Spatial Databases, The Final Frontier

This paper is divided into two parts. In Section 1 we discuss five different geomatic data models. All of them are intentional, i.e. they give a finite representation of the mostly infinite and even non-enumerable 2 set of points of the spatial objects that are described by the database. In the Raster Model an object is given by a finite number of its points. These points are equally distributed following an easy geometric pattern, which is normally a square. In the Spaghetti Model an object is intentionally deduced from its contour, which is a polyline. The Peano Model also uses a finite number of object-points, but here these points are distributed non-uniformally, according to the form of the object. This distribution method is based on the well-known Peano curve. In the Polynomial Model we use a calculus, extended with comparisons between polynomials. In the PLA-Model, finally, only some kind of topological information is handled without dealing with the exact position and form of the spatial objects. In Section 2 we try to focus on the typical geomatic operations and we introduce different kinds of spatial queries. Therefore we generalize the well-known concept of genericity of Chandra and Harel [Cha80]. A taxonomy of genericity-classes is given. We end with investigating two of these classes more deeply: the isometry-generic queries and the topology-generic queries. On many occasions we will introduce open problems, open areas or topics that have to be studied. We are convinced that a lot of research has to be done in the field of geomatic data types and geomatic operations and we hope that this text can motivate some of the young and intelligent researchers to pay more attention in the future to this unexplored forest that is irrigated by three main rivers: geo..