RE: [mawg] action-249: Ontology rev 5 available & call for competency questions wrt. to actor - role part of the ontology

Hi Pierre-Antoine,

Thanks for the feedback. 

1/ funny that title got messed up, we need to check this.  No need for a title for named fragment whose name is given by its URI. Point taken.

2/ Here I still believe we need to be semantically rigorous. TopBraid and Protégé have no problem with it and triples being generated avoid duration being attributed to pictures.

3/ This is a good question. I did it that way wondering at what level we were in the ontology, Doesn't change much anyway. This can be changed easily if you think it is more appropriate. 

Best regards,

Jean-pierre

-----Original Message-----
From: public-media-annotation-request@w3.org [mailto:public-media-annotation-request@w3.org] On Behalf Of Pierre-Antoine Champin
Sent: mercredi, 25. août 2010 16:10
To: Tobias Bürger
Cc: public-media-annotation@w3.org
Subject: Re: [mawg] action-249: Ontology rev 5 available & call for competency questions wrt. to actor - role part of the ontology

Hi Tobias and Jean-Pierre,

a few remarks:

1/ some properties have multiple domain declarations, e.g. title has
domaine MediaResource and NamedFragment, meaning that everything with a
title is *both* a MediaResource and a NamedFragment, which is obviously
wrong. What you need here is

  ma:title rdfs:domain [ owl:unionOf
      ( ma:MediaResource ma:NamedFragment ) ] .


2/ some properties have complex domains, such as e.g. duration applies
to MediaRessources which are not Images... I understand the rationale of
those constraints, and I agree they are probably *semantically* correct,
but on the *pragmatic* level, they may give wrong impressions, like:
- any media resource must belong to one of the 3 subclasses defined by
the ontology -- not the intent of the ontology, IMHO
- any media resource which is not an Image can have a duration -- as a
(quite farfetched) counterexample, think of a Smell resource


3/ I am a bit confused about Contributor being *equivalent* to
(contributorIs some Agent). I would rather make it a *subclass* of this
restriction; the equivalence semantically would follow from the domain
of contributorIs.

Since this is semantically equivalent, why would I like to change it? It
may be a pedantic distinction, but I find it a little disturbing, as it
creates a kind of cyclic definition with the domain axiom of
contributorIs, and makes it difficult to understand the intent of the
ontology author -- here again, I'm only talking about the pragmatic
level; the semantics of all this is perfectly well defined.

  pa



On 19/08/2010 09:44, Tobias Bürger wrote:
>   Dear all,
> 
> Jean-Pierre and myself had some discussions around the ontology recently 
> which resulted in an updated version of the current OWL version 
> (revision 5) which you can find here: 
> http://www.salzburgresearch.at/~tbuerger/ma-ont-rev5.owl

> 
> (essentiallly it includes some rules and restrictions for some elements)
> 
> We are still discussing around the actor-role pattern in the ontology 
> involving the agent (i.e. person) and its role (creator, actor) wrt. to 
> a media resource.
> (see the discussion which we started here: 
> http://lists.w3.org/Archives/Public/public-media-annotation/2010Jul/0034.html)
> 
> To somehow reverse engineer this problem we would like the members of 
> the group to contribute the type of queries they would like to pose wrt. 
> to this part of the ontology in order to decide which modelling option 
> we should go for.
> 
> Thanks a lot for your input!
> 
> Best regards,
> 
> Tobias
>

Received on Wednesday, 25 August 2010 14:20:20 UTC