A Daunting PREMIS: Implementing Preservation Metadata within the METS Framework

1
A Daunting PREMISImplementing Preservation
Metadata within the METS Framework

• Jerome P. McDonough
• Graduate School of Library Information Science,
  UIUC
• ICDAT 2006
• Inst. of Information Science, Academia Sinica
• October 19, 2006

2
One Great Loss for Mankind
3
One Great Loss for Mankind
Source Sarkissian, John M. (21 May 2006). The
Search for the Apollo 11 SSTV Tapes. Parkes,
Australia CSIRO Parkes Observatory. http//www.ho
neysucklecreek.net.nyud.net8080/Apollo_11/tapes/S
earch_for_SSTV_Tapes.pdf
4
Houston, weve had a problem here.

• Loss of data due to format conversions
• Need to insure viable access to playback devices
  for media
• Inadequacy of traditional archival practice for
  insuring item-level access to media
• Need to detailed event history to document
  life-cycle/provenance of information

5
History of PREMIS

• OCLC/RLG Preservation Metadata Framework Working
  Group (2001-2002)
• to define the concept of preservation
  metadataand evaluate the prospects for a
  community-wide, consensus-building activity.
• Final Report Preservation Metadata for Digital
  Objects A Review of the State of the Art
• to develop a framework outlining the types of
  information -- i.e., metadata -- that should be
  associated with an archived digital object.
• Final Report A Metadata Framework to Support the
  Preservation of Digital Objects

6
History of PREMIS

• PREservation Metadata Implementation Strategies
  PREMIS (2003-2005)
• Develop a core preservation metadata set,
  supported by a data dictionary, with broad
  applicability across the digital preservation
  community.
• Identify and evaluate alternative strategies for
  encoding, storing, and managing preservation
  metadata in digital preservation systems.
• Final Report Data Dictionary for Preservation
  Metadata Final Report of the PREMIS Working
  Group
• PREMIS Maintenance Activity at Library of
  Congress, including XML Schema

7
PREMIS Data Model
8
PREMIS Data Dictionary Object
An Object can be associated with one or more
Rights statements, can participate in one or more
Events, and can be related to one or more Agents

• Object Identifier
• Preservation Level
• Object Category
• Object Characteristics
• Creating Application
• Original Name
• Storage
• Environment
• Signature Information

• Relationship
• Linking Event Identifier
• Linking Intellectual Entity Identifier
• Linking Permission Statement Identifier

9
PREMIS Data Dictionary Event
An Event must be related to one or more objects,
and can be related to one or more Agents.

• Event Identifier
• Event Type
• Event Date Time
• Event Detail
• Event Outcome

• Linking Agent Identifier
• Linking Object Identifier

10
PREMIS Data Dictionary Agent
An Agent may hold or grant one or more rights,
may carry out, authorize, or compel one or more
events, and may create or act upon one or more
objects.

• Agent Identifier
• Agent Name
• Agent Type

11
PREMIS Data Dictionary Rights

• Permission Statement Identifier
• Granting Agreement
• Permission Granted

• Linking Object
• Granting Agent

12
ltpremisobjectgt ltpremisobjectIdentifiergt
ltpremisobjectIdentifierTypegthdllt/premisob
jectIdentifierTypegt ltpremisobjectIdenti
fierValuegtloc.music/gottlieb.09611lt/premisobjectI
dentifierValuegt lt/premisobjectIdentifiergt
ltpremisobjectCategorygtfilelt/premisobjectCateg
orygt ltpremisobjectCharacteristicsgt
ltpremisfixitygt
ltpremismessageDigestAlgorithmgtMD5lt/premismessage
DigestAlgorithmgt ltpremismessageDigest
gt36b0319..lt/premismessageDigestgt
ltpremismessageDigestOriginatorgtLocalDCMSlt/premis
messageDigestOriginatorgt
lt/premisfixitygt ltpremissizegt20800896lt/
premissizegt ltpremisformatgt
ltpremisformatDesignationgt
ltpremisformatNamegtimage/tifflt/premisformatNamegt
ltpremisformatVersiongtlt/premisformatVersion
gt lt/premisformatDesignationgt
lt/premisformatgt lt/premisobjectCharacte
risticsgt lt/premisobjectgt
13
ltpremiseventgt ltpremiseventIdentifiergt
ltpremiseventIdentifierTypegtLocalRepositorylt/
premiseventIdentifierTypegt
ltpremiseventIdentifierValuegte001lt/premiseventIde
ntifierValuegt lt/premiseventIdentifiergt
ltpremiseventTypegtingestionlt/premiseventTypegt
ltpremiseventDateTimegt2006-06-06T000000.001lt/p
remiseventDateTimegt ltpremislinkingAgentIden
tifiergt ltpremislinkingAgentIdentifierTy
pegtAgentIDlt/premislinkingAgentIdentifierTypegt
ltpremislinkingAgentIdentifierValuegtna12345
lt/premislinkingAgentIdentifierValuegt
lt/premislinkingAgentIdentifiergt lt/premiseventgt lt
premisagentgt ltpremisagentIdentifiergt
ltpremisagentIdentifierTypegtAgentIDlt/premisag
entIdentifierTypegt ltpremisagentIdentifi
erValuegtna12345lt/premisagentIdentifierValuegt
lt/premisagentIdentifiergt
ltpremisagentNamegtLC Repositorylt/premisagentNamegt
ltpremisagentTypegtorganizationlt/premisagent
Typegt lt/premisagentgt
14
Overview of METS

