Customer Reviews

Introduction to Automata Theory, Languages, and Computation (3rd Edition)

5 star:		(12)
4 star:		(9)
3 star:		(1)
2 star:		(4)
1 star:		(3)

 

 

This is a good book - but as a revision of a much-revered classic of
the field, it's a bit of a disappointment.

Hopcroft & Ullman wrote the classic text way back in 1969, and then
revised it in 1979. It was pretty much the standard text the world
over for an introduction to the theory of computation.

But over the last two decades, more and more people have been studying
Computer science, and many of them have no time for theory and
formalism and all the 'dry stuff' ..........

The authors point out that because of such reasons and also because
nowadays there's little research in the theory of computation per se,
and more in its applications, they've written a book to cater to today's
students.

Which, in other words, means they've simplified the presentation, tried
to provide intuition whenever possible, given lots more examples and
done away with some of the more difficult material.

This approach puts the book into direct competition with Michael Sipser's
excellent 'Introduction to the theory of computation', a contest it
cannot win, though it might be a respectable second.

Almost all topics are motivated by giving examples of how they're
related to applications in the 'real world', and similar to
Sipser's 'proof idea' approach, the authors first present a topic
informally and then formally, thus gently leading the reader to
the formal proofs.

This book sets out to do pretty much the same as what Sipser's book
does, ie to provide a readable, user-friendly introduction to the
theory of computation with lots of examples and intuitive approach
to problems wherever possible, but Sipser's already done an
'optimal' job.

Moreover, this book tries to be 'chatty', which i'm afraid is just
not the authors' style - the 'economy of expression', which has long
which has long been the hallmark of the legendary textbooks by
Aho,Hopcroft and Ullman, is sadly missing here.

Which means that this may not be the book for you if you're pressed
for time - but on the other hand, if you want to led gently to the
proofs and results with lots of examples and motivation, then this
might be just the book for you.

So all in all, it definitely worth a read - in fact, i'd say
it's still among the top textbooks around.

In fact, i would suggest that you read both this and Sipser, if you
have the time. Otherwise Sipser's the better choice for most of the
part, though it may not cover all the topics you need.

And if you're comfortable with a terse, concise & rigorous
presentation, then the earlier edition of this book is still
unbeatable - and you'll surely need it if you want to pursue research
in this area.

I've just passed my exam on Theory of Computation, and I've used both editions of this text. Frankly speaking, I couldn't choose one of the two should I keep only one of them.
Whereas the first was full of strict formalism, the second has traded this for a more discursive approach. Whereas the first reported theorems name (of their authors), the second has traded this for a richer bibliography at the end of the chapters. And more objectively, the first edition covered more "classical" topics with shorter treatments than the second, but this last treats survived topics with richer details (starting from the first chapter on mathematical basis for the course) and with updated examples of applications (XML and Markup Languages, e-commerce for DFA, etc).
This said, you know why I can't decide. A discursive approach is of course always desiderable, especially if you're completely new to a subject, but a strong notation is helpful in my mind because it improves communication and removes ambiguities. Hence, the best approach would probably have been a mix of the two, or halfway the two.
As a second matter, having a rich bibliography is surely helpful both for further studies and as a reference, but it's quite tedious to look at the index and be unable to find something like "Kleene theorem": you've to dive into bibligraphy to discover that "L is an L(DFA) if and only if it also is L(REG)" is something that has been studied by Kleene.
Finally, I surely can't question the removal of the complexity theory part since it is in the right of the authors to remove "optional topics" (if you use the book for a course on Theory of Computation only) and give a more focused target to the book, but removing stuff like the Myhill-Nerode theorem make things annoying since virtually every course on Automata theory and Computation includes it (like my one did, as well as the course on Languages and Compilers), so you have to look for it elsewhere if your only one book is this second edition.

I would give four stars, should I keep in heavy account the radical changes they made over the first edition and that includes the removal of some stuff, important on my opinion. But ... this is just my opinion, and since it is a very well written and informative book (rich of many details that other texts lack of) and surely one of the bests in the area (I've had 4-5 books in my hands for this course), that's why I gave it 5 stars.

