Access to distributed resources Lorcan Dempsey VP, Research Research Library Directors Conference OCLC Institute Post-Conference, "Building the Global Digital Library: Access, Transform, Create, Integrate," Wednesday,

Download Report

Transcript Access to distributed resources Lorcan Dempsey VP, Research Research Library Directors Conference OCLC Institute Post-Conference, "Building the Global Digital Library: Access, Transform, Create, Integrate," Wednesday,

Access to distributed
resources
Lorcan Dempsey
VP, Research
Research Library Directors Conference OCLC Institute Post-Conference, "Building the Global
Digital Library: Access, Transform, Create, Integrate," Wednesday, March 6
From portal to environment
• Unified discovery experience across
heterogeneous resources
• Focus on access rather than use
• Overview
– Retrospective look at portals (Z)
– OAI
– Towards an information environment
The portal/gateway/broker – mid 90s
Presentation
Mediation
Resources
Individual
presentation
Aligned
presentation
Flexible
presentation
Independent
applications
Independent
applications
United service
Monolithic
Monolithic
Components/data
Vertical and horizontal integration
Discover
Locate
Request
Z39.50
Collection/
item
Z39.50
Resolution
Z39.50
ILL
Deliver
Use
Use
Broker component examples
User access
Information
landscape
User
profiles
Authentication & authorisation
Environment intelligence
Service
Description
Collection
Description
Schema
Description
Distributed services access
Search
Interfaces
Request
Interfaces
Delivery
Interfaces
Some observations
• ‘interoperability’ a mantra
• Scepticism about ‘strong’ interoperability
• Limited targets
• Information landscape vs information brandscape
• Some emerging demand
– Reference linking
– Metadata sharing (OAI)
– Authentication/authorisation
OAI-based mediation
OAI Server#1
DC & EAD
DC
OAI
Harvester
OAI Server#2
DC & MARC
MARC
DC/MARC
Union
Catalog
HTML
DC
OAI Server#3
DC & VRA
Web
Browser
Z and OAI working together
• http://www.humbul.ac.uk
• http://www.rdn.ac.uk
• http://www.renardus.org
EEVL HUMBUL PSIgate
BIOME
SOSIG
OAI-PMH
Person
RDN
Z39.50
Person
Z39.50 client,
e.g. Renardus
Person
Towards an information environment
• No single resource is the sole focus of a user’s
attention
• In a distributed environment resources can be
specialized to provide particular function
• Leverage communication between resources to
add value to user’s experience
provision
content
infrastructure
brokers
and
aggregators
shared
services
portals
presentation
mediation
Presentation/Portals
• Surface resources, pulling together other
services available in the environment
• Learning management, library portal, ‘button’
services
• Tools for manipulation, analysis, assembly,
play, …
• No prescription
Infrastructure/
Shared services
• Add value to applications by being provided as
part of infrastructure
• Directory services: environment intelligence
– Collection description
– Service description
– Schema description
Infrastructure/
Shared services
• Resolution
– Resolution service
– Institutional profiling
– Personal profiling
• Authorization
• Knowledge services
– Gazetteer
– Terminology services
Applications/Mediation
• Services which mediate access to distributed
heterogeneous resources
– Search services (z39.50, ..)
– Gathering services (OAI, ..)
– Alerting services
Content
• Metadata and data of interest
A note on metadata
• Data which supports operations on entities
– Life cycle
• Creation, discovery, assembly, administrative, contextual,
preservation, rights, …
– Domain
• Learning, library, archives, etc
A couple of other ‘consensus’ issues
• Semantic diversity – common sense
– subject
– audience level (who is this resource aimed at?)
– resource type (what kind of resource is this?)
– certification (who has created this resource?)
• Collection description
Some issues
• Focus on standards and acronyms needs to
give way to a focus on architecture
– Clarifies roles and relationships
– Bridges between business and technical
perspectives by identifying specialised functions
• Our organisational and business models do
not well support such an approach
– E.g. How do we secure infrastructure
Thankyou
[email protected]
Acknowledgments
• Draws on work done over several years within
UKOLN models project and the DNER
– http://www.ukoln.ac.uk/models
– http://www.dner.ac.uk/arch
• http://www.oclc.org/research/staff/dempsey/
publications.shtm