Re: RDF and CIDOC CRM

Dominic,

The work done on the CIDOC CRM ontology is very impressive and it served as
a basis for my masters thesis, which focused on converting VRA 4.0 into an
RDF ontology.  My approach to the issue of developing a rich RDF ontology
for cultural heritage items differed slightly in that I focused on using
existing vocabularies (primarily schema.org and FOAF) whenever possible and
only use custom classes/properties where extra granularity was needed.  To
that extent, I also tried to nest custom classes/properties under existing
schema.org/foaf classes or properties.  My thinking focused primarily on
increasing search engine optimization while still retaining the granularity
that is found in detailed domain specific data models such as CIDOC CRM or
VRA.  In my opinion I did not think that "reinventing the wheel" for
properties such as name or title was necessary and I additionally I wanted
the foundation of the model to be based on ontologies that the major search
engines consume.  Below is a link to my website, where you can feel free to
download the ontology I created, the stylesheet that was used to map the
sample data as well as the thesis paper.

http://purl.org/jmixter/thesis/  <http://purl.org/jmixter/thesis/>

If you have any comments please feel free to email.  I know that the VRA
community is strongly considering developing a new version of their data
model that can be published as Linked Data and I would like to speak with
you more about strategies to do so as well as ways to leverage museum data
that is published as RDF.

Thanks,

Jeff Mixter
440-773-9079


On Wed, Jun 12, 2013 at 3:39 PM, Dominic Oldman <doint@oldman.me.uk> wrote:

> In response to today's conversations I would like to celebrate the virtues
> of RDF, particularly when used with a well engineered ontology.
>
> The ResearchSpace project has just completed a stage of work that
> demonstrates both the richness and practicality of RDF and the CIDOC CRM
> ontology. A working prototype shows how a collaborative research
> environment can be constructed exclusively using a triple store and which
> forms the basis for further development towards a production system during
> the year. It serves the British Museum's 2 million digitised records with a
> harmonised dataset from the RKD, with other datasets will be included in
> short course.
>
> The use of CIDOC CRM, the only ontology able to represent the full
> richness of cultural heritage data like the British Museum's collection
> and, at the same time, provide quality semantic data harmonisation over
> entirely different datasets, is achieved with minimal specialisation. This
> provides the basis for practical user applications that work across
> different institutional data sources - with institutional context (or
> knowledge) intact. The project is gradually adding more integrated apps.
>
> The approach to CRM mapping is to provide a choice of constructs that are
> portable (and non-contentious) for use by other organisations for different
> concepts like, production, acquisition, inscription, visual depiction and
> so on. It is the combination of RDF and a strong domain ontology (CIDOC
> CRM) that creates the opportunity for sustainable cross organisation user
> applications.
>
> A video of the search system using condensed CRM relationships for a
> general user interface is available on the home page of
> www.ResearchSpace.org. The search returns objects but could equally
> return bibliographical and biographical data.
>
> I guess my provocation to the list is this. Given the lack of useful,
> sophisticated end user applications that can robustly span different data
> sources, isn't it time to look seriously at ontologies like the CRM that
> provide a solid basis for highly practical solutions for wide ranging
> audiences?
>
> Dominic
>
> Sent from Yahoo! Mail on Android
>



-- 
Jeff Mixter
jeffmixter@gmail.com
440-773-9079