Over the last five years, the level of interest in Linux and open source software has risen dramatically. Not coincidentally, the number of books on Linux and open source have also risen. The number of titles available leaves the would-be buyer in shelf-shock.

If you want to learn the intricacies of Linux administration, should you go to a training class or get mentored, self-teach using the Google method and experimentation, buy a book, or a combination of all of them? Training and mentoring are valuable but only available at a high cost and for a limited time. Students can ask their friendly all-knowing sys-admin type the same question so many times. Formal training only lasts for a limited duration.

Another method, which I'll call the Google method, is interesting because it seems as though the amount of information available is limitless. Using the Google method, the learner queries for whatever information he or she needs, on demand. Naturally, the Google method isn't limited to the Google search engine; the combination of Google's main search along with Usenet searching provided through Google Groups (formerly Deja News) make Google the first choice of many searchers.

Unlike a mentor, Google doesn't get tired of answering the same question over and over and it's available anytime for free with an Internet connection. However, the Google method can't be wholly relied upon as a means for learning Linux and open source. One problem is quite obvious: What if the Internet connection isn't available and you need to query for information on how to fix the Internet connection? Sure, you could have a backup Internet connection but that isn't always the case.

A more subtle problem with the Google method of training is that there is a certain level or depth at which information is not readily available. Many times, finding the information is possible but not available to searchers using their query skills. Other times the information is simply just not there.

Anyone who has ever played in the annual World's Largest Trivia Contest in Stevens Point, Wisconsin, will attest to the limits of the Internet query. Books provide a noteworthy combination of both mentoring and the Google method all wrapped into one. Not only does the book contain information based on the experience of the author, but a book is also highly available, meaning a book doesn't go down for routine maintenance nor, barring natural disaster, does a book suffer from a denial of service attack. However, I do find that if I'm not within a certain proximity to a book, there is an effective denial of service.

A service like O'Reilly's Safari Bookshelf provides a unique solution to this form of DoS attack though. Using Safari I can read full versions of many popular titles, not just O'Reilly published works. I frequently find that not only do I purchase a book for reference, I'll also read the book on Safari, thus solving the proximity problem suffered by the physical copy of the book. As stated in the beginning of this article, these methods are not mutually exclusive.

Just because I receive formal training or bug a smart geek doesn't preclude me from using Google as well as a book. However, I find that when I need in-depth knowledge of a given subject, a book is almost always the best source. A key to obtaining this knowledge is finding the right title in a given subject area. With the explosion of titles in nearly every area of Linux and open source, it's easy to buy the wrong title. There are relatively few well-written titles on any given subject and there are even fewer essential books in a given area. You might not know that an essential title exists in a certain area and find you have to buy two or more books just to cover that subject.

Reliance on a feedback system like Amazon.com's is helpful but has its limitations due to the seemingly unmoderated nature of the feedback. Too often I see feedback on books that haven't even been written, much less released, feedback on older editions of titles or, worse yet, feedback that has nothing to do with the title at all. I've even seen a post where the apparent reviewer selected a single-star (the lowest rating) and posted that he was located outside the U.S. and looking to purchase this book but didn't have a credit card.

When reading some of the reviews, you have to wonder about the motivation of the anonymous reviewer. Was the review posted by a competing author or an estranged spouse or did the title truly deserve the treatment given? The subject of un- or barely moderated reviews aside, knowing what titles to purchase is a difficult task indeed. Some of the titles being printed are nothing more than rehashes of old material while some of the titles truly stand out.

The January issue of LinuxWorld Magazine is our "Best of..." issue where we will examine some of the best and most popular Linux and open source products and projects. For that issue, I'm putting together a list of some of the best books for learning Linux and open source software, tools, and administration techniques. I've already started the list and I'd welcome your suggestions. What books have helped you learn Linux and open source software? What books do you consider to be essential to a new Linux administrator?

Please take a moment to participate by posting a response in my blog at www.braingia.org/webnotes/. I've created a topic for this entitled "Essential Books." I'll examine as many of the books as I can for the article.