Slashdot Mirror

take a look at numerical methods. There are lots of algorithms for calculating a definite integral. Most revolve around the idea that a definite integral is a Reimann Sum. For more information see either http://en.wikipedia.org/wiki/Numerical_integration or
http://www.nr.com/.

If you are looking at calculating a general integral, symbolic computation is the place to start, for example http://www.symbolicnet.org/toc.html

If you are the individual owner of a copy of this book, we hereby authorize you to type into your computer, for your own personal and noncommercial use, one machine-readable copy of each program. You are not authorized to transfer or distribute a machine-readable copy to any other person.

In my experience in the academic world very rarely somebody uses (code, ideas, etc.) without citing. And if that happens you can name-them-and-shame-them.

My favorite book for understand the nature of mathematics, physics, etc. has always been Numerical Recipes: The Art of Scientific Computing.

I can think of lots of other scientific books I like better and might put on my list, of course. Not a single one places the kind of emphasis on cosmology or unified theory authors.

The parent, GP, and GGP are all pretty much confined to these areas. Why does popular science book = book about those things?

Perhaps that's the root of the problem. These things are areas science delving into realms of speculation that we can't verify or test, and for the most part, can't apply to anything, and that requires highly advanced training to really understand the concepts (rather than just the facts that people tell you). Unusable knowledge isn't exactly the most gratifying knowledge to be had.

Perhaps we need to find a way to promote the more attainable bits of knowledge.

Oops! My bad. I mixed up Numerical Recipes by Press Teukolsky, Vetterling and Flannery http://www.nr.com/ with the completely different (but still relevant and excellent) Handbook of mathematical functions by Abramowitz and Stegun http://en.wikipedia.org/wiki/Abramowitz_and_Stegun. Sorry about that chief!

Just because you bought a book that came with sample code doesn't mean you are allowed to use it:

Numerical Recipes (in C, C++, etc), has a restrictive license that only allows you to use the code for personal non-commercial uses. There doesn't seem to be any provision for using those samples in commercial products.

Just because you bought a book that came with sample code doesn't mean you are allowed to use it:

Numerical Recipes (in C, C++, etc), has a restrictive license that only allows you to use the code for personal non-commercial uses. There doesn't seem to be any provision for using those samples in commercial products.

(least-squares solution, that is), probably the best resource is "Numerical Recipes", they have been changing their location for as long as I remember, but Goodle says it is currently at http://www.nr.com/

I remember dealing with that particualr issue myself... ;-)

Paul B.

Yes indeed Fortran is still alive and kicking. Although I have heard that some of the physics libraries are being converted to C/C++.
As an aside, has anyone else noticed the lack of Fortran texts in brick & motor bookstores? I know Numerical Recepies in Fortran is online, anyone care to mention a good intro. text for a "n00b" like me?

Yea its not as if thats unprecedented either.. There are other "per eyeball" licenses. Numerical Recipes software for instance has a "single screen" license

The subject matter of this book is slightly different, since it has an emphasis on real-time. If you're just interested in crunching a large problem as fast as possible, then latency is not an issue.

BTW, if anyone wants to take a gander at Numerical Recipes in C/Fortran they are available here.

The Numerical Recipes handbooks are written in the same easy, non-self-consiously witty manner. Offhand, I recall a quote about an algorithm that "hunts down and corners the root like a frightened rabbit."

Numerical Recipies - Numerical Recipes in C, 2nd edition is the numerical methods book.

Autobook - GNU Autoconf, Automake and Libtool.

GGAD - GTK+/Gnome Application Development by Havoc Pennington. I'm not sure which is better, the book or the authors name!

WGA - Writing GNOME Applications by John R. Sheets. Not complete, which is a pity. I'm sure that will change though.

Docbook - The definitive guide to SGML.

CVS book - Open Source Development with CVS by Karl Fogel. It is not quite the complete book, but it is the interesting bits.

FreeBSD Handbook - FreeBSD documentation.

Maximum RPM - Documentation for the RedHat package manager.

Based on that list, can anybody suggest further online books that I may be interested in? (Don't bother telling me about the old O'Reilly books, I know about those)

Fortran already has most of the really basic objects used to solve engineering/numerical analysis problems: Integers, reals, arrays, matrices, subroutines, functions, etc.

Object-oriented languages are so popular at the moment because it's a way of adapting a language to a new problem domain (keep in mind that C was written to write an OS kernel, and it's only through bad luck that folks are using an extended version to do everything under the sun today).

But just as C++ and Java are (slowly) displacing COBOL for business applications, I think we'll see object-oriented code displace Fortran and C for number-crunching. (Remember, COBOL was written specifically for Business applications!)

I honestly don't know of a good book on numeric computation that is truly object-oriented. The current edition of Numerical Recipes in C++ doesn't really take advantage of everything you can get from an OO approach. Maybe the new edition of Numerical Recipes in C++ will do a good job at this. Still, the current version shouldn't be ignored, there are some good ideas about how objects can really help in large computational challenges.

And indeed, here it is at Amazon too:

Numerical Recipes in C: The Art of Scientific Computing

Amazon have a lot of hostile comments up, it seems, but ignore them. They're mainly from people who've used the example routines in mission-critical systems and are whining about the odd funny result. The guy who claims that the routines aren't robust is daft in my book -- they're examples from a textbook on numerical algorithms, not super-powered black boxes with a million special cases.

Of course, if you're feeling cheap, pick it up at http://www.nr.com/

jsm

(who once won a pint for managing to sneak a reference to "Tukey's biweight" into a memorandum)