On-line information resources for economics researchers:

RePEc database and RuPEc Web portal ^[1]

Thomas Krichel
Surrey University

V.M. Lyapunov, S.I. Parinov
Institute of Economics and OPP,
Siberian Branch of the Russian Academy of Sciences

Synopsis

The international community of economics researchers is currently implementing a unique project targeted at establishing a common decentralized database to store data about all basic publications on economic issues. Current achievements of the project include an operational database (RePEc) and a large number of services used to process user requests to the database. The services are developed independently of one another at various servers located in different countries. Among those servers one (RuPEc) is located in Russia. It hosts local versions of a number of international services. We started creating a Web portal for economic papers and began developing some other user interfaces.

1. Introduction

RePEc (Research Papers in Economics) is an international network of archives where electronic papers in economics are stored. Its history begins in 1993. By March 1999 RePEc comprised over 70 independent archives with over 13,000 downloadable full test files and over 50,000 document descriptions grouped into about a thousand series. Besides, the network includes data on over four thousand academic organizations involved in economic research. Thus, the network integrates several types of data (documents, published reports, software descriptions, personal data, and corporate data). RePEc's long-term goal is to set up a distributed on-line database covering all aspects of academic research in economics. The database will consist of interrelated electronic archives and databases belonging to different institutions.

In 1997 RuPEc – the Russian node of the RePEc network – was launched. It is a Russian on-line archive which offers Russian users open on-line access to the full-scale RePEc database with all its services. The RuPEc archive has a separate Russian language section with document descriptions in Russian. RuPEc enables Russian researchers and on-line economics archive administrators to add their archives and specific publications to a common database (RePEc) as well as to its Russian language section.

A group of Russian researchers develops modern on-line tools to visualize contents of document flows entering the RePEc database. These tools will eventually make up an economics papers Web portal with a customizable user interface capable of filtering the incoming document flow in line with specific user profiles.

Comprehensive international coverage, wide range of incoming documents, customization and filtering tools make RuPEc one of the potential leaders in providing services to researchers of economics.

In Section 2 below we describe the history of the project. Section 3 describes the most important features of RePEc. Section 4 is devoted to the Russian part of the project (RuPEc). Section 5 concludes the publication.

2. History of the project

RePEc is largely an extension of the NetEc project which was launched in February 1993 on the initiative of Thomas Krichel. The NetEc concept was to merge a number of projects to distribute results of economic research via the Internet. Economists have long ago been using a special system to exchange preprints (often called "working papers"). At the moment when NetEc was launched electronic working papers did not exist at all. At the initial stage of the project Fethy Mill from Universite de Montreal provided on-line access to bibliography data on 250 series of working papers. Some of the series contained data dating back to 1988. The first Gopher servers were capable of storing such data in a convenient search format. The Manchester Computing Centre provided disk space and processing time while Geoff Lane, who was their system administrator, developed an interface for WAIS requests to the database accessible through Gopher. This is the way the project began. However, the capabilities were seldom used because access to the database was provided via e-mail and a mailing list which included economic researchers from academic institutions. Thus, the first component of NetEc was the BibEc project related to distributing printed working papers.

In April 1993 the first electronic working paper was developed. It was the "MatGlass: a Matrix Glass for C++" paper by Chris Birchenhall. This isolated publication would have been lost among five thousand database titles, but for a separate WoPEc section to store downloadable electronic publications. Each publication of this type was represented by two lines on Gopher. The first line contained a name of the author. By selecting the first line a user started a bibliographic search (title; author(s); sometimes an abstract; JEL subject classifier developed in the Journal of Economic Literature, etc.). The second line contained the title of the publication. By selecting the second line a user could download full text of the publication.

