Community:Resource multiple attribution

From Odp

Revision as of 02:23, 12 April 2010 by MichaelUschold (Talk | contribs)
Jump to: navigation, search

Resource multiple attribution

Title: General Issue: RDF Resource multiple attribution

Description: General Issue: How to attribute a single resource with two values of the same property in one document

Diagram (this article has no graphical representation)

About

Users MichaelUschold
Domains Intellectual Property, General
Competency Questions
Scenarios
Proposed Solutions (OWL files)
Related patterns


Additional information

Issue: How to attribute a single resource with two values of the same property in one document

Example: publishing multiple licenses along with one rdf document.

Source: The thread: attaching multiple licenses in the W3C Linked Open Data Discussion List.

Conclusions:

  1. One way is to use RDF reification, but this has some issues.
  2. OWL 2 has Axiom Annotations designed to do this.
  3. Caution: approaches often won't work unless the community adopts conventions and follows them.
  4. Principles:
    1. two documents about the same resource should serve the same data
    2. data should not be replicated all over the place


Georgi Kobilarov posed the original question:

I'm looking for a best practice for publishing multiple licenses along with one rdf document. Say I publish one URI for an artist: http://example.org/resource/Madonna. I aggregate information from multiple sources about that artist, and those sources have different licenses. One triple comes from a source under GNU FDL, another triple from a source under Public Domain, and a owl:sameas link which I want to publish under Creative Commons License.

Any pointers to how to do that?


Toby Inkster suggests two options. One is reification, where you could create a statement resource which has the usual subject, predicate and object properties and an additional one for the licence. The other is to publish your data in a format that can make use of multiple graphs (e.g. N3). Unfortunately, this is not well supported in triple-consuming software.

However, you can fake named graphs in RDF/XML and Turtle, for example, by splitting the data into multiple documents and use rdfs:seeAlso links between the files to enable auto-dicovery. You could also duplicate that data - but that introduces other problems.


Georgi Kobilarov responds that this solution would run into the issue that "two documents about the same resource should serve the same data". The html output would probably show all triples, but the content negotiated / 303ed rdf document would contain only links to other rdf documents. This would have to be set as common practice for it to work.


Tom Heath suggests including the triples by reference to the source documents, rather than replicating external triples. In his view, "there's something that feels fundamentally wrong about replicating RDF data rather than just pointing applications to the places where they can find it... [it] seems fundamentally un-Web-like"


Alan Ruttenberg points out that OWL 2 provides a documented way of doing this using Axiom Annotations along with a mechanism for translating axioms with annotations. It will last, and avoids confusion caused by RDF reification.


Keith Alexander adds: I think that any publishers of linked data that want to place some restrictions on how their data is reused, should provide clearer guidelines on the technical solutions to fulfilling those requirements. Maybe if they thought about it and realised how confusing it all is, they'd decide they didn't really want to restrict reuse that much in the first place.

References

Add a reference


List of Modeling Issues | Post a new modeling issue | Add a comment in the discussion page
Personal tools
Quality Committee
Content OP publishers