Estimating software projects

There are several aspects of software project that can be estimated. Most commonly the target of estimation in software projects is the effort needed to complete the project. Based on the effort estimate, a schedule can be built for the project. In many cases, in order to estimate the effort needed for the project, it's size must be estimated.

The estimation techniques can be roughly divided into two categories: heuristic estimation methods ("expert estimation") and algorithmic models ("formal methods"). Heuristic methods are based on knowledge and experience of one or more individuals, while algorithmic models are more formally specified, and are based on quantifiable input on the project. Commonly it is considered useful to combine several methods, in the best case both expert and algorithmic estimation is used. When combining the methods, the underlying assumptions should be evaluated and assessed when comparing the results from different methods. While many techniques can provide fairly accurate results, the largest benefit can be indeed exploited by understanding how these methods come up with their respective estimates.

I have written extra material for algorithmic models. Hopefully this material, and its references, will help you to understand how function point analysis and constructive cost model should be applied to software project estimation. Also, the Wikipedia article on software development effort estimation is one starting point for more information. For recent research on software project estimation, I recommend to browse through the research articles by professor Magne Jørgensen from Simula Research Laboratory, Norway.