A model for estimating software size for large-scale business applications

Software measurement is fundamental to the software engineering discipline. In particular, software sizing at the very early stages of the software development life cycle has been one of the critical issues in the information system field. In order to plan and develop information systems successfully, it is necessary to obtain an initial estimate of the size of the system being undertaken. Even though the field of software size measurement and prediction has made progress in the last decade, it cannot yet provide a reliable approach for widespread applications such as client/server applications and object-oriented technology. This dissertation develops a methodology that gives a more reliable and accurate predictor of software size at an early stage of the systems development process.; The proposed model called Product Points Model (PPM) mainly applies to large-scale business applications. It is a two-staged model. In the first stage, Base Product Points (BPP) determine the sizing at a logical level that is independent of information technology. In the second stage, a secondary group of metrics called Enhanced Product Points (EPP) is measured at the physical level. Albrecht's Function Point analysis (FPA) and other FPA variations such as Symons' Mk II FPA are critically reviewed for the proposed model. An empirical investigation was conducted for validating this model using empirical project data.