Performance should include
- locating attributes within a graphical representation to ensure that there is no redundancy or misplaced attributes in the diagram. (There are many normal forms, but the first three, used in a systematic fashion, perform checks on the data.)
- comparing normal forms by using the following questions associated with each:
- First normal form: Will there be one and only one value for this attribute for each instance?
- Second normal form: Does the attribute depend upon the entire unique identifier for its identification?
- Third normal form: Are there any attributes that have attributes of their own? Are there any repeating groups?