CNI is an organisation with a wide representation - its membership is interdisciplinary, international and cross- sectoral. The cross-sectoral dimension was evident in the number of museums and museum consortia attending and participating in the meeting. CIMI, the Consortium for the Interchange of Museum Information, has been developing a Z39.50 profile for the search and retrieval of museum information and over the past year implementation of the profile has begun in several countries [17].
CIMI is a group of organisations which have been working together since 1990, studying the potential use and interchange of museum information. The group is supported by membership fees required from each participating organisation and the aim is to pursue an agenda of work in networked information that no one member could achieve alone: the strength in numbers philosophy. The approach has successfully provided a persistent community focus for efforts which, it is hoped, will foster museum leadership in standardisation activities.
The CIMI programme is based on testbed work. Most of the projects are short-term, much like a great deal of JISC's current development work which can be an effective way of experimentation and exploration. It is hoped that the continuity and focus which can be offered through the CIMI consortium will help to transcend these short-term initiatives and provide long-term benefits.
Current Z39.50 work in CIMI included building test databases from a number of museums and a Java application to facilitate searching of multiple databases including museum, bibliographic and image resources. The work is building on the CHIO project (Cultural Heritage Information Online) and over the past six months has been running an extensive testbed of the CIMI Z39.50 profile. Organisations involved in the testing include:
Alternative participants include representative companies from Taiwan and Canada.
Evaluation results from the participating sites were not available at the time of the CNI session, but Bill Moen, the project manager, felt that participants had been surprised by the amount of work required of testbed participants to effectively implement and test the CIMI profile. However, he also pointed out that the participants seemed convinced of the benefits and were therefore willing to put in hours over and above what was allocated in project planning. The testbed is an iterative process however, and participants seemed to be given a definite sense of ownership of the project. Early on it became apparent that the profile was unrealistic in that it expected participants to provide data coded in SGML. The profile was thus scoped back to a level with which the implementers could more easily cope.
Search and retrieval within the system focuses on three types of cultural heritage information:
The key accomplishment has been the use of Z39.50 to search and retrieve images as well as textual information. The Java-based client allowed for multiple searching across multiple servers simultaneously. The client was developed by a US company called Blue Angels Technology. At present the client is intended for in-house curatorial purposes and not for endusers.
When the evaluation from the testbed is complete, CIMI will revise the profile based on the experiences of the participants and provide implementation guidance. This guidance is viewed as critical for the museum world where many management systems are 'in-house' and not provided by one of several large systems vendors as is the case for libraries. This is an interesting difference between the museum and library worlds and in fact may effect the speed with which Z39.50 is implemented. CIMI work has suggested that implementation of Z39.50 requires considerable local work and resources - coupled with a knowledge of the local systems. It may be an advantage to have an in-house system where implementors know and fully understand the system they use. It may be the case that implementation in libraries is hindered by local lack of experience and/or intricate knowledge of the commercial system in use. Work on the Virtual Canadian Union Catalogue project [18], based at the National Library of Canada, suggests that sometimes vendors do not sufficiently understand or effectively communicate the intricacies of their systems. If changes are required in order to implement Z39.50, it will be much easier (and faster) to make changes to an in-house system than to a commercial product.
CIMI currently have an application with the National Endowment for the Humanities for resource discovery work using the Digital Collections profile.
In order to give an update on Dublin Core development, Cliff Lynch gave an overview on the Dublin Core meeting in Helsinki which was held at the beginning of October. He described the DC workshops succinctly as an interesting concurrent development of a solution and an understanding of the problem for which the solution is intended. An interesting perspective and one which may account in part for the seemingly painstaking and lengthy efforts devoted to the development of what will (hopefully!) be a very powerful standard. Paul Miller and Tony Gill represented the JISC at that meeting and have provided an excellent report which I will not attempt to reiterate here [19].
Following the update, Cliff was asked how he saw DC being used in the future in a research library context. He said DC would be effective in a number of ways including:
Material on this page is copyright Ariadne/original authors. This article last updated/links checked on 28-Jul-1998