A few months ago, my boss asked me to provide some estimates for our software work. I realized that not just me but also most of my colleagues did hard dealing with this issues. A structured approach was needed. Spending some time with Google, I found the book Software Estimation: Demistifying the Black Art by Steve McConnell.
In his book, Steve McConnell starts with building up a basic understanding about the backgrounds of software estimating (e.g. cone of uncertainty). Then he describes various approaches for effort, size, and schedule estimates (e.g. expert estimates, t-shirt sizing). McConnell handles all estimates as probability distributions. This approach allows you to describe the impact of a setting a milestone a few weeks earlier or later in a very confortable way. Finally, the book ends with a section about presenting estimates to managers. Especially this last chapter is a must read for every engineer who is going to be asked to provide estimates.
The book offered me an much improved view on how to deal with estimates. McConnell created a very pratical orientated book which is easy to understand and implement. Just a little background knowledge about statistics may be helpful for fine grasp.
Finally, after reading this book I got a very good understanding what must be done to create good estimates and were we can optimize our processes. Therefore, I want to recommend it to all my readers.
