Recently, I was asked to explain briefly what metadata is to a group of scientific researchers interested in fulfill the recent NSH’s decision that grants include a data management plan. Let me backup a little and explain this data management plan that is expected to be a part of NSF grant applications.
Back around May 10, 2010, the NSF announced that all proposals include a data management plan. This plan covers five sections in two pages. The second section concerns standards, which includes metadata.
Thus far, several institutions, including the NSF, have provided some explanations and examples. For the NSF, you can find their announcement and a FAQ link at: http://www.nsf.gov/eng/general/dmp.jsp. On the one hand, these examples provide a general direction and approach to formulating a data management plan. On the other hand, these examples are vague, providing only a brief overview of what is expected. Furthermore, the examples used are still only proposals. As a result, data management plans are in their infancy. More importantly, the role of metadata within those data management plans has yet to formalized.
My question are: how do help researchers write about metadata in section two of their data management plan? Is it sufficient to instruct research as to what metadata is and why it is important in general? Does the motto “data about data” have other uses than a strikingly empty tautology meant to move to a slide that presents a more meatier definition of metadata as “structure information” for instance? Should institutions move beyond the role of instruction and into consultation? Should institutions create repositories built specifically for their researchers’ needs? Or should institutions try to juggle a little of each of these aspects and perhaps something else?
Granted, not all institutions have the resources to create their own repositories. Moreover, a home grown repository might not be a suitable solution. There are a number of great subject repositories out there – check out OpenDOAR.
Providing workshops and offering consultations can provide help for researchers who are learning the same as we are about data management plans. This is made possible because the library (or similar institution) has specialists in the five sections (format standards, metadata standards, privacy/copyright/use/re-use, preservation, and types of data. All in all, libraries find themselves at a new and exciting frontier with this topic as shown by the ARL eScience institute. In this, metadata librarians have a crucial role to play. For research to be discoverable and searchable, metadata is going to be the driving force. But how do we do this? More than likely, there will be varying approaches because not one size fits all. How will you approach metadata in data management plans and eScience?