In some situations it is very difficult, if not impossible, to identify any of the requirements for a system at the beginning of the project. Theoretical areas such as Artificial Intelligence are candidates for using the Exploratory
Model, because much of the research in these areas is based on guess-work, estimation, and hypothesis. In these cases, an assumption is made as to how the system might work and then rapid iterations are used to quickly incorporate suggested changes and build a usable system. A distinguishing characteristic of the Exploratory Model is the absence of precise specifications. Validation is based on adequacy of the end result and not on its adherence to pre-conceived requirements.
Model is extremely simple in its construction; it is composed of the following steps:
Initial Specification Development. Using whatever information is immediately available, a brief System Specification is created to provide a rudimentary starting point.
System Construction/Modification. A system is created and/or modified according to whatever information is available.
System Test. The system is tested to see what it does, what can be learned from it, and how it may be improved.
System Implementation. After many iterations of the previous two steps produce satisfactory results, the system is dubbed as "finished" and implemented.
Problems/Challenges associated with the Exploratory Model
There are numerous criticisms of the Exploratory
It is limited to use with very high-level languages that allow for rapid development, such as LISP.
It is difficult to measure or predict its cost-effectiveness.
As with the Prototyping
Model, the use of the Exploratory
Model often yields inefficient or crudely designed systems, since no forethought is given as to how to produce a streamlined system.
© 2003 Center for Technology in Government