Databases Reference
In-Depth Information
Appropriate standard modeling notation for representing component and service
concepts—Different but isomorphic notation types can be proposed, such as textual
(human understandable), graphical (e.g., UML) or machine-readable (e.g., XML-based
grammar), that can serve the needs of different actors in the development process.
•
The mechanisms for defi ning the system models at different levels of abstraction (from
business to code) as well as the rules for performing transformations between them.
•
Traceability of the component and service concepts from business requirements to
software code and at the same time the integration of different viewpoints on the
system being developed using these concepts.
•
Model transformations, code generation and model reusability through an iterative
and incremental development practice.
•
A proper process management approach focused on components and their collabora-
tions as the main artifacts of the development process, together with measuring quality
parameters of the process.
•
High-quality, effective component-based and service-oriented tools that should provide
the necessary support for all the elements of the CBD/WS method.
•
CONCLUSIONS
A framework for effective CBD methodology support is introduced as a fi rst step
towards arriving at truly component-oriented systems development methodology. The
framework is based on the fi ve aspects of the system development process, namely the way
of thinking, modeling, working, controlling and supporting, each of them capturing truly
component-oriented requirements. A methodology sample was evaluated using the concepts
and requirements of the evaluation framework. The framework can be used in practice to
evaluate the true nature of available CBD methodologies and to elicit requirements. It can
also be used to guide the development process to focus on CBD and SOA principles and
concepts consistently throughout the development phases. At the same time, the confusion
between the OO and the CBD way of systems engineering can be eliminated by referring
to the framework.
Current CBD methods and approaches, such as Rational Unifi ed Process, Select Per-
spective, Catalysis and so forth, do not include full support for the component and, specially,
service concept. They propose handling components mainly at the implementation and
deployment phase, instead of throughout the complete system life cycle. The methods are
signifi cantly infl uenced by their OO origins, while trying to introduce the CBD concepts
using standard UML concepts and notation. The research presented here is an early call for
researchers to re-think the fundamentals behind the whole research area of CBD. CBD is
not another way of using old methodology structures for getting OO software technology
to produce functionality. It is a paradigm shift and an opportunity to tighten the loose ends
left dangling from the OO era. CBD should be considered from a market-based production
platform that will bring the whole demand-supply chain in line with future developments,
able to deliver time-to-market units of functionality. This opens a number of new opportuni-
ties for researchers as well as for practitioners. These new opportunities will include taking
the challenge to determine what are truly CBD and SOA methodologies, techniques, and
tools, and how to develop further CBD-based market models. Researchers undertaking these
