I'm just going to comment on one chapter of the
book: "Numerical Methods", the thirteenth.

Incredibly, Loudon presents a chapter on numerical methods without any discussion of the effects of the limitations imposed by the finite representation of numbers in C. No mention is made of the problems that occur when one tries to represent extremely large numbers or extemely small numbers (close to 0) in C, yet these are common situations when line gradients approach vertical or horizontal. Loudon presents C as though it was capable of representing numbers, when of course C is only capable of a very limited approximation to numbers.

Novices who enter the field of numerical methods without having been warned about the problem are going to risk producing very inaccurate results (and one presumes that some of the readers of this book may simply take the implementations of algorithms presented and apply them to the field they work in, be it analysis of biological experiment results, or investment performance analysis.)

The best Loudon does is to mention "Error Approximation" in the "Related Topics" section at the end of the chapter -- but please note, Loudon is referring to the problems of approximation in the algorithm itself, not the additional problems caused by implementing the algorithm in C where numbers are not represented exactly.

In summary, the problem here is that Loudon has written a useful chapter about some algorithms used in numerical methods, yet has completely failed to deal with the consequences of the implementation of those algorithms in C (surely the point of such a book as this?)
I'm really quite amazed that Loudon's editor at O'Reilly allowed such a serious omission to remain uncorrected. This part of the book is, in my opinion, extremely dangerous to the user, and must be corrected in the next edition.