This book does what it says... it's a synthesis of information. It's a scant 94 pages that feel more like 60, and can likely be read (slowly with understanding) in an afternoon. The first two sections of the book seem to be a history of information retrieval casually based around the topic of faceted search. As the other reviewer mentions, there is little actual practical information here... with the exception of a few suggestions to reduce information overload giving guidance on what metadata is most useful used as facets.

Again, there's nothing earth shattering, but it does point to a number of good pieces of literature and papers in the reference section, and it's much shorter than reading Marti Hearst's book entirely (though if you have, you can probably skip this).

This booklet is a short (70 pages) introduction to the topic of Faceted Search. The book reminded me watching an episode of "Lost" - I enjoyed every minute of it, but for every question it answered, it opened up two more...

The book starts with an insightful introduction to information classification and retrieval. It explains why the Faceted Search paradigm is so useful, and why it so much better than what preceded it. This was probably the best introduction to faceted search I have ever read, well written and fun to read (you can probably read the entire book in one afternoon).

But once you finish reading the introduction chapters, you realize that most of the book is over, and you still haven't read any details on how faceted search is actually implemented or best used in practice. These practical concerns are briefly surveyed in 20 pages, but this means that extremely interesting topics that each deserves a whole chapter, are often briefly mentioned in one sentence. If the author accompanied each of these sentences with a reference to a full research paper discussing this topic, it would also have been great, but a lot of statements are made in this booklet without being qualified 18: that ranked search is much better (and much more common) than set search, and yet faceted search only works well with set search. I waited throughout the book to see how the author proposes to treat this elephant in the room, and was disappointed when he only returned to this subject briefly in page 63, where he says that techniques which are useful for general search because they improve recall (e.g., query expansion), only hurt faceted search. Unfortunately, he doesn't offer any solution. Maybe there is no known solution - this is after all a young research field.

To summarize, this book is a fantastic *start* of a book about faceted search, but I would have loved for it to contain much more substance.