The OAI Object ReUse and Exchange ORE Initiative

1
The OAI Object Re-Use and Exchange (ORE)
Initiative
Jane Hunter, The University of Qld
2
OAI-ORE Background

• New initiative conducted under the umbrella of
  the OAI
• Supported by
• the Andrew W. Mellon Foundation
• Coalition for Networked Information
• Digital Library Federation
• the National Science Foundation
• October 2006 -September 2008
• http//www.openarchives.org/ore/

3
Organizational Structure

• Coordinators
• Carl Lagoze, Cornell University
• Herbert van De Sompel, Los Alamos National Lab
• ORE International Advisory Committee
• Universities, UKOLN, CNI, HPL, NSF, OCLC,
  Eduworks
• ORE Technical Committee
• Universities, Ingenta, Eduserv, Nature, NCSA,
  OCLC
• ORE Liaison Group
• JISC, DEST, Microsoft, DLF, DARE, DCMI, Google,
  DSpace

4
Objectives of OAI-ORE

• Develop standardized, interoperable,
  machine-readable mechanisms to express compound
  objects on the web
• Enable more effective and consistent ways
• to support the creation, management and
  dissemination of these objects
• to facilitate the discovery of these objects,
• to reference (link to) these objects (and parts
  thereof),
• to provide access to different representations of
  these objects,
• to aggregate and disaggregate these objects,
• to enable their re-use by repositories, agents,
  and services beyond the bounds of the holding
  repository
• Provide foundation for value-adding services
  (eResearch)

5
Why?

• OAI-PMH Protocol for Metadata Harvesting
• Exchange of (DC) metadata about simple atomic
  objects
• eResearch ?
• new forms of intellectual products requiring
  curation, management, publication
• New forms of multi-component learning objects
• Complex, multi-part, multi-media, distributed,
  service-oriented

6
Example of an eScience Compound Object
How to create, describe, publish, discover, re-use
?
External Database
Metadata
image_of
analysis_of
Title Creator Description Type Discipline Date.Pub
lished License
Average LE 1/T exp (A B/T)
derived_from
graph_of
refers_to
refers_to
refers_to
Slattery, O., Lu, R., Zheng, J., Byers, F., Tang,
X. "Stability Comparison of Recordable Optical
Discs- A study of error rates in harsh
conditions," Journal of Research of the NIST,
109, 517-524, 2004
Each component has software, OS, hardware
dependencies interdependencies
7
Compound Digital Objects

• Digital content with multiple components of
  variable
• Semantic type or genre
• Article, web page, documentary, photo, dataset,
  music recording
• Media type
• Text, image, video, audio, 3D, numerical, mixed
• Media format (PDF, XML, MPEG-1, AVI, SMIL)
• Network locations
• Institutional repositories, databases, web
  portals
• Fedora, DSpace, ePrints, SRB
• Relationships between components
• Lineage, versions, derivation, is_part_of

8
Examples

• An unordered set or bag of images from different
  web sites
• A multi-page HTML document where pages are linked
  via hyperlinks to next page etc.
• Overlay journal issue that aggregates separate
  publications, each in their own repository, in an
  issue
• Museum object page a splash page that links
  contextual text to multiple images, audio, video

9
Simple Compound Object
http//arXiv.org/astro-ph/061175/

• Set of unrelated objects in same, local
  repository
• Group together through containment node
• Only relationship is_part_of
• Attach metadata to containment node

Identifier
URI DOI PURL
10
Complex Compound Object

• OAI/ORE Named Graphs/Resource Maps
• Define set of components

• Relationships between components
• Relationships to external components
• Metadata attached to compound object
• Different views of the compound object

PDF
cites
is_derived_from
PS
HTML
hasRepresentation
MP3
View1.html
hasRepresentation
View2.smil
11
Components Distributed Across Repositories
http//arXiv.org/astro-ph/061175/
Identifier
DSpace
Fedora
SRB
PS
MP3
HTML
PDF
12
OAI-ORE model

• Named graphs
• - Represented by Resource Maps (ReMs)
• - graph of RDF assertions
• identified by a URI
• aggregation of resources (plus their types and
  relationships)
• multiple serializations

13
composite_object.ng

• Resources have URIs
• Nodes are typed
• Arcs/Links are typed
• Aggregation has a URI

14
Compound Object modeled according to ORE model
derived_from
15
Summary

• Enrich web graphs/hyperlinks by
• Defining a boundary for each compound object
• Adding meanings to links through typed
  relationships
• Decision use named graphs/resource maps
• Publish compound objects as ReMs
• A single Resource Map can have multiple
  serializations

16
Resource Map for Compound Object

• Resource Map representation of a named graph
• Resource Map must express the full set of
  resources that are part of a compound object.
• Resource Maps may also
• express the relationships between the resources
  referenced by the named graph.
• express the types of the relationships between
  the resources referenced by the named graph, i.e.
  label the arcs.
• express resources that are not part of a compound
  object but are referenced by the named graph that
  corresponds to the compound object. Distinguish
  between those resources that are part of the
  compound object and those that are external.
• express other information related to the named
  graph and to the resources that it references
  such as metadata, etc.

