Cornell Institute for Digital Collections Metadata and Cross-Collection Searching in Luna’s...

Preview:

Citation preview

Cornell Institute for Digital Collections

Metadata and Cross-Collection Searching in Luna’s Insight

Cornell Institute for Digital Collections

Problem: Integrating Access to Visual Collections

• Diverse visual resources and descriptions– Multiple repositories at Cornell, multiple digital

collections, distributed digital collections– Different discovery methods and metadata formats

• Searchers are on their own to be aware of collections, know how to link to them, and search different interfaces

Cornell Institute for Digital Collections

1st Solution: A Shared Union Catalog for Images

• Adopted MultiMIMSY 2000 from Willoughby Associates– Museum collections management software– Moved data from stand-alone applications

into it– For the past 4 years, have worked on

developing shared standards and practices

Cornell Institute for Digital Collections

MIMSY Demo

Cornell Institute for Digital Collections

Standards Issues

• No museum descriptive standard– CIDOC reference framework as a glue?

• We have tried to follow VRA 3.0

• Use AAT, ULAN, TGN, for data values

Cornell Institute for Digital Collections

Is VRA 3.0 too complex?

• [example]

Cornell Institute for Digital Collections

2nd Integration Solution: Insight from Luna Imaging

• Addresses issues of collection diversity– Can search multiple collections at once

• Addresses issues of metadata diversity– Maps data to a common standard– Allow searching across multiple

heterogeneous collections

Cornell Institute for Digital Collections

Insight Demo

• Selected features:– General search and display attributes– Cross-collection searching– Variable metadata displays– Annotation tool– Support for formats

Cornell Institute for Digital Collections

Insight’s support of descriptive complexity

• Controlled vocabulary lists and repeating values for fields;

• Hierarchical structures and values; • Groups of fields that should be treated

together, e.g., artist name, life dates, nationality;

• Display order of values, fields, and groups of fields

Cornell Institute for Digital Collections

Images

Intranet/Internet

ODBC

Image Server

Insight Application Server

HTML Server

HTML & Active Server

Pages

Active Server

Insight JVA Clients Insight Browser Clients

Database Server

Data

User Manager User Manager

Values

Insight v3 Data Structure

Objects People Location Location Hierarchy Events

Source Data Tables

FieldGroups

Mapping Tables

Tables Joins Fields

Terms

Inverted Index Tables

• Replicates source data in a format common to all Insight collection databases

Insight Virtual Collection Manager

CollectionManager

RepositoryB

RepositoryC

RepositoryA

VirtualCollection

A

VirtualCollection

B

VirtualCollection

C

VirtualCollection

D

Cornell Institute for Digital Collections

Standards Field Mapping Mapping collection fields to standard fields to allow

searching across separate collection databases

• Maps Artist Name to CDWA FieldID 102 (Creation-Creator-Identity-Names)

Field Mapping Results for “Artist Name”

StandardID StandardName FieldID DisplayName MappingStandard MappingStandardFieldID

1 ObjectID 9 Maker CDWA 102

2 DublinCore 6 Creator CDWA 102

3 VRA 6 Creator CDWA 102

4 VRA v3.0 16 Creator CDWA 102

4 VRA v3.0 19 Personal Name CDWA 102

5 CIMI 68 Creator Name CDWA 102

5 CIMI 79 Creator General CDWA 102

6 USMARC 10 Main Entry CDWA 102

6 USMARC 11 Added Entry CDWA 102

15 Dalton Museum 6 Artist Name CDWA 102

Cornell Institute for Digital Collections

“Built-in” Metadata Standards

• Dublin Core

• MARC

• VRA 2.0

• VRA 3.0

• CDWA

You can add whatever you like

Cornell Institute for Digital Collections

Implementation: Data into Insight

• Currently, export desired data as Text files, clean it up, and import into Insight

• This year – link tables between the two systems?

Cornell Institute for Digital Collections

What is ahead for Insight?

• Development of stand-alone cataloging tool (May?)

• Further support for hierarchical objects– Books, letters

• Links to LDAP and Kerberos authentication

• GIS support

Recommended