An Analysis of the Requirements Traceability Problem

Author(s): Gotel, O., Finkelstein, A.
Venue: 1st International Conference on Requirements Engineering
Date: 1994

Type of Experiement: Survey/Multi-Case Study
Sample Size: 100
Class/Experience Level: Professional
Data Collection Method: Survey



The authors of the paper surveyed over 100 practitioners in the UK to try to understand difficulties in requirements traceability. They study consists of five hour-long semi-structured sessions involving 37 practitioners, which were audio taped and transcribed; a two-stage questionnaire; and 2 informal interviews. The structured sessions were used to discover how requirements traceability problems are solved, get suggestions for improvement, and help design the questionnaire. The questionnaires were used to rapidly gather data from a large population of practitioners, and the interviews were to validate the answers. For the first round of questionnaires, 80 were distributed and 69% returned. For the second round, 39 distributed, 85% returned.

The results from their studies indicate that the main source of the requirements traceability problem is a lack of a common definition. The authors found many definitions in use. Each definition has a different emphasis, which is purpose-driven, solution-driven, information-driven, or direction-driven. They also identified a diverse group of problems that requirements traceability was attempting to solve: from safety analysis to permitting flexible process modeling. Practitioners were using "requirements traceability" as an umbrella to encompass many problems. Other findings include numerous techniques for providing and developing requirements traceability.

The authors provide a definition for requirements traceability and divide it into two sub-problems: pre-RS and post-RS, which are concerned with the requirements life before and after its inclusion in a requirements specification.