With the arrival of new metadata models, vocabularies and tools, cataloguing in libraries will step into a more multidimensional metadata world in the upcoming years. Record-based cataloguing will be replaced with the linking of metadata, which means creating relationships between entities. Through these relationships, the metadata are linked to each other resulting in a metadata network extending across organisational, sector or national borders.
Creating a global metadata network calls for compatible metadata, in other words abandoning old system- and organisationrelated policies and practices. In addition to the Statement of International Cataloguing Principles, cataloguing rules (e.g. Resource Description and Access (RDA), International Standard Bibliographic Description (ISBD)), conceptual models (e.g. Functional Requirements for Bibliographic Record (FRBR), Functional Requirements for Authority Data (FRAD), Functional Requirements for Subject Authority Data (FRSAD)) and various standards, the use of common element sets and value vocabularies makes production of metadata more economic and improves the quality of metadata.
National and international metadata registries containing terms related to description elements (e.g. the language of the material) and the values of the elements (e.g. English) are essential in creating a metadata network. The international Open Metadata Registry (http://metadataregistry.org) defines the elements and their standardised values of the vocabularies, such as ISBD, RDA, Dublin Core (DC) and FRBR, influencing the descriptions for library materials.
Virtual Authority File
In addition to the vocabularies associated with element values, there is the VIAF (Virtual Authority File, http://viaf.org/) for names related to corporate bodies and individuals as well as the Finnish general ontology YSO (http://finto.fi/yso/en/) for concept and place entities.
It would be worthwhile to integrate these types of vocabularies into a metadata production system to optimise the description process. Moreover, using common description elements facilitates the construction of information search services (e.g. for the part of faceted classification and narrowing searches). Using common values of elements then again expands search results.
Identifiers bring functionality
Namespaces have been created for each of the central frames of reference that guide the production of metadata (e.g. RDA, ISBD, FRBR, http://finto.fi/yso/en/). The namespaces are included in the URI identifiers given to the terms for the vocabularies (Uniform Resource Identifier). For example, The URI identifier for the RDA term mediatype is http://rdvocab.info/termList/RDAMediaType, while the URI identifier for the ISBD term for mediatype is http://iflastandards.info/ns/isbd/elements/P1003.
In addition to the frame of reference, the URI identifier is language neutral as regards the description term, which is an important aspect in the global world of linked metadata. For example, the Finnish term mediatyyppi can be linked to the URI identifier for the RDA term mediatype.
Likewise, proper names in different languages can be collated under a common global URI. See for example http://viaf.org/viaf/89612684/. URI identifiers are machine readable, which supports the automatic linking of metadata and more wide-ranging search results.
Making metadata compatible
To make metadata compatible, the terms of different vocabularies must be mapped to each other. Mapping reduces alignment needed in the application profile. The relationships between the different terms can be expressed in the RDF schema (Resource Description Framework).
For example, when harmonizing ISBD and RDA terms, the references to FRBR entities are removed from the RDA terms, because the ISBD terms have not been defined in relation to the FRBR schema. The FRBR-neutral (unconstrained) RDA terms are presented as RDF relationships to the corresponding ISBD terms.
The differences in the global identifier systems for defining the boundaries of entities, such as the ISTC (International Standard Text Code), ISNI (International Standard Name Identifier) and ISSN (International Standard Serial Number) as well as the FR conceptual models (FRBR, FRAD and FRSAD), influence the compatibility of metadata.
New types of data models
When developing identifier systems, together with the publishing sector, for example, the perspective is broader than that of bibliographic metadata. The boundaries and identities of the entities should indeed be defined more clearly and universally so that the linking, which is reliant on the identifiers, may produce sensible descriptive element sets as search results.
The structural alteration of metadata calls for a new type of data model for descriptions. Descriptions that steer away from the current formats will rely on the functional conceptual model (FRBR) and the defined elements of the chosen vocabularies for descriptions.
The ontological structures of vocabularies and functional conceptual model explicitly determine the relationships between the entities. The hierarchy of the entities helps to maximise the use of common elements for several entities. The hierarchic structure is also related to descent, whereby the lower levels in the hierarchy are descendant to the relationships created higher in the hierarchy.
Data model for metadata
The FRBR conceptual model as type of frame of reference for description helps the producer of metadata to approach the material to be described from the perspective of the global metadata network. The FRBR model assists in producing structured metadata which is free from the current record structure.
In addition to a conceptual model, a data model for metadata, which would be appropriate for, e.g. a linked data environment, is needed. Compatibility with RDF is essential because RDF as a semantic web “grammar” and global metadata platform makes it possible to open the metadata on-line in a format which can be utilised by different applications. This metadata, mainly found in libraries, can be obtained through web search services.
Bound to record structure
Although creating a new data model for metadata is based on the objectives for content description and the entity-relationship model, consideration must also be given to the possibilities to convert the current form of metadata (e.g. data in MARC21 format) and its’ compatibility with the new data model.
Unfortunately the fields in the MARC format do not always represent independent description elements; rather, they are bound to the record structure. Configuring the MARC metadata into a more structural, easier linked format is a challenge that the international metadata network of experts willingly takes on to ensure the existence of libraries in the digital data environment.