Re: Opinion on Numerical Recipes In Pascal?
Randy Carr
Joe,
I've gone through the 1st Edition (C,FORTRAN) and the 2nd Edition (C++).
I've probably actually read about 80% of the book, and skimmed the rest. I read the sections that I'm using code from so I can try to understand what they're doing in the code. Their explanations are good, although sometimes I've got to read and re-read to get it to sink in (but that's just me). They provide good references as well. A lot of their algorithms are based on "classic" numerical recipes that you might find on in FORTRAN code on netlib or the like.
I think the book is worth having as a desktop reference. You'll find opinionated reviews about the quality of the code itself, especially the C/C++ code. It's a bit convoluted at times, but it works just fine. I don't have a problem with it.
The pascal code from 1989 seems ok, but the arrays are all [1..whatever] based and seem to use a bunch of global variables. Same as the fortran and original C code. The C++ code changed things and the arrays are 0-based, so the code is just a bit different. Since I usually prefer 0-based arrays (dynamic), I will often just convert the C++ code to Pascal on-the-fly for routines that I might use.
Randy