It is difficult to choose a computer book for several reasons. Firstly, the market is saturated so there is a lot of choice. Secondly, there is huge variation in quality and a wide range of different audiences for the same subjects, both of which are difficult to assess quickly.
This buyer's guide identifies and explains several indicators that allow you to guess a book's quality and who it is aimed at. This is important because the majority of books are poor quality and aimed at people who know nothing about the subject. This makes it especially difficult to find good quality books on advanced technical subjects. In summary, you should
Some bookshops save you time by having an O'Reilly section
- choose an O'Reilly book, where possible
- avoid books more than 2 cm thick
- avoid books with lots of padding
- choose a book that focuses on a single topic
- avoid books written by idiots for idiots.
Beginners don't usually need books
If you only want an introduction to a subject, or you're a novice, then you can probably find what you need on the web. Google Is Your Friend; include "how to" or tutorial in your search. You will find many high-quality tutorials about IT (especially web-related) subjects. This is not a coincidence.
Once you've done a few Google searches for things like how to make photograph thumbnails you'll notice that as well as individual 'how to' pages and tutorials, you'll find specialist web sites that combine articles and tutorials on a particular subject; these are often far more useful than the average book. Good examples are W3 Schools for web development and Photo.net for photography.
O'Reilly books are the best for technical reference
If you're an IT professional, or full-time geek, then a good book may let you read a lot about a subject in one go. On the web it can be difficult to get past the introductory material and find enough coverage, detail and examples.
If O'Reilly publish a title that covers the subject you're interested in then you will save yourself lots of time and effort by buying an O'Reilly book. The O'Reilly books illustrate what is wrong with other books, as described on this page, by how good they are.
The O'Reilly books aren't for everyone, of course. If you just want to learn about a subject then you should probably look elsewhere for a suitably shallow treatment of the subject. If, on the other hand, you actually want to learn more about a subject than just what the best buzzwords are, then you need a proper book.
The best books focus on a single topic
Vague titles do not identify the book's subject; instead they suggest that the book lacks focus. For example, Internet: The Complete Reference, Milennium Edition (Osborne Publishing, 1999) has a title that is no more descriptive than 'Television: The Complete Reference'; it does not give any clues that the book is about using the Internet as an individual, say.
Thick books have less useful content than thin books
The worst books contain little or no useful content. Publishers usually try to make these books stand out by making them physically large. This, of course, exacerbates the problem of what to fill the book with, which makes the padding easier to spot.
You can find the most useless forms of padding quite easily; look for
- chapters that give basic introductions to topics that ought to be pre-requisites for the book, in case you have never used a computer before
- chapters containing instructions for installing software that the book refers to, in case you threw your software user manuals away
- chapters containing nothing but large amounts of unnecessary reference information
- several chapters covering the same topic for several different software packages, in case you are simultaneously using several different packages to do the same thing
- unnecessary screen shots that add nothing to the explanations in the text, in case your computer doesn't have a monitor
- a long section on 'how to use this book' that contains a set of 'icons' that are not at all self-explanatory, in case you have never read a book before or do not understand how a table of contents or an index works.
Unnecessary reference information: for example, HTML Complete (Sybex, 1999) boasts '1000 pages for only $19.95' on the cover, but 179 pages - a sixth of the book - are taken up by HTML and CSS reference pages. Reference information may be useful in a technical book, but in this case these two chapters are much less useful than the HTML and CSS reference information on the World Wide Web Consortium's web site.
Screen shots: for example, Teach Yourself the Internet (IDG Books, 1999) has 409 pages. In every chapter, every facing page contains nothing but screen shots. If you exclude these pages and the contents, index and chapter cover pages there are only 180 pages left for any other content - less than 45 per cent of the whole book. This wouldn't be so bad if the screen shots were all useful, but they are not. For example, one of the more ridiculous screen shots shows how to select 'Page Setup...' from the 'File' menu.
The worst icons: Building a Windows NT4 Internet Server (New Riders Publishing, 1999) has an appalling set of 'icons' that appear with its margin notes. The icons are ultra-low quality pictures of pliers, saws and clamps - good for nothing.
One exception, which proves (as in tests) this rule is Code Complete (Microsoft Press, 1993). Its 3.7 cm thickness represents an unusually detailed treatment of a large subject, with none of the blatant padding that so many other books contain.
Most of the computer books in your average book shop are aimed at people who don't understand or don't recognise any of the words in the title.
Most of these books are easily identified by the bright colours on the cover. Garish covers indicate a sacrifice of style and aesthetics for bookshelf visibility, suggesting that the publisher has no confidence that the book will attract any attention on the basis of its content. These books, designed for impulse buyers, do not have good content, reviews or reputation to recommend them.