Patterns in Scientific Software Development [article]

Boyana Norris
Scientific software development has increasingly adopted modern software engineering practices and tools, which collect a wealth of data. In this talk, we describe how we analyze some of this data to identify a number of development patterns in ECP and other scientific software as part of the DOE IDEAS project. For example, we use git commit history to identify domain champions, developers who have near-exclusive control over significant portions of a project's code over extended time periods.
more » ... nded time periods. Our overall goal is to enable an automated pattern-based analysis of the development processes in scientific software projects, hence providing actionable information to the teams on a regular basis.A talk presented at the SIAM CSE21 conference, MS187: Data-Driven Analysis of Scientific Software Quality, Availability, and Development Productivity.Authors: Boyana Norris and Bosco Ndemeye, University of Oregon, U.S.; Armando Acosta, California State University, Fullerton, U.S.; Stephen Fickas, University of Oregon, U.S.; Kanika Sood, California State University, Fullerton, U.S.; Anshu Dubey, Argonne National Laboratory, U.S.
doi:10.6084/m9.figshare.14151794.v2 fatcat:eabajmj6t5fznetpjalas4k6k4