A Probabilistic Method for Predicting Software Code Growth

Abstract

A significant challenge that many cost analysts and project managers face is predicting by how much their initial estimates of software development cost and schedule will change over the lifecycle of the project. Examination of currently-accepted software cost, schedule, and defect estimation algorithms reveals a common acknowledgment that estimated software size is the single most influential independent variable. Unfortunately, the most important business decisions about a software project are made at its beginning, the time when most estimating is done, and coincidently the time of minimum knowledge, maximum uncertainty, and hysterical optimism. This article describes a model and methodology that provides probabilistic growth adjustment to single-point Technical Baseline Estimates of Delivered Source Lines of Code, for both new software and pre-existing reused software that is sensitive to the maturity of their single-point estimates. The model is based on Software Resources Data Report data collected by the U.S. Air Force and has been used as part of the basis for several USAF program office estimates and independent cost estimates. It provides an alternative to other software code growth methodologies, such as Holchin's and Jensen's code growth matrices.