University of Illinois Experiences Using OAI Protocol for Metadata Harvesting

1
University of Illinois Experiences Using OAI
Protocol for Metadata Harvesting
Joanne Kaczmarek University of Illinois at
Urbana-Champaign
ALA ETEXT Discussion Group June 15, 2002
2
Outline

• OAI Protocol for Metadata Harvesting
• Illinois Project Objectives
• OAI Cultural Heritage Repository
• Research Issues

3
OAI Protocol for Metadata Harvesting

• Harvesting Approachto Interoperability
• Provider Services Harvesting Services
• Relies on HTTP XML
• Applicability BeyondE-Print Archives

Service Provider Metadata Provider
H A R VESTER
REPOSITORY
6 HTTP Requests
OAI
OAI
6 HTTP Responses
4
2. Illinois Project Objectives

• Cultural Heritage Resources
• Create Harvesting Provider Tools
• Heterogeneous Cross-Collection Repository

5
Project Objectives (continued)

• Focus on...
• CIC Resources EAD Finding Aids
• Interface for Scholars K-12 Communities
• Coordinate with University of Michigan OAIster
  Project

6
Native Metadata View

7
Views of Metadata OAI -XML

8
Views of MetadataIllinois Repository
9
3. OAI Cultural Heritage Repository

• 1,004,500 records from 34 different institutions
• 3 museums
• 12 academic libraries
• 4 historical or cultural societies
• 1 state library
• 14 digital libraries
• http//oai.grainger.uiuc.edu/search/

10
Participating Institutions

• Alliance Library System
• American Museum of Natural History
• American Numismatic Society
• American Philosophical Society
• Bentley Historical Library, University of
  Michigan
• Celebration of Women Writers
• Colorado Digitization Project
• CIMI
• Cornell University Library
• Formations
• Harvard University Library
• Ibiblio Digital Library Project
• Illinois State Library
• Indiana University Digital Library Program
• Library of Congress
• Lincoln Trail Libraries System

• Michigan State University
• Minnesota Historical Society
• National Library of Australia
• Northwestern University Library
• Ohio Historical Society
• Online Archive of California
• Open Video Project
• Perseus Digital Library
• Spurlock Museum, University of Illinois
• University of Chicago Libraries
• University of Illinois Library (Archives, DIMTI)
• University of Iowa
• University of Michigan Library
• University of Minnesota
• University of Pennsylvania, Schoenberg Center
• University of Tennessee Libraries
• University of Texas, Austin

11
Simple Search Interface
12
Advanced Search Interface
13
Brief Record Results Listing
14
(No Transcript)
15
(No Transcript)
16
4. Research Issues

• Variations in use of metadata
• Organization of resources
• Alternative data presentation
• Rights and permissions

17
Variations in use of metadata

• Mapping other schemas into DC
• MARC
• EAD
• Local
• Mapping variance
• Normalizing

18
(No Transcript)
19
Mapping variance
20
Normalizing DC Type element

• Review existing vocabularies used
• Dublin Core Type Vocabulary
• LC Thesaurus for Graphic Materials
• Select terms to add or replace
• Continuous application of vocabulary through
  filters
• Provide end-user options
• Drop down menus in interface

21
Use of Type element

• 1400 different values
• Used 1 to 873 times per collection
• Variety of controlled and local vocabularies
• Dublin Core Type Vocabulary
• LC Thesaurus for Graphic Materials II

22
(No Transcript)
23
(No Transcript)
24
Organization of Resources
25
Type of resource or institution

• Online access to text/image collections
• Personal papers, correspondence, manuscripts
• Visual materials (photos, paintings, etc.)
• Physical objects/artifacts
• Textual materials (books, letters, etc.)
• Institution type (museum, library, archive,
  historical society)
• Individual collection descriptions

26
Need collocation to enhance utility of
heterogeneous repository

• Difficult to provide efficient automatic
  indexing of DC subject element
• Testing with NCSA data mining tools
• Metadata representing resource, not actual
  resource

27
Alternative Data Presentation
Native EAD Records
http//oai.grainger.uiuc.edu/ead/search
28
(No Transcript)
29
(No Transcript)
30
(No Transcript)
31
Test Scalability and Sustainability

• 45 million catalog records
• Low cost for metadata providers
• Moderate cost for harvesting service providers
• Long-term funding