SCHEMAS: Project Objectives

Introduction

The rise of the World-Wide Web has created an urgent need to define standard methods and vocabularies for describing its contents in a consistent and orderly manner. Since 1995, a number of related initiatives have arisen in what has been called a Metadata Movement. "Metadata" is a broad term that covers many types of "structured data about data" -- from traditional resources such as library catalogues, subject indexes, book reviews and abstracts, to new forms of technical and descriptive data for Web resources ranging from digital signatures and digitised map co-ordinates to online mail-order catalogues. In its most familiar form, metadata may be used to list the Author, Title and Subject of resources in a collection. Other types of metadata may list the price or quality rating of those resources, specify the format of their computer files, name the administrators responsible for their preparation, or clarify the terms and conditions under which they may be viewed or copied. Some of these metadata types are meant to be read by humans, while others are designed to be processed directly by computers. No single type of metadata can suit every application, every type of resource, and every community of users. Rather, the broad diversity of potential metadata needs can best be met by a multiplicity of separate but functionally focused metadata packages, or schemas.

Schemas that are designed to cover the basic needs of users in a wide range of fields and applications are called "core" element sets. In the Web environment, the most important of these is the Dublin Core Element Set -- a set of fifteen broad categories such as Title, Creator, Subject, Publisher, and Date. This element set was designed to be simple and intuitive enough for ordinary people to use without special training. As of November 1999, the definitions of the Dublin Core elements have been translated into twenty-five languages, making it a truly multilingual standard. Alongside the Dublin Core, efforts are underway to standardise element sets for purposes other than simple resource discovery, such as digital commerce and rights management, and for use with specific domains, such as instructional materials, museum objects, archives and government information.

With the rapid development and consolidation of these new metadata schemas, implementers of information services on the Web today are faced with a complex new challenge. Wherever metadata is needed, schemas must be designed to be as compatible as possible with existing and emerging standards. However, many of the standards needed are still under development and lack good documentation or user guides. Moreover, it is rare that all the requirements of a particular project or Web site can be met by an existing schema "straight from the box." The basic Dublin Core elements, for example, are often too generic for a particular project and must be refined with "qualifiers" (such as to distinguish Authors from Composers). Implementers often find that they need some core elements, perhaps with refinements, along with parts of domain-specific element sets. For special local needs that have not become the object of standardisation activities, implementers may need to invent new elements of their own. Ideally, implementers will design schemas that meet the local needs of a project or service with appropriate precision, yet follow existing standards and formats closely enough to allow their sites to be integrated with other resources on the Web ("interoperability").

Purpose

The purpose of SCHEMAS is to look at this diverse and often confusing landscape of new and emerging metadata standards from the viewpoint of project or service implementers who must use these standards to design their own interoperable schemas. We picture these schema designers as institutions and companies who want to publish on the Web or provide information services at a high level of access quality. This is precisely the audience that is of interest to developers of software tools for managing metadata and to the developers and maintainers of metadata standards. To delimit the audience in a practical and operational sense, we will focus on institutions and companies that are participating in projects under the IST Programme and in related initiatives on a national level in Europe. Many of these projects will need schemas that are customised but interoperable. Indeed, SCHEMAS could help improve access to European project results while contributing to the standardisation of metadata schemas for EU-sponsored projects in a more general and lasting sense. Such an effort would provide valuable input to the metadata-related standards activities of CEN/ISSS.

Project Goals

These general goals can be addressed with two key project objectives:

To guide and educate metadata schema implementers about the status and proper use of new and emerging metadata standards. Meeting this objective presupposes an understanding of the characteristics and requirements of our target audience. Training activities will take the form of bi-annual workshops, and annotated Web-based databases of ongoing initiatives, roadmaps of standardisation activities, user documentation, FAQs and schema registries. Using SCHEMAS materials, a reasonably experienced systems builder with no prior experience with metadata schemas should be able to attain a sophisticated understanding of the problem and guidance on possible solutions.
To promote good-practice guidelines for adapting multiple standards or metadata modules for local use in customised schemas. Efforts to standardise such modules inevitably overlap somewhat, and the consequences of this overlap for good schema design are still poorly understood. The implementation experiences tracked in the SCHEMAS knowledge base and discussed in SCHEMAS workshops will provide valuable feedback to standardisation activities in contexts such as CEN/ISSS and the Dublin Core Metadata Initiative.

Knowledge Base and Training Material

To meet these objectives, SCHEMAS will build a knowledge base with several components: Metadata Watch is a database of people, projects and programmes related to metadata schemas and implementation. This database will also cover existing tools, guidelines and documentation of relevance to the Web-based training materials and to the workshops. A network of correspondents in research, industry and not-for-profit organisations will aid the survey of existing projects and standardisation efforts. These correspondents will also be invited to contribute their knowledge and expertise from their own background. In addition, the database will cover the target audience of SCHEMAS: potential schema implementers among the partner organisations and companies of Fifth Framework projects. A related activity, Standards Watch, will focus specifically on mapping the scopes of various standardisation activities.

Registry

Both of these activities will feed into a Web site at UKOLN, the focus of which will be a database of metadata schemas per se – in other words, a Metadata Registry. The registry will hold links to elements and their definitions, in whatever form they may be available. These schemas will include both recognised standards and the less formally published schemas of specific projects -- whether they have been published as simple Web pages, RDF schemas or other types of databases, and whether they have been formatted for simple printout or following more structured conventions such as ISO 11179. It will cover schemas across a broad range of functional requirements, from resource discovery to rights management and digital preservation. In designing this registry, SCHEMAS will review best practice for managing registries within a distributed infrastructure. The registry itself will serve as a good-practice example of registry use and benefits; we plan to make configuration details of the registry available as a technology baseline for other implementers of registries.

RDF Registry

As a part of this more general registry, a Multilingual RDF Registry will focus specifically on the use of the Resource Description Framework (RDF), a new standard for supporting the exchange of metadata on the Web that is up for recommendation by the World-Wide Web Consortium (W3C). One aspect of RDF, called RDF Schemas, is a format for encoding metadata schemas using explicit Web links to related parent schemas. This capability is particularly useful, indeed crucial, for linking translations of schemas in multiple languages – hence the focus of this registry activity on multilinguality. The Dublin Core Element Set, which is already available in twenty-five languages, provides a convenient and manageable starting point for this registry. Data-model conventions for expressing Dublin Core schemas in RDF are being discussed and agreed in the Dublin Core Metadata Initiative. Our focus in SCHEMAS will be on explaining and documenting for the benefit of implementers how to use RDF schemas for linking locally designed schemas (in any language) to the standards on which they are based (in English).

Maintained by: UK Office for Library and Information Networking (UKOLN)
Last updated: 11 April 2000