Can Duplicate Questions on Stack Overflow Benefit the Software Development Community?

Abstract

Duplicate questions on Stack Overflow are questions that are flagged as being conceptually equivalent to a previously posted question. Stack Overflow suggests that duplicate questions should not be discussed by users, but rather that attention should be redirected to their previously posted counterparts. Roughly 53% of closed Stack Overflow posts are closed due to duplication. Despite their supposed overlapping content, user activity suggests duplicates may generate additional or superior answers. Approximately 9% of duplicates receive more views than their original counterparts despite being closed.In this paper, we analyze duplicate questions from two perspectives. First, we analyze the experience of those who post duplicates using activity and reputation-based heuristics. Second, we compare the content of duplicates both in terms of their questions and answers to determine the degree of similarity between each duplicate pair. Through analysis of the MSR challenge dataset, we find that although duplicate questions are more likely to be created by inexperienced users, they often receive dissimilar answers to their original counterparts. Indeed, supplementary textual analysis using Natural Language Processing (NLP) techniques suggests duplicate questions provide additional information about the underlying concepts being discussed. We recommend that the Stack Overflow's duplication policy be revised to account for the benefits that leaving duplicate questions open may have for the developer community.

Ethics in the mining of software repositories

“…As might be expected given that collective recognition of the breadth of potential ethics issues is only relatively recent, we found some, but not extensive, discussion: only one mining challenge (Wilkie et al 2018) discussed ethics issues in some detail, and two papers (Soto-Valero et al 2018;Abric et al 2019) mentioned the anonymity of the underlying datasets in the research that built on them. 35 papers contained discussions of threats to validity (perhaps a section of a paper best suited to the discussion of ethics issues).…”

Looking for related discussions on GitHub Discussions

“…The literature characterize duplicate posts on PCQA forums as: (1) questions that address the same topic, but are not necessarily identical copies (Silva et al, 2018;Wang et al, 2020); (2) questions conceptually equivalent to other questions previously posted (Abric et al, 2019); (3) questions that were already asked and answered before (Zhang et al, 2017;Wang et al, 2020); (4) questions asked to solve the same problem (Ahasanuzzaman et al, 2016); and (5) questions that 'express the same point' (Zhang et al, 2015). We can note the duplicates conceptualization is not a rigid definition; it can depend on human subjective criteria.…”

“…The authors adopted an official Stack Overflow dataset dump to train, optimize, and evaluate the approaches that predict and classify questions as duplicate or not, respectively. Wang et al (2020) Conversely, Abric et al (2019) analyzed how helpful duplicate questions are to the software development community. The authors argue that duplicates on Stack Overflow help the developer community by providing different formulations of the same problem or solution.…”

“…We used the score to set the similarity value between pairs of discussions and to rank sentences with similar meaning (similarity(master, target)). The cosine similarity value outcome is bounded in [0,1] (Abric et al, 2019). Cosine similarity value 1 refers to identical discussion contents, value 0 refers to dissimilar discussions (Abric et al, 2019).…”

“…The cosine similarity value outcome is bounded in [0,1] (Abric et al, 2019). Cosine similarity value 1 refers to identical discussion contents, value 0 refers to dissimilar discussions (Abric et al, 2019). There is no hierarchy or priority relationship between pairs of masters and targets discussions.…”