Re: Metadata about single triples

Hi Carsten,

On 2/22/2012 12:02 PM, Carsten Keßler wrote:
> Dear LODers,
>
> we are currently working on a project for the United Nations Office
> for the Coordination of Humanitarian Affairs (OCHA) in Geneva to
> develop a Humanitarian Exchange Language (HXL). Some information about
> the project is available at https://sites.google.com/site/hxlproject/.
>
> One of the core components of HXL will be an RDF vocabulary to
> annotate the data that are exchanged between humanitarian
> organizations. The current draft is available at
> http://hxl.humanitarianresponse.info. It is far from complete, but I
> think it already shows where we want to go with this. Any feedback on
> the vocabulary draft is very welcome, of course.

At a first glance, your ontology looks very interesting and well designed.

>
> The aspect we are currently working on is a metadata section that will
> include classes and properties to state who has reported a certain
> piece of information, when it was reported, whether it was approved
> (and at which level), and so forth. The current idea is to create
> named graphs that can be described by these metadata elements. I'd
> like to hear your comments on this approach, since this will lead to a
> situation where we can have the same triple in several named graphs
> For example, graph A with all data reported on Januar 20, 2012 by an
> OCHA information officer in Suda, graph B with all data approved by
> the OCHA regional office on January 21, and graph C with all data
> approved by OCHA in Geneva on January 22. The rationale is  to be able
> to query based on these metadata elements via SPARQL, e.g., "give me
> all figure about refugess in Sudan from January 2012 approved by OCHA
> Geneva". Note that the regional office may only approve some of the
> triples originally reported, and OCHA Geneva may only approve a subset
> of those approved by the regional office. So basically we need to be
> able to attach those metadata elements to every single triple.
>
> We will probably run into a situation where we can have the same
> triple in 10–20 graphs at the same time. Likewise, we will have a
> pretty large number of named graphs in our store, and I'd like to know
> whether you think this approach is problematic (e.g. in terms of query
> performance), and whether you see an alternative approach?

I investigated some thoughts on this topic as well in the past. This is 
also a topic of the current RDF WG Graphs TF (See [1]).
I think, you exactly pointed out the problems with duplicated triples 
and single triple named graphs. So there might be the (rather old) need 
for statement identifiers, i.e., a URI (or maybe also a bnode) for 
identifying a  single triple and to be able to describe external context 
information. You can find my proposal at the RDF WG comments mailing 
list, see [2].

Cheers,


Bo


[1] http://www.w3.org/2011/rdf-wg/wiki/TF-Graphs
[2] 
http://lists.w3.org/Archives/Public/public-rdf-comments/2011Jan/0001.html

Received on Wednesday, 22 February 2012 11:21:58 UTC