In my first essay I mainly focused on emphasizing the impact of sharing raw data and codes in repositories, the importance of clear documentation and benefits of preprints, from the perspective of both individual scientists and communities. In the field of signal and image processing, we rely heavily on experiments that require high quality datasets and algorithms. To carry our research further, we have to have access to the data, as well as the algorithms in previous researches to make fair comparisons and justify improvements. Aside from this practical aspect of our research, it is crucial for us, as it is for the rest of the scientific community, to be able to access recent publications related to our field to advance in the theoretical aspect. Moreover, easy access to publications builds communities faster by increasing chances of interactions, hence strengthening the motivation and development of research, and researchers themselves.
For sharing algorithms and data, I knew only of GitHub, which is an open access platform allowing users to share their projects and provide documentations. GitHub also helps to maintain a version control, and though I had heard about this many times before I was still hesitant to give it a try and have first hand experience on what this "version control" is. Unfortunately I have been one of those people who have many versions of codes, and therefore, troubles retrieving the last version. During workshops I had the chance to have a GitHub account and start using it, and I am now noticing which of my colleagues have shared their repositories along with their research. I do want to become one of them, as now I am very much convinced about the importance of reproducible research. As pointed out in our workshops, we all have hesitations about sharing codes. We worry about being ridiculed, losing time on preparing documentation and cleaning up, not knowing exactly what to share, and to what extent. However, the essence of our work is to develop science. We have to put such personal worries aside and regard the time spent on data sharing as an investment to our practice. As for what to share, we have to inform ourselves on what is within our rights to share. The algorithms we hold the rights of, we should not be hesitant to share. After all, what good is science for if it is carried out only within closed circles?
The fear of being scooped goes hand in hand for sharing data and publishing preprints. Current practices in open science for data sharing and publication have time stamps and license options to protect the authors, which prevent attempts of scooping. Learning about the different types of licenses providing distinct levels of openness was very beneficial, and discovering the rights held by different publishers was as interesting. Other benefits of preprints have been emphasized, such as the difference between the time it takes for one paper to go through a peer-reviewed journal versus a preprint server, as well as the speed of interaction between communities and hence the development of research \cite{polka}. The "buts" of preprints were successfully cleared, giving way to a more efficient way of conducting research.