primarily by external users around the world.
The Office’s Mission Statement calls for it to remain a global leader in the provision of patent information. MIMOSA Software and the ESPACE® products play a crucial role in this. They enable the Office to disseminate patent information products easily and cheaply to other patent offices, their libraries and the general public.
For thousands of users outside the Office, MIMOSA Software and the ESPACE® products are essential archiving, document delivery and search systems. They permit full-text searching in a local environment through a single multi-lingual interface with a very complete range of features for complex searching, display and printing facilities that make the content readily accessible and rapidly retrievable.
The ongoing development of the MIMOSA Software ensures that products around the world remain compatible, which supports their use as documentation exchange media.
ESPACE® BULLETIN Online
ESPACE® BULLETIN Online is a fully-searchable edition of the Bulletin available on-line.
During 2004, the Office started to make available PDF copies of the European Patent Bulletin on the Internet on the date of publication free of charge from the Office’s web-site. At the beginning of 2005, this new service replaced the publication of the Bulletin on paper.
Subscribers to the ESPACE® BULLETIN series are also able to access an Internet-based version of the database.
ESPACE® BULLETIN Online is a fully-searchable edition of the European Patent Bulletin available on-line. A user-interface similar in appearance to that of MIMOSA facilitates access for those already familiar with other ESPACE® patent information products. Having found the information required, a user can print out the relevant section of the PDF file.
With the enhanced online functionality of the MIMOSA application software, it is planned to replace the present ESPACE® BULLETIN Online by the GTI V5 database used for CD/DVD-ROM production; users would then access the BULLETIN Online with the MIMOSA application software through the Internet. User registration and administration that could be used with other ESPACE® series offered for Internet access is being developed.

MIMOSA
MIMOSA Software covers the software package used by content providers to author ESPACE® series as well as the user application software used to access the series.
This software creates a common authoring software and corresponding user application interface for both mixed mode ESPACE® CD-ROMs and earlier “Patsoft” products. It greatly facilitates access to patent documentation and other intellectual property data publications (e.g. trademarks) thereby ensuring compatibility between products worldwide and supporting their use in documentation exchange.
The MIMOSA Authoring software is a UNIX-based data preparation chain suitable for both large-scale, mass production of CD/DVD-ROMs and creation of update files to be distributed through other channels.
In addition to patent information products of the Office, the MIMOSA software has been successfully implemented for the production of MIMOSA compatible intellectual property information products within the framework of EPO co-operation activities in a number of member states (AT, CH, LI, BE, NL, LU, DK, ES, FI, GR, IE, IT, MC, PT, SI, UK, PL, RO, BG, HU) and other patent offices, including Australia, the CIS, Canada, Cuba, Japan, Korea, Russia, Taiwan, United States. The software is also used by commercial data providers to prepare patent information products e.g. Bundesdruckerei, Micropatent.
The MIMOSA Authoring Software has already been in use for many years and through its widespread application become the de facto worldwide standard for the publication of patent information on optical media. With the MIMOSA Software being amended to cover, in the course of the coming years, bigger databases and volumes of records, publications in accordance with WIPO ST. 36, IPC 8 it will assist the BPA to meet its strategic objective of “publishing” the contents of applications according to technologies of the 21st century.
For certain projects the MIMOSA Authoring V3.2, which is based on GTI V.3 technology, has reached its limits with respect to index size, intermediate file size or number of records (e.g. ESPACE® ACCESS, ACCESS EP-B, BULLETIN and Globalpat). A number of the EPO’s ESPACE® production chains have already been converted to a newer database technology (GTI V.5); the remaining series will follow in the course of 2005.
The MIMOSA User Application software is a Windows based search-and-retrieval software which is distributed free of charge with the various ESPACE® series, on the ESPACE® Demo Disc and via the MIMOSA website. A beta-version of the latest release – labelled version 5 – was revealed at the EPIDOS Annual Conference 2004. It contains a number of enhancements over the previous versions, including the ability to display, print and export XML data according to WIPO ST. 36 and to access to GTI V5 databases across the Internet / Intranet. A final release version will be available in May 2005.
The remaining EPO ESPACE® production chains will be migrated to GTI V5. Following this conversion, it will be necessary to adapt the authoring software to the new standards ST. 36 and IPC 8 only once. Contractual maintenance can then be limited to only one version of the authoring software. A new version of the Mimosa Authoring Software based on GTI V5 will be compiled and distributed to users of the current authoring software. Windows and Linux versions of the new Authoring Software will be developed.
The user application software will be upgraded to handle IPC 8 in 2005. Other improvements will enable the MIMOSA application software to offer completely transparent data-base consultation on a range of media, as well as even bigger databases.
The Office’s Mission Statement calls for us to remain a global leader in the provision of patent information. MIMOSA Software and the ESPACE® products play a crucial role in this. They enable the Office to disseminate patent information products easily and cheaply to other patent offices, their libraries and the general public.
For thousands of users outside the Office, MIMOSA Software and the ESPACE® products are essential archiving, document delivery and search systems. They permit full-text searching in a local environment through a single multi-lingual interface with a very complete range of features.
The ongoing development of the MIMOSA Software ensures that products around the world remain compatible, which supports their use as documentation exchange media and thus safeguards investments made in past.
MIMOSA Website

