This was my textbook for my algorithms class at UW-Madison.
I like it better than Cormen's book, (introduction to algorithms) I think it's more readable and easy to follow.

A couple of caveats:
1.) This book is more of a survey. It does a deep-dive of some algorithms, but a lot of them only get a high-level description. If you are interested, you will have to look elsewhere for details, etc.

2.) The exercises are very well-written, and very interesting, some of them have been interview questions that I have encountered. The thing is, there is no answer key, and the hints section is not very useful at times.

Still, this book opened my eyes to the beauty of algorithms and math, and the elegance of solutions that are possible for "hard" problems.
After you read this book, you will have an idea of what is possible and it will give you enough ammunition, enthusiasm, and background for digging deeper.
A great collection to pair with is Sedgewick's collection on algorithms, Bundle of Algorithms in Java, Third Edition (Parts 1-5): Fundamentals, Data Structures, Sorting, Searching, and Graph Algorithms, Third Edition which may fill in some of the detail that this book may lack.

Even though this book is only a few years old, it seems like a well-worn classic. That is not to say that it is outdated and irrelevant; it is exactly the opposite! While my knowledge of algorithms textbooks isn't encyclopedic, I can say that this book was way ahead of the curve in it's layout. Before it, most books presented a bunch of algorithms with some loose affiliations such as being search algorithms or dealing with graphs. This book groups algorithms into major themes and makes it much easier to learn the fundamentals of algorithm design. There have been a new set of algorithms textbooks coming out as late (Kleinberg and Tardos, Dasgupta et al, etc) that have the same approach as this book because it works so well. Don't miss out!

The author claims that the common used classification of algorithms have serious drawbacks and that not all important algorithms, such as Euklid algorithm, can be classified using that classification. Because of that author suggests a new classification of algorithms, containing:

* Brute force
* Divide-and-Conquer
* Decrease-and-Conquer
* Transform-and-Conquer
* Dynamic Programming
* Greedy

The proof of completeness of the classification is not given in the book, but it is obvious that more algorithms can be classified by this classification than by standard one. However, some problems built in the standard classification are not eliminated from the new one: this classification has no unique property which would be crutial property of the classification.

Book is written very good. It is interesting for reading, methodically corectly written, so it can be used as a textbook for the primary course on algorithms.

I find proposed classification very interesitng and usefull, regardless all problems that are built in it. This will, for sure, make many people dealing with theory of algorithms think about all drawbacks of the existing classifications of algorithms and maybe they will produce even better and more sound classifications.

For those of you who've always wanted to throw the Cormen text out the window rejoice you finally can! This book is an excellent introduction to algorithms. Though not as thorough as the Cormen text it's orders of magnitude more readable!

I recently finished using this book for a course. I found it to be very well written and the presentation was excellent. Most books on algorithms are hard to read and overly mathematical. This book emphasizes more the textual explanation of algorithms for practical applications and not as much on theoretical proofs etc. The exercises at the end of the chapters strongly reinforce the contents and insightful hints to every exercise are given at the end of the book. I thoroughly enjoyed using this book and found it challenging. Other books I used for reference such as the MIT classic Introduction to Algorithms was also useful but much more difficult to understand initially.