Evolution - Programming GPS and OpenStreetMap Applications with Java - page 225

(from the ROApp) can be interesting to simulate social behavior. In arti-
ficial models, simple theses can be implemented to simulate a population
and observe the effects in the community. In the end, we only see other
people externally no matter how close they are.
RealVehicle
The programming of a RealPerson seams pretty vague at this point and
should be kept in mind for future study. There are much more promising
physical objects to work with, like vehicles. A vehicle is 100% man-made,
and a blue print exists for every one. In OO, the blue print is mentioned
often as the initial idea of a class.
While people are gradually beginning to digitize their identity, most
vehicles are publicly described in detail. Similar to personal profiles in
social networks, most car vendors supply a configuration tool for different
models. The consumer can choose a vendor, then the model, fuel type, and
so on. At the end of the process, a RealCar generator could create the RO
composed of a well-defined body, wheels, windows, doors, and engine. This
process is comparable to the composition of a Java Bean with dropdown
lists of predefined values. Also, images of the car can be added to depict
it from different perspectives in a scenario.
In contrast to a real person, most cars are perfectly predefined and the
ROs could be created for every release of a new car. As described in the
development of GPS traces with a Buell motorcycle, the external observer
can validate the actual behavior. With all of the information about cars,
the server can classify a car (top-down) step by step: A high acceleration
excludes cars with little power; a narrow street between buildings and a low
bridge restricts the car to smaller models. The SO reflection of the car can
temporarily store these values as they are determined. A configured car
can serve as the root class to personalize (i.e., tune) it and as a reference
implementation on the ROApp server side.
Traffic and interaction. Once instantiated, the car can be traced and saved
to a history. In the ROAF, even the history files could be validated against
a road network and trac rules. A little ROApp server scenario could be
used as a driving school and a more complex one could require a certificate
by this server as a driver's license.
Driving a car in an isolated environment is good for learning how to
drive. Yet the actual challenge is to guide a car through the trac at rush
hour. In Germany, the server can make sure that the car is always driving
on the right side of the road. In order to automate the vehicle to drive
programmatically drive, the car needs to be equipped with sensors (i.e.,
Listeners in Java terminology) and has to react to events propagated by
the server.