17
Different Serializations of ORE Models

• RDF Resource Description Framework
• ATOM Syndication Format (standardized RSS)
• YADS http//nurture.nature.com/yads
• Simple RDF data model for describing collections
• TriX way of expressing RDF triples in XML

18
Comparison of Serializations

• Bag of resources
• Sequence of resources
• Semantic relationships between nodes
• External relationships and alternate
  representations

19
Example 1 Bag of objects
20

• ltTriX xmlns"http//www.w3.org/2004/03/trix/trix-1
  /"gt
• ltgraphgt
• lturigthttp//example.org/graph1lt/urigt
• lttriplegt
•   lturigthttp//example.org/FrogCollectionlt/urigt
•   lturigthttp//example.org/relations/has_partlt/urigt
  
•   lturigthttp///frog_photo1.jpglt/urigt
• lt/triplegt
• lttriplegt
•   lturigthttp//example.org/FrogCollectionlt/urigt
•   lturigthttp//example.org/relations/has_partlt/urigt
  
•   lturigthttp///frog_photo2.jpglt/urigt
• lt/triplegt
• lttriplegt
•   lturigthttp//example.org/FrogCollectionlt/urigt
•   lturigthttp//example.org/relations/has_partlt/urigt
  
•   lturigthttp///frog_photo3.jpglt/urigt
• lt/triplegt
•  

21
Example 2 sequence of objects
PNG
JPG
GIF
22

• ltTriX xmlns"http//www.w3.org/2004/03/trix/trix-1
  /"gt
• ltgraphgt
• lturigthttp//example.org/graph1lt/urigt
• lttriplegt
•   lturigthttp//example.org/FrogCollectionlt/urigt
•   lturigthttp//example.org/relations/has_partlt/urigt
  
•   lturigthttp///frog_photo1.pnglt/urigt
• lt/triplegt
• .
• .
• .
•  lttriplegt
•   lturigthttp//example.org/frog_photo2.jpglt/urigt
•   lturigthttp//example.org/relations/is_derived_fro
  mlt/urigt
•   lturigthttp///frog_photo1.pnglt/urigt
• lt/triplegt
• lttriplegt
•   lturigthttp//example.org/frog_photo3.giflt/urigt
•   lturigthttp//example.org/relations/is_derived_fro
  mlt/urigt

23
Example 3 alternative representations
24

• ltTriX xmlns"http//www.w3.org/2004/03/trix/trix-1
  /"gt
• ltgraphgt
• lturigthttp//example.org/graph1lt/urigt
• lttriplegt
•   lturigthttp//example.org/FrogCollectionlt/urigt
•   lturigthttp//example.org/relations/has_partlt/urigt
  
•   lturigthttp///frog_photo1.jpglt/urigt
• lt/triplegt
• .
• .
• .
•  lttriplegt
•   lturigthttp//example.org/frog_photo3.giflt/urigt
•   lturigthttp//example.org/relations/is_derived_fro
  mlt/urigt
•   lturigthttp///frog_photo2.jpglt/urigt
• lt/triplegt
• lttriplegt
• lturigthttp//example.org/graph1lt/urigt

25
Controlled Vocabularies

• Types of links (arcs) between resources (nodes)
• Is_part_of, is_derived_from, references
• Properties of resources (denoted by nodes in a
  named graph)
• semantic type chapter, photo, song
• media type text, image, audio
• media format PDF, JPEG, MP3
• ORE -gt Top-level minimum to bootstrap
  implementation and adoption
• Communities define own extensions/subclasses

26
Problematic Issues

• Do compound objects only contain information
  objects or other abstract objects (e.g.,
  processing events?)
• Do named graphs only reference resources that are
  part of the compound object, or also other,
  related ones? How to define internal/external
  nodes? Is boundary constrained by type of
  relationship?
• Do named graph resources always have a
  containment node?
• How should the containment node be named?
• How to reference, the named graph, the compound
  object, a component of the compound object?
• For a given resource, how to find all compound
  objects to which it belongs?

27
Compound Object Authoring
28
SCOPE Architecture
29
Link Provenance Trails with Web Objects
30
Metadata for the current digital object
31
(No Transcript)
32
(No Transcript)
33
ORE Services

• URI that represents an ORE compound object is the
  access point for service requests
• Metadata associated with named graph author,
  date.creation
• Classes of services
• Publish submit a named graph to a repository
  (add metadata)
• Harvest a request for a batch of instances that
  correspond to the ORE model from a repository
• Discovery find instances that correspond to the
  ORE model
• Obtain A request for an instance that
  corresponds to the ORE Model from a specific ORE
  Resource URI.
• Enhance A request to add new nodes or
  relationships to an ORE aggregation.
• Usage A request for the number of compound
  objects that contain a given resource (or part of
  a resource)
• Presentation services for compound objects
  application-specific rendering

34
Activities

• First Meeting OAI-ORE Technical Committee,
  Columbia University, Jan 11-12, 2007
• Second f2f Meeting May 29-30, Google, NY
• ORE Panel at JCDL 2007
• White paper posted on web site
• Alpha specs released October 15, 2007
• Undergoing review by technical and liaison groups
• Data Model
• Top level/Minimum Vocabulary
• Resource Map Implementation (ReM)

35
Acknowledgements

• Carl Lagoze and Herbert Van de Sompel
• Kwok Cheung, Ronnie Schroeter, Imran Khan
• DEST/RUBRIC

36
References

• http//openarchives.org/ore/
• http//www.itee.uq.edu.au/eresearch
• Contact j.hunter_at_uq.edu.au