Open-source practice in Computer science / Astronomy research
Open Research objectives/practices
- We provide the code and also synthetic datasets to produce the same results as our paper on Github, thus everyone around the world can download and reproduce the same results as our paper.
- To write the code, I used python and C++ programming and some other publicly available toolbox, and I also wrote the code in a way that could be integrated into the same toolbox. Therefore, everyone who can use this open source toolbox also will appreciate our code and openness.
- One of my journal papers and one conference paper are totally open access.
- I also put the other journal paper on Arxiv, thus although it is under review, everyone can read about our work and learn about new developments even before its publication.
Introduction
As I started my Ph.D. I know that I like to share my research with the world and I don't like to keep it only in an academic atmosphere. Thus, I first start to use GitHub and GitLab as a version control tool but always have this opportunity in mind that I will also have time to share my code, data, and paper with everyone as far as it is allowed by my contract and also my employer (university). This was the reason to code in Python and C++ instead of MATLAB (commercial) and use open source libraries. Besides, as I described before, I also tried to keep my research output either open access or on Arxiv, thus it will reach a greater audience.
Furthermore, by presenting many talks and attending workshops I also spread our recent work verbally to the aimed society. I investigate astronomical datasets as a computer scientist, hence it is crucial to introduce these new tools to the astronomy society, and make it frictionless for them to use our newly developed tools.
Consequently, presentation is also vital to reach these scientists.
Motivation
The motivation for open research is four folds for me:
- Reaching a larger audience by providing the scientific output as open access papers, open-source codes, and also datasets with known ground truth.
- By uploading papers on Arxiv, I also make sure during the process of article review no one else can claim my idea.
- Providing the code and dataset for a paper on a repository is advantageous to advertise the scientific work even more since many people can download and run the code in a short time. Besides, I get attention also from researchers or programmers from the outside academy and I can present it as a sample of my coding skills.
- Extra materials such as code, dataset, talks alongside the open access papers help the new researcher in the same field to get the idea more straightforward. It is specifically important for me because I learned several new techniques and methods using these open repositories provided by others, hence I feel I should pay my tax to the community.
Lessons learned
As a young researcher, I was thrilled when I realized that the University of Groningen has an open-access contract with many journals related to computer science, and this improved the number of views for my work. I am also privileged that my supervisors support my idea of sharing code and materials on GitHub. However, at the same time, I should spend extra time to make the code and other materials clean and readable for every outside project user and sometimes it is not appreciated by my supervisors. Moreover, sharing a paper on Arxiv may introduce some risks because even now some journals don't accept papers previously uploaded there as new work.
URLs, references and further information
| Last modified: | 16 March 2022 11.23 a.m. |