Customer Reviews

The Algorithm Design Manual

5 star:		(20)
4 star:		(3)
3 star:		(3)
2 star:		(0)
1 star:		(1)

 

 

This book fills a nice niche -- it is practical enough to be useful and accesible to professional programmers (rather than algorithms researchers or academics) but is build on solid theory as well.

Aside from this, the book has several features to recommend it:
(1) There are "war stories" scattered throughout the text. These are special sections that describe the author's experience working with algorithmic problems "in the field". These are particularly interesting because the present false starts and failures along the way to the final solution. This is a nice change from the standard model of simply presenting algorithms and proving them correct.

(2) Unlike others, such as the popular Cormen, Lieserson, Rivest and Stein text -- which is a bit advanced for beginning study -- this really is an introduction to algorithms. It is quite suitable for, say, undergradutes who have taken a couple of basic CS courses.

(3) Chapter 8 is a 250-page "Catalog of Algorithmic Problems". Here, you will find descriptions of hundreds of standard formulations for algorithmic problems along with the basic solution approaches, and -- this is what really sets it apart -- pointers to implementations of these algorithms. This is part of the overall emphasis of the book: that of understanding standard algorithms in order to avoid "reinventing the wheel".

Overall, I recommend this book for people with some programming experience that would like to take their work to the next level. Hardcore computer scientists (especially those with a more theoretical bent) may also benefit from this book, but should consider it a supplement to one of the more advanced texts, such as the above-mentioned CLRS.

Apart from beeing really useful, this book has two more interesting characteristics: 1. it is very readable 2. it is very amusing/interesting at times. The book is very good if you are a beginner in the subject of computer algorithms, but note that it is not a stand-alone book. You also need a good algorithms textbook to cover the subject of algorithms. Skiena reviews some aspects of the algorithmic theory he regards especially useful/important and presents his interesting perspective, but some times he uses algorithms or techniques he has not presented. What makes this book great is the "War stories", where the author describes real problems and the process that led to the algorithm that solved the problems. Something like Bentley's magnificent book "Programming pearls" (get this one too!). And, of course, the categorized catalog of algorithmic problems, which comprises the bigest part of the book, is almost a guarantee that once you come across a real problem you will shortly know where to look for a solution. Moreover, that solution may already be coded, waiting for you in the CD that comes with the book (which also includes a one-semester lectures of the author in shockwave-audio format).

In my line of work I am, many times, in need of some algorithm or resource to do some job. Sure I learned many during my B.Sc. studies, and during my work, but some are hard to remember, and some I never knew - and sometimes you just need to know how to call the problem in order to locate resources about it from the internet.

Well - this book solves it all!

One part contains "war stories", which I found very useful, and amusing - After reading them I felt like I learned many lessons.

The second part, which is the reason I bought this book at the first place, is a very impressive catalog of algorithms and problems - you just need to know something about the problem you need to solve, and most probably, your problem will be described here with clear definitions, and some suggested algorithms for solutions and with resources from other books/articles and web resources.

I tell you - so many times this book was the first and only stop in my persue of finding the algorithm I needed.

I'd recommend this book to anyone studying and working with algorithms. This is a must in your bookshelf - even more - it's a must on your table!

There are many reasons I like this book more than other Algorithms books I own (e.g. Sedgewick). One is that Skiena's book comes with an HTML version, so it's easy to keep it with you at all times (e.g. on your laptop, at work, and at home). Another is the "war stories". I found that I wanted to read the war stories first, rather than the technical content. Another is that the book points you at existing implementations, which really is what you'll eventually need to look at if you're going to use or write any code.

Overall, I found that when confronted to real world problems, the "Algorithm Design Manual" was a better resource than other Algorithm books. This is why I'm recommending it to software engineers out there. It seems to be written for people working with algorithms to solve problems, rather than as a support for an academic course.

On the negative side: I find the resource catalog to be exhaustive but somewhat shallow. As I'm getting older and slower, some things are not as obvious as they used to be. For example, I was quite unable to derive how to use Voronoi diagrams to perform nearest neighbour search, although it is supposed to be "a simple matter".

