Problem

How to improve the performance of design patterns when they are involved in intensive computing? Context Implementation of algorithms that have to be both efficient and reusable with a language that features genericity, i.e., parameterization. We give here C++ snippets but this pattern also ap-plies with other languages such as Ada or Eiffel. Example Let us consider a very simple algorithm: the addition of a constant to each element of an ag-gregate. We aim at having a single implementation of the algorithm, which means that this procedure should accept various aggregate types and data types (the types of the aggregate elements). A general implementation is possible thanks to the ITERATOR pattern [6], parame-terized by the data type: template< typename T> void add ( Aggregate<T> & input, T value) Iterator<T> & iter = input.CreateIterator()