ARROW institutional repositories and discovery services: presentation to the CORDRA workshop, Melbou

1
ARROW institutional repositories and discovery
services presentation to the CORDRA workshop,
Melbourne, 4 Feb 2005

• Geoff Payne,
• ARROW Project Manager

2
Presentation structure

• ARROW objectives and strategies
• ARROW software
• ARROW and CORDRA

3
ARROW acronyms

• Australian Research Repositories Online to the
  World (ARROW) is a Federated Repositories of
  Digital Objects (FRODO) Project funded by the
  Australian Commonwealth Government Dept of
  Education, Science and Training (DEST)
• DEST Funding of A3.66M over three years
  (2004-2006)
• ARROW Consortium Partners
• Monash University (Lead Institution)
• University of New South Wales
• Swinburne University of Technology
• National Library of Australia

4
What is an Institutional Repository?

• A managed collection of digital objects
• institutional in scope
• with consistent data and metadata structures for
  similar objects
• enabling resource discovery by the Communities
  of Practice for whom the objects are of interest
• allowing read, input and export of objects to
  facilitate resource sharing
• respecting access constraints
• sustainable over time
• facilitating application of preservation
  strategies

5
ARROW Branded Services Profile
 Internet
                                               
National Library of Australia
        ARROW Web Site   Project Information  
National Library of Australia   ARROW Resource
Discovery Service   Using TeraText to index
metadata harvested by OAI PMH
        ARROW Open Access Journal Publishing
System     Using OJS from Public Knowledge
Project
        Internet Search Engines         Capture
text exposed by ARROW Repositories
Swinburne                
UNSW
Monash   ARROW Repository   Digital Object
Storage using Fedora VITAL
Members only area Meeting Minutes etc

6
Repositories Technical Issues Metadata
Exchange

• Dublin Core insufficiently granular for many
  purposes
• Learning Object Metadata not good for
  bibliographic metadata
• Need to preserve metadata relevant to categories
  of objects as decided by the community of
  practice that produced the object
• Open Archives Initiative Protocol for Metadata
  Harvesting (OIA-PMH) can gather Dublin Core
  metadata to establish resource discovery services

7
ARROW Metadata Strategy

• Supports metadata schemata to suit individual
  data models
• No requirement to shoehorn all metadata into one
  schema
• Each stored object can retain metadata developed
  for it by the community of practice which
  generated the object
• Maintains flexibility to store many types of
  digital objects in the repository
• No need to anticipate every object type now

8
OCLC Metadata Interoperability Core
From Godby, Smith and Childress. 2003. Two
paths to interoperable metadata p. 3 at
http//www.oclc.org/research/publications/archive/
2003/godby-dc2003.pdf
9
ARROW Branded Services Profile
 Internet
                                               
National Library of Australia
        ARROW Web Site   Project Information  
National Library of Australia   ARROW Resource
Discovery Service   Using TeraText to index
metadata harvested by OAI PMH
        ARROW Open Access Journal Publishing
System     Using OJS from Public Knowledge
Project
        Internet Search Engines         Capture
text exposed by ARROW Repositories
Swinburne                
UNSW
Monash   ARROW Repository   Digital Object
Storage using Fedora VITAL
Members only area Meeting Minutes etc

10
(No Transcript)
11
ARROW Persistent Identifiers

• Repositories need to offer a preferred form of
  citation for their content
• Which does not break as URLs do when files are
  moved or web sites are restructured
• Handles from CNRI seem to be becoming widely
  adopted
• DOI (Digital Object Identifier is a Handle)
• UK Stationery Office adopting Handles
• DSpace uses Handles

12
ARROW Repository Persistent Identifiers

• ARROW Handles Format adopted
• http//arrow.monash.edu.au/hdl/1959.1/nnnn
• 1959   ARROW handles naming authority
• 1959.n one sub number for each ARROW repository
• nnnn running number
• ARROW will assign a handle to each datastream in
  a digital object to ensure that
• individual parts of the digital object can be
  cited and re-used independently
• Internal data models in the repository can be
  reworked and the datastream can still be reliably
  retrieved
• http//www.handle.net/index.html

13
ARROW - Summary of design criteria

• A generalised institutional repository solution
• Initial focus on managing and exposing
  traditional bibliographic research outputs
• Expand to managing non-bibliographic research
  outputs
• Design decisions are being taken with the
  intention of not precluding management of other
  digital objects such as learning objects and
  large research data sets

14
ARROW technology software selected

• Flexible Extensible Digital Object Repository
  Architecture -Fedora
• Software implementation of architecture by
  Cornell University and University of Virginia
• VTLS Inc www.vtls.com as development partners
• ARROW / VTLS partnership to take the Fedora
  engine and construct a working repository to
  meet ARROWs functional requirements
• ARROW licensing VITAL repository product
• VTLS doing ARROW-specified development
• Ongoing sustainability through vendor support

