Using Test Cases to Size Systems: A Case Study

摘要

Productivity, quality and speed are the three key areas that all technology organizations strive to understand. Yet despite the keen interest, the ability to quantitatively measure these aspects of performance often eludes us. Central to the issue is that software development varies significantly in size from project to project making comparisons challenging. Measurement of the size of software is a key component to comparing dissimilar projects. Without a means to normalize for size, drawing conclusions between projects is a nearly impossible task. While solutions have existed to size systems in both lines of code (KLOC) and function points (FP), this paper explores a successful alternative approach to sizing that exhibits the benefits of existing methods with less incentive to manipulate the measurement system and a significantly lower cost. This paper proposes that the number of test cases is a viable measure of system size, available early in the process.