RE: [Fwd: Re: [DC] Implementing DC in OWL DL]

Hi Dan

I think this issue is more general than the DC example. In many use cases, people want
properties to be migrated to OWL without having to decide if they become Object or
Datatype, because they need both, typically the former when values are controlled
(reference to known individuals), and the latter when they are not.

IMO the original DC properties should keep this flexibility and stay as they are as
unspecified rdf:property, but more specific subproperties could be defined to use in OWL,
one Object and one Datatype, such as


Certainly better names are to be found, but that's the idea.

How does that sound?

Bernard Vatant
Mondeca Knowledge Engineering
(+33) 0871 488 459

> -----Message d'origine-----
> De :
> []De la part de Dan Brickley
> Envoyé : mercredi 6 juillet 2005 19:45
> Objet : [Fwd: Re: [DC] Implementing DC in OWL DL]
> OK this is getting towards FAQ territory (see fwd'd
> msg on DC and OWL DL).
> The DC properties can all be used either with
> resources or literals, and there's 5+ years practice
> of doing so.
> Users who are trying to exploit DL facilities seem to
> suffer pain because of this.
> I wonder if SWBP has any advice (a) for them, or
> (b) for the Dublin Core initiative (noting that
> Tom Baker is a critical link between SWBP and DCMI).
> There is a DC conference coming up in mid-Sept,
> and we have a SemWeb session
> in the schedule that needs fleshing out, as well as
> a  meeting of the Dublin Core architecture WG. I'd like
> to start gathering perspectives from this WG re the
> costs (in terms of exploiting SW infrastructure) of DC
> continuing to focus on properties that can be used in
> both styles. If there is a lot that can be shown to be
> gained from making DC more DL-friendly, perhaps
> (Tom, any comments?) some proposals could be
> sketched here (and on dc-architecture) for a way forward...?
> Thanks for any thoughts,
> Dan

Received on Wednesday, 6 July 2005 18:10:23 UTC