Four years ago, I bought a book entitled, "Mac OSX: The Missing Manual" and noticed it had reached the #1 best sellers slot at Amazon. I remember wondering how an operating system with 3% of the PC Desktop market could sell enough books to rank #1. Then, I realized there I was buying one too. I didn't use a Mac, but my wife bought one and needed to learn this new fangled UNIX desktop. The point? The Missing Manual served a big need - big enough to warrant a #1 best seller.

Back in late 2004, Andy Oram told me that a need existed for a Linux system administration book, similar to AEleen Frish's "Essential UNIX Administration". He asked me to write a proposal. Now, writers hate writing book proposals. If you're published, then you think a publisher should simply say "go for it". That's not how it works.

Shortly after discussing a proposed book with an editor, an author needs to kick it in gear and get the proposal back to that editor in a couple of days. Consider putting the proposal in an editor's inbox like you would a couple of all niters in a data center.

A book proposal consists of several components that ultimately must convince a publisher that a need exists for a work. Take a look at O'Reilly's Proposal Guides. That should give you an idea of the complexity of what it takes to assemble a proposal.

Publishers want to know what purpose the book servers. That means it has to be useful to some audience. The publisher also wants to see a market study to see if similar books exist and if your proposed book differs enough that people will find it unique.

The publisher will also want writing samples, an outline and your resume. Why a resume? Because you're applying for a job.And that job entails throwing a lot of resources at the project.

A publisher takes a risk of financial loss every time he or she accepts an author's proposal. A couple of successful books can cover the costs of many marginal sellers during a single year. Unfortunately, one never really knows if a title will sell, much less become a Best Seller.

The system administration book Andy and I discussed in 2004 took two years to complete. In the technical market, two years is a life time. But then, you could write a book that never ends. Fortunately, Andy found a co-author for our project, Bill Lubanovic. Bill played a valuable role in bringing closure to a book that could have become a never ending story.

Today, we have asked ourselves how did Linux System Administration become a Best Seller? Did a single Slashdot book review do it? Or did a real need exist?

My vote will stay the same as it was in 2004. I needed a different kind of book than the ones available. I needed a Linux system administration book that did for me what "The Missing Manual" did for Mac users. So, I wrote it for me.

Today, before I got word of the book's status, I had it open on my desk using it as a reference to help me upgrade a web server. Maybe Mick Jagger had a good point about getting what you need.