website paiken:research

Data Reengineering: CASE Tool Usage

by

Peter Aiken & Bill Girling

We used Visible Advantage from Visible Systems Corporation of Waltham, MA because of its support for this type of data reengineering and its ability to directly generate the Oracle schema from the various versions of the project data model. In addition, development of the physical model was facilitated by its support for data model decomposition. Data models are comprised of model views each containing subsets of the entire data model. Although other mappings are possible, for this analysis, the model views corresponded largely to the functional system decomposition of the systems. So the four entities contained in Figure 5, are also contained in the superset of entities comprising the entire PMIS data model (Figure 7) and all these entities are contained in the physical, 'as is' model - shown in Figure 8. These features provide an easy mechanism for both data engineers and functional participants to comprehend the model in bite-sized chunks instead of dealing with the entire model as once. The entire unnormalized model consists of 126 entities and more than 2,800 attributes. Several screen shots illustrate CASE tool features including the following:

• Four entities comprise the 'Classic PMIS' model view
• Business rules maintained using the CASE tool facility.
• Interim version of the PMIS data model.
• A portion of the logical 'as is' the project data model.