Database and Expert Systems Integration - Information Systems Reengineering, Integration and Normalization - page 314

Fig. 8.5 An example of an
attribute in the frame model
metadata

Object Identifier: Hector_Person
Attributes:
Name="Hector"
Date_of_Birth="06/02/65"
Sex="M"
Address="63 Chester Road, Sunderland, SR2 7PR"
Age= Method(age) +
Father= Object(Andrew) ++
Mother= Object(Anne)
+ The syntax to represent a virtual value in an object is Method(<method>)
++ The syntax to represent an object value in an object is Object(<object>)

• Attributes

These represent the properties of a class. A particular object will have a value for
each of its attributes. The attribute values that describe each object become a major
part of the data stored in the database. An attribute that is composed of several more
basic attributes is called a composite attribute. Attributes that are not divisible are
called simple or atomic attributes. An attribute value can also be derived or calcu-
lated from the related attributes or objects, for example, the Age and Date_of_Birth
attributes of a person. For a particular person object, the value of Age can be deter-
mined from the current date and the value of the person's date of birth. This type of
attribute is called a virtual attribute in the frame model metadata, and is the result of
a deductive rule or an active rule. For example, an attribute Generation of a person
class can be deduced from the following rule:

If age >40thenold person;
if age<16 then child;
if 16 <age <40, then young;
on theevent dead then dead person.

Most attributes have a single value for a particular object; such attributes are called
single-valued. In some cases an attribute can have a set of values for the same ob-
ject; for example, a College_Degrees attribute for a person. A person can have two
or more degrees. A multivalued attribute may have lower and upper bounds on the
number of values it can store. For example, the Colors attribute of a car may have
between one and five values. Figure 8.5 is an example of attributes of an object
Hector_Person:

• Methods

Rules extend the semantics of the data. The specification of rules is an important
part of semantic data modeling, since many of the facts in the real world are derived
rather than consisting of pure data (Gray et al. 1992 ). It is increasingly important
to integrate rules into data models in new information systems. A crucial charac-