Thumbnail Image

Function Point Metric Calculation and Sensitivity Analysis

Tantikul, Tongchit
Managing software during development and maintenance is one of the most troubling aspects of modem information technology. Documented historical records indicate that poorly managed projects can run into cost overruns and schedule slippages and, in some cases, may be abandoned. Careful attention to quality control by using trained specialists, modem software estimation tools, and effective planning tools can immunize software development companies against some of the common sources of software disaster. The software development process can thus be managed more thoroughly or even controlled. The size of a software document at varIOUS stages of development can be measured using a number of alternative methods. Lines of code is probably the size measure most widely known, but it has no universally accepted definition. Function points is one ofthe better-known and widely-used metrics for measuring software size. The objective of this thesis was to study and analyze the sensitivity of the function point metric. The work done consists of collecting the size in terms of the number of Lines of Code (LOC) of various programs, converting the lines of code to function points, calculating function points from the program documentation independently of the previous calculation, comparing the estimated function points and the actual one, and investigating the sensitivity of function point parameters by plotting a graph. The Line of Code was measured by the CSlZE software tool, and the function points was calculated by the COSMOS software tool. It was found that among all the function point parameters (i.e., external inputs, external outputs, internal logical files, external interface files, and external inquiries), the internal logical files is more sensitive than the rest.