Product Description

Product Details

• Amazon Sales Rank: #53184 in Books
• Published on: 1965-06-01
• Original language: English
• Number of items: 1
• Binding: Paperback
• 1046 pages

Features

• ISBN13: 9780486612720
• Condition: USED - VERY GOOD
• Notes:
• Click here to view our Condition Guide and Shipping Prices

Customer Reviews

Excellent, but with limitations
This book, originally published in 1964 by the National Bureau of Standards, is the result of a project started in 1954. It provides information on most of the functions then widely used in numerical computation in engineering and the physical sciences, including many formulas, and numerical tables of values for most of the functions.

In 2001 it has two drawbacks. First, because algorithms for computing numerical values of mathematically functions have improved dramatically over the 37 years since this work was published, you will not find suitable algorithms for computing values of the various functions discussed. To write a program for a computer or programmable calculator to produce values of any of these functions, you should use algorithms obtained from more modern works.

Second, and for much the same reason, you should not assume that all the numerical values given in all the tables are completely accurate; in 1964 calculations of some of these values with then-known algorithms pushed the state of the art to the limit. For example, in Table 7.3, "Complementary Error Function", two of the values attributed to a 1951 table by O. Emersleben are slightly incorrect in the last digit tabulated. This is not a criticism of this book, or of Emersleben; accurate calculation of values of the complementary error function for large arguments is tricky, and I have found similar errors in tables compiled more recently. However, good algorithms are now known, and should be used by anyone who desires reliable values.

These days I find this book still useful for refreshing my memory on various of the many formulas it contains, but for numerical values I prefer to rely on more recent sources, or on programs that derive values using the better algorithms known these days.

Once Great
I probably would never have gotten my PhD without this book, and it is a stupendous classic. Nowdays, though, my first resort is always Maple or Mathematica, with their manifold capabilities. I still find it useful for trying to understand what those programs are doing by way of simplification (or, more commonly, not doing). Eventually Maple and Mathematica will figure out that they need to couple a powerful explanatory capability to their marvelous algorithms, and this book will become truly obsolete - but that date is not yet here as of 2004.

Four stars only because it has been partly overcome by history. 5 Plus for its historical importance.

A & S is one of the true "MUST HAVES"
A & S's Handbook of Mathematical Functions is an absolute must have for any lover of mathematics. The explanations are top rate. With the abilities of computers to tabulate numbers rapidly, the value of the tables is questionable. But the mathematical coverage is absolutely without equal. Any function you can think of, and some that you probably can't are covered in steamy, lurid detail. The paperback is such a great value that it makes me whinge to imagine that any lover of mathematics could possibly pass it by.