15
ARROW architecture component software
VITAL Access Portal, OAI/PMH, SRU/SRW, Web
Exposure
VITAL, OJS Fedora
Fedora
16
Resulting VITAL Application Stack
Vital Closed Source Management Client(Windows)
Access Portal (Web)
ARROW-Funded Open Source Web Services
Fedora Open Source Repository
17
ARROW stages

• Demonstration (2004)
• Developing architecture, selecting, testing and
  developing software
• Deployment (late 2004 end 2005)
• Populating the ARROW Partners repositories
• Distribution (mid 2005 end 2006)
• Enabling others to participate
• Under review for earlier participation by others

18
ARROW partnerships

• Established
• VTLS
• Fedora
• Google, to test indexing of research materials
• Thomson ISI Web Citation Index
• Being negotiated
• OCLC, to test their metadata interoperability
  core
• Open Journal System, to enhance the OJS Software
• Research Master, to test integration between RM4
  and ARROW

19
ARROW FRODO Partnerships

• MAMS Meta Access Management System
• Access control through eXtensible Access Control
  Markup Language (XACML) metadata
• Needs development of a FRODO profile of XACML for
  access control interoperability
• APSR Australian Partnerships for Sustainable
  Repositories
• Interoperability through consistent metadata for
  similar data objects
• Needs FRODO Metadata schemata for object
  exchange, export and ingest into new repository
  environments as part of sustainability and
  preservation initiatives
• ADT Australian Digital Theses
• Interoperability through harvestable Dublin Core
  metadata
• Supporting e-theses online which are pointed to
  from ADT
• Role for an overarching FRODO Web services
  strategy?

20
Repositories and Middleware

• List of possible open source repository software
• http//www.soros.org/openaccess/software/
• Regardless of software selected, need to deal
  with same issues
• authorisation/authentication
• object processing on ingest
• object workflow on ingest
• metadata consistency
• providing search exposure
• identifying OA status of deposits
• collaboration with other repositories and
  repository initiatives
• These may or may not be handled in the software
  selected

21
Search exposure

• Requirements
• standards-based way for information gateways, as
  well as other repositories, to query repository
  contents directly
• way to harvest from repositories to support
  single search gateway (preferable to federated
  search)
• ARROW will be supporting OAI-PMH and also
  commissioning development of SRU/SRW web service
  on top of Fedora
• Google has agreed to work with ARROW to expose
  repository content via Google Scholar
• Middleware opportunity
• Extending OAI-PMH to harvest content as well as
  metadata
• modOAI project already looking at this
• Agreement to use SRU/SRW for searching across
  different repositories
• Moderately coupled with repository

22
Collaboration with other repositories

• Requirements
• need for interoperability
• avoiding wheel re-invention
• learning from each others progress so far
• ARROW working with Fedora Development Consortium,
  National Science Digital Library and APSR
• Middleware opportunities
• Registry of standard content models
• In the absence of existing practice, influencing
  the emergence of de facto standards
• Agreement on standard framework for re-usable web
  services

23
ARROW Content Committee

• Unfortunately it is not as simple as build it and
  they will come
• Publisher and Library/Learning Solutions (PALS)
  Pathfinder research on web-based repositories ,
  Final Report, January 2004
• We find that IRs are currently rather small,
  with an average (median) of 290 records per
  institution (smaller but comparable to the median
  size of other OAI data providers). (Page 33)

24
Incentives are needed for academics to submit
their materials to repositories

• Substantial advocacy is required to achieve
  participation
• Mandatory deposit of e-Theses
• Credits towards promotion
• Funding linkages
• Demonstrable additional exposure such as in Web
  Citation indexes and search engines

25
ARROW and CORDRA

• ARROW
• Relies on exposing metadata for harvesting by
  discovery services
• Exposes content through search engines
• Not exclusively SCORM compliant learning objects
• Exposes basic elements for use in building
  learning objects etc
• Need to minimise human effort in the registration
  process
• CORDRA as one of many pathways to ARROW content

26
Questions

• What is in it for authors?
• Protocols for exposing content in scope for
  CORDRA registries via automatic harvesting or
  push mechanisms?
• Boundaries for CORDRA instances
• National?
• Community of interest
• Medical images for diagnostics
• Medieval manuscripts
• Content eligibility
• ARROW fundamental building block objects
• ARROW learning objects
• What is the difference between harvesting and
  CORDRA registration
• Would CORDRA relate to ARROWs federated resource
  discovery service, or to individual repositories?
• Business model reciprocal access or real ?