Combining Perceptions and Prescriptions in Requirements Engineering Process Assessment: An Industrial Case Study

  • Authors:
  • Nannette P. Napier;Lars Mathiassen;Roy D. Johnson

  • Affiliations:
  • Georgia Gwinnett College, Lawrenceville;Georgia State University, Atlanta;University of Pretoria, Pretoria

  • Venue:
  • IEEE Transactions on Software Engineering
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

Requirements engineering (RE) is a key discipline in software development and several methods are available to help assess and improve RE processes. However, these methods rely on prescriptive models of RE; they do not, like other disciplines within software engineering, draw directly on stakeholder perceptions and subjective judgments. Given this backdrop, we present an empirical study in RE process assessment. Our aim was to investigate how stakeholder perceptions and process prescriptions can be combined during assessments to effectively inform RE process improvement. We first describe existing methods for RE process assessment and the role played by stakeholder perceptions and subjective judgments in the software engineering and management literature. We then present a method that combines perceptions and prescriptions in RE assessments together with an industrial case study in which the method was applied and evaluated over a three-year period at TelSoft. The data suggest that the combined method led to a comprehensive and rich assessment and it helped TelSoft consider RE as an important and integral part of the broader engineering context. This, in turn, led to improvements that combined plan-driven and adaptive principles for RE. Overall, the combined method helped TelSoft move from Level 1 to Level 2 in RE maturity, and the employees perceived the resulting engineering practices to be improved. Based on these results, we suggest that software managers and researchers combine stakeholder perceptions and process prescriptions as one way to effectively balance the specificity, comparability, and accuracy of software process assessments.