Object-Oriented Implementation of Numerical Methods

En savoir plus sur le livre

Numerical methods align well with an object-oriented approach, as mathematics builds complex ideas on simpler foundations. Once a property is established for a concept, it can be applied to any new concept sharing the same basis, akin to reuse and inheritance in object-oriented (OO) methodology. However, few resources on numerical methods focus on teaching developers how to design and build effective code. Good computing routines are tailored to specific problems, requiring insight and understanding rather than mere recipes and black box solutions. Developers must be equipped to create new programs for varied applications. This work presents a comprehensive OO design methodology in a clear, systematic manner. Each method is introduced with a brief explanation, followed by a description of the general OO architecture for the algorithm. The code implementations are discussed alongside real-world examples from the author's extensive experience in commercial applications. Key features include an exploration of design methodology and patterns rather than just canned solutions, implementations in both Java and Smalltalk to enhance understanding of OO programming languages, a step-by-step guide for programmers transitioning from procedural languages like C or Fortran, and a chapter dedicated to data mining, a crucial application of numerical methods.