원문정보
초록
영어
The design of an information system for a business process can be implemented either in the object-oriented paradigm or in relational database management system as per the client’s choice. In object oriented paradigm, the design starts with abstraction of objects, classes, attributes, functional dependencies, interrelationships and use cases but in relational database management system (RDBMS) the abstractions may be in the form of entity, attributes, interrelationships, functional dependencies and functionalities. These abstracted view elements are analogous to each other with nomenclature change e.g. view element entity of a RDBMS is analogous to object/class of an Object Technology. The pragmatic practice of designing a relation or class/object of an information system is modeled with Unified Modeling Language (UML) by human intervention. Further, the structuring and grouping of an attributes for a class/object or entity in a design is made possible by the intuition of a designer, which may lead to the error discretions. This paper proposes a realizable methodology that accepts the attributes and the functional dependencies abstracted from the Software Requirements Specifications (SRS) and trims them using minimal cover of functional dependencies to the form of a dependency matrix. The representation is used to refine the functional dependencies that exist among the attributes of each class/entity and grouping them for a relation or object/class by using the normalization process. The procedure is automated to eliminate possible human errors discretions and the correctness and completeness is authenticated through the amalgamation of axiom in the methodology.
목차
1. Introduction
2. Background
3. Frame Work
3.1 Candidate Key Attribute(s)
3.2 Design Algorithm
4. Case Study
5. Conceptual Results
6. Conclusion
References