Future Directions for Geolibraries

Future Directions for GeolibrariesFuture Directions for Geolibraries

Michael F. Goodchild

University of California

Santa Barbara

GeolibraryGeolibrary

A library that is searchable by geographic location– "What have you got about there?"– impossible in the physical library

• space is continuous and multidimensional

– enabled in the digital world

From the UCSB Map and Imagery Laboratory to the Alexandria Digital Library

From the UCSB Map and Imagery Laboratory to the Alexandria Digital Library

NSF funding 1994-2004 Universal access

– www.alexandria.ucsb.edu Searchable catalog

– a subset of FGDC– MARC compliant– approaching 3 million records

Downloadable objects Map, coordinate, and gazetteer interfaces

1994-19991994-1999 Proliferation of warehouses, geospatial data

libraries, clearinghouses– user interfaces, catalogs

How to know where to look?– the knowledge of the SAP– by data type

• DOQs, DEMs, DLGs, DRGs– by regional emphasis– by thematic emphasis– the collection-level metadata problem

How to achieve interoperability?

CLM of the Alexandria Digital Library

http://www.thesalmons.org/lynn/wh-greenwich.html

Geoportals: 2000-2005Geoportals: 2000-2005

A single point of access– Geography Network– Geospatial One-Stop

Offering both data and services "Live data"

– using OGC standards

Providers "publish" to the site Automated metadata harvesting

– from collaborating sites– GOS at 76,000 data sets in 1/05

GOS coverage, 1/05

There Will Always Be More Than One ShopThere Will Always Be More Than One Shop

Competition– between agencies, levels of government, public

and private sectors Segmentation

– not all providers of geospatial information are content to be part of a single whole

Disciplines– not all disciplines see the geospatial content of

data sets as their primary characteristic Jurisdictions

– GOS is national

CLM revisitedCLM revisited

Regional partitions– a data set is most likely to be found on a

server located within the region Thematic partitions Data type and format partitions Level-of-government partitions

– a data set is most likely to be found on a server maintained by an agency whose footprint most closely matches that of the data set

FunctionsFunctions

The library model– search and retrieval of information objects– legacy granularity– OGC standards support subset but not

superset– gatekeeper

Towards an information model– answers to queries– independent of the source– which is the world's highest capital city?

Geolibraries as GISGeolibraries as GIS

Focus on queries rather than data acquisition

Fully functional servers Accurately georeferenced

The future of geolibrariesThe future of geolibraries

Mixed models– libraries serving information objects– information sources responding to queries

The CLM problem– SAPs have an assured future

Multiple sources will not agree– conflation– "A person with one watch always knows

what time it is; a person with two is always uncertain"