Re: PROV-ISSUE-259: hadTemporalExtent domain and range [Ontology]

Hi Stephan,

Prov-dm does not define temporal relations for everything, since time is 
associated with instantaneous events, and we have only 4 types of events.

Second, from an interoperability viewpoint, while this may look 
convenient to have a domain which is owl:Thing, it breaks 
interoperability with other representations.

Finally, it was already indicated by Paolo that some subclasses of 
Involvement (e.g. Association) do not have time information associated 
with them in prov-dm.

So, on the contrary, the domain of hadTemporalExtent has to be aligned 
to what prov-dm specifies.

Cheers,
Luc

PS. Can't we simplify the name of this relation. e.g. hadTime

PS2. Why isn't it a data property to xsd:dateTime?

On 22/02/2012 18:06, Provenance Working Group Issue Tracker wrote:
> PROV-ISSUE-259: hadTemporalExtent domain and range [Ontology]
>
> http://www.w3.org/2011/prov/track/issues/259
>
> Raised by: Stephan Zednik
> On product: Ontology
>
> hadTemporalExtent has rdfs:domain prov:Involvement, which means if anything has a prov:hadTemporalExtent that it will be inferred to be a member of the class prov:Involvement.
>
> Since a temporal extent describes the time (interval or instant) over which something occurred, I would prefer to not restrict the class of things with temporal extent to Involvement.  For example, I think it would make sense in practice for Activities to use hadTemporalExtent.
>
> Also, I would like the range of hadTemporalExtent to be relaxed to include both TimeInstant and intervals of time (which we currently do not model in prov-o).  In OWL-Time this would be time:TemporalEntity (union of time:Instant and time:Interval).
>
> --Stephan
>
>
>
>
>

Received on Wednesday, 22 February 2012 22:19:56 UTC