The text (usually in PostScript format) was not necessarily located on the Manchester server. It could be hosted by any other Internet server. Although WoPEc started as a collection of documents (the first collection of the kind), its basic feature is that it became more a collection of metadata about papers than a collection of papers as such. This shift became clearly visible from September 1993 when the Economics Working Paper Archive (http://econwpa.wustl.edu/) was opened at a site belonging to Washington University, St. Louis. The site was positioned as a central archive for all working papers on economics. The idea was taken from the community of researchers dealing with high energy physics who had a XXX archive (http://xxx.lanl.gov/) storing all relevant publications on a central site. It was more difficult, however, to implement the idea in the economics research community. By 1999 we have several hundred Web servers and FTP servers with on-line working papers [2]. The inability to set up a central site might be explained by the fact that economists have much less respect for monopoly than physicists. Economists, more than anybody else, realize the risks and hazards that monopolies bring about. (in this particular case, they are cautious about the monopolistic position of a person who controls the hard disk with all the results of economic research). It looks as if the same reason prevented establishment of a centralized system in WoPEc-type archives where metadata about economics publications are stored.

From the very beginning the project included a distributed database protocol. This made it possible for many organizations to place their archives into WoPEc without any special coordination efforts. Each organization had to keep its publications at its own server while users had to send their requests to many servers. The protocol started working in 1994 on the basis of the whois++ system. In mid-1995 the database was converted (by Jose Manuel Barrueco Cruz and Thomas Krichel) into the Internet Anonymous Ftp Archive (IAFA) format. Meanwhile, the database retained its compatibility with whois++. The database entries were transformed into static Web pages and indexed by the WAIS system. Later, feeling the necessity to separate user services from the distributed database management system we eliminated whois++ from the system.

In the spring of 1995 in the USA (Washington University of St. Louis) BibEc and WoPEc wirror sites were established. In the fall of the same year a mirror site was set up in Japan (Hitotsubashi University, Tokyo). In 1996 the WoPEc project won a grant from the Joint Information Systems Committee (JISC) of the UK Higher Education Funding Councils. Jose Manuel Barrueco Cruz - who had managed the project on his own free will - became the official project manager. The JISC grant had the following goals: 1) expanding the number of documents in WoPEc (by that moment the archive contained over a thousand documents); 2) extending participation of British universities. The first target was reached with considerable success: by the middle of the next year 4500 documents were stored in the archive. However, the project failed to attract interest and participation from other British institutions. In 1996-1997 the project team got in touch with various organizations from a number of European countries (Royal Library of Sweden, Dutch DEGREE working paper publishing consortium). By that time participants of the project agreed to use a common data exchange protocol to simplify coordination in the common archive of working papers. Introduction of the common protocol has another goal as well, i.e. separation of data collections from data processing and data usage to process user requests.

At that moment we felt that we need a software demon who would trace changes made in the archive sections hosted by different servers and automatically upgrade the common archive on a regular basis. A draft protocol for such a demon was proposed by Т . Krichel and approved by the project participants in Guildford on May 12, 1997. The protocol – with minor modifications – is used today as a fundamental basis of RePEc.

3. The RePEc project

The RePEc project is essentially a product of activities of many distributed participants. Therefore, it escapes clear definitions.

The following three components of RePEc are commonly recognized:

• A collection of archives containing economic research data.
• Data stored in separate archives.
• Data about organizations and individuals who own separate archives or publications.

The projects does not have any formal management structure.

Aims of the RePEc project:

• provisioning a comprehensive description of economics in the basis of papers accessible via Internet ("library purpose");
• provisioning free access to Internet resources devoted to economics ("publishing purpose") [3].

In this case "free access" means publishing expences are largely paid by the provider rather than a user. RePEc does not intend to review publications. However, its databases can be used for reviewing.

RePEc is based on the following principles:

1. many archives;
2. one database;
3. many services.

А . Data collections in RePEc are decentralized. Each particular service provider opens access to his collection (or archive) on http server or FTP server which, as a minimum, contains bibliography of economic research publications and, sometimes, hosts the publications themselves. The archive structure is described in greater detail in Subsection 3.1 below.

