[chaboud]

I've used Numerical Recipes in C heavily since 1996, but I've never used the code (which isn't great). Instead, I've used it as a reference to understand approaches and algorithms. Any code I've made using that book has been my own.

If you know how to code and want a leg up on understanding an algorithm, NR is a very convenient structural cheat sheet.

[owlbite]

But its algorithms are often incorrect, inefficient and numerically unstable. Just use a library written by an expert in whatever your area of interest is.

[toast0]

Isn't this book essentially a library written by experts[1] in numerical processing?

You can't just use a library by experts. You have to know or learn how to evaluate said library for correctness, efficiency, and stability (numerical and operational). If you know how to do that, you may be able to use a flawed library as a tool to get the shape of your processing quickly and fill in the rest of the details yourself.

[1] They've published a major book through three editions and the book cover calls them leading scientists.