| Author: | Thomas Zurek |
| Date: | December 1993 |
| Type: | Diploma Thesis, Fakultät für Informatik, Universität Karlsruhe, Germany |
| Note: | in German |
In this project a component for an EIS was designed to process the data and information given by several national and international laws and guidelines. The goal was to design a database to hold the rules and limit, guideline, warning values given by the laws and guidelines and to process the informations due to the needs of the various user groups.
As the logic data model for the designing process the context model OPAQUE was chosen. OPAQUE is based on the object-oriented paradigma offers features to model contextsensitive aspects apart from the advantages of usual object-oriented modelling. Contextsensitivity is especially useful to cope the various user groups and to design a database that can be adapted to changes caused by political decisions and new scientific aspects.
Firstly the complex static aspects of the limits scenario were modeled using the object modeling technique (OMT). The result of this designing process was an object model that was transformed to the OPAQUE-EM notation -- an OPAQUE-specific design method.
Based on the static object model a first dynamic model was designed to describe the dynamic behaviour of the objects identified by the object model. This approach was derived from the object model nearly automically. This lead to a simple and straightforward but large and procedural dynamic model. As in similar (e.g. network-oriented) data query languages the responsibility of coping with the complex structure of the scenario was given to the user or the application programmer respectively. The advantage of this approach is the simpleness of the design and implementation process.
A second dynamic model was designed to counteract the disadvantages of the first model. In contrast to the procedural approach a descriptive query model was designed to incorporate a large bulk of responsibility of query processing in the database system. Additionally the query model was designed in a way to automatically cope changes of the static model if these changes go along some minimum demands for incorporating them into the query process.
Finally the database design was evaluated and compared with a database designed by the FZI Karlsruhe (Computer Science Research Centre Karlsruhe). The latter was based on the relational data model. Neglecting the aspects caused by different design decisions it was evident that a contextsensitive object-oriented approach leads to a more flexible and adaptable database design.