Reproducible Computational Science - a Guideline
More than 70% of researchers have tried and failed to reproduce another scientist’s experiments, and more than half have failed to reproduce their own experiments. [1]
In the field of modeling and simulation, we can and should ensure reproducibility with a fully automated workflow [2] from the input scripts to a finished document. This allows for external verification, greater trust in your results and comparative ease in reconstructing details of your own work in the future.
We would like to establish a working group within BAM that develops a wiki/handbook of possible approaches to tackle these challenges. Everyone should be able to access these documents and will be able to use these as a basis for future projects. The competence center modeling and simulation will coordinate the efforts towards this goal, provide guidance and the infrastructure (e.g. a wiki, repositories). For completing the tasks itself, however, we mainly rely on the work of the interested participants.
Contents: