Databases for GIS

Abstract

The problems of discussing Geographic Information Systems GIS, for short begin with defining this term. There are countless definitions for GIS, each based on the type of user and application domain [43]. The more general definition [24] would be “a digital information system whose records are somehow geographically referenced”. For more precise definitions, one may emphasize their functional capabilities (e.g., that GIS capture and process spatial data) or the applications supported (linking the definition to the type of problem solved). Others stress the fact that GIS are ultimately tools to be used for decision support. Last but not least, a GIS is database-dependent (“a database system that supports management of spatial data”). In the database community, GIS are primarily associated with spatial databases, and therefore a large amount of the research effort in databases for GIS is related to spatial structures and access methods (see section 7). Research has also considered data modelling [69, 25, 561, query language constructs and optimization (see sections 5 and 6), temporal data support [64, 38, 541, and experimenting with existing DBMS to manage georeferenced data (section 8). Database specialists often ignore many concerns that underly endusers’ applications, such as data accuracy evaluation, feature generalization, fuzziness of geometric boundaries or need for versatile data analysis tools.