¹Faculty of Engineering and Business. UCINF University. Pedro de Valdivia Avenue 450, Providencia, Santiago, Chile.
²Department of Computer Engineering. University of Santiago, Chile. Ecuador Avenue 3659, Estación Central, Santiago, Chile.

---

Abstract—The development of adequate scientific software within the framework of a research project plays a key role in the success of the research itself. However, not all research teams complete the development of a specific software within the deadlines and with the necessary quality standards. These difficulties have been studied for a lot of years and we can conclude that these applications are difficult to create because defining their requirements is a complex task and usually developers are not fully skilled in construction of these. The following systematic literature review characterizes this kind of development and exposes its difficulties. A number of methodological solutions as well as a series of widespread practices are presented, aimed at improving the development of such kinds of software. A combination of elements is offered which would allow software development teams to choose the most adequate solution according to their specific requirements. The research results show the difference between engineering and scientific disciplines from the type of problems that it solves. Because of this, it is difficult to use a software development technique to work well in both cases and for this reason, the proposals are very useful for the future of scientific software development.

Index Terms—Scientific software development, software development practice.

[PDF]

Cite: Gerardo Cerda Neumann, Héctor Antillanca Espina, Víctor Parada Daza, "Development of Scientific Software and Practices for Software Development: A Systematic Literature Review," Journal of Software vol. 12, no. 2, pp. 114-124, 2017.