Functional size measurement for real time and embedded software

Abstract

Measures can be used to quantify software products as well as the process by which they are developed. Once these measures are obtained they can be used to build cost estimation models and productivity models. Through this perspective, a key measure is the size of a software product. There are basically two kinds of size measures: Technical size measures are used to quantify software products and processes from a developer's point of view. These can be used in efficiency analysis to improve the performance of the design. Functional measures are used to quantify software products and services from a user's perspective. Being independent of technical development and implementation decisions, functional measures can thus be used to compare the productivity of different techniques and technologies. In this context, organizations frequently use functional size measurement methods to quantify software products included in their outsourcing contracts. Function Points Analysis (FPA), first introduced in 1979 by Allan Albrecht of IBM (Albrecht, 1979), is an example of a functional size measure. FPA measures the size of software in terms of its delivered functionality to the end-users by quantifying such objects as inputs, outputs and files. Since FPA was developed for an MIS 1 environment, its concepts, rules and guidelines are adapted to software typical of that environment. FPA has been criticized as not being universally applicable to all types of software (Conte, 1986; Galea, 1995; Grady, 1992; Hetzel, 1993; Ince, 1991; Jones, 1988; Jones, 1991; Kan, 1993; Whitmire, 1992). A problem with the function point approach is that it assumes a limited band of application types: typically, large file-based systems produced by agencies such as banks, building societies and retail organizations, and is unable to cope with hybrid systems such as the stock control system with a heavy communication component. (Ince, 1991, page 283) When FPA is applied to such software, the measurement results do not constitute an adequate size measurement.In 1997, a new extension to Function Points (Full Function Points - FFP) was developed for measuring the functional size of real-time software to address weaknesses of IFPUG's Function Points. The International Software Benchmarking Standards Group recently accepted FFP as a new measurement standard for real-time software.This mini-tutorial will present the key concepts of this extension as well as the results of the measurement field tests in different organizations. The ability of FFP to adequately capture the functional size of real-time software will be illustrated by FFP and FPA measurements taken on the same software products.