If your programming environment allows text export of code, you can verify and refine the initial estimates by sampling code for a few completed size elements and comparing these code counts to your estimates.
Remember: the goal is not perfection, but the creation of a practical, consistent, and repeatable process that simplifies future calibration and estimation of similar projects. A line of SQL query code might be more complex than a line written in Visual Basic but when development work is reduced to the smallest identifiable and practical unit of work, these differences will be minimized. The important thing to realize is so long as you are consistent in your measurement approach, differences in complexity will be reflected in the Productivity Index. Complex projects of the same size will take longer to build and will exhibit lower average PIs, but when these projects are used to estimate future projects of a similar nature, their calibrated PIs will automatically “build in” the right amount of time and effort going forward since their increased complexity is reflected in the final project PI.
Measurement is never a perfect endeavor but we learn more about the interplay between various project metrics when we measure all dimensions of a software project than when we count on incomplete measurement, intuition, expert judgment, or rules of thumb for guidance. Armed with a few simple questions and the power of completed project data, project managers will be in a much stronger position when it comes time to negotiate a few extra weeks of schedule flexibility or a reduction in delivered functionality.
Kate Armel is a technical manager with Quantitative Software Management, Inc. She has 12 years of experience in technical writing, metrics research and analysis, and assisting Fortune 1000 firms with software estimation, tracking, and benchmarking. Ms. Armel was the chief editor and a co- author of the QSM Software Almanac.
© Copyright 2011 Quantitative Software Management. Inc. All Rights Reserved.