Increasing Precision of Automatically Generated Trace Links

1Citations
Citations of this article
9Readers
Mendeley users who have this article in their library.
Get full text

Abstract

[Context and Motivation] In order to use automatically created trace links during a project directly, the precision of the links is essential. Our interaction-based trace link creation approach (IL) utilizes the interactions of developers recorded in an integrated development environment (IDE) while working on a requirement. For this, developers need to indicate the requirement they are going to work on before coding. This approach worked well in an open-source project with developers who were interested in the interaction logs, but did not work well with students who were not particularly motivated to trigger the interaction recording. [Question/problem] Developers often create trace links themselves by providing issue identifiers (IDs) in commit messages. This causes little effort and does not require the awareness for interaction recording. However, as confirmed by recent research, typically only 60% of the commits are linked. In this paper, we study whether and how IL can be improved by a combination with links created by issue IDs in commit messages. [Principal ideas/results] We changed our approach so that interaction logs are associated with requirements based on the IDs in the commit-messages. Thus, developers do not need to manually associate requirements and interaction logs. We performed a new student study with this approach. [Contribution] In this new study, we show that with this new approach and link improvement techniques precision is above 90% and recall is almost 80%. We also show that for our data this is better than using commit-messages only and better than the often used information retrieval-based approaches.

Cite

CITATION STYLE

APA

Hübner, P., & Paech, B. (2019). Increasing Precision of Automatically Generated Trace Links. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 11412 LNCS, pp. 73–89). Springer Verlag. https://doi.org/10.1007/978-3-030-15538-4_5

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Save time finding and organizing research with Mendeley

Sign up for free