В . A software demon connects different archives into a joint database. Both local and remote archives are merged in a separate site. If the site accumulates all archives registered in the RePEc database, it provides users with access to a full-scale version of that database. RePEc contents constitute an independent database which defines each bibliographical component. Basic logical structure of the connections model is described in Subsection 3.2.

С . Various sites containing copies of the RePEc database are used to build different sets of software services to process user requests. There is no official user interface in RePEc. Some examples are given in Subsection 3.3.

3.1 Structure of the archive

RePEc is based on the following:

1. IAFA-borrowed format called ReDIF (Research Documentation Information Format). ReDIF defines a number of templates describing various objects of the archive (documents, series, archives, etc.). Each template describes a set of authorized fields. Some of the filed names are mandatory. Others can be defined with a greater degree of freedom.
2. The Guildford protocol defines the rules of ReDIF storage in the archive. In principle, ReDIF can be retrieved from an archive without using Guildford. However, for convenience we shall use this kind of separation between the structure and the contents of archives in the examples below.

RePEc defines each archive with a simple identifier (handle). We can regard, as an example, the RePEc:sur archive which is stored at ftp://www.econ.surrey.ac.uk/pub/RePEc/sur. The root directory of the archive should contain two mandatory files. The first one (surarch.rdf) contains a filled ReDIF template for one archive.

Template-type: ReDIF-Archive 1.0

Name: University of Surrey Economics Department

Maintainer-Email: T. Krichel@surrey.ac.uk

Description: This archive provides research papers from the Department of Economics of the University of Surrey, in the U.K.

URL: ftp://www.econ.surrey.ac.uk/pub/RePEc/sur

Homepage: http://www.econ.surrey.ac.uk

Handle: RePEc:sur

This file contains basic information about the archive. The other mandatory file is surseri.rdf. It should contain one or more filled templates for series (an archive may consist of several series, each devoted to its own subject matter).

Template-Type: ReDIF-Series 1.0

Name: Surrey Economics Online Papers

Publisher-Name: University of Surrey, Department of Economics

Publisher-Homepage: http://www.econ.surrey.ac.uk

Maintainer-Name: Thomas Krichel

Maintainer-Email: T.Krichel@surrey.ac.uk

Handle: RePEc:sur:surrec

Documents for the RePEc:sur:surrec series are contained in the surrec directory. It can contain files of any type. Any file with a .rdf extension is considered as containing ReDIF formats. Let's take as an example one of such files: surrec/surrec9601.rdf [4]

Template-Type: ReDIF-Paper 1.0

Title: Dynamic Aspect of Growth and Fiscal Policy

Author-Name: Thomas Krichel

Author-Email: T.Krichel@surrey.ac.uk

Author-Name: Paul Levine

Author-Email: P.Levine@surrey.ac.uk

Author-WorkPlace-Name: University of Surrey

Classification-JEL: C61; E21; E23; E62; O41 File-URL: ftp://www.econ.surrey.ac.uk/pub/RePEc/sur/surrec/surrec9601.pdf

File-Format: application/pdf

Creation-Date: 199603

Revision-Date: 199711

Handle: RePEc:sur:surrec:9601

Two values (File-URL:, File-Format:) can be repeated as many times as needed to indicate different types of files with full text of the document. It is possible to add a File-Function: field to show status of the file for the given document. By the way, in the given example full text of the document is located within the directory structure of the archive. That it why it is mirrored together with bibliographic data. Since the document URL shows the complete path, a reference to the full text will be valid regardless of the server to which this information has been copied.

WE've got a central archive (RePEc:all) to which all ???arch.rdf and ???seri.rdf files are copied from all archives registered in RePEc. The central archive contains software which mirrors archive sites and supports reading procedures as well as template verification together with verification of RePEc's general documentation. The RePEc:all central archive is located at ftp://netec.mcc.ac.uk/pub/RePEc/all.

