Research into the usage of collection level descriptions is currently being done within ETB, the results of which will be presented in ETB-deliverable 3.1. This article aims to clarify the concept of collection level descriptions, and present examples of its application in a few relevant projects (RSLP, Renardus and Agora).
What is a collection level description?
Describing collections is nothing new, it has however recently become an increasingly topical and important issue. The reason being the growing amount of digital resources, the gathering of these resources into collections and the interest in the aggregation of these collections. In addition to the enhancement of information retrieval, the owner or curator of a collection can, with the creation of a collection level description, reveal information about its existence and availability to users in a standardised manner.
By using a machine-readable format that is structured and standardised, instead of the unstructured textual documents sometimes used when describing collections, the collection level description permits: 
- users to discover and locate collections of interest,
- users to perform searches across multiple collections in a controlled way,
- the refinement of distributed searching approaches based on the characteristics of candidate collections,
- software to perform such tasks on behalf of users, based on known user preferences.
It also simplifies production of a standalone catalogue of participants - this could feature general information but also information specific to the service in question, and where catalogues do not exist for collections, a collection level description may provide information for the remote user of content and coverage.
Even a metadata format originally constructed for usage on a resource level contains elements which are well suited for description on a collection level. All the elements in the Application Profiles  we have looked upon, consist of both home-made elements and elements taken from existing schemas. With additions and adjustments required to fit the specific environments, these elements ensure the users´ ability to find, identify, select and obtain relevant and correct information about the collections.
The project has developed a model of describing collections and their catalogues, which has formed the basis of a collection description metadata scheme. The implementation is done using RDF/XML. To ease the creation of CLDs a tool has been developed within the project, see RSLP Collection Description Tool. Element sets used in the application are derived from formats such as the Dublin Core, GILS and ROADS.
The UK Research Support Libraries Programme (RSLP) funded the RSLP Collection Description Project with the aim of enabling all projects in the programme to describe collections in a consistent and machine-readable way. The work undertaken by RSLP is not restricted to the description of digital collections, but should be applicable to various types of physical and digital collections. The RSLP collection description model is aimed in the first instance at those responsible for the development of collection descriptions, serving collection management purposes, particularly in discharging an institution's curatorial responsibilities.
RSLP collection description schema includes:
- descriptive attributes concerning the collection,
- descriptive attributes concerning the location (or locations),
- identification and/or description of three kinds of related agents, the collector and owner of the collection and the administrator of the location,
- external relationships - identification of collections and other resources that are related to the collection being described.
Renardus, a collaborative project that aims to improve academic users' access to a range of existing Internet-based information services across Europe, intends to use collection level description. Renardus will offer a distributed search and browse broker for European Academic Subject Gateways, allowing the integration of information from several suppliers into one single user interface.
The Renardus pilot broker service will be based on existing quality-controlled subject gateway services such as:
- DutchESS - Dutch Electronic Subject Service,
- SSG-FI - Special Subject Guides (SSG Fachinformation),
- DAINet - Deutsches Agrarinformationsnetz,
- RDN - Resource Discovery Network,
- Finnish Virtual Library/Virtuaalikirjasto,
- EELS - Engineering Electronic Library, Sweden,
- NOVAGate - Nordic Gateway to Information in Forestry, Veterinary and Agricultural Sciences,
- EULER - European Libraries and Electronic Resources in Mathematical Sciences.
The Renardus architecture will consist of a number of interoperable and simultaneously searchable databases. This requires the individual partners to extract, normalise and export information from their existing databases into a local Renardus database. The partners are also expected to create and offer administrative information i.e. a collection level description of their services. This information, human and machine readable, is intended to support subject selection and to simplify the production of, access to and integration of background information about participating Subject Gateways.
At present Renardus has defined an Application Profile that will be used by Renardus' partners to describe their services. The Renardus CLD schema is a simplification of the RSLP model where, with the exception of one field (DC:Relation i.e. an element which provides a reference to a related collection), all elements are mandatory. Information will be produced by the CLD-tool in RDF/XML syntax that has been adapted for use within the Renardus project.
Agora, one of five hybrid library projects that began in January 1998, is forming part of phase 3 of the e-Lib programme investigating issues surrounding the integration of digital and traditional library resources. Agora was in last year subject to a case study concerning, among other things, collection level descriptions. At the time of the study, there were 58 CLDs in Agora.
The study revealed mixed reactions to CLDs: 'It was acknowledged that CLDs were useful as a way of learning more about unfamiliar resources and in assessing their relevance for inclusion in different landscapes. However, there was a consensus that more work will need to be done to enrich the CLD content to provide more information about the different collections. Further enhancement and refinement of the controlled language to aid searching is also felt necessary with special mention being made of the lack of a developed subject schema.' 
In ETB CLDs should facilitate the discovery and organisation of collections on many levels: CLDs can be valuable for curators of repositories when exchanging metadata, for visitors to the ETB site in search for resources and also for the participating collections as a means when presenting themselves.
The element set will follow the semantics used by RSLP and Renardus use, with appropriate changes to this particular educational collection of collections, to ensure future interoperability.
 Powell, Andy, Heaney, Michael and Dempsey Lorcan, 'RSLP Collection Description', D-Lib Magazine, September 2000
 Baker, Thomas, 'First Step Strawman', ETB NEWS, October 2000
 Palmer, David and Robinson, Bridget, 'Agora: the hybrid library from a users perspective', Ariadne, issue 26, 2001
Annakim Eltén <http://www.lub.lu.se/~annakim/>
Jessica Lindholm <http://www.lub.lu.se/~jessica/>
Lund University Libraries
URL to this document: http://www.lub.lu.se/ETB/wp3/cld_intro.html.
Last update: 2001-01-29