Why a Geodatabase?
The creation and subsequent improvement of ESRI's geodatabase structure provides an excellent method for generating living GIS documents. This format was introduced by ESRI in the ArcGIS 8.1 release in 2001. There are numerous advantages to structuring GIS data in a geodatabase environment, compared to doing so with a mass of shapefiles.
The main benefit is the number of system files created. Whereas a complicated archaeological GIS may use dozens of shapefiles - which will translate to hundreds or even thousands of system files - the geodatabase exists as only one system file, thus reducing both space requirements on the hard drive and relaxes strain on system resources.
However, this is not the only advantage in migrating information to a geodatabase. Ormsby et al (2001) outline three advantages to using a geodatabase. These advantages deal with the structure of GIS data. As already mentioned, vector data (points, lines, and polygons) is stored in shapefiles, otherwise known as feature classes. A feature class is defined by Ormsby et al as “a group of points, lines, or polygons representing similar” geographically-related objects (2001: 351). A geodatabase allows for the creation of feature data sets. The creation of a feature data set enables coordinated relationships between feature classes. Therefore, if a feature class representing artifact points is moved, then the subsequent artifact polygon and line feature classes are also moved, ensuring their continued relationship.
Another advantage of using the geodatabase structure is that it allows for the creation of domains. A domain assigns valid values or ranges for the attribute table that forms part of the information contained within a feature class. This helps to reduce errors in data entry (by eliminating invalid entries) and also reduces data entry time by creating a series of drop-down menus.
The construction of a geodatabase also has advantages for future research, specifically work that uses GPS receivers. The structure of a geodatabase, with its domain settings in place, mirrors the data dictionary structure used by Trimble GPS units. In other words, a data dictionary can be created using the same field and attribute data used by the geodatabase. This means that data collected with a GPS unit using a data dictionary derived from the geodatabase records data in a format that is immediately translatable into a usable GIS file. These GIS files can be quickly incorporated into the geodatabase. This fulfills the requirement that an information system be updatable by future work, making it a true information system.
Finally, a geodatabase relaxes the system requirements needed to run the GIS software on a computer. A GIS accessing a few dozen shapefiles is, in actuality, accessing several hundred system files simultaneously. This creates an enormous drag on system resources and results in unsteady performance, even crashing some computers (the author’s included). The geodatabase is one system file, and the GIS software, no matter what is asked of it, only has to access this one file, freeing up system resources.