3.2. Link structure

From the RePEc's viewpoint, economics is a set of four mutually linked basic elements. The basic model can be described by the following table:

A "document" can be a preprint, a published article, a book, a piece of software, a data set, etc. The most widely spread types at the moment are preprints and published papers, but we already have examples of software pieces in the RePEc database.

A "collection" can be a set of documents assembled into one subject group. At present a collection usually consists of series of preprints and magazines with published articles. Each document, by the way, is from the very beginning part of a certain series. In principle, the collection concept can also be used for grouping reviewed articles (e.g., into a separate collection). However, it is possible to add one more additional field into a template to reflect a new status of the reviewed document.

At the moment of writing this article, personal information integrated in a document template. However, we try to specify personal data in a separate way. Soon it will look as follows:

Template-Type: ReDIF-Person 1.0

Name: Thomas Krichel

Email: T.Krichel@surrey.ac.uk

Author-Paper: RePEc:sur:surrec:9404

Author-Paper: RePEc:sur:surrec:9601

Homepage: http://gretel.econ.surrey.ac.uk

Handle: RePEc:per:1965-06-05:thomas_krichel

As a result, we shall be able to replace information about the author (as shown in the first example from the RePEc:sur:surrec:9601 document template) with the following:

Author-Name: Thomas Krichel Author-Person: RePEc:per:1965-06-05:thomas_krichel

Advantages of this system are obvious. It brings down system administration workloads. For example, when the author changes his telephone number an appropriate modification can be done in a single point within the system. RePEc service users will be able to find the author even if his personal data in the document title has become outdated.

In conclusion, an "organization" can be represented as a set of individuals (much like a collection which is a set of documents). When registering an author, his personal data will automatically be complemented with data about his organization if the latter is loaded in the RePEc database.

3.3. User request processing services

A key feature of RePEc is its inherent capability to support numerous services related to user requests. The negative side of this approach is a certain vagueness of the RePEc concept in comparison with an XXX-type archive where user data and services are merged together. However, our approach has one more benefit: a potential provider realizes that sending his data to RePEc means a simultaneous connection of the data to all user services which were developed independently and continue operating on different RePEc servers in different countries.

Below you can see a list of user services placed in the historical order:

1. BibEc at http://netec.mcc.ac.uk/BibEc.html – statis HTML pages with information about working documents available on paper only.
2. WoPEc at http://netec.mcc.ac.uk/WoPEc.html – static HTML pages for all working documents available in electronic format.

Both of the above databases use a common search mechanism. Three search options are available: a full text WAIS search; a field search based on mSQL; and a field search based on the ROADS system. Both databases are mirrored in the USA and Japan as part of the NetEc project.

3. EDIRC at http://ideas.uqam.ca/EDIRC – offers data and search tools to economics-related academic institutions and research centers all over the world. This service is also mirrored in the NetEc project servers.
4. IDEAS at http://ideas.uqam.ca/ – offers an Excite-type index of static HTML pages represented for all "document", "article", and "software" templates from the RePEc database. This site is one of the most popular user interfaces providing access to RePEc data.
5. NEP: New Economics Papers at http://netec.wustl.edu/NEP – is a set of reports on newly arrived documents stored in the RePEc database. Each report is reviewed by an expert in the given subject matter. The system defines a few dozen subject matters which enables users to choose only those new papers that are related to their sphere of interest. The reviewing experts are usually PhDs, students and young researchers who work on a voluntary basis.
6. INOMICS at http://www.inomics.com/query/search – offers an index of RePEc data as well as enables parallel search in indices of other Web pages related to economic research.

A concluding observation: a search server of the Z39.50 format for all documents stored in RePEc is available at dbiref.kub.nl:9997. The database is called repref. The attribute set is Bib-1. The entry syntax supports USmarc, SUTR.S, GRS-I (line tags only, Type 3 tags).

