The logical model created at the end of Step 2 is then used as the source of information for physical database design described in Steps 3 to 8. Again depending on the complex- ity of the database systems being design and/or the functionality of the target DBMS, some steps of physical database design may be omitted. For example, Step 4.2 may not be applicable for certain PC-based DBMSs. The steps of physical database design are described in detail in Chapters 17 and 18. Database design is an iterative process, which has a starting point and an almost endless procession of refinements. Although the steps of the methodology are presented here as a procedural process, it must be emphasized that this does not imply that it should be per- formed in this manner. It is likely that knowledge gained in one step may alter decisions made in a previous step. Similarly, it may be useful to look briefly at a later step to help with an earlier step. Therefore, the methodology should act as a framework to help guide the designer through database design effectively. To illustrate the database design methodology we use the DreamHome case study. The DreamHome database has several user views (Director, Manager, Supervisor, and Assistant) that are managed using a combination of the centralization and view integration approaches (see Section 10.4). Applying the centralization approach resulted in the iden- tification of two collections of user views called Staff user views and Branch user views. The user views represented by each collection are as follows: n Staff user views – representing Supervisor and Assistant user views; n Branch user views – representing Director and Manager user views.