Chapter in Systems Theory and Practice in the Knowledge Age pp 239–246. [cost to download £19.95]
The multiple dimensions of the factors affecting software development productivity represent a challenge for an interdisciplinary research to their nature. Because of the significant role played by people, software engineering is already one of the computer science disciplines that is closest to the social sciences (Kemerer, 1998:66). Lederer and Prasad (1998:138) report that past published research in that area includes experiments on cost estimation using algorithmic techniques, without formal techniques but only using expert judgment, using analogy or case based reasoning, expert systems, simulation and in a few cases action research and case studies. The majority of these approaches are of a technical nature. Pfleeger (2001) concludes that the statistics for most models are disappointing, indicating that no model appears to have captured the essential characteristics and their relationships for all types of development. Similar concerns about the accuracy of normative models for software development effort estimation have been expressed by Subramanian and Breslawski (1993) and others.
