Customer Reviews

The Art of Computer Programming, Volume 2: Seminumerical Algorithms (2nd Edition)

5 star:		(2)
4 star:		(2)
3 star:		(0)
2 star:		(0)
1 star:		(0)

 

 

To an ordinary computer scienist, an algorithm is polynomial-time. To Knuth, it takes 3*n^2 + 17*x + 5 steps on MIX, not counting the time required to display the output, but there might be a way to reduce the number of steps to 3*n^2 + 17*x + 4. For precision and rigor, the Art of Computer Programming books are hard to beat. But, at least for an undergraduate CS student like me, they are slow going. Their greatest value seems to be as a reference for mathematical ideas needed in analysis of algorithms: recurrence relations, combinatorial identities, etc. Like the Bible, the TAOCP books are good to have around even if you don't plan to read them.

Knuth's presentation is obscure and difficult, but he's awfully comprehensive. If you want to learn algorithms, or even if you're looking for a reference, there are many better choices (especially Introduction to Algorithms, the CLR book). That said, I can't point to a more thorough book. Among other things, The Art of Computer Programming series is a great source of problems if you're teaching, learning, or just looking for fun. It might be worth the price just for that.

These volumes are considered first-class references on the subject of computing and algorithms. And they are... but these aren't really books to read. Nobody has time for that. However, for those computer scientists that really are hard up for something to read one of those rainy nights, Knuth has a personal flair and humor that spills into his algorithm summaries and exercises that is unsurpassed, at least in the field of CS handbooks. An excellent source for solutions to commonly asked questions and problems.

Coo