The ADL Testbed

1
The ADL Testbed

• Greg Janée
• gjanee_at_alexandria.ucsb.edu

2
Outline

• ADL
• an architecture for georeferenced digital
  libraries
• a digital library
• Library contents
• ADL architecture ramifications for
  interoperability
• Status
• Recent usage
• Summary

3
Library contents qualitative summary

• Catalog
• georeferenced items aerial photographs, scanned
  paper maps, GIS layers, scientific datasets,
  text documents, videos, etc.
• FGDC, USMARC, and many other types of metadata
• Gazetteer
• place feature names, earthquake epicenters,
  county boundaries, USGS quadrangle names, etc.
• ADL Gazetteer Content Standard metadata
• GeoRef
• bibliographic records
• USMARC metadata
• Z39.50 gateway
• BIB-1 or GEO attribute sets

4
Library contents quantitative summary

• Series Number Size (GB)Geodex
  335,000 0Landsat 1,200,000
  56DOQQ 12,000 684DRG 124,000
  3,000 52DRG 1100,000 ?
  2.6NASA/Ames 430,000 0Aerial
  photos 25,000 344SPOT
  200 5.9DEM 3,000 ?NIMA
  Geonames 5,600,000 0GNIS 300,000
  0TOTAL 7,908,200 1,144.5

5
ADL architecture
JIGI
Web interface
GeoWorlds
Catalog
Z39.50
Gazetteer
GeoRef
6
ADL architecture search buckets

• Abstract, searchable indexes
• Similar to other umbrella metadata schemes such
  as GILS and Dublin Core, except search buckets
  are
• optimized for retrieval, as opposed to
  description
• optimized for geospatial data and geospatial
  searching
• Searchable collection metadata is mapped to
  buckets
• Users phrase queries in terms of bucket
  constraints (only)

• Geographic locations
• Dates
• Types
• Formats
• Originators
• Assigned terms
• Topical text
• Identifiers

7
ADL architecture metadata

• Flow of metadata and metadata semantics
• Collection metadata is mapped to search buckets
• metadata semantics are lost
• Metadata is also mapped to ADLs standard
  metadata reports
• XML DTDs for the most part describe syntactic
  structure only, so again, metadata semantics are
  lost
• Conclusion ADLs bucket system works well, but
  it makes interoperability in larger contexts
  difficult
• Conjecture should interoperability be transitive?

8
Status

• The ADL library will become an operational
  component of the California Digital Library in
  January 2000
• currently rewriting software to increase
  reliability, performance, portability, etc.
• reloading all collections
• Research will continue under the Alexandria
  Digital Earth Prototype (ADEPT) and InterLib
  projects
• emphasis on intelligently using library holdings
• library will contain both data and programs
  (e.g., geophysical models)
• system will be used in educational settings
• We are interested in exploring how ADL
  (particularly, the gazetteer) can be used as a
  service

9
Recent usage

• Internal
• preliminary analysis of user/usage correlation
  was presented at the ASIS mid-year meeting
• continued informal usage by UCSBs Map Imagery
  Laboratory to help service its customers
• External
• Goodchild and Zhou, Finding Geographic
  Information Collection-Level Metadata, will
  appear in Geoinformatica
• an analysis of the role and value of
  collection-level metadata in finding appropriate
  collections to search
• relied on the spatio-temporal characterization of
  collections that ADL provides
• anticipating significant usage under CDL

10
Summary

• ADL both an architecture and a library
• Close to becoming an operational component of CDL
• anticipating significant usage
• Future research will focus on use of the library,
  particularly in educational settings
• Conjecture should interoperability be
  transitive?
• For more information
• Website http//www.alexandria.ucsb.edu/
• Email gjanee_at_alexandria.ucsb.edu