Like other reviewers, I find that the first ("classic") edition of H&U is an excellent reference when one is already familiar with some Theory of Compuation material. It was indispensible in at least the first half of the graduate Theory of Computation course that I took. H&U1 featured more detailed descriptions of automata construction than any other text. It did, however show its age.

In the second edition, the authors have added a few chapters near then end of the book on topics that simply did not exist twenty years ago (eg, there is a treatment of randomization).

At the same time, I find that the new edition is more readable for an undergrad. The introductory chapters expect less from students coming into contact with CS Theory for the first time. There are far more diagrams and sidebars and the overall tone of the book is far less formal.

On the one hand, this book has the potential to become the canonical undergrad text on Theory of Computation, I find that it has the feeling of a book that would appeal to undergrads much more readily than Kozen (which tends to intimidate students by the density of the material it manages to pack per page).

On the other hand, somehow I still prefer H&U1. One gets the feeling from H&U2 that it tries to hide something from students, whereas H&U1 pulled no punches.

And the cover art on H&U1 made it really distinctive (ala the cover of the Dragon Book), whereas H&U2 looks pretty much like any other modern textbook.

It's sad that H&U2 is a second edition of the book, rather than an entirely new book. It would have been wonderful to have both books in print as they serve somewhat orthogonal roles.

The first edition is one of the best book in its field. A classic. A reference for many advanced courses in computer theory.

Sadly, the second edition misses a great deal of the first edition. Many chapters were removed. Important lemmas and theorems are missing.

I would gladly exchange my second edition for the first one, if it wasn't out of print.

J.

I had to use this for a Formal Models of Computation class last semester. It's okay but can be hard to follow. It is often hard to learn from the examples. The formalism and proof gets in the way of intuition. It would make a better 2nd book or reference than a first book on the subject. I supplemented the book with Sipser and found that a much better book for learning from. Hopcroft (this book) is more mathematical in nature but the explanation is harder to follow. If you have a choice, go with Sipser.

As near as I can tell, the big improvement in the 3rd edition over the 2nd is the inclusion of some online practice problems. If your class isn't going to be using these, can you save money by going with the older copy.

This second edition of the classic text tries desperately to be more accessible to beginners than the first edition was, but the results are disappointing. In terms of clarity, the Sipser book still beats this one hands-down.

The authors seem to have assumed that the more words they use to explain a concept, the easier it is to understand. Unfortunately the reverse is often true.

For instance, in the beginning of the chapter on Turing machines is an extremely long, drawn-out example of an undecidable problem. The example is based on a C program that tries to solve Fermat's last theorem and prints "hello, world" when it finds an answer. If this sounds weird to you, you are not alone. I dare anyone who doesn't already comprehend undecidability to be enlightened by this example as it drags on for nearly 10 pages.

That said, overall the book does a respectable job of presenting the material, and if you are patient enough to muddle through the murkier sections you will gain a solid understanding. I especially like the fact that solutions to selected exercises are available on the authors' web site.

But if you have a choice, start with the Sipser book as an introduction, and pick up a used copy of the first edition of Hopcroft and Ullman to use as a reference. Use this second edition as a backup, for alternative explanations of key concepts, or for a broader set of exercises and problems.

Having studied the Formal Language and Automata Theory (FLAT) using the first edition of the text, I think this 2nd edition is much easier to read and to understand. This is a very good thing because the 1st edition was a bit difficult to follow, especially for those who aren't familiar with the subject already. (As stated in the preface of the 2nd ed. that the 1st was written for graduate students who already have the idea of what the subject is all about).

Also, I found the approach "Example->Informal Def->Formal Def->Proof" very good, and should be used in more textbook. This way, the readers will have the idea of what they're going to study, what for, and why, instead of pointlessly studying something, forget it, and realize how important it is later (this happened to a lot of people, though). A lot of useful and interesting Theorems were also introduced.

