Scientific and Engineering C++: An Introduction With Advanced Techniques and Examples

Barton and Nackman is probably the best advanced C++ book I have read. It covers the basics of C++ programming for FORTRAN and C programmers, then dives into more advanced material. The book discusses many important design issues, such as how to represent and take advantage of name, structure, and function commonality. Several of the later chapters build extended examples making use of advanced template techniques. One chapter develops in detail a family of array classes, another presents classes for smart pointer functionality. Though this book predates the standard template library, it discusses the use of iterators to provide sequential access to data stored in classes. My favorite chapters build a family of template classes to implement algebraic categories, e.g. groups, monoids, rings, and fields, and a set of classes to implement some rudimentary symbolic computation capabilities through function objects. The use of templates in the later chapters is truly advanced and it is fascinating to see many of the tricks templates allow you to perform.My only gripe with the book is its age. It predates the ISO/ANSI C++ standard and the standard template libraries. It would be interesting to see how the authors would incorporate these recent advances into their treatment of advanced C++ programming.Despite being a book for scientists and engineers, Barton and Nackman should be read by everyone doing serious C++ programming. Highly recommended.

I was first interested in this book for its implication of the title, scientific computing (whatever it is). However, I must say I was pleasantly surprised to learn that the book is not just relevant for the implied target audience, but also for those who need a good introduction to OOP with C++.What makes this book great for learning OOP is that it is rich in concrete examples that are relevant for people who have worked in science or engineering. Many OOP books take more abstract approaches, and often the relevance of discussions for the practical situations is not clear. It is not the case with this book.Another thing I like is the quick and concise introductions of C++. I think the authors do a great job of summarizing "do's and don'ts" of C++ by extracting the information in indented, bullet style lines throughout the book. Other C++ books are just enormous and often it is hard to find relevant information.I would guess the drawbacks of learning from this book is that if you do not have any background in science or engineering, you might have harder time understanding some examples in it. Also, the book is somewhat outdated in terms of coding practices (pre-Standard Library, etc.).In any case, the superb introduction to C++ OOP is in this book, which mostly remains unmentioned when people talk about good C++ books authored by other prominent authors. This one definitely deserves better recognition.

I found this book to be very good, except for a few unnecessary parts. A lot of what the book covers (i.e., "scientific & enginnering") doesn't overlap my problem domain, and the Fortran/C intro sections were skipped since I already knew C++. It would have been better if the Fortran/C sections were put into different books. However, the rest of the book was excellent; even the stuff I didn't use right away was intellectually stimulating.This book should be bought by every serious C++ programmer. Barton and Nackman show several *excellent* examples of more advanced template usage than the boring old "generic container".For those of you who are more familiar with advanced template programming, the "B & N trick" gets its name because it's in this book.

Enumerates reasons for choosing C++ as an object-oriented programming language for scientific applications. Provides an excellent introduction to C++ concepts for FORTRAN programmers. Gradually explains advanced techniques such as templates, abstract interfaces, and class hierarchies. Develops new C++ data types for arrays, matrices, smart pointers, strings, and Albelian groups. Employs relevant code examples in areas such as finite element grids, solving systems of equations, GPIB controller, and automated dimensional analysis. Discusses techniques for creating C++ wrappers for legacy C and FORTRAN libraries. Concludes with a program for data modeling that is built from the preceding chapters. The data-modeling chapter illustrates bottom-up program design.

I waited two years before buying this book - and it was a mistake I find very hard to understand in retrospect. The book is recommended by Scot Meyers, by James Coplien, and even by Bjarne Stroustrupp (the inventor of C++.) Don't let the title of the book fool you: it's almost essential reading for anyone doing serious work with C++, and probably should be for designers and programmers using other object oriented languages too. Skip past the first 199 pages (which provide an incredibly rapid introduction to C++ for Fortran programming particle physicists) and you'll find one of the finest discussions of the fundamentals of object oriented design ever written - an essential antidote to the "finding the nouns" silliness of too many books. B & N is the perfect intermediate to advanced level object oriented design book, and might even be the best choice for a first-timer who really wants to get to grips with the subject (or who has to find out enough about object oriented design to use it in the real world rather than just pass an exam.) If you're interested enough to want know more, but still not certain you should read the book, I'd suggest picking up a copy and reading the design study in chapter 8 - the best I've seen in any book, including Design Patterns.