A review of metadata: a survey of current resource description formats
Work Package 3 of Telematics for Research project DESIRE (RE 1004)
Table of Contents
The Museum Computer Network (MCN) first proposed an initiative to investigate standards for interchange of museum information in 1988. The outcome of this effort was the CIMI Committee, which was active from 1990-92 and produced A Standards Framework for the Computer Interchange of Museum Information, by David Bearman and John Perkins (MCN, May 1993. Available from <URL:http://www.cni.org/pub/CIMI/www/framework.html>, and also published as a double issue of the MCN quarterly SPECTRA, vol. 20, no 2 and 3). This framework encompasses interchange protocols, interchange formats, and lower level network and telecommunications building blocks as well as content data standards.
The CIMI Consortium continues the work of the CIMI Committee that resulted in the Standards Framework <URL:http://www.cimi.org/>. The Consortium's current major project is CHIO (Cultural Heritage Information Online). The main aim of the project is to offer in a demonstrator at least 10,000 records of objects and information about Folk Art as a searchable online resource, including the full text of exhibition catalogues, wall texts, as well as images and more traditional museum database records. The CHIO project consists of two parts: CHIO Structure, which explores the use of SGML, and CHIO Access to explore the utility of Z39.50.
Museum communities. The CHIO project addresses information needs relating to cultural heritage information in a broad sense: information held by museums, archives and special collections in libraries. As Project CHIO aims to provide information to the average paying museum visitor, an over-academic approach has to be avoided.
The Standard Framework and especially the choice for SGML and Z39.50 seem to offer ample opportunities to reach a high degree of standardisation within the museum community, but much will depend on the results of CHIO and future projects. The advantage of SGML is that it offers a framework for the encoding of any type of document. Any tags can be used as long as a description of the tags that are used is given at the start of the document, as a DTD (Document Type Definition).
For future developments CIMI has decided that the CHIO DTD should be compatible with HyTime (ISO/IEC 10744), a standard for Interactive Open Hypermedia which is based on SGML and fully SGML-compatible.
At the conclusion of Project CHIO, CIMI expects to see a community endorsed system of encoding museum information using SGML; a methodology for searching texts and collections data using Z39.50; and a demonstration system that will show the power of a standards based approach to electronic interchange.
The museum/cultural heritage communities offer a broad range of information of different types and structures (e.g. structured texts, full-text documents, images). One of the aims of CHIO is to provide different DTDs for different types of information with different requirements, for the marking up of this information with SGML.
The first type of information to be analysed were exhibition catalogues (as a testcase for any text-based museum information resources). The following is based on the DTD for this type of information.
The CIMI Standard Framework can be implemented at two levels. The first is specification of hardware and software so that it supports the standards defined in the CIMI Standards Framework. This will ensure that the data can be interchanged even if all the institutional meanings cannot be. The second level adresses the standardization of data content (the fields of information), and data values (what goes in the fields).
CIMI is using SGML (and will eventually use other standards too like HyTime and Z39.50), to make museum information available.
SGML is used to express the structure and content of exhibition catalogues and as the foundation for a data interchange format for collections records. CIMI has developed DTDs along with stylesheets and a navigator which can be seen in action in the CHIO Demonstrator.
CIMI decided to develop a comprehensive set of museum DTDs, one for each genre of museum information, rather than one over-generalized DTD for all museum information. It was decided that the DTD should allow for all the significant features of the source document to be marked up. For exhibition catalogues and wall texts the TEI Lite DTD (a cut-down version of the TEI prose DTD) offered a good starting point. Based on the CIDOC Data Model the TEI Lite framework was modified and extended with access point tags to support the access needs of the CHIO project (this was done following the TEI Guidelines).
It was agreed that database records have different requirements and need a separate DTD.
The need for a robust linking mechanism to connect the parts of the distributed public information resource is acknowledged. It will depend on future developments relating to FPI (formal public identifiers). For more information on the proposals for the FPI resolution scheme see James K. Tauber, FPI-URN Resolution on the Internet: Notes <URL:http://entmp.org/fpi-urn>.
The foundation of the original MCN Standards Framework was the Open Systems Environment (OSE) Reference Model. CIMI addresses the interchange aspects of the OSE model that includes how data is represented, how various data types are identified, and how data content objects are presented.
Transport services can be provided by OSI or an appropriate alternative such as TCP/IP. CIMI uses high level OSI protocols such as FTAM for file transfer, X.400 and X.500 for messaging and directory services, and ISO 9040/41 for terminal access but these can be run over any appropriate lower layer transport protocols. CIMI recommends EDI for business transactions and ISO 10162/10163 for information retrieval. At first no consensus was reached on one standard for the building of collections databases and reference files, only rationales were given for using either ISO 2709 MARC, ISO 8879 SGML, or ISO 8824 ASN.1. But since then a positive choice has been made for ISO 8879 SGML.
The CHIO Structure project database can be accessed over the Internet using generic WWW browsers. CIMI is developing the CIMI Profile: Z39.50 Application Profile Specification which is a specification of what features of Z39.50 should be implemented in what way to provide the required functionality.This profile has now been issued as a draft for comment <URL:http://www.cni.org//pub/gils/profile/final.report/>.
The Profile is reaching a point of stability and has the consensus of the working group that has been working on it since September 1995. After the comment period, and necessary revisions, the revised version will become the initial version of the draft Profile upon which implementors will build support for search and retrieval in Project CHIO.
The CIMI Profile requires the following basic services of Z39.50-1995:
The CIMI Profile will include a CIMI Attribute Set, which will enable the expressions of queries for searching cultural heritage museum information resources. The Profile will use the Z39.50 Generic Record Syntax (GRS) for packaging retrieved records for presentation to the client.
The initial Z39.50 implementations of the CIMI Profile will support access to a demonstration CHIO Information Resource. The CHIO Information Resource can be modeled as a digital library comprised of hierarchical, distributed collections of digital information. The CHIO Information Resource consists of a number of physical and/or logical datastores of museum information, and the datastores may consist of one or more databases. A user may search the CHIO Information Resource to retrieve digital information objects in several possible data types (SGML, MARC, other structured records, Image, Audio, Video).
The CIMI Profile is being developed as a companion profile to the Z39.50 Profile for Access to Digital Collections (the Collections Profile) <URL:http://lcweb.loc.gov/Z3950/agency/profiles/collections.html>. The focus of the Collections Profile is the access and navigation of digital collections. The focus of the CIMI Profile is the search and retrieval of specific information resources contained in the digital collections.
The CIMI Z39.50 Profile which is being developed is intended to
support a demonstration project rather than working implementations.
A pragmatic approach has been agreed so that implementation experiences
will guide subsequent extensions, and existing implementations
that will serve up museum information for Project CHIO will be
the basis on which a number of Profile decisions will rest (e.g.
attribute sets, element sets, record syntax, etc.). During the
period Summer 1996 through Spring/Summer 1997, interoperability
testing will provide input to revisions and changes to the CIMI
Profile. By the end of the Project CHIO (Fall 1997), a final,
revised CIMI Profile will be completed.
Note that the CHIO demonstartor requires the use of an SGML browser such as Panorama, in addtion to a generic web browser, in order to view the SGML encoded documents.
|Next||Table of Contents|