Originally reviewed by Francis Glassborow in C Vu 4.5I have the book. It is mostly composed of code listings for Mr. Baker's attempts at coding all of Abramowitz&Stegun's "Mathematical Functions with Charts, Tables..." though there is a small amount of discussion.
The code [which is the lion's share of the book] is positively vile. Not very structured, and his constants are abominable. You are likely to find
a = (1.0/6.0)*bto be coded as:
a=.66666*b;and things of that ilk. While the enormous coverage of functions is commendable, and they do "work" for the most part, the code is really inferior. On the other hand, I don't know of any C book for mathematical functions that does a good job in the code department. "Numerical Recipes in C" (for instance) is well known for its icky[tm] code.
All of your other reviews are spot-on, but I don't think I would recommend this book. That's what puzzled me. You seemed to like it and we agree on all of your other reviews. I do find the book useful, since I like to write math functions as a hobby, and for hard to find functions, it does give me something to bench against.
I wish Steven L. Moshier would write a book. His Cephes code collection from NETLIB is positively sparkling by comparison.
- Dann Corbit Dcorbit@SolutionsIQ.comNote that I wrote the original review eight years ago, not that that is an excuse but my understanding of coding has progressed a little since then. My thanks to Dann for taking the time to write the above. - Francis Glassborow