Because the MIMOSA Authoring Software is used around the world (the Office is aware of users in Europe, Australia, the USA, Canada, Russia and Korea, among others), the MIMOSA home-page was developed to offer a very practical way for registered developers to upload configuration files for new MIMOSA based CD-ROM series they have created. They can then obtain verification for them and Jouve can make them available for distribution to MIMOSA users with future official releases of the MIMOSA user software. The latest version of the authoring software is always available for download from the home-page to registered developers too.

Through the MIMOSA software contractor the Office now offers a MIMOSA home-page. Users of the MIMOSA Application Software can download the latest updates from the home-page. There are currently around 50 updates available, including new configuration files for ESPACE® series and patches.
Update procedures, content management and the layout have been aligned to procedures being implemented for the EPO website in order to facilitate the transfer of the MIMOSA website to an EPO server. The new website also provides a forum for direct exchange of information between users and/or producers.
Update procedures, content management and the layout have been aligned to procedures being implemented for the EPO website in order to facilitate the transfer of the MIMOSA website to an EPO server. The new website also provides a forum for direct exchange of information between users and/or producers.
Advantages of the home-page are also to be found in the product registration offered, which enables the EPO to ensure proper version control and maintain a full product inventory.

European Publication Server

Following the decision of the President of the Office dated 22 December 2004 to discontinue the publication of the European patent documents on paper, the European Publication Server became the official medium for their publication of European patent applications (A-documents) and European patent specifications (B-documents) on 1 April 2005. Via a very simple interface, users can view or download the documents in various formats on the date of publication.

European patent documents can be accessed very simply via the application number, the publication number, the main classification and the date of publication. The documents can then either be displayed and processed in PDF format or downloaded in SGML (in the future XML) format. All types of documents are supported, including corrections.

Accessible via a secure connection to the public free of charge, the European Publication Server is a big step in the direction of a modern way of making public documents available, thereby promoting the "knowledge based society" in Europe. Last but not least, this service avoids the unnecessary use of paper.

The European Publication server is available under:

https://publications.european-patent-office.org
OPEN PATENT SERVICES
The Open Patent Services (OPS), offering access to family, legal status and bibliographic data, have been available in production since July 2003. Since then, the usage of OPS increased steadily. Currently, an average of 450 000 requests are made per day, serving well over 400 distinct hosts.

The number of clients developed by third parties since the introduction of the OPS is as well increasing constantly, although by its very nature many of these developments remain unknown to the EPO.

Following the growing demand from current OPS users and other interested parties, the EPO is planning to extend the OPS to include other data now accessible via esp@cenet®. The next extension is planned to be implemented until the beginning of the year 2006. This extension will allow the users to access the facsimile of the first page image, drawing images and full document images available already via esp@cenet®.

The successive developments planned for the years 2006 and 2007 will include the addition of bibliographic and textual data to the currently available fields and the possibility to perform a generic search based on selected criteria.

