4.1 COST MODELS AND BEYOND

 < Day Day Up > 



4.1 COST MODELS AND BEYOND

Why are we still getting it wrong? Part of the answer lies in the type of work that has been done. For many years, cost estimation has been seen in the context of parametric or mathematical models derived from sets of collected data. Examples of such models include COCOMO II, Boehm et al (1) , and SLIM, Putnam (1) . The premise has been that such models can be applied to different organizations and project types without any loss of accuracy, provided you calibrate, of course. Unfortunately, the term "calibration" is often misunderstood or the process of calibration is misapplied. I will discuss calibration in more detail later.

Another problem with the various cost models is that they have, mainly, been driven by a size estimate in terms of Lines of Code. Now more than one project manager has said to me that they find it as difficult to estimate the Lines of Code to drive a cost model as it is to estimate the effort directly. Despite this, cost models are widely used and many find them beneficial.

Within the industry we are starting to realize that cost estimation is a process in its own right and that optimizing that process depends on more than simply adopting a model or cost estimating tool.

Of course the attraction of the model-based approach is obvious. To derive a cost estimate from a mathematical model is inherently satisfying as it smacks of a scientific approach. Such models are also mechanistic in use and their application is repeatable. These are two of the criterion for the automation of a process, so many people quickly realized that such models could be packaged within software and an industry has grown up around this idea.

I am loath to list a set of cost estimation packages as it may be seen as a recommendation when in truth I only have experience of some half dozen such tools. It is also true to say that there is little to choose between the various tools in terms of accuracy. Also, any such listing would be out of date before this material is published, so frequently do new packages arrive in the market place.

The underlying theme of this chapter is to take a different viewpoint. I claim that cost estimation is a discipline and a process in its own right and I hope to give a coherent view of that process. I firmly believe that the commercially available cost estimation tools have a place within that process and I also hope to show where that place is.



 < Day Day Up > 



Software Metrics. Best Practices for Successful It Management
Software Metrics: Best Practices for Successful IT Management
ISBN: 1931332266
EAN: 2147483647
Year: 2003
Pages: 151
Authors: Paul Goodman

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net