Semantic Object Modeling (SOM) Methodology for modeling data requirements Similar to ERM: represents entities of interest (called semantic objects) represents characteristics of these objects represents relationships uses graphical language
SOM Concepts Semantic object (collection of attributes) Attribute simple attribute formula attribute group attribute semantic object link (SOL) attribute consists of a semantic object also called paired attribute Identifiers unique, nonunique, surrogate
SOM Concepts Subtype semantic objects Cardinality Domain represent a specialization subtype object inherits attributes from supertype object exclusive and nonexclusive Cardinality minimum, maximum relationships between attributes and objects or attributes multivalued, single-valued attribute Domain describes the kinds of values an attribute can have physical properties, semantic description
Domain: Physical Properties and Semantic Description
Semantic Object Diagram Simple Attribute Unique Identifier SOL Attribute Group Attribute Nonunique Identifier
Semantic Object Diagram (cont.) Group Attributes Formula Attributes
Modeling Subtypes
Object View Window Displays all possible paths from a selected object to other objects to which it is related. This is essentially SALSA’s way to display relationships but we can only see relationships of one entity at a time.
Model-Building Process Examine application’s attributes and semantic objects; sketch a rough first draft Start SALSA Build a profile list Build the semantic object diagram Customize the attribute properties Validate the model Generate a database schema
Comparison of SOM to ERM Similarities: Represents conceptual view of data Supports subtypes Differences: Presents attributes, more detailed Cardinalities between an attribute and another attribute Domain specification Formula, group and SOL attributes “Implicit” presentation of relationships, objects self-contained