A controlled experiment to assess the impact of system architectures on new system requirements

  • Authors:
  • Remo Ferrari;James A. Miller;Nazim H. Madhavji

  • Affiliations:
  • University of Western Ontario, Department of Computer Science, N6A 5B7, London, ON, Canada;University of Western Ontario, Department of Computer Science, N6A 5B7, London, ON, Canada;University of Western Ontario, Department of Computer Science, N6A 5B7, London, ON, Canada

  • Venue:
  • Requirements Engineering - RE'09 Special Issue; Guest Editor:Kevin T Ryan
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

While much research attention has been paid to transitioning from requirements to software architectures, relatively little attention has been paid to how new requirements are affected by an existing system architecture. Specifically, no scientific studies have been conducted on the “characteristic” differences between the newly elicited requirements gathered in the presence or absence of an existing software architecture. This paper describes an exploratory controlled study investigating such requirements characteristics. We identify a multitude of characteristics (e.g., end-user focus, technological focus, and importance) that were affected by the presence or absence of an SA, together with the extent of this effect. Furthermore, we identify the specific aspects of the architecture that had an impact on the characteristics. The study results have implications for RE process engineering, post-requirements analysis, requirements engineering tools, traceability management, and future empirical work in RE based on several emergent hypotheses resultant from this study.