Customer Reviews

Prolog Programming for Artificial Intelligence

5 star:		(7)
4 star:		(2)
3 star:		(0)
2 star:		(0)
1 star:		(0)

 

 

Prolog is not an easy subject, especially for someone not well familiar with mathematical logic. Thus it is very important how the foundation would be laid down. Other books I had read on Prolog tended to two extremes. They were either too condensed for such a complicated language, or too broad and mathematically intensive. The author of this book had managed to find the optimal style of presenting both the essence and the practical aspects of the language. The book makes learning Prolog easy and fun. It covers many practical applications of the language and manages to convey the basic concepts of Prolog without overwhelming the beginner with too abstract passages. This book gives you the taste of declarative programming, which could be a very challenging shift in the way of thinking, especially for programmers used to procedural techniques. The book of Ivan Bratko will soften the learning curve and make the experience much pleasant.

The book is well structured, which makes it also a good programming reference. The material is very well illustrated and supported by a plenty of easy to understand examples. One can open it at any page and easily understand the material without tracing back and forth multiple chapters. This is a great book. Buy it if you don't want to stop learning Prolog before you understood it.

This book is clear and well written. Its an excellent first book on Prolog programming. It also takes you far into the subject, up to a point where you should be able to write useable Prolog programs.

Much more clearly written than "Programming in Prolog" Prolog by Clocksin and Mellish, and it covers the same ground.

Prolog is not for everyone. You have to be quite intelligent to be able to use it effectively. And its not a general-purpose language. But Prolog is excellent for some applications, such as writing Expert Systems, Natural Language Parsers, and other A.I. applications.

This book is best used in combination with a prolog compiler/interpreter, so you can try out Prolog as you read about it. There are some free ones available on the Web.

... a horrible Prolog tutorial.

This is not a good first book on Prolog. If you are new to Prolog and Logic Programming, you should read 'Art of Prolog' first.

Prolog is quite different from other languages, and you'll need some time to get it. This book doesn't give you that time: after briefly introducing the basic concepts, Bratko dives at breakneck speed into recursion and list processing.

Don't get me wrong, this is a magnificent book on how to do AI with Prolog, but it shouldn't be your first Prolog book. It's an excellent second book.

Although this text is always mentioned in the same breath as other introductory textbooks on prolog, I don't think I've ever seen it described as "the best."
The book which usually takes the palm in such comparisons is "Art of Prolog." While "Art of Prolog" is an outstanding book, I think that now, in 2006, it has been eclipsed by the 3rd edition Bratko's book. Why?

Simply this: Bratko's textbook is (as far as I'm aware) the _only_ textbook on prolog which treats the language as a living, developing language! Other textbooks are great for their time, but they are unfortunately stuck in their time. Its as if nothing has happend to the prolog language since February 16, 1987. But this isn't true at all!

The biggest case in point: constraint logic programming! Bratko's text is the only introductory prolog textbook to even acknowledge the existance of CLP. And Bratko gives very lucid descriptions of it, along with very helpful examples and challenging exercises.

Another case in point: inductive logic programming! An entire new branch of machine learning theory has risen, based on logic programming, and NONE of the other introductory prolog textbooks cover it? Come on guys!

I would love to see a 4th edition of this book, because since this one has been published, logic programming has moved even further ahead. Constraint handling rules (CHR), logical functional languages (like Curry), using prolog for the semantic web, etc etc etc. It might be the best kept secret in computer science, but logic programming is really still one of the most exciting areas of programming, and Bratko's book does the best job of staying abreast of, and conveying the excitment of, this living and dynamic field.

Having just read a book on automated reasoning which covered topics in logic, such as unification and resolution, I was very pleased with the prolog programming language, in that it takes concepts from logic, and uses them in ways that can be applied to a number of interesting areas of AI, such as NLP, planning, and expert systems. I thought Bratko succeeded in making this connection with logic, without burdening the reader with notation and concepts that are really not needed for writing or understanding prolog programs.

The book also worked for me on the level of providing a good introduction to the syntax and sematics of the language. The first 200 pages succeed to this end.

Finally, the last 13 chapters can be summarized as representing an introduction to AI from a prolog perspective. On the positive side, he shows how to apply prolog to all of the modern, main streams of AI study. However, on the negative, the slant towards prolog in these chapters tends to oversimplify these disciplines.

My impression of the language is that it seems good as a prototying language, since it is declarative in nature, but from my experimentation (using SWI Prolog) I think I could write better implementations using c or java.

In closing, Bratko's book represents a very good place to start learning prolog and the world of AI.

Professor Bratko has done a tremendous job of putting all the fundamental concepts of Prolog and its applications in various areas of AI. Although this book is focused on Prolog, the concepts that he has discussed are so fundamental that they can be implemented in other languages like Java as well.

I recommend this book to everyone who wants to learn Prolog. I would also recommend the readers to use a Prolog system to work out the examples and exercises as s/he goes through every chapter. A DEC10 Prolog system (like SICStus Prolog) would probably be the best companion for this book.

This book is extremely well-written. It's a very good first book for learning Prolog. My only complaint is that in the later, more difficult chapters, there are less answers to the exercises. Often, these were the exercises with which I needed the most help. This isn't a big complaint and I would have given this 4 1/2 stars if I could have. The topics covered in the AI section were presented in a way that made them relatively easy to understand. The early chapters did an excellent job of teaching the fundamentals of the Prolog language. I highly recommend this book if you want to learn Prolog.

I find the book does not adequetly explain the more complex code examples. First of all the code is not adequetly commented. Secondly, it does not explain the code well for programmers. First when introducing a program like in the expert systems shell chapter it should first define an interface for the program, and explain each goal listed. It should adequetly explain what each goal and clause should hope to achieve. Also, for the more complicated programs it should draw some type of diagram, maybe a flow chart or something that explains the concepts involved. It leaves too much figuring out and guessing for the reader. It is not very user-friendly!
On the positive side, it does an adequate job of explaining concepts when complex code is not involved. I found that I could follow along on even the more advanced chapters mostly everything at least until code was suddenly introduced. Then it became a guessing game as to what it was trying to do.
The author does not seem to realize that it is more difficult to try to understand somebody else's program than it is to write your own program from scratch. As a consequence the reader wastes a lot of time trying to guess what his program is doing.
Note: this review is of the 2nd edition and does not necessarily reflect the 3rd. But, then again, every other review on this page prior to mine is about the 2nd edition as well!

A new edition will be out soon, in 2010. This is an excellent book on Prolog *and* on AI.

For example, chapter 19 is an introduction to inductive learning in first-order logic, an advanced topic rarely found in introductory books. The example program HYPER is a very powerful learner as compared to other "propositional" machine learning methods such as decision trees, neural networks, or support vector machines. I have ported HYPER to Lisp and am still exploring it.

Prolog is not a very popular language nowadays, but basic knowledge of it is still essential to learning logic-based AI.