4. The RuPEc project

In 1997 the Russian Virtual Laboratory for Economists and Sociologists (RVLES) won an RGNF grant (See http://www.ieie.nsc.ru/). The project involved experts from three organizations located in Novosibirsk: IEOPP SO RAN, GPNTB SO RAN, and ISI SO RAN. One of the trends within the project was to establish conditionsfor developing electronic working paper archives among Russian research institutions dealing with economics and sociology. The subproject was titled RuPEc, because it was technically based upon RePEc standards and protocols. The initial goals of RuPEc were as follows:

1. bulding a Russian mirror of the full-scale RePEc database and its main services;
2. developing RuPEc's own search procedure for the RePEc database (SWISH-E - freely distributed text indexing software - was used);
3. setting up a Web interface for adding new documents to the RePEc database;
4. creating an independent Russian database of working papers with a Web interface to support remote loading of new documents.

Implementation of the last two points gave Russian researchers a choice between placing their papers in the common international database (this required, as a minimum, bibliographic data in English) or in the Russian section of the database only (in the latter case all data should be provided in Russian).

Form early 1998 all the four parts test operation of all the four components started at http://www.ieie.nsc.ru/r-archive/. An independent Russian database was called RAWPES (Russian Archive of Working Papers in Economics and Sociology). To avoid misunderstanding let's point out that RuPEc means a family of services designed to process user requests to the RePEc database located at the RVLES server (http://www.ieie.nsc.ru/) while the term "RAWPES" relates to the Russian language database and to the Russian language user services associated with that database.

The goals of the next stage of RuPEc development include the following:

1. popularization and promotion of RePEc concept and standards of data exchange between electronic archives of Russian economics research organizations in order to create a Russian language network of interlinked archives in line with the international RePEc network;
2. development of various services to process requests of Russian and international users of the mentioned databases.

In Subsections 4.1. and 4.2. below we describe capabilities of individual researchers and administrators of existing electronic archives in adding new data about papers and entire collections to the RuPEc database. Subsection 4.3. describes the concept of a user interface which looks like a Web portal and makes it possible for users to keep personal sections at the RuPEc server with personal profiles, selections from the database matching the user profile, customized methods of visualizing new entries in the database, etc.

4.1. The rules of adding new papers to archive

The paper adding procedure works in the following way. At http://www.ieie.nsc.ru/r-archive/add.html there is a form where one has to indicate a mandatory minimum of data about the document to be added. The form has an English part and a Russian part. If a user fills the Russian part only, information about the document remains in the RAWPES database and is not sent to RePEc. If the English part or both parts are filled, the information is added to both databases (in any case RePEc captures only the English part of the form). See current contents of RAWPES at http://www.ieie.nsc.ru/cgi-bin/ar-search.cgi. Part of the papers which were also sent to the international RePEc database can be seen at

http://www.ieie.nsc.ru/~rupec/data/noseconom.html.

We have to mention that the international database should contain paper descriptions in English while the bibliographic data can give a reference to full text of the paper in Russian. When filling an archive entry form, a user – among general bibliographic data – should indicate directory and file name of the paper on his local computer which will be automatically copied to the RAWPES server for loading into the database. If full text of the paper has already been published on the Web and there is no need to keep it on the server (because a user can simply indicate its URL), a special version of the form is used which can be seen at http://www.ieie.nsc.ru/r-archive/add-url.html.

By the middle of 1999 the procedure had the following peculiarities:

1. Only one method (mentioned above) was used to place full text papers on a server. However, limited bandwidth of Russian communication channels requires an off-line mode for that process.
2. Availability of a full text file is mandatory for all archived papers (to successfully complete the process of placing the paper on a server a user has to indicate a file name to be uploaded or a URL address if the paper is already on the Web).
3. A RAWPES administrator has the right to delete a document from the archive if he reasonably believes that thematic or other rules of the archive have been violated.
4. After being placed on the RVLES server, a file with the paper and its bibliographic data can be modified only by a RAWPES administrator.

In the near future the project is to expand capabilities of the archive in all areas mentioned above.

