Database Reference
In-Depth Information
to the creation of the fi eld known as Method Engineering (Brinkkemper & Joosten, 1996;
Hofstede & Verhoef, 1997).
Conceptual models, however, can be said to cause greater method and technique
specifi city. This is because the conceptual models defi ne the problem domain aspect on
which the methods and techniques focus (Davis, 1993). Accordingly, for example, the basic
conceptual model of the structured methods and techniques is the data fl ow diagram, which
can record transformations that occur in the problem domain through its basic constructors:
processes, data fl ows, etc. On the other hand, the dominant model in the object-oriented
methods and techniques is the class diagram, which has constructors that can account for
objects, classes, attributes, etc., related to the problem to be solved.
When a particular problem does not fi t the conceptual model used by a method and
technique, this model should be adapted. Two main, partially overlapping, lines of research
have been pursued in this respect. The fi rst is characterized by the conception of increas-
ingly richer conceptual models, which can express a greater diversity of problem domain
aspects and are, therefore, less specifi c. Models like
i*
(Yu, 1995), KAOS (Lamsweerde,
Dardenne, Delcourt & Dubisy, 19991) or EM (Kirikova & Bubenko, 1994) are within this
line. A second line of research is characterized by the explicit defi nition of metamodels,
which, in some cases, can extend the representation capabilities of the conceptual models.
One of the most signifi cant examples of this line of work is the ConceptBase tool (Nissen,
Jeusfeld, Jarke, Zemanek & Huber, 1996), based on the TELOS knowledge representation
language (Mylopoulos, Borgida, Jarke & Koubarakis, 1990).
It is clear then that there is profound interest in adapting the different methods and
techniques to each particular problem addressed. As yet, however, no criterion or formal
metric has been defi ned that can be used to identify when a method or technique is suitable
for a particular problem
(Glass & Vessey, 1998). This is largely due to the diffi culty of
comparing problems with methods and techniques. This comparison is equivalent to estab-
lishing a correspondence between two sets: a set P of problems and a set M of methods and
techniques, as shown in Figure 1.
There are two main problems that need to be solved to be able to establish the above-
mentioned correspondence:
Figure 1: Procedure for determining method and technique fi tness
