Abstract:
Merging agile with more traditional approaches in software development is a challenging task, especially when requirements are concerned: the main temptation is to let two opposite schools of thought become rigid in their own assumptions, without trying to recognize which advantages could come from either side. Mind mapping seems to provide a suitable solution for both parties: those who develop within an agile method and those who advocate proper requirements engineering practice. In this paper, mind mapping has been discussed as a suitable technique to elicit and represent requirements within the SCRUM model: specifically, we have focused on whether and how mind maps could lead to the development of a suitable product backlog, which in SCRUM plays the role of an initial requirements specification document. In order to experimentally assess how effectively practitioners could rely on a product backlog for their first development sprint, we have identified the adoption of mind maps as the independent variable and the quality of the backlog as the dependent variable, the latter being measured against the "function points" metric. Our hypothesis (i.e., mind maps are effective in increasing the quality of product backlogs) has been tested within an existing SCRUM project (the development of a digital library by an academic institution), and several promising data have been obtained and further discussed.
Full Text Link: http://doi.org/10.1109/ICCITechn.2011.6164775
