Re: PROV-ISSUE-378 (clarifyHadActivity): clarify hadActivity [Ontology]

Daniel,

On May 21, 2012, at 6:31 PM, Daniel Garijo wrote:

> Hi all,
> following up the discussion we had today in the teleco, 
> we could add a "Derivation subclassOf [ on prov:activity max 0 ] ."

Thanks for making this explicit. I agree with this one.

> That would mean that prov:activity should not be used in Derivations 
> (or that if you have something with the prov:activity relationship, then it is not
> a derivation).

Correct. the "prov:involvee" on Derivation is the Entity (in line with the "rdf:object rule" of the qualification pattern)

> This would also be applicable for Invalidation, Responsibility and Start.
> 
> By the way, I've realized that Invalidation is in the domain of prov:activity
> and prov:hadAvtivity.
> If the activity is optional, then it is not an ActivityInvolvement.

I'd have to double check the constraints document, but an Invalidation is always invalided by _some_ activity, you just don't have to say which.
(this is a min 1 cardinality, which we have been avoiding saying for RL "lightness"; instead, we're relying on the rdfs:domain to say it).

If you do cite the Activity, you use prov:activity; You never use hadActivity on Invalidation.

Do you agree?


> If the activity is not optional, then it shouldn't be in domain of prov:hadActivity.

This is a _very_ good catch.
I now think that Invalidation should be removed from the domain of hadActivity.
(I think this is a reasonable issue to keep under the current ISSUE).

Do you agree?

Thanks!

Tim



> 
> Thanks,
> Daniel
> 
> 
> 
> 2012/5/18 Timothy Lebo <lebot@rpi.edu>
> 
> On May 18, 2012, at 6:51 AM, Provenance Working Group Issue Tracker wrote:
> 
> > PROV-ISSUE-378 (clarifyHadActivity): clarify hadActivity [Ontology]
> >
> > http://www.w3.org/2011/prov/track/issues/378
> >
> > Raised by: Daniel Garijo
> > On product: Ontology
> >
> > We should clarify the difference between prov:activity and prov:hadActivity (so people don't use prov:hadActivity in qualifiedGenerations).
> >
> > We could add a restirction on Generation:
> > Generation subclassOf [ on prov:hadActivity max 0 ] .
> >
> 
> ^^ This is within RL and states the restriction that would prevent the confusion between activity and hadActivity.
> 
> > And, since the difference between activity and hadActivity is that the former is not optional:
> > Generation subclassOf [ on prov:activity min 1 ] .
> 
> ^^ min 1 goes against RL, which is why we've been avoiding them.
> 
> >
> > Also, we should add an rdf:comment explaining this decision.
> 
> ^^ Do you have a proposed comment to put in?
> 
> THanks,
> TIm
> 
> >
> >
> >
> >
> >
> 
> 
>