Apresentado na 8th IFPUG International Software Measurement & Analysis Conference (ISMA) entre 1o de outubro de 2013 por Carlos Eduardo Vazquez no Rio de Janeiro / RJ
The high risks on engaging on a business relationship based solely on benchmark data; the importance to have a statistical perspective to the software production variables behavior and how they correlate to each other; When to prescribe the value of a software project using function points is an economically efficient choice and when it is better to use function points to estimate with an upper and lower band.
Painel
Rules of the Game: Establishing Initial Conditions for Price-per-FP Contracts / As Regras do Jogo: Estabelecendo Condições Iniciais para Contratos com Preço-por-PF.
Learning Objectives / Objetivos de Aprendizagem
To understand the use of statistical tools like regression using the minimum square method, confidence intervals and prediction intervals as a solution to better use of functional size measurement for business purposes. To emphasize the “average” nature of productive indexes and to point out when it is appropriate to count on it to prescribe the value of a software project in a determinist fashion and when a heuristic component is appropriate.
Topic / Tema
Whenever someone presents a price for a software development or maintenance in terms of moneys for function point, he is presenting an average productivity. In the long term he expects to have a balance between demands which costs where underestimated and overestimated.
In many occasions we realize this average productive (price) is derived from market pressure or benchmarking data. As a result, we have been witnessing a series of cases where the contract becomes insolvent and/or poor quality products are delivered.
The use of some simple statistical methods and test may give further direction complementary to market pressures and benchmarking data. Furthermore, it may point out to scenarios where a business model prescribing the amount to be paid is far too risky and a model where the uncertainty is considered as a essential step lowers this risk for all the parts involved.
