Solving Pdes in C++: Numerical Methods in a Unified Object-Oriented Approach

By Yair Shapira

This text provides a comprehensive guide for the numerical solution of PDEs using C++ within an object-oriented approach. The high level of abstraction available in C++ is particularly useful in the... This description may be from another edition of this product.

• Format: Paperback
• Language: English
• ISBN: 0898716012
• ISBN13: 9780898716016
• Release Date: January 1987
• Publisher: Society for Industrial and Applied Mathematic
• Length: 184 Pages
• Weight: 2.00 lbs.

Customer Reviews

Good but could be Better

I've had the book for several weeks now and have worked through most of the book and tried to use several of the example classes. From the perspective of an engineer who writes a lot of code for engineers to use and understand I'd say the book is definitely written from a computer scientists point of view with heavy use of templates, inheritance and operator overloading. Several of the code examples are missing parts such as a default constructor in the template for the point class needed to run the ultra simple example. Also, there are several inaccuracies such as claiming in Fortran you must declare array sizes at compile time. This hasn't been the case in over 14 years. I still give the book 4 stars for being the only book of its kind plus it isn't all bad. There are some good ideas presented especially in the use of templates and how to use objects to build up a hierarchy from points to meshes. This was presented in a well detailed way which was easy to understand and really improved my understanding of templates, classes and OOP.

Excellent for C/C++ Programmers who need this type of Math, but ...

The book does give a good introduction to C/C++ if that is what you need. If you already are an expert at C/C++, this book is excellent introduction for applying PDE's with C/C++, though I could have asked for more examples. I bought the book to help me write better PDE's in my work with Image Processing, Fuzzy Logic, Neural Networks. The Author does explain thing in a very concise way, but be prepared to do research and work through the exercises. However, there is some noticable code in the book that just does not work, but can be easily corrected. There is also some misinformation with the STL library and it's capabities. This can be a problem for those not well versed in the STL. Also some of the functionality is not as effiecent or well written as he claims to be. This is especially true the operator overloads, but it not bad. Otherwise I would have given it a five star.