4.2. The rules of adding new collections to the archive

Many of the Russian economics research organizations (e.g., TsEMI, IEPPP and others) have their own electronic archives to store working papers. Some of the Russian economics magazines (e.g., EMM, The Problems of Forecasting, etc.) review Web publications and refer to their bibliographic data, academic institutions and publishing houses announce new books on the Web, etc. In most cases the collections (archives) contain descriptions of papers only in Russian which effectively prevents their immediate loading into RePEc. Availability of a large Russian-speaking community and some other national features require a purely Russian network of mutually linked Russian electronic archives and a common database to integrate metadata about their contents. Protocols and templates developed for the RePEc project could become a convenient methodical and technical basis for such a network.

In fact, the RAWPES server already has all the tools needed for organizing metadata about all electronic archives devoted to economics. To add a new collection to the database its administrator should perform the following actions:

1. Send a request to rupec@ieie.nsc.ru asking to attach a unique symbol identifier for the new archive to be placed in the common meta-database (for example, RAWPES has a nos identifier).
2. Having obtained an identifier, one should make files in accordance with the templates described in Subsection 3.1 above. The only addition to those templates is as follows: it is necessary to add a Charset: field and indicate there the type of Cyrillic codes used in the files, e.g. Charset: Windows-1251.
3. Create a server directory accessible via FTP or HTTP. Place ???arch.rdf and ???seri.rdf files in that directory. Instead of ??? put a symbol identifier obtained at Step 1.
4. In the URL field of ???arch.rdf it is necessary to indicate full Internet address of the directory where the remaining .rdf files are located with bibliographic information about the archived publications (there is a separate .rdf file for each publication).
5. To show that the archive is ready for addition to the meta-database one should send a message to rupec@ieie.nsc.ru indicating the URL directory created at Step 3.

When the RuPEc administrator receives the message the address indicated is added to the list used on a daily basis by the software demon which checks if the source archives have been changed and transfers the changes to the common database.

The user interface of this database is fully identical to the interface used by the RePEc international database located at the RVLES server.

Classification codes used to define subjects of archived papers are to be taken from the JEL classifier. This enables inclusion of Russian publications into a uniform subject classification common to all databases.

To simplify translation of large volumes of bibliographic data from user format to REDIF format a customizable converter has been developed. The converter can operate with data source files in local and remote communications modes. Using a user-prepared description of bibliographic data format structure (the description is based on a language with a PERL-type syntax), the converter can, for instance, with a user-defined regularity check the contents published at a user Web site, convert the captured bibliographic data into the REDIF format and send a message about the conversion results to the user via e-mail. This service radically simplifies the work of a collection administrator aimed at supporting the collection's parallel descriptions at one's own site (in any format) and in a meta-database (in the REDIF format).

4.3. Web portal of publications in economics and other trends of development

One of the trends for RuPEc services aimed at processing user requests is linked with developing more convenient means of visualizing RePEc database contents and especially new entries coming to the database. The problem of convenient visualization of new entries stems from gradual growth of the number of incoming papers. The incoming paper flow consists of daily collection of new publications in a great number of archives (over 60 archives by early 1999), regular (monthly or quarterly) publications in electronic archives, and of new archives added to the RePEc database on a nearly weekly basis.

Orientation in the flow of new papers and monitoring new publications in the subjects of interest require, on the one hand, convenient and, on the other hand, customizable tools to filter the incoming paper flow and to display the resulting data on a PC screen. This problem is not a new one. The concept used for its solution is called "a Web portal". The sires which demonstrate successful implementation of the Web portal concept are Excite (http://www.excite.com/), MyYahoo (http://my.yahoo.com/), InfoArt (http://www.infoart.ru/) and some others.

In this particular case, the problem of convenient visualization of incoming data flows on a computer screen has the following aspects:

1. compact reflection on a screen of basic categories of information contained in the RePEc database;
2. personal customization of the list of categories to remain on the screen; and definition of the database contents filtering rules through a predefined subset of categories.

As a result, all these capabilities create an individual user-defined Web portal for publications in economics customized to specific user interests. A software demon will regularly search the incoming data flow contents and build a Web page with references to all new papers matching the defined user profile.

