This paper presents an empirical study on how self- organized software teams could attain high performance using agile and lean practices. In particular, the paper qualitatively examines characteristics of high performance and self- organization in one project team. The case under study is a customer-driven student project, carried out to develop an alpha-version prototype. The paper also studies how certain agile software practices aid in initialising self-organization in the team. The main results indicate that self-organization as supported by certain Agile and Lean practices helps teams in achieving higher performance.
Vaibhav Oza, Petri Kettunen, P. Abrahamsson et al.· arXiv.org· 7 citations
Quality managers, change agents and researchers are often troubled in defining and demonstrating the level of success achieved in software process improvement (SPI) initiatives. So far, there exist only few frameworks for identifying the level of success achieved in SPI. Analysis shows that these frameworks do not provide a comprehensive view from all relevant stakeholders involved in SPI. Early results from an ongoing research effort to discover and operationalise success dimensions are reported. Adapted from the project management literature it is suggested that five dimensions characterise the level of success achieved in SPI: (1) project efficiency, (2) impact on the process user, (3) business success, (4) direct operational success and (5) process improvement fit. Results from an empirical analysis are reported where 23 change agents evaluated the relative level of importance of each dimension. Early results indicate that change agents valued the process user satisfaction the most and the process improvement fit the least. This finding confirms the need of having various stakeholders and dimensions acknowledged in a framework that is used to measure the overall success of an SPI initiative.
This study presents the results of a systematic mapping study on software size estimation metrics and methods. The methods are investigated based on the entity type(s) measured by the method, size attribute type, size measure(s), and theoretical/empirical validation status of the method, and restriction with respect to functional domain type(s) and software development methodology applicability. Based on the results of the survey, we provide a discussion on the current state of the art in software sizing.
Sohaib Shahid Bajwa, Çigdem Gencel, P. Abrahamsson· Joint Conference of Internat...· 10 citations
A new method for surgically removing training examples from a model reveals that as datasets grow, the link between what a model learns and what it produces dissolves.
