Abstract

Paper
Representing Temporal and Historical GIS Data Using ILE DBMS
Track: Education
Authors: Vitit Kantabutra, Daniel Ames, J. Owens, Charles Burns, Barbara Stephenson

Thus far, GIS has been using RDBMS, OODBMS or ORDBMS to model spatiotemporal data. RDBMS is well known to be awkward for use in modeling complex relationships such as those needed in historical GIS. OODBMS and ORDBMS are better, but still not good at modeling relationships with multiple arity (roles), which arise naturally in the modeling of complex spatiotemporal data found in HGIS.

We introduce a new type of DBMS (a modern implementation of E/R model), called the Intentionally-Linked Entities (ILE) DBMS, that uses relationship objects and pointers to implement all the relationships in a native fashion using dynamically-allocated linked data structures.

Using Harvard's CHGIS as an example, note that each place has no single thing representing it through changes, including boundary, administrative seat, and name changes. In ILE the same data can be represented so that each place persists as a single object through time.

Vitit Kantabutra
Idaho State University
921 South 8th Avenue
Pocatello, Idaho 83209-8060
United States
Phone: 208 251-4368
E-mail: vkantabu@computer.org

Daniel Ames
Idaho State University
1776 Science Center Dr.
Ste. 115
Idaho Falls, Idaho 83402
United States
Phone: 208 533-8141
E-mail: amesdani@isu.edu

J. Owens
Idaho State University
ISU-Meridian Health Science Center
1311 E. Central Drive
Meridian, ID, Idaho 83642
United States
Phone: 208 373-1816
Fax: 208 282-1795
E-mail: owenjack@isu.edu

Charles Burns
Idaho State University
921 South 8th Avenue
Pocatello, Idaho 83209-8060
United States
Phone: United States
Fax: 208 282-3405
E-mail: vkantabu@computer.org

Barbara Stephenson
Idaho State University
921 South 8th Avenue
Pocatello, Idaho 83209-8060
United States
Phone: 208-282-0211
E-mail: stepbarb@isu.edu