Reproducibility in computational science

The AAAS meeting going on right now includes a symposium on The Digitization of Science: Reproducibility and Interdisciplinary Knowledge Transfer. Mark Liberman is speaking at it, and has posted the symposium abstract as well as some remarks at Language Log.

It is interesting to me to think that digitization has the potential to lead to less reproducibility, but once the point is raised, it's easy to see how that could come about. Clearly as we talk about making data available along with research reports (papers, etc), we should include with the data the scripts, software, and "recipes" (how to use the software) to reproduce the results.

In the comments on today's Language Log post, Mark also provides a link to a previous LL post about a 2008 event in Berlin on a similar topic: Open Data and Reproducible Research: Blurring the Boundaries between Research and Publication That symposium page includes links to presentations.