期刊
IEEE ACCESS
卷 7, 期 -, 页码 61145-61169出版社
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/ACCESS.2019.2914429
关键词
Non-functional requirements (NFRs); quality requirements; iOS; Latent Dirichlet allocation (LDA); Stack Overflow
资金
- National Science Foundation of China (NSFC) [U1636203]
- National Key Research and Development Program of China (NKRDPC) [2017YFB1002101]
Context: Mobile application developers are getting more concerned due to the importance of quality requirements or non-functional requirements (NFR) in software quality. Developers around the globe are actively asking a question(s) and sharing solutions to the problems related to software development on Stack Overflow (SO). The knowledge shared by developers on SO contains useful information related to software development such as feature requests (functional/non-functional), code snippets, reporting bugs or sentiments. Extracting the NFRs shared by iOS developers on programming Q&A website SO has become a challenge and a less researched area. Objective: To identify and understand the real problems, needs, trends, and the critical NFRs or quality requirements discussed on Stack Overflow related to iOS mobile application development. Method: We extracted and used only the iOS posts data of SO. We applied the well-known statistical topical model Latent Dirichlet Allocation (LDA) to identify the main topics in iOS posts on SO. Then, we labeled the extracted topics with quality requirements or NFRs by using the wordlists to assess the trend, evolution, hot and unresolved NFRs in all iOS discussions. Results: Our findings revealed that the highly frequent topics the iOS developers discussed are related to usability, reliability, and functionality followed by efficiency. Interestingly, the most problematic areas unresolved are also usability, reliability, and functionality though followed by portability. Besides, the evolution trend of each of the six different quality requirements or NFRs over time is depicted through comprehensive visualization. Conclusion: Our first empirical investigation on approximately 1.5 million iOS posts and comments of SO gives insight on comprehending the NFRs in iOS application development through the lens of real-world practitioners.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据