- From: Jodi Schneider <jodi.schneider@deri.org>
- Date: Mon, 13 Dec 2010 16:50:36 +0000
- To: Jakob Voss <jakob.voss@GBV.DE>
- Cc: "public-lld@w3.org" <public-lld@w3.org>
- Message-Id: <A7F23AE7-F64E-41E8-8DE0-579C87E61622@deri.org>
Hi Jakob, Thanks for this -- and especially for delineating the entities of interest. On 13 Dec 2010, at 16:22, Jakob Voss wrote: > Hi, > > I quickly hacked an RDF wrapper to some of our bibliographic databases. just a few million records. There is no SPARQL, content negotiation etc. because I first want to fix the model. > > Maybe you can have a look at this entry (in Turtle syntax) > > http://uri.gbv.de/database/opac-de-7:ppn:609060929 > > I found there are at least four entities to model, but the existing ontologies are pretty disappointing for library data: > > > 1.) The record in a library catalog: > http://uri.gbv.de/database/opac-de-7:ppn:609060929 A few comments: dc:creator "Michael Waitzmann; Roland Heinzmann" ; I'd expect two dc:creator lines here, one per creator. dc:publisher "Karlsruhe : Landesanstalt für Umweltschutz" Ideally, the publisher's name should ideally not be prefaced with the location. Similarly, I suspect that GUK is a separate term (short title) here: dcterms:title "Göttinger Universitätskatalog (GUK)" ; > > > 2.) The publication that is described by the library record. I used a blank node to model this, because known identifiers (such as ISBN) are not as unique as they seem to be. I hesitated to make this a bibo:Document from the Bibliographic ontology, because the latter says "A document (noun) is a bounded physical representation of body of information". This is useful feedback; I hope you'll share it with BIBO! > But we also have non-physical publications. The right class could be frbr:work, frbr:expression, or frbr:manifestation, but also skos:Concept, depending on the record. > > > 3.) The library catalog as whole: > http://uri.gbv.de/database/opac-de-7 > > The best classes I found to models library catalogues are: > > cdtype:CatalogueOrIndex as described at > http://dublincore.org/groups/collections/collection-application-profile/ > > and void:Dataset from voiD as described at http://vocab.deri.ie/void/ > > > 4.) The library as institution: > info:isil/DE-7 > http://lobid.org/organisation/DE-7 > > ISIL is an *international* system to uniquely identify libraries, so we should use this instead of local systems like the American OCLC- or WorldCat identifiers. I had to look up ISIL, so I'll save others the effort! The standard is closed [1] and I found useful info via the Danish Agency for Libraries and Media: http://biblstandard.dk/isil/ -Jodi [1] http://www.iso.org/iso/iso_catalogue/catalogue_tc/catalogue_detail.htm?csnumber=52666 > > > Note that the pysical holding is *not* included in this model yet - some people think that physical objects are the core of library data, but they are only the most visible part. Apparently current ontologies, that aim at library data, have a narrow focus on physical documents. > > Jakob > > -- > Jakob Voß <jakob.voss@gbv.de>, skype: nichtich > Verbundzentrale des GBV (VZG) / Common Library Network > Platz der Goettinger Sieben 1, 37073 Göttingen, Germany > +49 (0)551 39-10242, http://www.gbv.de >
Received on Monday, 13 December 2010 16:51:11 UTC