Human factors and analytical models in software estimation: an integration perspective

In the field of software estimation, supporters of a rather pragmatic and individualistic approach based mainly on intuition and personal evidence – usually undertake endless discussions with supporters of a more general and theoretical approach based on the collection and statistical analysis of actual data from past finished projects. The present paper supports the idea that the utilization of both the approaches represents a better answer to the ever challenging forecasting needs. The author starts from the definition of a classification framework for estimation methods and for the relative advantages and disadvantages. It then goes through the presentation of a real case of integration between a human approach, based on structured analogy, and an analytical model for the Function Point estimation of a software application. Finally, it ends with the clarification of the interaction of such tools as Early Function Point Analysis, team based estimation, case oriented training, Experience Data Bases and AHP analysis. The final goal which had been set up was the gain of the advantages of the two distinct approaches without suffering the correspondent disadvantages. This is what has been actually achieved in some real environments.