Header

UZH-Logo

Maintenance Infos

Keeping Evolving Requirements and Acceptance Tests Aligned with Automatically Generated Guidance


Hotomski, Sofija; Ben Charrada, Eya; Glinz, Martin (2018). Keeping Evolving Requirements and Acceptance Tests Aligned with Automatically Generated Guidance. In: 24th International Working Conference on Requirements Engineering: Foundation for Software Quality (REFSQ 2018), Utrecht, 20 March 2018 - 22 March 2018, 247-264.

Abstract

[Context and motivation] When a software-based system evolves, its requirements continuously change. This affects the acceptance tests, which must be adapted accordingly in order to maintain the quality of the evolving system.
[Question/problem] In practice, requirements and acceptance test documents are not always aligned with each other, nor with the actual system behavior. Such inconsistencies may introduce software quality problems, unintended costs and project delays.
[Principal ideas/results] To keep evolving requirements and their associated acceptance tests aligned, we are developing an approach called GuideGen that automatically generates guidance in natural language on how to modify impacted acceptance tests when a requirement is changed. We evaluated GuideGen using real-world data from three companies. For 262 non-trivial changes of requirements, we generated guidance on how to change the affected acceptance tests and evaluated the quality of this guidance with seven experts. The correctness of the guidance produced by our approach ranged between 67 and 89 percent of all changes for the three evaluated data sets. We further found that our approach performed better for agile requirements than for traditional ones.
[Contribution] Our approach facilitates the alignment of acceptance tests with the actual requirements and also improves the communication between requirements engineers and testers.

Abstract

[Context and motivation] When a software-based system evolves, its requirements continuously change. This affects the acceptance tests, which must be adapted accordingly in order to maintain the quality of the evolving system.
[Question/problem] In practice, requirements and acceptance test documents are not always aligned with each other, nor with the actual system behavior. Such inconsistencies may introduce software quality problems, unintended costs and project delays.
[Principal ideas/results] To keep evolving requirements and their associated acceptance tests aligned, we are developing an approach called GuideGen that automatically generates guidance in natural language on how to modify impacted acceptance tests when a requirement is changed. We evaluated GuideGen using real-world data from three companies. For 262 non-trivial changes of requirements, we generated guidance on how to change the affected acceptance tests and evaluated the quality of this guidance with seven experts. The correctness of the guidance produced by our approach ranged between 67 and 89 percent of all changes for the three evaluated data sets. We further found that our approach performed better for agile requirements than for traditional ones.
[Contribution] Our approach facilitates the alignment of acceptance tests with the actual requirements and also improves the communication between requirements engineers and testers.

Statistics

Downloads

3 downloads since deposited on 27 Mar 2018
3 downloads since 12 months
Detailed statistics

Additional indexing

Item Type:Conference or Workshop Item (Paper), refereed, original work
Communities & Collections:03 Faculty of Economics > Department of Informatics
Dewey Decimal Classification:000 Computer science, knowledge & systems
Language:English
Event End Date:22 March 2018
Deposited On:27 Mar 2018 08:35
Last Modified:13 Apr 2018 11:48
Publisher:Springer
OA Status:Closed
Other Identification Number:merlin-id:16279

Download