• Digital Library Federation Initiative launched in
  2001 as successor to Making of America II project
• Goal Create a single document format for
  encoding digital library objects which can
  fulfill roles of SIP, AIP and DIP within the OAIS
  reference model
• Scope limited to objects comprised of text,
  image, audio and video files (or combination
  thereof)
• METS Maintenance Activity at Library of Congress,
  including XML Schema

15
METS Framework
METS Document
Header
Admin. MD
Link Structure
Behaviors
Descriptive MD
File Section
Structural Map
16
METS Structure

• Object modeled as tree (e.g. movie is composed of
  scenes, which are composed of one or more shots)
• Every node in tree structure can be associated
  with content files and descriptive
  administrative metadata
• Every content file can be associated with
  descriptive administrative metadata

17
METS Administrative Metadata

• 4 Types Technical, Rights, Source Document,
  Digital Provenance
• Non-prescriptive/Multiple instances
• may be internal (XML or binary) or external
  (XLink) to METS document
• Internal XML reliant on extension schema (e.g.,
  PREMIS) for support

18
METS PREMIS
19
OAIS Information Package
20
On-going Issues

• Architecting objects for performance, or the
  Metadata that Ate Cincinnati
• Organizing successful complete representation
  networks
• Enabling trustworthy metadata
• Supporting non-generic Event, Rights Agent
  metadata
• Creating metrics methods for evaluating digital
  preservation activities

21
The Metadata That Ate Cincinnati

• Add a 300 page digitized book with TIFF page
  images, a TEI encoding and a METS wrapper to your
  repository
• 302 PREMIS Object Records, 302 Other Technical
  Metadata Records, 1 Descriptive Metadata Record,
  1 Rights Record, 1 PREMIS Event Record (Ingest),
  1 PREMIS Agent Record (Ingesting Agent), 302
  PREMIS Event Records (JHOVE Validation)
• Migrate TIFF to JPEG2000
• Add 300 PREMIS Event Records, 300 Additional
  Event Detail Records, 1 PREMIS Agent Record 300
  PREMIS Object records, 300 Technical Metadata
  Records
• Run Fixity Check on Content Files
• Add 302 PREMIS Event Records, 1 PREMIS Agent
  Record
• Continue ad infinitum.

22
Representation Networks
Partial (first layer) representation network for
Digital Cinema System Specification

• ISO/IEC 15444-1 2004/PDAM 1 (JPEG 2000
  Amendment 1/profiles for Dig. Cinema)
• SMPTE 384M (MXF)
• W3C XML 1.1
• SMPTE 372M
• EBU Standard N22 1997
• AES3-2003
• SMPTE 196E
• ISO/IEC 159482004 (PNG)

• Unicode version 4.0.01
• SMPTE 12M (auxiliary file format)
• SMPTE 336M (KLV)
• ISO 15706 (ISAN)
• SMPTE 330M-2004 (UMID)
• ITU-T Recommendation X.509
• ISO 3166 (language code)
• TIA-442 (RS-422)
• IEEE802.3

23
Trustworthy Metadata

• Metadata from a known (and trusted) source
• Metadata that has not experienced unauthorized
  change
• Metadata that is accurate
• Metadata that is sufficient to need
• Metadata that is transparent

24
Generic vs. SpecificEvents, Rights Agents

• Event Example -- Migrate SD DTV to HD DTV. You
  may want to know
• De-interlacing technique (motion-compensated or
  not, linear or non-linear)
• Colorspace conversion (gamma correction, luma
  equations for source and destination, primary
  chromaticities and white points for source and
  destination)
• Aspect ratio conversion technique
• Similarly, we may want to know more about Rights
  and Agents than the minimal generic information

25
Evaluating Digital Preservation Programs

• What does it mean to preserve digital content?
  Does the meaning of preservation vary with
  context?
• What metrics should we employ to evaluate the
  success of a digital preservation program?

26

• ??!

Jerome McDonough Graduation School of Library
Information Science University of Illinois at
Urbana-Champaign 501 E. Daniel Street,
MC-493 Champaign, IL 61820 jmcdonou_at_uiuc.edu