I am a scientist who does a lot of numerical analysis. I have been using a lot of commercial apps (e.g. Excel) as well as a lot of high-level interpreted languages to do my work. However, I recently discovered that these packages were actually returning the wrong answers. This is because of how variables have to be reset and other software requirements which, frankly, I would have never discovered if I hadn't had some colleagues who were CSE majors.

So I binned the packages, downloaded a FORTRAN compiler for about $200, and wrote my own code. But to save myself a lot of trouble I turned to this book that I had used in graduate school. All of the routines are there - matrix operations, transforms, numerical integration - and shortly I was back to crunching data.

Why FORTRAN? First, it's sort of the Latin of scientific computing. It might be old but there is a lot of legacy code (as evidenced by this book) that has been tested over and over. Second, it is incredibly easy to write. Third, you KNOW what it is doing, you're not trusting some algorithm written by somebody else who wasn't trying to solve the problem you're trying to solve. And this code runs quickly. Data analysis that took two to three minutes in Excel takes a few seconds with these programs and subroutines.

It's like driving a car with a manual transmission - once you get the hang of it you have a great deal more control and can make your code go like you want it. And so much of that code is right here, waiting for you.

I love this book! I had never written Fortran but this book made learning very enjoyable. This is not a "learn Fortran" book so don't get that idea. You will probably want to already have some Fortran experience before you tackle this book. I have a Computer Science degree and 20+ years of experience as a Software Engineer so I can learn a new language from a technical book. I love learning new languages even if they are old like Fortran. I wish I had discovered the Numerical Recipes books back when I was a young college student. But they are great, even for old men like me.

Fortran 77 is easier than C but almost as powerful (no pointers)

Book arrived in good conditions

It gives a good description of numerical algorithms. What interested me with the discussion of Lanczos coeffiecients to evaluate the gamma function was that the solutions are not unique and a different set of coefficients might be found which makes the algorithm more efficient. They don't mention that in the book. Maybe this applies to other algorithms in the book?

I used Fortran for 30 plus years and adapted to Object Oriented languages for the final 15 years of my career. I purchased this book for a Lagrangian interpolation routine which I quickly converted. Has many code examples which I am sure I will draw from again.

A useful handbook of programming routines, if one does scientific work.

Written for FORTRAN but also bought C version great solution book for implementation of numerical methods in systems that do not have advanced mathematical libraries

This book features about every numerical standard algorithm you will ever need to do scientific, performant computing in fortran. I use it to get an idea to code in C++, and this book is totally worth it (in contrast to the C++ version of it, which is just bad).

If you do numerical calculations on a daily basis, you should read this book!

The go-to book for numerical methods. The code is great, well-tested and easy to use, and easy to adapt to e.g. F90, but one of the most useful things about the big thick volume is the description of the methods, especially the stats. If you haven't got a (very) well-thumbed version of numerical recipes in your language of preference you can't call yourself a numericist in any discipline.

This is an excellent book, contains 'tons' of hardworking, it is beautiful!

一時代前のデータ処理の実務書だが，データについての考え方を，今のデータ・クラウドの時代に照らして，再検討するいい材料になる。

新品ということで、注文をいたしましたが、表紙のすれや色あせが酷く、誰が見ても「古本？」と思うような状態でした。