Scientific Software Engineering: Mining Repositories to gain insights into BACARDI

Abstract

For Space Situational Awareness, the German Aerospace Center (DLR) develops the software system "Backbone Catalogue of Relational Debris Information" (BACARDI), which allows for keeping track of resident space objects. BACARDI's key features are automated processing services which produce orbit information and products like collision warnings. We present how we applied new methods of software analytics to the BACARDI project. BACARDI is an example of a complex software system with large development effort carried out by a team of various specialists. Our goal is to design and implement an efficient software development process, balancing the explorative character of a research project and operational requirements (i.e. tailored from official standards in the aerospace domain). Therefore, we established a software development process for the project where we focus on software quality. We applied methods to structure, communicate, and utilize the diverse skills, knowledge, and experience in the team concisely and precisely. After one year of practical utilization, we analyzed the process based on the repository data. By analyzing these data, we assess and prove the effects of the introduced process on the development of a software, which is used in the aerospace domain.