Computer Algorithms: Introduction to Design and Analysis (3rd Edition)

The author is to be commended for producing such a fine work. Note that a new edition is supposed to come out sometime in the fall (of 1999).


A pretty good book in my experience   October 11, 2001
Philip Machanick (Johannesbureg, South Africa)
8 out of 12 found this review helpful

Note that the Preface starts with "This book is intended for an upper-division or graduate course..." and should be seen as such. In my School, we have more emphasis than average on algorithm analysis, so we are able to use this book at 3rd-year level (final year of our degree), but without a strong background, the book would fit a graduate course better. Some of the criticisms relate to use of pseudocode and inadequate diagrams. The latter is maybe a criticism of an introductory book, but is not applicable to an advanced book. Eye candy fools the reader into thinking they understand more than they do. Pseudocode as used in the book should not be a serious problem for more advanced courses. Detail is not left out to the extent that is sometimes the case in pseudocode. In any case, I worry about students who expect to be spoon fed. If you have to derive your own algorithm, the details aren't all provided for you -- you need to work them all out. I do however find it strange that in the 3rd edition, the authors changed to a Java-like pseudocode but didn't adjust array indexing to 0.
My only reason for not giving the book a 5 is that I have found a number of errors in the string matching algorithms. But even this is not unusual in algorithms texts; I would hope all readers would test (if not prove correct) any textbook algorithm before using it. Another thing I do not quite agree with is the relatively informal treatment of solving recurrences (using recursion trees) -- I would prefer that more formal approaches like the characteristic polynomial method were used.
On the whole, this book has a good balance between breadth and depth. Other than the bugs introduced in the string matching algorithms and the inadequate conversion to a Java-like notation, it is a significant improvement on the 2nd edition.
I recommend it strongly for courses where students already have a good algorithms background, and are ready for a more advanced course.
The last book I used, Fundamentals of Algorithmics by Brassard and Bratley, was much harder for my students. It developed a lot of mathematical material that was not used much in subsequent examples, and they found explanations hard to follow.


Not bad could be better   December 1, 2000
Jim Bogan (Easton, PA USA)
7 out of 8 found this review helpful

I am a CS major and I was required to get this book for my Algorithms class, I wasn't too dissapointed with it but I didn't see much in it either. The major flaw is this book would have to be the code, I know it is a algorithms book and code shouldn't matter. I like the idea of psuedo java code. But the fact that the code is much more complicated than it has to be can make it tough for some people to follow it. Also throughout the book there are a lot of "refer to blah", and blah is a couple chapters back. So you spend a minute looking for blah and when you finally find it you forget why you are looking at it.

The book does have a good but quick intro to the math needed throughout the rest of the book, but it might be a little too quick for some. For almost every algorithm there is a thorough explanation, proof, and whatever else is needed.

All in all this isn't a bad book but you could probably find better.


the result is poor   May 12, 2001
B. Etin (USA)
11 out of 13 found this review helpful

This is one of two books I must say sorry to. I hope readers may find my words helpful. I must say this is a poor book, although I feel the authors did put efforts to make it nice.

There are a lot of classical and excellent books on this subject. But here's the reason my department chose it as our textbook: Because the other books are relatively hard and deep for the students.

But here is the response from the studnets at the end of the semester, no matter it is an A student or C student: They hate this book, since they can not get much information after spending hours and hours on it. And they eventually found those "harder" books in lib, and loved them. The key reason is, those books explained everything clearly. (In almost the same number of pages.)

The key problem as I see is, the authors just understood the materials in a certain way, but not thoroughly, and not able to explain it in a clear way. Only a person who has already know all the stuff can figure out what are the authors talking about in some part of the book.

Now I believe, in order to write a good textbook for students, at least you should be a master in this area.

If some of my words hurts, I am sorry. But I am talking about my feeling and most students' feeling.


Considerably worse than the second edition.   May 6, 2005
G. C. Rhoads (NJ)
5 out of 5 found this review helpful

The second edition was arguably the best undergraduate text on algorithms. But unfortunately the 3rd (current) edition is a major disappointment. On the positive side, the coverage of graph algorithms was expanded; on the negative side, is everything else! The explanations are consistently worse than in the second edition. The description and explanation of Quicksort is particularly bad, in fact it's so horrible it should be ripped out of the book and thrown away. About the only use the book now has is as a source of exercises for professors to assign to their students -- fortunately, the exercises are almost entirely the same as in the second edition and are still the best set in any algorithms text.

I have only one thing to say to the author; if it ain't broke, don't fix it!