Metadata definition at NLW
Most of the work relating to metadata at NLW is derived from its digitisation programme. In 2001 NLW adopted | METS as the metadata schema it would use to deliver the results of its work in digitisation on its website.
Gradually as the Library became more aware of the demands and challenges of digital preservation it also recognised the possibilities for the use of METS not only for its digitised materials, but also as a general framework for metadata for all of its digital material – both born-digital from outside producers and that digitised by the Library itself as a producer.
Since its adoption of METS the Library has followed its development closely and on the whole has tended to follow the lead set by the METS Editorial Board such as the adoption of the extension schemas endorsed by the board.
Where we are now
When the Library began to look at the possibilities of establishing a repository for its digital material in 2004-2005 it naturally saw METS as the standard for encoding metadata, and some important decisions were made at this time. One of these was to build a system which would reflect the OAIS Reference Model. As the Library acquired its new information management system at the same time, which included a digital repository component, these decisions were reviewed. The review did not result in great changes and so the Library's approach to metadata remains within a METS framework.
Preparation for ingest into the new digital repository has meant detailed work on the preparation of submission, archival and dissemination information packages and the Library's metadata definition consists of exemplar METS documents created for this work. This work was also informed by the OAIS Reference Model.
NLW metadata framework
Type of materials
Most of the work on metadata definition therefore has been in the context of material digitised by the Library as part of its digitisation programme. Therefore the main type of material we have been working with is digital still images. But we have been working with both simple and complex digital objects. The practices adopted to deal with these objects and materials will be adaptable to other types of materials.
Simple/Complex digital objects
Within the digital repository metadata for single digital objects (consisting of one file) are contained within in one METS document. Metadata for complex digital objects (consisting of more that one file) are contained in a series of parent/child METS documents enabling referencing of single parts of complex objects.
Most of the metadata are contained within the METS envelope, but the exception is the descriptive metadata. It was decided that the main description for digital material would be bibliographic or archival and held within the Library's main catalogue. From the bibliographic or archival records a digital holdings record will point to the object within the digital repository. Within the METS document the bibliographic or archival record as the descriptive metadata for the object represented by that METS document will be referenced
For identification purposes a very limited subset of elements will be included in the METS documents so that if any manual intervention is required. These elements will be encoded using MODS. They will be derived from the MARC21 encoded record.
The main type of material we are dealing with at the moment is digital still images and the standard we have adopted for metadata for this material is the NISO Technical Metadata for Digital Still Images encoded using the MIX schema. These metadata are encoded in the METS document.
At present the Library is working with the Schema for Rights Declaration developed by the METS Editorial Board. This will be a short to medium-term implementation while more comprehensive Rights Expression Language schemas are being developed. But we have developed rights profiles mainly based on the types of users enumerated in the schema. These metadata will be referenced by the METS document.
The bulk of the Library's digital material has been digitised from analogue originals and so the Library includes source metadata in its METS documents. This is usually in the form of reference to the MARC21 encoded record for the original material. Therefore the standard adopted for source metadata is the same as for the main descriptive metadata, that is AACR2 encoded in MARC21 for bibliographic descriptions and ISAD(G) encoded in MARC21 for archival descriptions.
Digital provenance metadata
The Library has adopted elements from the PREMIS metadata schema that are mainly Digital provenance metadata and specifically focus on preservation matters. It has not adopted PREMIS in its entirety as some of the elements included in the schema are encoded using other schemas.
Lyn Léwis Dafis 2007-06-14