Structures for Storing Geographic Data - Introducing Geographic Information Systems with ArcGIS - page 247

Subtypes
Geodatabases go even further in promoting data integrity. Within feature type, say roads, you can define
subtypes of roads. Perhaps your planning agency has classed all the roads as freeway, major, or minor.
You could set up the database so that “concrete” was the default material for freeways, “asphalt” for
major roads, and “macadam” for minor roads. Also you could set the domain for minor roads so that
the number of lanes could only be one or two. A subtype is basically an attribute of the feature that gets
special attention from the software.
Objects—First Acquaintance
A more profound difference between the depiction of features in the coverage data model and that of the
geodatabase is that features are not just geometric entities with attributes, but objects , in the computer
science sense of the word.
The study of electricity and magnetism is customarily divided into two general areas: fields and circuits.
The term fields refers to the characteristics of the invisible forces that are caused by magnetic material
or by current flowing in a wire. The term circuits refers to the study of electricity where electrons are
confined to wires and other elements. You might think of a loose analogy between fields and areal
features, on one hand, and circuits and linear (network) features, on the other. In networks, entities such
as trucks and gas molecules are confined within physical structures, as the electrons are confined within
the wires of a circuit. While Esri products have had a network capability for a long time—mainly to deal
with transportation systems—the geodatabase takes this capability to new heights. Using the networking
features of ArcGIS 10 geodatabases, you can represent and simulate complex and extensive linear,
human-built infrastructure—loosely: pipes, wires, and roads.
This new networking capability is facilitated by storing geographic features and their attributes in
a database system that is “object-oriented.” Each row represents an object. Objects are described by
attributes. But objects can also have “behavior.” For a human example, you might describe a person's
characteristics with attributes, such as weight, hair color, and irritability. However, if in addition, you
“allow behavior,” then the person might be instructed to drive to the store for a jar of pickles. This
process could involve other objects: a particular automobile instructed to allow the person to drive it,
a cashier who would accept money for the pickles, and so on. To bring this closer to GIS, a road object
might be allowed to connect to another road object, but not to a freeway object. For another example,
a high-pressure gas line could be connected to a high-pressure valve, but not to a low-pressure valve.
Objects in geodatabases bring us one step closer to integrating the various ideas and components of GIS.
A detailed discussion of geodatabase objects is beyond the scope of this text.
The Shapefile Data Structure
Geodatabases are powerful and sophisticated data structures. In addition to topology, you get for free the
area and perimeter of delineated areas and the lengths of linear features. However, Esri also supports a
much less complex data structure: the shapefile.
A shapefile combines the same two essential major elements that geodatabases do: a (geo)graphic
component and an attribute database. The database software is a relational database management system
named dBASE.