Home // International Journal On Advances in Software, volume 15, numbers 1 and 2, 2022 // View article
Authors:
Daniel Adorno Gomes
Pedro Mestre
Carlos Serôdio
Keywords: computational environment; infrastructure-ascode; open science; computer programming; containerization; virtualization; reproducible research.
Abstract:
Nowadays, computational resources are vital practically in all areas of science. At the same time, science's dependence on computing is pointed to by experts as one of the main causes of the reproducibility crisis. Many factors have contributed to the low reproducibility of scientific research. They are related to the cultural aspects of the scientific software's development, the behavior of the scientist-developer, and technical issues. Based on these factors, the authors presented the Environment Code-First (ECF) framework to guide researchers on increasing the reproducibility of their works by developing computational environments that can be easily recreated without manual intervention. The framework's foundation is the Infrastructure-as-Code approach, and it intends to permit other researchers to recreate an environment only by executing a script. A real case is presented, demonstrating the provision of a bioinformatics environment by using the Prokaryotic Genomics and Comparative Genomics Analysis Pipeline (PGCGAP) protocol, and the ECF framework. The paper shows a comparison between these two methods in terms of time-consumption, manual intervention, platform-agnosticism, and portability. The tests perfomed on three different machines demonstrated that there are many benefits on using the ECF framework such as independency of platform, total portability, and practically any manual intervention. Of course, there is a cost, and it is related to the hard work on developing the code that generates the environment. Another point that needs to be highlighted is the time spent and efforts on achieving the necessary knowledge to create those programs.
Pages: 1 to 13
Copyright: Copyright (c) to authors, 2022. Used with permission.
Publication date: June 30, 2022
Published in: journal
ISSN: 1942-2628