Measuring the Effects of ObjectOriented Frameworks on the Software Process, submitted for.pdfVIP

Measuring the Effects of Object-Oriented Frameworks on the Software Process 1 Simon Moser, Oscar Nierstrasz University of Berne2 Abstract. A field study of over thirty projects using Object Technology has shown that the availability (or absence) of reusable frameworks has substantial productivity impacts. This can make it more difficult to reliably estimate the size and cost of such projects early in the software process. The newly proposed System Meter method tack- les this problem by distinguishing functionality to be implemented from functionality supported by reusable components. It therefore yields more uniform and predictable productivity measures. Moreover, it can also be applied already after a preliminary analysis phase, in contrast to the more traditional Function Points approach. 1 Introduction “How much will it cost?” “How long will it take?” These questions are notoriously difficult to answer accurately and early in the software process. Software metrics tackle this problem by assuming a statistical correlation between the “size” of a software project and the amount of effort that is typically required (in a given context) to realize it. A metric of size that will be useful for cost estimation must be based on the inherent complexity of the system to be devel- oped (rather than on, say, lines of code, since no lines of code exist yet). Such metrics exist, and have been applied with varying degrees of success in the past, but the nature of software development has been changing, and some of the assumptions behind the established cost es- timation techniques are slowly being invalidated. How can these techniques keep up? The trend towards “open syste