Database Reference
In-Depth Information
For some years now, however, results have been published that contradict the supposed
universality of development methods and techniques. Accordingly, the survey by Hardy,
Thompson and Edwards (1995) concerning the use of structured methods and techniques
in the UK, indicates that 88% of methods and techniques undergo some sort of adaptation
before being used in each particular development project. Russo, Wynekoop and Walz
(1995) offer similar fi gures (88.6%) for the likewise structured methods and techniques in
use in the USA. Things do not appear to be much different as regards object orientation, as
notation adaptation is routine practice. Accordingly, UML, the most popular object-oriented
modeling language today, has three built-in extension mechanisms—stereotypes, tag defi ni-
tions and constraints (Fowler & Scott, 1999)—which can be used to adapt the representation
capabilities of the language to better represent particular domains and problems (Fowler
& Kobryn, 2002).
From percentages like the above, we can deduce that almost all the methods and tech-
niques should be adapted before being used in practice, which means that the generality of
these methods and techniques is merely a guise. The need for method and technique adaptation
refl ects the fact that they are in some measure specifi c for particular problems. Specifi city
should be taken to mean that the methods and techniques are primarily oriented to solving
a paradigmatic problem type, namely, the problems for which they were designed. As the
problems addressed in practice move away from the paradigmatic problem, the methods
and techniques become less effective and need to be adapted before being used.
Method and technique adaptation is a fi eld in which various results have been published
over the last ten years, as indicated. However, no criterion has yet been proposed that can be
used to decide to how well suited a method or technique is for a given problem P (Glass &
Vessey, 1998). Subjective assessment by developers usually fi lls in for the missing formal
criteria. However, this procedure is neither systematic nor repeatable, and therefore can be
qualifi ed as not very engineering-like.
Moreover, such a criterion could be used to establish a methods and techniques hierar-
chy with respect to their fi tness for a problem P, making it possible to identify and select the
best-suited method or technique. Additionally, a fi tness measure would be equally important
for adapting the method or technique, as it could be used to formally assess the situation
before and after adaptation.
Therefore, in this chapter, we propose a method that can be used to assess how well
suited a method or technique is to a given problem P. For this purpose, we will proceed as
follows. The following section will review the different alternatives that have been proposed
in this respect in the scientifi c literature. This is followed with a description of the proposed
assessment method. Next, the chapter will discuss possible extensions and future improve-
ments. The chapter will end with some conclusions.
BACKGROUND
There are two factors that lead to method and technique specifi city: (1) the process;
that is, the prescribed procedural steps for their use and (2) the conceptual models used
by the methods and techniques (Glass & Vessey, 1998). The process is usually adapted to
particular problems on the basis of experience (Hardy et al., 1995), although work on the
systemization of this type of adaptation has been published over the last ten years, leading
