Abstract :
[en] [Context and Motivation] User Stories (US) are often used as requirement representation artifacts within agile projects. Within US sets, the nature, granularity and inter-dependencies of the elements constituting each US is not or poorly represented. To deal with these drawbacks, previous research allowed to build a unified model for tagging the elements of the WHO, WHAT and WHY dimensions of a US; each tag representing a concept with an inherent nature and defined granularity. Once tagged, the US elements can be graphically represented with an icon and the modeler can define the inter-dependencies between the elements to build one or more so-called Rationale Trees (RT). [Question/Problem] RT and their benefits have been illustrated on case studies but the ability to easily build a RT in a genuine case for software modelers not familiar with the concepts needs to be evaluated. [Principal ideas/results] This paper presents the result of a double exercise aimed to evaluate how well novice and experienced modelers were able to build a RT out of an existing US set. The experiment explicitly forces the test subjects to attribute a concept to US elements and to link these together. [Contribution] On the basis of the conducted exper-
iment, we highlight the encountered difficulties that the lambda modeler faces when building a RT with basic support. Overall, the test subjects have produced models of satisfying quality. Also, we highlight these necessary conditions that need to be provided to the lambda modeler to build a consistent RT.
Scopus citations®
without self-citations
13