Canberra.
The final speaker in this session at DHA 2012 [2] is Antonina Lewis, who begins by highlighting the question of entities in describing linked data. Along with additional issues, such as data storage and retention, they raise a range of key questions for the creators, custodians, and curators of linked data.

Importantly, interpretation of data requires context; this is especially true for collections of coded data, where coding schemes and the various provenance of data sources are also crucially important for meaningful interpretation.

Antonina herself works on a rich archive of the source material used for Australian opinion polls from the 1970s to the 1990s, existing across a wide range of print and digital formats (where ‘digital’ means floppy disks!) and containing a very diverse range of data which can hardly be understood without the support of the original opinion researchers (which the project – whose name I didn’t catch, unfortunately – does have, happily).

Michael Jones now takes over, and notes that the aim of the archive is to be valuable for non-scholarly users, too. It is an initiative of the eScholarship Research Centre at the University of Melbourne, and builds on the archival infrastructure available there, using a range of standards for digital archives.