However, there're things missing in this edition. For example, the Greibach Normal Form (GNF) for the Context-Free Grammars, which was covered in the 1st edition. I'd talked to the professor who taught the FLAT class I took, and he doesn't seem to like this edition either. He said something like, it is of course easier to understand, but something had been left out. So, he still prefered the 1st edition. (Well, he is a very good researcher in Computation Theory and is now writing a book on Term Rewriting System, another model of computation, which will be published by Springer-Verlag. Don't know when, though. So, this edition is probably too easy for him :-)

I, on the other hand, think that this 2nd edition is good on its own ground. It made the subject of Computation Theory easier and more accessible to wider range of people. (You still have to "think" and have Math knowledge anyway).

Also check out Michael Sipser's book on Computation Theory. I think that one is somewhat better than this one. It uses lesser Math (don't get me wrong, I like Math), and clearer explanations.
Also look for the first edition in Library, used book stores, or wherever you can. That one is definitely a classic, and cannot be replaced by the 2nd edition. I agree with one reviewer here, so bad that this was a new edition, rather than a new book....

This was my textbook for an introductory course on Finite Automata and Languages - I enjoyed it a lot and I think that the chapters until the Turing Machines are covered very well, along with good examples. As one previous reviewer has already mentioned, the exercises can get very hard as compared to what's actually presented - this I found not too good.

The topics of complexity classes and NP-Completeness, as well as the chapter on Turing Machines are rather succint and do not cover the full depth. Papadimitriou's "Computational Complexity" does a better job in this respect, even though it is not at all flawless. Some might say that there is a reason why this book is introductory, but I argue that instead of doing a poor job, the authors should have maybe just made another book dealing with the above-mentioned topics.

PS: My professor told me that the first edition was much better - maybe you could find it somewhere in the library, if interested.

I started to learn this course at the beginning of this semester and I just brought this book from Amazon in August.
I had no introductory course before but I was so curious about this subject so I am taking this graduate level course.
Now, I am in chapter 10, and I would like to give a review of this book.
This book is well organized, from the beginning to the end.
I have read almost each word in this book(including the extra ones in the box), and I would like to say: It is worth to do that.
Although sometimes the sentences are not very clear(maybe because I am an international student), but almost all the ideas are precious. So, please be patient when you are reading.
Trust me, if you do not have any related course before, you need time for it. but if you can understand all the contents in this book, and if you are more energetic, finishing most of the exercise with excalmatory marks, you will find your mind becomes so clear that is beyond your imagination.
For the tests, if there are some in your class, is only a half piece of cake. you will feel 100 points is just for the left hand(given the condition that you are a right-hander). :)
If you buy an international version, prepare to visit the book's website. and I will say this second edition seems to me the -1th edition because it contains all the errors listed on the website. Prepare you pen and become a co-auther of the book.
If you feel you need to improve your mathematics, take it, because reading this book can improve your mathematical thinking and proof ability tremendously.
If you feel all the course in your university is too easy and can not match your intelligence, take it, then you will find something interesting.

Hopcroft's book is a very good introduction to the theory of computing, from finite automata to undecidability. He introduces the text with a crash course in proofs, which is useful for a text of this nature. They have several examples with illustrations to facilitate quicker learning of deterministic finite automata, pushdown automata, and Turing machines. These illustrations proved very helpful for me, a visual learner. The book itself is chock full of examples and theorems with proofs. Problems with the book: more explanation on Homomorphisms would be nice. The exercises can get very much harder than the simple material the book teaches, so running through them takes considerable amount of time often. Overall it's a good book, and a lot easier to understand than their first edition in 1979. The material can at times seem a bit outdated since the computing world has changed by several orders of magnitude since their original work, but it still provides a solid foundation in the philosophy and mathematics of computing. Perhaps if you're a Cornell student you'll get the privilege of taking this theory class with Hopcroft as your instructor; he's very nice and willing to help students understand the material.