INPADOC Services

Traditionally the INPADOC databases collect together bibliographic, family and legal status data for patents from all around the world. In fact the bibliographic part of the database has been harmonised with the bibliographic master database of the EPO, DOCDB, and since October 2003 the latter one is the only physically existing bibliographic database of the EPO. Currently the legal status database covers 47 countries and patent issuing authorities while the bibliographic database, extended in January 2005 to include the patents from the Gulf Cooperation Council, currently covers 75 countries and patent issuing authorities.

The principal targets for the coming decade remain the extension of coverage, in particular that of the legal status database in order to include more countries of the Pacific rim and of Latin America, the improvement of the documentation and an easier, more understandable presentation of the database.

Document Consultation / Production Service

The Document Consultation / Production Service of Vienna has with a XEROX DT 6100 a highly sophisticated print facility at its disposal. Documents and brochures can be provided fast and efficiently. The data resource of the Document Consultation / Production Service is made available to support other tasks within the patent information area like training and support as described below.

The Document Consultation / Production Service has been upgraded using the Global Index Database (GID) which catalogues the contents of a DVD-network and provides a searchable document database. The GID is connected to the Web Order Module (WOM) which is developed for document delivery e.g. from esp@cenet® and provides access to documents from BNS.

The Document Consultation / Production Service supports as well the implementation of the New Search Tool (NST), a new development which provides fast and sophisticated search and retrieval access to the data stored in local and remote resources. The NST user interface for professional use is based on the JViewer / JSearch components of the EPOQUE system. Retrieved documents may be printed via the WOM/GID system.

The technical concept of the Document Consultation / Production Service has proven its flexibility, reliability, robustness and economic competitiveness for the production of copies of patent documents. Documents can be sent to clients within one hour by e-mail, fax or within 24 hours by normal mail.

This technical concept for the document delivery of patent documents have been established as standard document production systems in many national offices of member and non member states.

Furthermore, the resources of the Document Consultation / Production Service serve next to helpdesk, demonstration and training purposes also as a reference system for support and maintenance purposes for similar installations at national offices. In addition the document resource will be used to provide backlog collections on different high capacity storage media (discs, HDDs) for easy delivery and upload to national offices, patent information centres and end-users.

In summary the resources of the Document Consultation / Production Service serve for the different tasks / functions of PD4.5 as follows:
- Regular Document Production (Document Delivery on Demand)
- Demonstration and Training
o DVD network
o New search tool
o Document Delivery on Demand
o MIMOSA search and retrieval application software
o Global Index Database (library function)
- Reference system for support and maintenance purposes
- Backlog production on high capacity media
- Helpdesk

New Storage Solutions

In the framework of New Storage Solutions, storage solutions based on DVD technology have been implemented by the EPO in a number of national offices of member states (e.g. Greece, Portugal, Italy, Belgium, Liechtenstein, Finland, Sweden, Spain and Ireland) within the framework of the EPO co-operation programmes. EPO non member states have also installed such systems (e.g. Ukraine, Cuba, Brazil and Mexico), the work normally being done under the co-ordination of EPO and WIPO.

With New Storage Solutions the EPO follows the trend towards sharing costly peripheral devices such as disc storage, virtual discs on HDDs or high capacity disc media and multiple drives over a network. Integrated index databases provide for fast access.

Compatibility testing

The Compatibility of new devices in view of fast access to virtual disc storage and integrated index databases is investigated under this topic.

East Asian Patent Information Services

Over the past years, the EPO has seen an increase in enquiries about patent information services for Japan and the Far East. In particular, Japanese patent information has gained in importance through a broad range of English-language services on the Internet. At the same time, other countries in the Far East are gaining in importance for patent searchers, while data from these countries is still difficult to get by. The EPO's role in this field is to provide a European service to help users access patent information from the Far East and interpret this complex data.

In Vienna in 2004, with the assistance of the Japan Patent Office and the State Intellectual Property Office of the People's Republic of China and with the support from the Korean Intellectual Property Office, the EPO for the third time organised its meeting for European users of East Asian patent information, which again offered the opportunity to find out about developments in patent information products and services. In the meantime, this annual "Far East meets West in Vienna" meeting has developed into a true forum for discussions on Asian patent information, with special support from the guest speakers from East Asian countries and experts from the Chinese, the Japanese and the Korean patent office.

VII. Matters concerning mutual exchange of patent documentation and information

...

IX. Other relevant matters

...