Data Formats for GIS: A Survey of Popular Systems

Abstract

The field of Geographic Information Systems (GIS) is still quite young and there are many incompatible systems on the market. Due to the high cost of data acquisition, the process of collecting data, it is desirable for any system to be able to accept data ported from other systems or sources. However, the conversion of data from one system to another is no simple task. Each system invariably uses its own set of graphic entities (points, lines, polygons, etc) as the building blocks for its model of the world. The requirement for rapid access to the data in the system's internal representation has resulted in each system having its own different (usually binary) file format. Sometimes the specifications for a system's internal file format are kept confidential. While it is often possible to port a system's internal files from one platform to another (for example the VAX, UNIX and Prime versions of ARC/INFO store data internally in the same format) it is usually necessary to rely on one of the many interchange formats to move data between systems. The general interfacing problem is very complex, and beyond the scope of this report. The information we are concerned with may be divided into three broad categories (the term "object" is used to refer to any geographical feature being modelled): • spatial information - information concerned with an object's location, coordinates. • topological information - information concerned with the physical relationships between objects. Topological relationships might be "next to", "to the left of", "comes before" or "comes after". • attribute information - other information that needs to be stored about an object. Examples of this could be "colour", "ownership" or "date of construction". Any good CAD package should have no difficulty handling spatial information. Likewise, a database manager (such as Ingres, dBASE III or Oracle) is designed to handle attribute information. A GIS is a combination of the two, displaying spatial information as maps and holding attribute information in database tables. The ability to manipulate attribute information is what separates a true GIS from a CAD or desktop mapping system.