Object‐Oriented Programming for Scientific Codes. II: Examples in C++

Abstract

This paper illustrates some specific numerical code applications using an object‐oriented programming (OOP) paradigm. It is the second of two papers describing OOP issues for scientific‐code development. A brief description is given of an object library as well as some simple finite difference and finite element user codes based on that library. The key to efficient OOP implementation in any environment is to conceive an effective object‐oriented design of the physical problem and its mathematical abstraction in the first place. This design places a premium on understanding the physical problem from the perspective of class structure, inheritance, and communication. Some of the issues that must be addressed prior to the development of a truly successful OOP environment for large‐scale scientific code development and employment are discussed. Optimization issues of OOP in general and the object‐oriented programming language C++ in particular are discussed. Our research suggests that there is a tremendous potential for OOP as a development environment for scientific codes.