Web Advisory Document
PREMIS Data Dictionary for Preservation Metadata

 
Page 1 of 3


Repositories Support Project March 2009






PREMIS Data Dictionary for Preservation Metadata
Sarah Higgins – Digital Curation Centre

Introduction
The Repositories Support Project Briefing Paper on ‘Metadata’1 examines how metadata standards can be
used to make it easier to find, use and manage digital objects stored in an institutional repository. Long-term
preservation of these objects can also be handled using metadata standards. This Web Advisory Document
introduces The PREMIS Data Dictionary for Preservation Metadata, v2.02, the current authoritative metadata
standard for digital preservation, and looks at how it can be used within an Institutional Repository.


Using a repository for preservation
Institutional repositories, which aim to preserve their digital objects, will need managerial support to put in
place policies and guidance. Adequate preservation ensures that the Essential Characteristics – the things
which give the object meaning – survive both time and technical development. Preservation metadata can
help to maintain these Essential Characteristics:

• Viability – an object is able to persist over the long-term;
• Renderability – an object can be correctly interpreted and displayed using computing equipment;
• Understandability – an object’s purpose and context can be understood when rendered;
• Authenticity – an object has not been altered, adapted or substituted;
• Identity – an object can be clearly distinguished from other objects.

PREMIS
PREMIS was specifically developed to help with long-term preservation. It builds on the OAIS Reference
Model (ISO 14721)3, which defines a framework for a successful repository. PREMIS is maintained by the
Library of Congress4 and consists of 3 parts:

1. The PREMIS Data Model defines 5 preservation activities or Entities, and how they relate to each
other:

• Intellectual Entity – the digital object or the parts which make up a complete digital object, eg the
digitised pages of a book, or the complete set of files which make up a web page. An Intellectual
Entity may have more than one Representation (see below) eg a scholarly work may be
maintained in both PDF and Microsoft Word format.
• Objects – a discrete digital information unit which can be described in 3 sub-types:

Web Advisory Document
PREMIS Data Dictionary for Preservation Metadata

 
Page 2 of 3


Repositories Support Project March 2009
o Bitstream – the bit set embedded in a file;
o File – a named and ordered sequence of bytes known by an operating system;
o Representation – the file set needed to render a complete Intellectual Entity.
• Events – an audit trail concerning changes made to a digital object throughout its lifecycle.
Examples of Events are changes in custodianship, format migrations, and the creation of new
relationships between objects.
• Agents – persons, organisations or software responsible for preservation Events throughout a
digital object’s lifecycle.
• Rights – rights and permissions statements for both the digital objects and their Agents.

2. The PREMIS Data Dictionary defines elements and sub-elements (called Semantic Units and
Semantic Components) to describe all of the Entities. The exception is the Intellectual Entity (the
actual digital object). A repository can continue to describe their digital objects using their usual
metadata standard. PREMIS includes well over 100 elements, but only a selection are mandatory.

3. The PREMIS schema, version 2.05, allows the Entities and their Semantic Units to be expressed
consistently in XML.

Using PREMIS in a repository
An established repository will have already decided which metadata standards, and which elements of these
to use, to make it easier to find, use and manage digital objects, on a day-to-day basis.

PREMIS should be used in addition to existing metadata to manage preservation. A repository would need to
establish which of the elements included in the Data Dictionary are essential to ensure adequate long-term
preservation of its digital objects. There may be some cross-over between existing metadata creation and the
needs of PREMIS, reducing the amount of new metadata collection required for compliancy. Mandatory
PREMIS elements are already frequently collected by a repository eg: objectIdentifier is a unique identifier;
objectCategory is the type of object being preserved. Information already collected about those who manage
the digital objects, and the rights associated with them, may be sufficient to satisfy the Agents and Rights
Entities of PREMIS. PREMIS assumes auto generation of most of its elements. The Preserv Project is looking
at ways external service providers can help in this process6.

Many PREMIS elements mandate the use of controlled vocabularies. Few are actually specified, so relevant
existing ones may need to be identified, or new ones defined. Guidance on data content and structure will
also need to be developed, as this is not included in the standard.

Help is available through supporting documentation7 and the PREMIS Implementors’ Group (PIG)8. The latter
share experiences and feed them into the ongoing revision process. Their website includes: a wiki to share
documents (the pigpen); an implementation registry; and a listserv.

Conclusion
PREMIS is currently the best metadata standard for managing preservation within a repository. Using
PREMIS, in conjunction with appropriate metadata for description and access, can help to ensure that digital
objects remain available for the future. Implementing PREMIS can be a daunting task, but there is good
documentation, an XML schema and peer support available. Furthermore, many repositories may find that
they are already collecting information needed by PREMIS, making an implementation easier to achieve.

Web Advisory Document
PREMIS Data Dictionary for Preservation Metadata

 
Page 3 of 3


Repositories Support Project March 2009
References & further information
1 Repositories Support Project Briefing Paper on ‘Metadata’ (April 2008)
http://www.rsp.ac.uk/pubs/briefingpapers-docs/repoadmin-metadata.pdf

2 PREMIS Data Dictionary for Preservation Metadata, version 2.0 (March 2008)
http://www.loc.gov/standards/premis/v2/premis-2-0.pdf

3 ISO 14721:2003 Space data and information transfer systems — Open archival information system —
Reference model. A publicly available copy with identical text is available at:
http://public.ccsds.org/publications/archive/650x0b1.pdf

4 PREMIS: Preservation metadata maintenance activity
http://www.loc.gov/standards/premis/

5 PREMIS: Preservation metadata schema, version 2.0 (March 2008)
http://www.loc.gov/standards/premis/premis.xsd

6 Hitchcock, S et al., Preservation metadata for institutional repositories: applying PREMIS (draft) (2007)
http://preserv.eprints.org/papers/presmeta/presmeta-paper.html

7 Introduction and supporting materials from PREMIS data dictionary for preservation metadata, version 2.0
(March 2008)
http://www.loc.gov/standards/premis/v2/premis-report-2-0.pdf

8 PREMIS Implementors’ Group
http://www.loc.gov/standards/premis/pig.html

















Author:
Title:
Version:
Date:
Sarah Higgins
PREMIS Data Dictionary for Preservation Metadata
2.1
March 2009