Databases Reference
In-Depth Information
CBD Methods and Approaches
The CBD paradigm is seen by many as a further step above the OO paradigm, resulting
in many similar concepts, principles and ideas. The similarities of objects and components
have become the focus of many discussions and studies (Szyperski, 1998). This has caused
the present introduction of course-grained objects as components in object-oriented methods
and techniques. On the other hand, the natural fi rst candidate for WS and SOA methodol-
ogy practice is using CBD methods and techniques. The question is whether current CBD
methods provide necessary concepts and mechanisms to support that. Components have
been for a long time treated mainly as binary packages of code infl uenced by the versions
1.x of the standard Unifi ed Modeling Language (UML) (Booch, Rumbaugh & Jacobson,
1999). This suggests handling components at the implementation and deployment phases of
a development lifecycle, while still following classical object-oriented modeling, analysis
and design. During the last few years, advanced CBD approaches have been proposed that
provide more sophisticated support to component concepts and mechanisms. However,
the identifi cation and specifi cation of components are still done mainly in an entity-driven
fashion, by closely matching the underlying business entities such as Customer, Product,
and Order. In this way, components are treated more in the form of business objects than
business services. For the purpose of developing modern business-driven service-oriented
systems, it is necessary to defi ne coarser-grained business components that potentially en-
capsulate several business objects and provide real world business services of a measurable
and perceivable value to the user. After the original implementation defi nition of components,
a more logical view on components has been introduced in the UML standard 1.4 and the
latest version, 1.5. The major revision of the UML (version 2.0), which is scheduled for
this year, promises further improvements in representing components as both design-level
and implementation-level artifacts.
A sample of well-published and widely used CBD methods has been chosen for analy-
sis and evaluation of the state-of-the-art of CBD methodology practice. These methods are
documented in topics, on web sites, and in companion papers, in parallel with opportunities
for training and consultancy. They have been already used in practical projects and are sup-
ported by software development tools. The methods show a clear structure and guidelines
for the development lifecycle through a sequence of process steps. The following methods
will be presented and analyzed:
Rational Unifi ed Process (Jacobson, Booch & Rumbaugh, 1999);
•
Select Perspective method (Allen & Frost, 1998; Apperly et al., 2003);
•
Catalysis approach (D'Souza & Wills, 1999);
•
KobrA approach (Atkinson et al., 2002);
•
UML Components (Cheesman & Daniels, 2000);
•
Business Component Factory (Herzum & Sims, 2000).
•
Rational Unifi ed Process
Rational Unifi ed Process (RUP) (Jacobson et al., 1999) is a software engineering pro-
cess developed by Rational Software (now part of IBM). RUP is the direct successor to the
Rational Objectory Process (version 4), which resulted from the integration of the Rational
Approach and the Objectory process (Jacobson, Christerson, Jonsson & Overgaard, 1992)
in 1995. RUP was the fi rst process to use UML from its origin (version 0.8). RUP includes
