Introduction

This chapter introduces the metadata standard that forms the centre of the architecture on which this strategy is built. METS, which stands for Metadata Encoding and Transmission Standard, is an XML schema designed to provide a framework for packaging all of the metadata (and potentially data) for a complex digital object. This includes its descriptive, administrative and structural components. METS allows all of this to be combined in a single XML instance, each type clearly delineated by an unambiguously defined place in its structure. As its name implies, packaging all of the metadata associated with a digital object in this way allows for its easy transmission and sharing with others.

METS was designed and is maintained by an international body of experts from the digital library community and is based at the MARC Standards Office at the Library of Congress. For the most part it does not supply semantics for the metadata that it houses; instead, it allows instances conforming to other XML standards, technically known as extension schemas, to be either embedded within its structures or held externally and referenced from within them. Any XML metadata can be incorporated into a METS instance in this way, although some schemas are specifically recommended for this purpose; using these enhances the interoperable potential of a record. The only excep - tion to this, where METS itself provides a mechanism for recording metadata directly, is the type that records the internal structure of a digital object.

Why use METS?

METS has established itself as the predominant packaging schema within the digital library community, as might be expected from its provenance in the central bodies responsible for the definition of library metadata standards. For this reason it immediately fulfils Principle 7 (‘Use standards whenever possible’). More pertinent than this, however, is its potential for meeting the requirements of Principle 2, ‘Support the long-term preservation of the digital object’, because of its congruence with the OAIS model.

As was shown in Chapter 3, the core of OAIS lies in its three information packages, the SIP, AIP and DIP. These are intended to bring together data objects and their associated metadata in discrete packages with clear boundaries which can then be used for their submission to an archive, its storage and preservation there and its dissemination to users.