Owen, this is absolutely great! On Thu, Jul 15, 2010 at 1:49 PM, O.Stephens <[email protected]> wrote:
> > > Might be worth having a look at this recent post from the Names project > about their Institutional URIs and recommendations from the NISO working > group on Institutional Identifiers? > http://namesproject.wordpress.com/2010/07/06/identifying-institutions/ > > > * URL/URI of its repository > > * table of contents of theses (this does not exist in many > > repositories). It might be machine extracted if we can get > > the whole repopsitory TOC > > * annotation of each entry's bibliographic metatdata to > > include Openness of each thesis. > > > > The UKETD DC Application Profile uses the dc.rights to store rights > metadata - but I guess we may need something more structured for machine > parsability? What might be useful from the UKETD DC AP is the relatively > well structured 'embargo' metadata. See > http://ethostoolkit.cranfield.ac.uk/tiki-index.php?page_ref_id=47 for more > detail. > > Clearly this is UK based, and it isn't clear how many repositories have > implemented this fully, but it is what EThOS recommends is used, and I think > (from memory) the Embargo fields werer intended to be functional for EThOS. > > Owen > We'll certainly have the challenge of different countries, hierarchy of institutions, disambiguation, etc. I am personally happy to go with DC-level stuff as it can be micro-semantified later and I'd settle for an uncontrolled text string which can be classified by machine-learning. I am generally a fan of folksonomies with machine classification and refactoring, rather than predefined structure unless it's in regulated use. Although all types of bibliographic metadata is complex, I think institution identity is much easier than author identity. P. -- Peter Murray-Rust Reader in Molecular Informatics Unilever Centre, Dep. Of Chemistry University of Cambridge CB2 1EW, UK +44-1223-763069
_______________________________________________ okfn-discuss mailing list [email protected] http://lists.okfn.org/mailman/listinfo/okfn-discuss