For those computer science students and programmers who are put off by the style of "Introduction to Algorithms" by Cormen et al., this book is a good alternative. Rather than just being a catalog of algorithms with code and little explanation, this book covers a much needed middle ground and contains a good deal about the analysis and design of algorithms in general along with discussions of common algorithms themselves in an accessible style. In fact, the author makes a point of mentioning in his preface that you will not find a single theorem in this book, and that the purpose of his book is to get working programmers up to speed quickly on both the generalities and specifics of algorithm design.
The first part of the book is on techniques, and covers the basics of modeling algorithms along with "Big Oh" notation, data structures and sorting, dynamic programming, graph algorithms, combinational search, and the concept of intractibility. What makes this section of the book particularly interesting are the author's "war stories" that talk about real world applications of the ideas discussed in each chapter.
Part two of the book, "Resources", is an extensive catalog of algorithmic problems organized by type. For each problem mentioned, the book includes a problem description and discussion, possible implementations, and other algorithmic problems that are related to this one. In the implementation section for each algorithm, the author demonstrates the kinds of questions that the reader should ask when designing his or her own implementation. The C, C++, Fortran, and Pascal code for all algorithm implementations mentioned is on the accompanying CD rather than in the book itself, which helps make the book more compact.
This book is a very good introduction to the methods of algorithm analysis and design, and an encyclopedic reference on many different types of algorithms. Highly recommended.

The Catalog was my main reason for buying the book. It's an invaluable reference base for people whose boss 'needs an answer by tomorrow'.

+ : The War Stories are fun reading, and do a good job of explaining how theory relates to practice. - : Restating the obvious at times, while deliberately vague elsewhere. Net : if you use a greedy heuristic to select your reading, this book probably comes ahead of the pack.

I have bought this book for "on the job" use and it is coming in very handy. We had a confusing problem concerning a matrix and the solution was here with examples. This sort of thing really impresses me about a book because it is paying for itself.

This is the most approachable book on algorithms I have. Reading should not be a chore, and yet with some of the other algorithms texts, it very tiresome just to find what you are looking for... and you can just forget about understanding those on the first read-through!

What I like about this book is that things are explained clearly without "dumbing down". When I have a quick question about a particular data structure or about which of two solutions would be better for a problem, I always reach for this book first. Money well spent.

This is a good attempt at an algorithm design manual, but from my perspective (as a professor of mathematics working in combinatorics and combinatorial optimization) the writing is simply not precise enough. I'm not saying the book is too "chatty" or "informal". A technical book can be either or both while imparting its content in a clear and unambiguous fashion. I compliment the author on the idea for this book, which I think fills a gap and is at about the right level. However, if a second edition is printed (and I hope it will be), the author should read every (technical) sentence of his text and ask the question "Am I saying exactly what I mean to say?". Unfortunately, despite knowing more often than not what the author _intends_ to say, often I have trouble convincing myself that this is what is actually said.

That having been said: this is a good reference and provides a good "meta-overview" of the basics of algorithms. The Catalog of Algorithmic Problems is a particularly useful tool for quickly locating problems that might be similar to one under consideration and determining (as of 5 years ago, anyway) the approximate state of the art.

I recommend this book, but not unreservedly.

This is a great book on the design of algorithms, emphasizing design over analysis. With perhaps 1/4 of mainframe machine cycles spent sorting, this is an important topic. The first part introduces those topics we forgot as computer science majors; "The Big Oh notation"; data structures, and heuristic methods. I got a lot from the section on dynamic programming, as well as the simulated annealing heuristic. He perhaps too quickly dispenses genetic algorithms. The discussion on "the theory of NP-Completeness" left me behind, but I would return to that section as a reference. The second part of the book is an impressive catalog emphasizing non-numeric algorithms. The accompanying CD adds great value to the book, with the complete book, a web site of algorithms, and audio class lecture.