This dissertation presents the design and implementation of a graphical user interface (GUI) to display spatial data in a web-based environment. The work is a case study for a web-based framework for distributed applications, the Web Computing Skeleton, using a distributed open spatial query mechanism to display the geographic data. The design is based on investigation of geographic information systems (GISs), GUI design and properties of spatial query mechanisms. The purpose ofthe GUI is to integrate information about a geographic area; display, manipulate and query geographic-based spatial data; execute queries about spatial relationships and analyse the attribute data to calculate the shortest routes for emergency response. The GUI is implemented as a Java applet embedded in a web document that communicates with the application server via generic GIS classes that provide a common interface to various GIS data sources used in the spatial query mechanism to access a geographic database. Features that are supported by the distributed open spatial query mechanism include a basic set of spatial selection criteria, spatial selection based on pointing, specification of a query window, description of a map scale and identification of a map legend. The design is based on a formal design process that includes the selection of a conceptual model, identification of task flow, major windows and dialog flow, the definition of fields and detailed window layout and finally the definition of field constraints and defaults. The conceptual model characterises the application and provides a framework for users to learn the system model. This model is conceptualised as a map that the user manipulates directly. Unlike a typical map, which just shows spatial data such as roads, cities, and country borders, the GIS links attribute data like population statistics to the spatial data. This link between the map data and the attribute data makes the GIS a powerful tool to manipulate and display data. To measure the performance of displaying spatial data, two main factors are considered, namely processing speed and display quality. Factors that affect the processing speed include the rate of data transfer from the generic GIS classes, the rate data is downloaded over the network and the speed of execution of the drawing. Two factors that influence the spatial data display quality are pixel distance and bitmap quality. The pixel distance set in the geographic database is represented by two pixels on the display screen, which affects the display quality since the pixel distance is the upper limit for display granularity. This means that setting the pixel distance is a trade-off between the processing speed and the display quality. Bitmaps are raster images that are made up of pixels or cells. To improve the raster image quality, the bitmap resolution can be adjusted to display more pixels per centimetre.
Dissertation (MSc (Computer Science))--University of Pretoria, 2007.