Lost Science – protecting research data through improved archiving
A recent article in Eos, the journal of the American Geophysical Union, reports that considerable NASA data has been lost, or is inaccessible, because insufficient documentation was kept about various missions over the past 50 years. Key challenges highlighted include the lack of a Mission Document Archive, Hardware & Software Migration complications, inaccessibility of key science observation design documentation and Metadata or lack of it! It illustrates the expense and complexity of data management. It also indicates that neither non-repeatability nor high costs associated with the collection of data (with NASA missions costing millions) are an indication of whether researchers will have access to good digital curation and preservation infrastructure for their research data collections. The clear message is keep good quality documentation with your research data so that it can be understood and stand the test of time. Download and read this free article for more detail: http://onlinelibrary.wiley.com/doi/10.1002/2013EO370006/abstract
The other message(s)for me are ‘lack of funding’ and ‘archival expertise on the team’. Karen E. Simmons advocates adequate resources to curate post-mission data, and who are dedicated to follow-up its preservation and timely availability. She addresses the funding problem in suggesting a post-mission archiving grant, which is an attractive proposal that can be applied to many research data collection ‘archives’.
For example, ‘living’ resources based on archival / research data that is dynamic, i.e. changing over time, for example a biographical encyclopedia, require ongoing maintenance (curation of data) to keep the data set up to date, ensuring its validity and high quality as an educational resource. Unfortunately, funding for such ongoing curation is most often neglected,and not necessarily limited to small projects.
What also struck after reading the article was that the 20/20 hindsight vision in the description of the lack of processes in place that lead to the loss of the data is, on a positive note, a good case of ‘learning from mistakes’. The admission of the errors made is a first step in acknowledging the need (and desire) to look for better ways.