HEIRPORT makes use of the Z39.50 communications protocol, which has been widely used in libraries and is an ISO standard that is designed to support communications between remote databases (Miller 1999; Z39.50 maintenance agency 2002). It does this by specifying rules that govern the way in which information is searched for and retrieved from databases. These rules allow searches to be performed on data that are held in different database management systems with differing information structures residing on servers at various geographic locations.
The Z39.50 protocol makes use of standard profiles that provide a common framework against which the information structure, search and retrieval functions of a target database can be mapped. Such profiles have been developed for use in different sectors and there are two profiles that are particularly relevant to historic environment datasets. The Bath profile was developed to provide the basis for communications between library catalogues, union catalogues and other electronic resource discovery services (Lunau et al. 2001) and is an ISO Internationally Registered Profile of Z39.50. The CIMI (Consortium for the Interchange of Museum Information) profile is being developed to provide the basis for searching cultural heritage information held in information systems maintained by museums, image banks and libraries. Z39.50 allows for the use of multiple profiles, thus enabling interoperability between sectors.
Both the Bath Profile and the CIMI profile have been implemented in HEIRPORT; this means that its target databases are mapped against both profiles. Although this may seem to complicate the process of establishing a target database there are real benefits. The Bath profile implements Dublin Core metadata (Dublin Core 2002), specifies use of XML (W3C 2002) and complies with the requirements of the e-Government Interoperability Framework (Office of the e-Envoy 2001; 2002a and 2002b). On the other hand, the CIMI profile has been developed to support searches of museum collection databases and supports the type of searches that are particularly useful to those who consult historic environment datasets. By specifying both profiles, HEIRPORT ensures that its target databases are interoperable with library databases and other information systems maintained by public sector bodies, as well as providing the who, what, when and where searches that are particularly valuable to its users.
Z39.50 is a well-specified solution to the problem of linking remote datasets in a flexible way, but it has been seen by developers as being difficult to implement (JISC 1999). As a result, projects seeking to implement Z39.50 portals are increasingly developing tool-kits to ease the process for web-developers, for example the Jafer tool-kit project (Jafer 2002) funded by the Joint Information Services Committee (JISC). The project to develop HEIRPORT also included the development of java applications (by the computing department of the University of Kent at Canterbury with the Archaeology Data Service), which are designed to support the Z39.50 communication functions between the portal and its target databases (Pinto et al. 2001). Those wishing to make their databases available through HEIRPORT in future will benefit from both this tool (now available from the Archaeology Data Service) and also by development work by System Simulation Ltd.
Z39.50 is one way of supporting technical interoperability between systems but there are other methods, for example the Open Archives Initiative protocol for metadata harvesting (OAI 2002 and UKOLN 2002). The OAI protocol developed as a way of setting up e-prints archives and supports the regular collection of metadata from a content provider by a service provider. It does this by using common Web standards (HTTP, XML and XML schemas) and established metadata standards (Dublin Core, IMS etc.). OAI is now being used in the Higher Education Community in the UK to create subject gateways and it may have potential for the development of HEIRPORT in future. Whichever method is used to support technical interoperability, web-portals also need to achieve semantic interoperability between databases to return useful sets of results to its users. In other words, searches of multiple databases return more useful results if they contain information that is conceptually similar. HEIRPORT is able to achieve semantic interoperability because it targets databases that are used to record the same sort of information about objects and structures in the historic environment and use similar terminology a search for a 'barrow' results in a list of burial mounds and not wheel-barrows. In this respect, HEIRPORT is benefiting from work to develop information content and terminology standards for historic environment databases (RCHME 1999a; RCHME 1999b; FISH 2002), which is not only increasing the consistency with which records about historic environments are being created but also how they can be searched and manipulated.
Communications protocols, standards and consistent data are all key to achieving technical interoperability. But the development of services like HEIRPORT is as much about people working together as it is about standards.
© Internet Archaeology
Last updated: Tue Feb 18 2003