Clever Recordkeeping Metadata Project Outcome Area 2

Recordkeeping metadata standards

Abstract modelling techniques need to be employed in the development of recordkeeping metadata standards to ensure semantic precision. Standards need to be expressed in robust canonical machine processable versions as well as in human-readable forms.

Brief Discussion

Australian recordkeeping metadata standards have lead the world as exemplars. Close and more sophisticated analysis in the research environment identified the inherent paper based and human centric bias of these standards, perhaps not surprising given their age and the speed of technology change. The failure to make canonical encodings in XML as part of the process of standardisation has led to the limited uptake of the recordkeeping metadata standards. At the same time, the research uncovered various inconsistencies, loose terminology and semantic imprecision in their published statements. Working within a records continuum framework, and seeking to leverage recordkeeping into business environments, the research clearly identified the limitations of a data models for recordkeeping metadata standards that cannot represent multiple recordkeeping entities and their relationships.

Summary Findings

  • Recordkeeping metadata standards are too human oriented, without robust canonical machine processable versions, and at the same time, they are not abstract enough.
  • Infrastructure issues in the technology to implement recordkeeping metadata in the service oriented architectures highlight problems with our industry metadata standards. We require:-
    • Canonical and authoritative encodings of recordkeeping metadata schemas in representational languages (e.g. XML Schema).
    • Encoding schemes within elements/element refinements also having canonical and authoritative encodings in representation languages.
  • The entity and relationship models for recordkeeping metadata schemas need to address the complexity of implementation, in order to support the re-use of metadata.
  • Technological issues, including obsolescence need to be taken into account in the construction of metadata schemas. For example, we were constrained in building interoperability into the prototype because a number of tools we needed to employ were written pre XML (e.g. X500). We need the capacity to render conceptual and standard metadata in the most up to date/applicable technologies. This is a tension, because at the same time we must guard against being constrained by the technology to write the standards in ways that privilege or only deal with the constraints of the prevailing technology.
  • Some semantic expression in our core conceptual recordkeeping metadata articulation require greater clarity in definition – for example, the relationship of business transaction (object) and records transactions (object); or the distinction between functional classification and title, in some schemes. Rigorous models are needed to assure semantic precision.
  • Logic all points to the implementation necessity of relationship as a recordkeeping entity in its own right, but the metadata associated with this entity is quite different from that of the other entities and it has yet to be defined clearly.
  • Recordkeeping processes need better definition and greater standardisation. This becomes a critically important encoding scheme for event history.