MIXED is a project of DANS, Data Archiving and Networked Services. MIXED is to contribute to digital preservation, by dealing with the problem of file formats. Over time, file formats become obsolete. When that happens, the information in such file types is no longer accessible. MIXED follows the strategy of converting files to XML as soon as possible, preferably when data is ingested into the archive. MIXED also converts these XML files to formats of choice by the archive user.

Ben O'Steen's blog post describing the DISC-UK DataShare project's approach at Oxford to incorporating a dataset type (a phonetics database) into a Fedora repository.

Gail Steinhart, co-chair of the working group, forwarded me a link to this paper during the summer, and I’m very pleased to have read it. The group, formed in 2006, has been investigating issues, current activities, and opportunities for the Library to get involved in “digital research data curation.” Thus, it serves as a very useful US equivalent to our DISC-UK State of the Art Review, but also hones in on the specific issues within a given institution, which is what I’d like to help the Information Services do within the University of Edinburgh.

The white paper begins with an environmental scan beyond Cornell, before turning to the strengths and potential areas of collaboration within the University. It looks at the actual and potential role of the academic research library, international organisations such as CODATA, activities in the UK including the importance of Liz Lyon’s 2007 report on roles and responsibilities, the EU DRIVER project, The Australian National Data Service and the activities at Monash University (“noteworthy in terms of utilizing institutional repositories for research data”), and developments in the US including the formation of the federal Interagency Working Group on Digital Data and the DataNet initiative funded by the NSF, as well as recent commercial activities by Sun, Google, and Microsoft. Institutions within the US mentioned for moving forward the state of the art include the San Diego Supercomputer Centre (for SRB, iRODS, and Data Central), Purdue University (for its Distributed Data Curation Centre, D2C2), University of Washington and Johns Hopkins University.

Four US universities are named as pursuing educational opportunities in data curation – Indiana University’s School of Informatics, University of Illinois at Urbana-Champaign, University of North Carolina at Chapel Hill, and Syracuse University.

A section on data curation issues covers financial sustainability, appraisal and selection, digital preservation, intellectual property, confidentiality and privacy, and participation by data owners. The recommendations made by the group include the need to seek out and cultivate partnerships, and the need to develop new services for Cornell researchers.

Neil introduces the report as a whole. The part about research data is extracted below:

The activities of the Alliance Initiative are directed to three areas: First, the partners wish to formulate a common data policy in order to promote both the need for action and to demonstrate the usefulness of primary data infrastructures for scientists and scholars.

Secondly, the partners wish to foster cooperation between scientists and information specialists and to offer funding for pilot projects. Such projects should develop subject-specific standards and methods of data curation and archiving; they should also define the division of labour required in the process.

These steps have the overall goal of establishing a reliable system of digital archives for primary research data, and to ensure that these remain accessible internationally and their data reusable in various interdisciplinary contexts.
Finally, the third and ultimate aim is to establish a system of discipline specific, internationally networked data repositories for primary research data. However,
this task can and should only be tackled when sufficient experience has been acquired from the funding and evaluation of pilot projects. This is to ensure that
the new structures respond to the requirements of the individual subject disciplines and are embraced by them.

Interesting item about how the volume of generated astronomy data has surpassed the ability of scientists to analyse it, and so how they have enlisted the public (amateur astronomers) to help.

An interview in CILIPS Update with DISC-UK members about the field of data librarianship in today's academic environment.

The Social Science Data Archive, DANS, Netherlands, has published quality guidelines for data archiving and "making data future-proof." An innovation is that the 17 guidelines apply in turn to data archives, but also data producers and data users.

JISC has organised a grouping of projects interested in data curation in and out of repositories, including DataShare and the Data Audit Framework. One of the key areas of shared concerns is training.

A report-back on the recent Research Data Management Forum organised by DCC with RIN, with links to key presentations.