Best Practices for Computational Science: Software Infrastructure and Environments for Reproducible and Extensible Research

Scholarly dissemination and communication standards are changing to reflect the increasingly computational nature of scholarly research, primarily to include the sharing of the data and code associated with published results. This paper presents a formalized set of best practice recommendations for computational scientists wishing to disseminate reproducible research, facilitate innovation by enabling data and code re-use, and enable broader communication of the output of digital scientific research. We distinguish two forms of collaboration to motivate choices of software environment for computational scientific research. We also present these Best Practices as a living, evolving, and changing document on wiki.