Re: AW: Literals representing people?

Patrick Stickler wrote:

> Right. Use URIs, yes, but make sure not to introduce any
> ambiguity. So don't use the URI of an email account or home
> page directly to denote a person. Either use another URI
> that explicitly denotes the person, or use a blank node
> to indirectly denote the person.

I'd go further than this, and only use URIs (well, URIrefs) to identify 
people, simply because bNode identifiers are scoped to a specific graph.
This causes problems if you're trying to smush or combine multiple graphs.
If you use bNodes to represent people, and you have no identifying 
properties (IFPs or FPs), you can't explicitly state that two people in 
different graphs are owl:sameAs.

[if the bNodes that you're trying to assert equivalence between do have 
identifying properties (an IFP that is not also an FP, say), then you could 
state that a thing identified by one property value is owl:sameAs a thing 
identified by some other property value]

I've got some strong use-cases for this approach, based on the expression 
in RDF of bibliographic sources which don't give any information about 
authors beyond their names. As previous posters have remarked, identifying 
people by their names alone is defeasible (similarly, there are very few 
true IFPs), so the ability to directly state equivalence is important in 
non-toyworld SW applications.

-- 
Nick Gibbins                                            nmg@ecs.soton.ac.uk
IAM (Intelligence, Agents, Multimedia)             tel: +44 (0) 23 80598347
Electronics and Computer Science                   fax: +44 (0) 23 80592865
University of Southampton

Received on Tuesday, 30 September 2003 06:05:09 UTC