Software architecting without requirements knowledge and experience: What are the repercussions?

  • Authors:

  • Remo Ferrari;Nazim H. Madhavji

  • Affiliations:

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

  • Venue:
  • Journal of Systems and Software
  • Year:
  • 2008

Quantified Score

Hi-index 0.01

Visualization

Abstract

Whereas the relationship between Requirements Engineering and Software Architecture (SA) has been studied increasingly in recent years in terms of methods, notations, representations, tools, development paradigms and project experiences, that in terms of the human agents conducting these processes has not been explored scientifically. This paper describes the impact of requirements knowledge and experience (RKE) on software architecting tasks. Specifically, it describes an exploratory, empirical study involving 15 architecting teams, approximately evenly split between those teams with RKE and those without. Each team developed its own system architecture from the same given set of requirements in the banking domain. The subjects were all final year undergraduate or graduate students enrolled in a university-level course on software architectures. The overall results of this study suggest that architects with RKE develop higher-quality software architectures than those without, and that they have fewer architecture-development problems than did the architects without RKE. This paper identifies specific areas of both architecture design as well as the architecture-development process where the differences manifest between the RKE and non-RKE architects. The paper also describes the possible implications of the findings on the areas of hiring and training, pedagogy, and technology. The empirical study was carried out using the ''mixed methods'' approach, involving both quantitative and qualitative aspects of the investigation. A bi-product of this study is an architectural assessment instrument (included in the Appendix) for quantitative analysis of the quality of a software architecture. This paper also describes some new threads for future work.