Ambiguity in natural language software requirements: A case study

Abstract

[Context and motivation] Ambiguous requirements are often seen as a cause for project failure, however there is little empirical data to support this claim. [Question/problem] In this research we study the effect of a highly ambiguous requirements document on project success. [Principal ideas/results] The studied project was a complex data processing system that took about 21 man year to develop. First, we determined the level of ambiguity by three independent tests. Next, we did a root cause analysis on a selection of the main issues to establish if ambiguous requirements were a significant cause. Surprisingly, this case study shows that only one of the examined failures was caused by ambiguous requirements. Both the independent test team and the third party development team found ways to cope with the high level of ambiguity. For the development team this required a substantial investment to clarify requirements. [Contribution] The main contributions of this paper are the counterintuitive findings, the collected empirical data and the method used to collect these data. © 2010 Springer-Verlag.