Um estudo para identificar fatores de influência na produtividade de desenvolvedores em projetos de software

Abstract

Looking to increase competitiveness, many companies invest in improvements in software processes. The application of improvements in software processes brings several benefits (Silveira et al., 2013; Simões et al., 2011; Tonini et al., 2008; Bouer & Carvalho, 2005), such as improvements on process prediction. In literature (Hamdan & Madi, 2011a; Jorgensen, 2011; Lopez-Martin et al., 2012) several negative aspects are identified due to inaccuracies in software projects estimates. Given this scenario, this master work describes procedures to identify factors that affect productivity in modern software projects, which is fundamental for the elaboration of time, cost and effort estimation in software projects. By identifying the factors that affect software productivity, it is possible to understand the behavior of developers and perform actions to reduce productivity instability and consequently reduce the problems faced by companies in this context, which are related in literature by authors such as Hamdan & Madi (2011b), Jørgensen & Sjøberg (2004) and Lopez-Martin et al. (2012). The elaborated procedures use statistical process control, linear regression, Pareto graph and are based on the framework proposed by Florac & Carleton (1999) and Montoni et al. (2007). The procedures were applied in current projects of a software development company in Alagoas and the results brought benefits to the development process used by the company.