Dear members of the OGF Digital Repositories Research Group,
thanks to all who participated at the discussions at OGF26 last week.
Please find a quick write-up of my impressions from the session. It is
hard to squeeze all the issues into just a few sentences, but please do
not hesitate to add / correct the lines below. We intend to publish this
text, so please make sure to contribute.
We hope to see you and all those who could not participate this time at
future events. Also, we are still collecting metadata scenarios and are
looking forward to your contributions. These scenarios are the basis for
the OGF repository work.
best wishes,
a
--------------------------------------
The latest meeting of the Digital Repositories Research Group (DR-RG)
took place at OGF26. The conference saw a much lower participation than
earlier OGF's, yet there were 14 participants from various backgrounds
contributing to a lively discussion. In fact, we had to stop
mid-sentence, but will continue the discussion at upcoming events. As
future venues for the DR-RG were suggested
* Digital Curation Conference, London, 1st week of December
* IEEE e-Science, Oxford, 2nd week of December
The OGF26 session featured two presentations from Jane Mandelbaum
(Library of Congress) and Richard Marciano (UNC). A common theme were
the interoperability requirements for metadata (e.g. on item and
collection level) to enable interaction between different communities,
organisational contexts, and systems. Herewith just a few points that
were picked up during the discussion. Slides can be found
http://forge.gridforum.org/sf/docman/do/listDocuments/projects.dr-rg/docman…
Jane Mandlebaum from the Library of Congress gave an impression of the
developments in one of the most experienced fields in information
science. For decades discussion among cultural heritage institutions
were dominated by the idea of the "perfect set" of metadata that is
globally applied, by information managers and users alike. This notion
of global unification by and large failed. Jane described how future
information environments are likely to allow for variations in metadata
standards and encourage participation rather than enforce it. In these
environments, infrastructure provides the tools to mediate between
different data, to simplify metadata creation and to raise its quality.
(Seemingly simple things can turn out utterly complex: e.g. how to
convert existing long titles of history videos into short titles as
required by YouTube.)
This point resonated with other participants who also agreed that
automatic metadata creation/conversion, and encouraging early metadata
creation were amongst the key opportunities in their communities.
Richard Marciano from the iRODS team DICE focused on policies for
collection management. In a talk seeded with the multiplicity of
experiences from iRODS installations, Richard identified standardised
mechanisms for describing collection policies (e.g. replication
criteria, legal constraints) as a key challenge for the repository
community. Without collection policy standards, important contextual
information may be lost when migrating or exchanging collections between
federated repositories.
All participants - cross-community activities in national grids (e.g.
D-Grid), at national service centers (e.g. NGS), as well as in
enterprise environments - agreed that requirements of users or user
communities differ greatly and need to be addressed individually (i.e.
dealing with a requirement once it's there, not creating the perfect
set, or the perfect service upfront). It is the information environment
between dedicated repositories that enables the necessary flexibility
over time through mediation, interoperation, added-value services, etc.
OGF Digital Repositories Research Group
http://www.ogf.org/gf/group_info/view.php?group=dr-rg
Hi,
enjoyed the talks at this mornings session at OGF.
I thought the following piece of work might be interesting to add to the
discusssion:
The CARMEN project from the UK has created something called MINI:
Minimum Information about a Neuroscience Investigation.
"This framework represents the formalised opinion of the CARMEN
consortium and its associates, and identifies the minimum reporting
information required to support the use of electrophysiology in a
neuroscience study, for submission to the CARMEN system. These
guidelines are a work in progress and will evolve."
http://www.carmen.org.uk/standards
They identified that metadata collection was an issue, and have proposed
a way this might be mitigated for their community.
chheers,
neil
--
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.
