In December 2012 Brown University hosted a fascinating workshop (which I was unfortunately unable to attend) looking at issues in the reproducibility of computational and experimental mathematics, which has many connections to broader questions about scientific data management and cyberinfrastructure requirements to support reproducibility. The workshop web page, which can be found at
http://icerm.brown.edu/tw12-5-rcem
links to many of the presentations and position papers, as well as extensive reference resources.
Clifford Lynch
Director, CNI