The RePEc database contents makes it possible to select the following data categories which a user cab, fully or partially, include in or exclude from the list of data to be displayed:

1. A list of organizations (universities, research centers, publishing houses, etc.) whose collections of electronic data are included in the database. A user can include only part of those organizations in his display list.
2. A list of archives and collections devoted to a specific subject (each organization may have several archives or collections of this type). The list can be further divided by types of publications (working papers, electronic magazine publications, annotations of books, software, etc.). In his Web portal a user can leave handles pointing to electronic contents of certain magazines, certain collections of working papers in his Web portal, annotations of new books, etc.
3. Thematic sections of the JEL classifier. With the help of handles pointing to publications marked with specific JEL codes by the authors, one can obtain new publications in specific areas of knowledge.
4. Publications of leading researchers. In principle, names of the authors which a user would like to monitor on a regular basis can be selected from the complete list of authors published in RePEc. Having marked those names, a user will obtain references to their publications (if any) in his Web portal.
5. Key words. Having selected sets of key words, a user will obtain references to corresponding search results in his Web portal. A software demon will update search results after each modification of the database.

As soon as the user completes his Web portal, the system will be able to identify the profile of his scientific interests on the basis of user-defined parameters.

Such a profile provides a number of benefits:

1. consolidated displaying of the profile facilitates its modification by the user;
2. the system can compare profiles of different users and generate recommendations about setting up interest groups; about availability of publications on similar subjects which might interesting to the user, etc. (the structure of this kind of service has been developed in the "recommendation systems" concept).

The above capabilities of the Web portal are an extension of one of the user services (a "reader" service). The other trend is to develop tool that could support a user in the process of generating publications (a "writer" service).

In the framework of the "writer" service RePEc can, in principle, simplify the process of quoting. For this purpose it is necessary to use an algorithm which provides a reference (a direct tab transition) to the quoted text contained in the file with full text of the quoted document. If such quotation references are inserted in an electronic document, their selection would display the required section of the quoted text directly on a user screen. In fact, the electronic quotations algorithm should make it possible to create bookmark tags which should be external to the file with the quoted document (since it is impossible to insert the tag physically, it operates at a logical level). The algorithm creates an intermediate reference database. Selection of an electronic quotation reference in a document launches a procedure on a server which logically inserts a bookmark tag in the required place of the quoted document on the basis of a unique context defined n the process of the reference generation. A browser is offered a version of the quoted document with inserted logical bookmarks, and the browser scrolls the document on the screen to the required bookmark.

There is a similar algorithm for HTML files (see a description of Web page section comments at CritSuite http://crit.org/). An algorithm of this type can be developed for any file format.

Massive use of electronic quotations and quotation references in the RePEc database creates an opportunity to set up another user service, i.e. electronic quotation index. The electronic index can be obtained by direct counting of database-stored electronic quotations for each document contained in RePEc.

5. Conclusion

Broad international coverage, wide spectrum of incoming documents, availability of personalization and filtering tools make the RePEc database and RuPEc user interfaces an important working instrument for researchers in the filed of economics and sociology.

References:

[1]. Creation of the RePEc database was supported with a grant provided by the Joint Information Systems Committee of the UK Higher Education Funding Councils in the framework of the Electronic Library Programme (RePEc/WoPEc). Creation of RuPEc user services r elated to the RePEc database and creation of the Russian RAWPES archive was supported by the Russian State Research Fund (№ 96-02-12039в ). The authors are thankful to Mrs. Yevgenia Stupina for comments which helped us to improve this text.

[2]. Some of them are home pages of the authors, others have been established by academic institutions or research centers. In rare cases we found joint directories belonging to several organizations. For example, the US Federal Reserve has a Fed in Print directory which includes publications of all its territorial branches.

[3]. Note that sometimes these aims are in conflict with each other.

[4]. We omit the Abstract field to save space.