Conceptual Data Modelling for Digital Preservation Planets and

Conceptual Data Modelling for Digital Preservation Planets and PREMIS Angela Dappert

PREMIS – Preservation Metadata: Implementation Strategies A de-facto standard, but still developing The PREMIS Data Dictionary defines preservation metadata q q … that supports long-term digital preservation … that most preservation repositories need to know … that is implementable … that is technically neutral

The PREMIS Data Model Data model Relationships between entities Semantic Units (“properties”)

Planets Data Dictionary q q A data model A specific vocabulary to describe concepts used across digital preservation processes Use to model organizations’ preservation policies Capture institutional preservation requirements Reuse and share requirements and vocabulary Informed digital object property ontology Develop machine-interpretable models (as added value)

Methodology q Top-down approach: Create a model § from first principles § from literature analysis q Bottom-up approaches (validate and simplify): § Document analysis § Interview decision makers § Planets work-packages extract concepts, vocabulary compile a requirements base for re-use categories of requirements

Planets Conceptual Model • Risk specifying • Preservation guiding • Significant Characteristics

Planets and PREMIS Models - Different scope PREMIS Planets

Planets and PREMIS Models q PREMIS: Preservation Risks and Requirements outside the scope – non-dynamic q Planets: Events, Agents and Rights are re-used from PREMIS Planets

Preservation Actions q Planets: Preservation Actions are a special case of PREMIS: Event that is richly modelled PREMIS Planets

Objects q Representations, Files / Bitstreams § PREMIS Bitstream : restricted to one file. § Planets Bitstream : sets of Bitstreams which can span several files q Components as subclasses of objects q PREMIS: file Planets: logical file and physical file q § logical file: expected checksum § physical file: actual checksum

Objects q q PREMIS: Intellectual Entities currently not fleshed out Planets: Intellectual Entities a subclass of Preservation Objects. PREMIS Planets

Environment q PREMIS: Environments subordinate to objects q Planets: Environments parallel concept to objects § Identify and describe environments § Model data carrier refresh, emulation as easily as migration PREMIS Planets

Properties q q PREMIS: specific properties that statically describe digital objects for preservation repositories Planets: rich, general property concept that dynamically describes the preservation environment for preservation processing PREMIS Planets

Properties Planets: § meta-level on which to describe the properties of Properties • value origins • data constraints • units • etc. § relationships to other Properties e. g. image. Aspect. Ratio = image. Width / image. Height Þ Property ontology Þ Resolve property clashes between preservation services and file formats

Properties § Planets: Supports dynamic preservation processes Use to represent characteristics and requirements Property and Vocabulary Description Properties Controlled Vocabulary Metadata Storage Service

Properties § Planets: Supports dynamic preservation processes

Significant Characteristics q q PREMIS: Value equivalence of a property Planets: Rich requirement / business rule with tolerance or importance factors, context under which it applies PREMIS: applies to and subordinate to one object Planets: expresses constraints on Environments or combinations of Environments and Preservation Objects. Primary entity PREMIS Planets

Planets and PREMIS Interoperability q q Next generation PREMIS is being informed by Planets. Priscilla Caplan (The Florida Center for Library Automation) and Angela Dappert (The British Library) have been asked by the PREMIS Editorial Committee to consider how the PREMIS model can benefit from concepts developed in Planets. They analyzed and documented the relationships between the Planets and PREMIS data dictionary. The PREMIS Editorial Committee is currently considering changes.

Planets and PREMIS Interoperability q PREMIS improves its understanding of its own scope. q Different scope makes complete alignment unnecessary. q Planets PP 2 data dictionary more granular than PREMIS Implementation flexibility and extensibility of PREMIS facilitates embedding Planets features.

Contributions of the Planets Model q Comprehensive model – everything you need to capture fits into the model. q Risks, requirements, and actions are first class objects within the model. q Different requirements categories play different roles in preservation planning q The model lines up actions against the risks they mitigate.

Comprehensive model Everything you need to capture fits into the model q q q full range of preservation processes technical as well as organizational properties full range of preservation actions full range of entities full range of organizational types

Thank you