Please use this identifier to cite or link to this item: http://hdl.handle.net/1843/31230
Type: Tese
Title: Identifying and characterizing unmaintained projects in GitHub
Other Titles: Identifying and characterizing unmaintained projects in GitHub
Authors: Jailton Junior de Sousa Coelho
First Advisor: Marco Túlio de Oliveira Valente
Abstract: Open source projects are key components of modern software development. Due to the appearance of novel platforms (e.g., GitHub and GitLab) for developing public code, developers has created thousands of open source projects. As a consequence, a significant number of open source projects is also unmaintained. To tackle this problem, in this thesis, we reported a set of quantitative and qualitative studies to help developers to maintain their open source projects. First, we surveyed the owners of open source projects that are no longer actively maintained, aiming to reveal the reasons for stop the maintenance of their projects. As result, we provide a set of nine reasons that motivated them to abandon their projects. Second, we conducted a survey with developers who recently became core contributors of popular GitHub projects. We reveal their motivations to contribute to these projects, the projects characteristics that mostly helped to contribute, and the barriers faced by them. Our key results show that the surveyed developers contributed to the projects because they are using them and need some improvements. The participants also answered that the lack of time of the project leaders was the principal barrier faced by them. Finally, the project characteristic which mostly helped them to contribute was the existence of a friendly community. Finally, in our third study, we propose a quantitative and datadriven model to identify GitHub projects that are not actively maintained. We train the model using a set of 13 features about project activity (e.g., commits, forks, and issues). The model achieved a precision of 80%, based on the feedback of 129 real open source developers and a recall of 96%. We also showed that the model can be used to identify unmaintained projects early, without having to wait for one year of inactivity, as commonly proposed in the literature. Finally, we defined a metric, called Level of Maintenance Activity (LMA), to assess the risks of projects become unmaintained. We provided evidence on the applicability of this metric, by investigating its usage in 2,927 active projects.
language: eng
metadata.dc.publisher.country: Brasil
Publisher: Universidade Federal de Minas Gerais
Publisher Initials: UFMG
metadata.dc.publisher.program: Programa de Pós-Graduação em Ciência da Computação
Rights: Acesso Aberto
URI: http://hdl.handle.net/1843/31230
Issue Date: 30-Aug-2019
Appears in Collections:Teses de Doutorado

Files in This Item:
File Description SizeFormat 
Thesis-JailtonCoelho.pdf2.63 MBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.