Database Reference
In-Depth Information
Table 1: Variety of CBD support provided by the methods (continued)
Compo-
nent repre-
sentation
UML sub-
system
Service pack-
age, UML
subsystem
Stereotype of
the UML class
Not specifi c
Stereotype
type
Stereotype
of the UML
class
Compo-
nent imple-
mentation
UML Com-
ponent and
Deployment
diagram
Component
package,
Deployment
diagram
Package,
Software
components
Realization
component
Not specifi c
Software
components
Defi ned
design pat-
terns
No
Yes
Yes
No
No
Yes
Com-
ponent
repository
No
Yes
No
Yes
No
No
Reusability
Software
components
Components,
patterns
Components,
patterns,
frameworks
Design-level
and software
components
Design-level
and software
components
Design-
level and
software
compo-
nents,
patterns
Incre-
mental &
Iterative
Yes
Yes
Yes
Yes
Yes
Yes
ner, often combining and using concepts from the selected methods. Their support for the
component concepts varies greatly in nature and extent. Therefore, these approaches were
not selected as representatives for the purpose of CBD method evaluation.
CBD METHODOLOGY
EVALUATION FRAMEWORK
Most visible high level categorization of the evaluation approaches for information
systems engineering methods fall into two main categories, namely the evaluation of the
whole methodology and the evaluation of the modeling approaches within a methodol-
ogy. A complete and a comprehensive review of the latter can be found in Siau and Rossi
(1998). In this study, a methodology as a whole entity is the subject of evaluation. Many
such evaluation approaches exist (see e.g., Hong et al., 1993; Blank & Krijger, 1983). All
these approaches present a framework with interest areas of concerns and recipes for con-
ducting the evaluation process. In this research, the evaluation of the whole methodology
is considered in terms of making a judgement as to whether a methodology truly supports
CBD and SOA. Kumar and Welke's (1992) evaluation of shortcomings of methodologies
introduces the concept of methodology engineering and argues that the contingency factors
given within the analytical framework of Sol (1988) provide a promising way to identify the
content of a methodology. This fairly standard framework for method evaluation proposed
by Sol (1988) was chosen because of its generic character, which makes it suitable for adapt-
ing to CBD and SOA issues. The experiences of many methodology evaluation researchers
