Access to Scientific Literature </td><td>on the WWW: the RePEc concept

Access to Scientific Literature on the WWW: the RePEc concept

Thomas Krichel


1: The purpose of RePEc

RePEc stands for "Research Papers in Economics". The word RePEc has three different meanings

  1. It is a collection of web or ftp archives that provide structured data (something like author: Karl Marx \\ Title: Das Kapital) about electronic and printed documents in economics. Each archive is contained in a subdirectory on the web or ftp server. Some archives also make the full text of the papers available. Others contain URLs for the papers from another site. Others have no option to downloads papers. The collection of the data on these archives make up the first aspect of RePEc.
  2. It is a group of people that provide these servers. This is a mixed group comprising academics, departmental computing support staff, librarians and other volunteers. By working in a coordinated way we can achieve a level of service that could not be reached by each of us working individually.
  3. The last aspect of RePEc is its rôle as naming authority. Essentially this means that each server, each series and each paper is identified by a unique handle. Each handle starts with the keyword RePEc.
The aim of RePEc is to construct a database about all aspects of research in economics. The RePEc aims to describe and identify paper, (working papers and journal articles), collections of papers (i.e. journals and working paper series), persons who act as authors of papers and editors of collection, and institutions (which are collections of persons just like the journals are collections of articles). By identifying each element, we can update the data easily. An example:
Template-Type: ReDIF-Paper 1.0 
Author-Name: Thomas Krichel 
Author-Handle: RePEc:per:05-06-1965:thomas_krichel 
Handle: RePEc:sur:surrec:9801
is part of a description of a paper authored by Thomas Krichel. Somewhere else in the dataset, we find
Template-Type: ReDIF-Person 1.0
Name: Thomas Krichel
Handle: RePEc:per:05-06-1965:thomas_krichel
This means that when the author Krichel moves from one place to another, all we need to update is his email address and his homepage in the second template, and the information will be updated in the description of the all papers that Thomas Krichel has written. Similarly, the handle tut(RePEc:sur:surrec:9801) will be a unique identifier for this particular paper. A wide deployment of this identifier will mean that automated citation analysis will become possible. Of course these particular features are not yet fully developed, but I hope that the example will give you an idea what RePEc wants to archive. Basically what we are aiming for is a relational database system for economics research information. This relational structure will be one distinctive advantage of RePEc over existing systems like the ISI Citation Index and EconLIT. The other advantage is that it is freely available to anyone in the world who is interested in Economics and has internet access.

Skeptics have suggested that such a system will ever emerge. It should however be noted that the majority of people that push RePEc forward are young(ish). We know that we have a long way to go, but we also know where we are going and we know that we have the time to move on. Up until now we have found working on RePEc an enjoyable and worthwhile experience.

2: A look at an archive

RePEc is founded on two sets of guidelines. "ReDIF" is a template metadata format. It describes templates for "Paper", "Article", "Series", "Archive" etc. The Guildford protocol is a convention on how to store the template files on a publicly accessible computer system using either ftp or http. The idea is that the data is not maintained by a single site, but by a number of archives all working independently. Let us show an example. The National Bureau of Economic Research (NBER) provides a RePEc archive at tut(RePEc:nbr) is the handle of the NBER archive. In the archive, they provide information about the NBER working papers, handle tut(RePEc:nbr:nberwo) at This directory contains plain ASCII files, that contain data that describe research documents. For example there is a template

Template-Type: ReDIF-Paper 1.0 
Title: Costs of Equity Capital and Model Mispricing 
Classification-JEL: G12; G31 
Author-Name: Lubos Pastor 
Number: 6490 
Creation-Date: 1998-04 
File-Format: application/pdf
File-Restriction: Access to the full text is restricted. Look up http://www.nber.
 org/wwpfaq.html for details, or write to If you have no access
 to the full text you will be shown an abstract page instead. Anyone browsing the 
 NBER working paper database from a site with a TLD in a non-OECD, non-OPEC
 country will be offered full text downloads for any paper for which the full text
 is available  online, but only if their DNS does reverse name lookup.
Handle: RePEc:nbr:nberwo:6490
Price: $5.00 per paper (plus $10.00 postage for orders outside U.S.)
This record also contains an abstract that has been suppressed here to conserve space. It illustrates that the collection is not necessarily limited to documents that are free. Since the access arrangements vary, the text of the restriction is free. If the resource is composed out of several files, then a construction like File-URL:, File-Format: and File-Restriction: is used for each file, and an extra field File-Function: may be added.

The most important feature of the ReDIF data is the handle structure. In our example, the handle tut(Handle: RePEc:nbr:nberwo:6490) contains both information about the archive, the series as well as an identifier for the paper. This hierarchical structure is evident in this case. In other situation that structure may be different. In fact RePEc is planning to build a handle structure for authors and institutions. These data could then be kept up to date. For example if the author moves, her new email address would only need to be changed at one place.

RePEc does not follow an existing or emerging internet standard. The advantage is that the RePEc metadata format it very flexible and can be altered as new needs arise. The metadata is simple enough that support staff working for departments and research institutions can create and amend ReDIF records with minimum difficulties. If the format would be more general it would also make it more difficult to encode.

3: Data and services

The RePEc data sits in archives, but that is not the form the user would normally access it. The user of the data can address one or more user services that render the data. There is no official user service for RePEc, but a variety of user services. They include IDEAS in Canada, WoPEc who operate in the UK, US and Japan, as well as the Russian RuPEc service. A Z39.50 service is being prepared in the Netherlands. There is also a current awareness service called "NEP: New Economics Papers". All these services use the same data from all archives. Each site that provides services regularly runs a piece of software that will update their local holdings of remote archives. The services all run independently but the providers of RePEc data will enjoy simultaneous exposure in all the services. This is a significant advantage.

4: Contact

Please mail any further question/suggestions.

href=>Thomas Krichel <>>