Decision Support Systems (DSS) are distinguished by the fact that they aid in taking decisions about problems that are semi-structured in their definition. However, they do not replace the decision maker. A DSS have capabilities for handling data, analyzing data and provides muti-dimensional views to help highlight the different aspects of the problem.
One may notice that GISystems are already dealing with the some of the things mentioned above. Hence, it may be said that a complete GI suite is quite close to a DSS. The paper by Densham rightly points out that there are however some aspects in which the GISystems lacks from being a complete Spatial Decision Support System.
GIS systems are traditionally meant to handle only spatial data. For a GISystem to be useful as a Spatial DSS, it should have more flexibility in how it handles non-spatial data. Moreover, the outputs of GISystems are usually only cartographic in nature and might not provide some insights about the problems. It is necessary for the system to be able to generates reports, charts and use other data visualization methods to supplement the cartographic maps, thus ensuring a 360 degree view of the situation. A further challenge for simultaneously handling spatial and non-spatial data is to model the complex relationships between them and to come up with algorithms which are able to leverage these relationships.
The paper also proposes a framework for the development of SDSS. The framework leverages the modular approach of building softwares. This approach enables maximum flexibility in terms of re-use of components in building different systems. SDSS toolboxes can be combined into generators, a combination of which can be further configured to produce specific SDSS. This approach not only provides the ease of component re-usability but also facilitates addition of new capabilities to an existing system without disruption.
Densham also emphasizes on the importance of incorporating research results from the fields of DBMS to have a high performance system. The UI of the system needs to be built keeping in mind the fact that the system is going to be used by decision makers who may not be GIS experts. Both the spatial analysis and non-spatial analysis components should be intuitive to use and a variety of outputs ranging from maps to charts to tables must be available in order to highlight all the aspects of the problem.