Annex B: Design Patterns

B7.Iterator

Description

When we want to mask in the client the mechanics of course of an aggregate, this pattern allows the encapsulation of the iteration in a class hierarchy, either for polymorphic iterations, or to offer various types of tranersing for the same type of aggregate.

Structure and Components

Structure and Components of the Iterator design pattern

• Iterator defines the interface of iteration,
• Iterator1 and Iterator2 establish two mechanisms of traversing,
• Aggregate defines the interface of an aggregate,
• Aggregate1 and Aggregate2 establish two types of concrete aggregates,
• IteratorClient launches the iteration on the aggregate.

Collaborations

• the client requests the creation of an object of iteration from the aggregate or another object responsible for the instantiation of the iterator,
• the client controls the iteration (external iterator) by handling the iterator.

Strengths and Weaknesses

allows to traverse an aggregate without knowing its internal structure,

allows us to traverse in various manners the same aggregate,

the aggregate can be obliged to export methods only so that the iterators can function.

Use in CBR*Tools

The Iterator pattern is used to traverse the recordings ( Record class) in the indexing by behavioural situations. Each type of recordings defines a traverse by defect adapted to its internal representation (properties of the chronicles of which it is made up). Moreover, the templates of potential cases ( PotentialCaseTemplate class) define their own traverse of a recording following their conditions of instantiation.

Brigitte Trousse
Last modified: Fri Sep 28 13:52:10 MEST 2001