GIS Design Approaches

Published September 13th, 2007 in GIS

If you're new here, may be you'd like to subscribe to my RSS feed. Thank you for visiting !

As a system, GIS design could be done by using some approaches. These approaches come from a view of how the GIS data are managed on a certain GIS. GIS data contains at least two kinds of data: the spatial data and the non-spatial attribute data. How to manage them such as store, maintain, integrate them could be done differently, and the practical way of doing this has introduced several methods or approaches.

There are at least four approaches to do GIS regarding on how to manage two kinds of different data on a GIS, according to Aronoff (1989) there are:

Four Approaches to GIS System Design

1. File Processing Approach

The File Processing Approach, develops a proprietary system providing the individual data management services required by the different application modules. This first category includes most existing GISes. Each data is stored as separated file. The user can invoke a separate analysis functions to manipulate one or more of these data files. Results are produced in the form of new data files that can be output using suitable routines.

2. Hybrid System

Hybrid System approach, develops a hybrid system using a commercially available DBMS (usually a relational one) for storage of the non-spatial attributes. Develop separate software to manage the storage and analysis of the spatial data, using the services of the relational DBMS to access the attribute data.

Example: ARC/INFO system from ESRI, where ESRI store the non-spatial attribute data in the INFO data base management system. While the ARC system provides for the storage and manipulation of the spatial data. Spatial analysis functions are provided using the "toolbox" approach. Individual modules are provided for such functions as data entry, editing, network analysis, and so on. These modules are then used sequentially to perform the required geographic analyses.

The GIS from Kork System also uses a hybrid approach. Where the non-spatial attribute data are maintained in a relational DBMS, and the spatial data are maintained in an object-oriented data base system called PANDA.

3. GIS Extension

GIS Extension approach, use an existing DBMS, usually a relational one, as the core of the GIS. Then develop extensions to the system where needed. Although the spatial and attribute data may be managed by the DBMS, a significant amount of software is generally added to the DBMS, to provide the spatial functions and graphic display used in geographic analysis. In this approach, GIS is built around an existing DBMS.

Example: System 9 developed by Wild Heerbrugg and Prime Computers. Where both the spatial coordinate and topological data as well as the non-spatial attribute data are stored using the relational data model. The relational DBMS supports variable length fields (to store spatial data), and extensions were added to the SQL query language to handle spatial referencing (such as the points and lines that comprise a geographic feature) and spatial query functions (e.g. overlay, connectivity, and neighborhood operations).

ORACLE (with its extensions of Oracle Spatial) included in this approach, where all data are stored using relational tables and queries are done using an extended SQL language. Such system is also developed by open source ORDBMS PostgreSQL (as object-oriented relational DBMS to manage the non-spatial data), and also adds an extension called PostGIS to handle spatial data. Some relational databases (DBMS) are doing research on developing DBMS extensions to handle spatial data, such as MySQL DBMS (free open source software).

4. Building from Scratch

Building from the scratch approach, starting from scratch and develop a spatial data base that is capable of handling the spatial and non-spatial data in an integrated fashion. A number of GISes under development at research facilities have incorporated artificial intelligent (AI) techniques, so that the system itself can create new information about objects as the system is used (i.e. to "learn" about objects from experience).

Example: Knowledge-based GIS, KBGIS-II has been developed at the University of California at Santa Barbara to explore these possibilities. Also the MAPS system being developed at Carnegie-Mellon University has taken a somewhat different approach than most other GISes.

Each of the four approaches has its own advantages and disadvantages, as well as differs in complexity. The later approach has the more complexity. The GIS Extension approach –for example- is more complex than the Hybrid System, and the Hybrid System has more complexity than the File Processing approach. As the GIS develops, the more advance approach should be used in a GIS, although it much more depends on the organization needs. It is true that the GIS Extension approach is better than the Hybrid System in data integration, but it has more difficulty on doing data analysis, so that it is not suitable for every organization using GIS.

Some of the research and project mentioned here are based on Aronoff's book (1989). It is possible that some of the researches have been done or even stopped. Further exploration on these researches needed for future information. In other parts of the world, many researches are also being done to develop GIS in a such more advance. So, it would be any possibilities that the examples stated here have undergoing major development to improve from one approach to enter the more advance approach respectively.

[Based on Geographic Information System: A Management Perspective. By Stan Aronoff, 1989]