Database theory washes ashore
Does all this sound like database theory applied to spatial data? It should. Geospatial data modeling is actually the union of traditional geospatial data analysis and spatially enabled databases. Even today s bountiful spatial business market is dwarfed by the enormous world of relational databases. Cross-pollination between spatial and database specialists benefits both communities, as the increase in GIS/database partnerships by leading vendors in each field attests. GIS softwaremakers, in partnership with commercial database companies such as Informix, IBM, Microsoft, and Oracle are beginning to promote their own version of the spatial data model. For example, ESRI unveiled its geodatabase concept at the 2000 ESRI Users Conference along with a well-organized book, Modeling Our World, that unites the knowledge of both spatial and database communities in an explanation of spatial data modeling.
A sea change
Using simple data structures and leaving the modeling to the applications that display them or the experts who manage them has been standard practice for a long time. Why hasn't modeling already replaced this time-honored practice? Spatial data modeling has its challenges and disadvantages as well as benefits. For one thing, building the model takes time, forethought, collaboration, and money. Many domain experts are not also computer experts, and may even feel justifiably threatened by a solution designed to replace them by capturing their precious wealth of knowledge in a database.
Commercial databases also present challenges to the implementation of a spatial data model. Quality databases are expensive and, compared with files, relatively complex. File-based GIS is not going away anytime soon, but the long-term gains of a spatial data model may justify the upfront investment to some forward-thinking companies.
Realizing that there may be initial resistance to a migration from files to spatial data models, GIS and database vendors alike are wooing their communities with user-friendly interfaces designed to facilitate the creation and maintenance of spatial data models. Although options exist for managing spatial data in a database without ever having to display them as a map, the GIS vendors hope that their graphics-intensive interfaces will win over database professionals who until now have worked exclusively in a command-line relational environment.
The technology for storing spatial data as just another column in a database table is a fairly recent development. Databases once sold themselves, but that market has leveled out. Now database companies sell databases in conjunction with solutions to specific problems. Extending a database to store and search for video clips exemplifies this shift in strategy from products to services. Database companies are eager to tap the spatial vertical markets for new customers and spatial software makers are eager to make inroads into the existing database community. Prepare yourself for increasingly aggressive cross-marketing and database/GIS package deals as geospatial data modeling sails into port in your industry segment.