W3C home > Mailing lists > Public > public-lld@w3.org > October 2011

Re: Linked Library Holdings/Items

From: Adrian Pohl <pohl@hbz-nrw.de>
Date: Wed, 19 Oct 2011 09:21:29 +0200
Message-Id: <4E9E9699020000140004264C@agrippa.hbz-nrw.de>
To: <public-lld@w3.org>

in a recent blogpost I tried to line out what is actually  catalogued, that is: gets a record, in traditional library catalogues. For if we want to migrate legacy data to the semantic web we have to deal with this question and not with the question of how concretely to apply the FRBR model to biobliographic description. The background problem of this post is the representation of journals and journal holdings in RDF.

See https://wiki1.hbz-nrw.de/display/SEM/2011/10/17/Serials+and+FRBR

As you can see we also leave out frbr:Expression and frbr:Work as these have no analogy in traditional cataloging.

Regarding "institution authority data", I just published my master thesis on representing libraries and related organisations, their services and collections in RDF.[1] There will be much overlap with DAIA but I actually only use existing vocabularies like vcard, the Organization Ontology or Good Relations in this approach. The thesis is in German but I started to give a summary at [2]. I would really like to establish some best practices for describing institutions and their sites, services, collections, events at the W3C. So, Karen I would definitely be part of a Community Group providing these.

All the best

[1] http://hdl.handle.net/10760/16175

[2] https://wiki1.hbz-nrw.de/display/SEM/Howto+-+Describing+libraries%2C+their+collections+and+services+in+RDF

 >>>Lukas Koster <l.koster@uva.nl> schrieb am 19.10.2011 um 0:43:
> I wrote some blog posts about this issue (FRBR, RDA and Linked Data) 
> recently. Maybe they are of some use here:
> - http://commonplace.net/2011/09/frbr-outside-the-box/
> - http://commonplace.net/2011/03/missing-links/ (I mentioned this one on 
> this list before)
> Jakob is right when he says we don't have Expressions in our catalogues. 
> Nor do we have Works. We only have Manifestations (=bibliographic 
> records) and Items (holdings/items). Works and Expressions are implicit 
> and repeated. It is virtually impossible to add these FRBR entities to 
> existing systems, also with RDA.
> Karen is right when saying that library linked data is about holdings 
> information (unique data) that should be linked to web based 
> bibliographic metadata.
> This means we also need to replace existing introverted catalogue 
> systems with new extroverted systems.
> I will say something like this in my talk about RDA, FRBR and Linked 
> Data at the Dutch RDA seminar November 2.
> Lukas Koster
> On 18-10-2011 18:41, Karen Coyle wrote:
>> Quoting Jakob Voss <jakob.voss@gbv.de>:
>>> Another reason why I do not care about frbr:Expression is that we just
>>> don't have this distinction in our catalogues. We only have
>>> bibliographic records and holdings items. The latter map to frbr:item,
>>> but I am not sure about the former. So I better wait until we have
>>> real world data that covers both expressions and manifestations.
>> IMO, there will ALWAYS be real world data that does not adhere to FRBR's
>> view of resources, in part because that view is overly divisive --
>> essentially, it doesn't allow you to have creators or subjects unless
>> you have a Work entity defined; nor to have a language of text unless
>> you have an Expression. This is overkill for most bibliographic
>> applications.
>> It might be simpler, since your interest is in holdings, to create a
>> non-FRBR item/holdings entity. This entity could include links to
>> institutions along with the usual item identifiers (call numbers, bar
>> codes, and such).
>> What we then need is a good file of institution authority data with
>> identifiers, locations, and contact information.
>> I think that holdings information is absolutely vital to linking
>> libraries to Web resources, at least as important as bibliographic data.
>> After all, what's the use of linking bibliographic data if you can't
>> then find what libraries can provide access? Linking libraries to the
>> Web really means linking library holdings to the Web, with the
>> bibliographic data as the linking method but the goal being the library
>> location.
>> For that reason I am very glad that Jakob has brought up this topic -- I
>> am chagrined that we did not give this topic much importance in the LLD
>> report, but it could become a first project for a community group on LLD
>> if we can find volunteers to lead that effort.
>> kc
Received on Wednesday, 19 October 2011 07:22:38 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:27:44 UTC