During the April 2003 Paris meeting of the committee for the development of databases on history of science sources Go to the main document, the possibility of building a distributed database using the Z39.50 protocol was discussed. This document provides additional information about this issue.

    The Z39.50 is a USA national and international (ISO 23950) standard defining a protocol for computer-to-computer information retrieval. It is not a database, and it is not a software. It is a protocol (a convention, that is, a set of agreed rules) which specifies data structures and interchange rules that allow a client machine (called an "origin" in the standard document) to search databases on a server machine (called a "target" in the standard) and retrieve records.

    Using the Z39.50 protocol, one computer can consult the databases of servers in several different places. It is possible to make individual searches, or to make simultaneous searches to different databases. This is only possible if
• the databases that are searched comply with the Z39.50 protocol; and
• the searcher computer uses a special software, called a “Z39.50 Client”.

    The formal home of the standard is the Z39.50 Maintenance Agency, hosted by the United States' Library of Congress:

    If several databases comply with the Z39.50 protocol, than it is possible to use them at the same time, building a “virtual database”, or “distributed database”. Using this approach, it is possible to maintain physically separate databases at several places (for instance: different countries) and to use them as if they were a single database, at a single place.

    To consult the databases, the user must either:
• install a Z39.50 client software in his/her computer, with suitable information about the databases he/she wants to use; or
• use an Internet site where there is a Z39.50 client connected to these databases.

    In the first case, each user must have the client software. This might be a problem, because there are several very expensive commercial products. In the second case, the user need have no special software (except for a common Web browser, such as Internet Explorer of Netscape). However, the server of the site where the Z39.50 client is housed must work for several users at the same time. This will lead to slow services. Besides that, if this intermediate site is down, for any reason, nobody will be able to consult the databases, although they may be fully operational.

    The best solution could be the use of individual Z39.50 client software (that is, each user will have the software installed in his/her computer), if a suitable public domain software can be used. We strongly suggest that the history of science database project should sponsor the development of a suitable software, and provide free copies to all historians.


    General information about the Z39.50 protocol can be found in these documents:

Z39.50: A Primer on the Protocol, authored by Leigh Watson Healy with augmentations from William E. Moen, May 2002.

Z39.50: Part 1 – An overview
Z39.50: Part 2 - Technical Details
By Peter Evans, in Biblio Tech Review (both parts updated April 1999)

National Information Standards Organization Z39.50 Information Retrieval Protocol (Z39.50/ISO 23950), with many links:

ANSI/NISO Z39.50-1992: American National Standard Developed by the National Information Standards Organization (Revision of ANSI/NISO Z39.50-1988), Approved July 28, 1992 by the American National Standards Institute


    Two successful examples of distributed library databases using the Z39.50 protocol are the Virtual Canadian Union Catalogue, and the California State University Unified Information Access System.

Virtual Canadian Union Catalogue (vCuc):

California State University Unified Information Access System

    The National Library of Canada provides additional information about distributed databases, and also a report on the Canadian project:
Distributed Library Databases:
Final Report:

    See also a technical discussion of distributed databases by Paul Miller, here:


    There is a nice (but outdated) general survey of Z39.50 client software at this address:

See also a comparison between three commercial products and one public domain software here:


    The best commercial client software is probably BookWhere. It can be easily used to find bibliographic records in hundreds of public databases worldwide. The problem is its price: US$ 395.00 (individual user).

Another commercial product is EndNote. This is a bibliographic software with several different tools, including a Z39.50 client. Price: US$ 240.00 (individual user)

ZSearcher is a Z39.50 client that provides Internet search and retrieval access to bibliographic databases available through any current Z39.50 server. Price: US$ 190.00 (individual user)

SLC PC Browser


    There are several free Z39.50 client products, but it seems that they have all been discontinued, and they are not as fine as the commercial ones.

Icone2, produced by Project One:

ZNavigator, produced by CaseLibrary:

Willow: the Washington Information Looker-upper Layered Over  Windows

Strategies for the Development of Databases Go to the main document
Roberto de Andrade Martins
Group of History, Theory of Science and Teaching

Document version 1, 23 April 2003