W3C home > Mailing lists > Public > public-prov-wg@w3.org > November 2012

Fwd: Comments on Prov-O

From: Timothy Lebo <lebot@rpi.edu>
Date: Fri, 9 Nov 2012 09:25:24 -0500
To: Provenance Working Group WG <public-prov-wg@w3.org>
Message-Id: <FB3B9D7E-364F-4FAD-9FA6-9DF33535E19B@rpi.edu>
prov-wg,

Some comments on prov-o.

Regards,
Tim


Begin forwarded message:

> From: <Laurent.Lefort@csiro.au>
> Subject: Comments on Prov-O
> Date: November 9, 2012 9:12:01 AM EST
> To: <lebot@rpi.edu>
> 
> Tim,
> 
> I’ve looked at Prov-O with one eye on the documentation and one eye on the prov-o OWL file as it appears in my favourite Ontology Engineering tool (CMAP Ontology Editor)
> 
> After trying to extract the core classes and properties (to get my own “automatically derived”, "small, scruffy provenance profile", I have found three properties which are not well described in the documentation and described in the wrong place (I have attached the figures I’ve done with COE - I’m not sure the public comments mailing list is the best place to send them).
> 
> I used COE plus an Ontology Module extractor based on the OWL API to limit the number of classes to visualise to focus on the core classes and properties as they appear in the documents under review (PROV Model Primer and PROV-O: The PROV Ontology). I have used the prov-20120724.owl file.
> 
> I’ve looked at the description of the top classes and properties in section 2 2 “Intuitive overview of PROV” in the PROV Model Primer and in Section 2 “2. PROV-O at a glance” in PROV-O: The PROV Ontology.
> 
> The classes and properties I have selected to create my “Prov-O core” module are:
> 
> -          Activity
> 
> -          Agent
> 
> -          Entity
> 
> -          wasDerivedFrom
> 
> -          wasGeneratedBy
> 
> -          endedAtTime
> 
> -          startedAtTime
> 
> -          used
> 
> -          wasAttributedTo
> 
> -          wasAssociatedWith
> 
> -          actedOnBehalfOf
> 
> My tool also extracted three additional “upper properties”:
> 
> -          generated
> 
> -          influenced
> 
> -          wasInfluencedBy
> 
> See the attached CMAP figures:
> 
> -          prov-o-core is the (almost complete) one. I cut the “less OWL-significant fields”
> 
> -          prov-o-starting point is the skeleton (with the super properties)
> 
> -          prov-influence explains the role of the super-properties
> 
> I’m surprised to find that the Primer and the Prov-O documentation do not highlight the role of these specific properties.
> 
> I found http://www.w3.org/2011/prov/wiki/ResponsesToPublicComments#ISSUE-519_and_ISSUE-523_.28Influence_Inheritance.29 which is more about the Prov-DM document and the UML representation
> 
> -          http://www.w3.org/2011/prov/track/issues/519
> 
> -          http://www.w3.org/2011/prov/track/issues/523
> 
> In the Primer and in the Prov-O documents, I think there should be more emphasis on these properties and their roles, but …
> 
> -          They are not described in the Primer
> 
> -          Two of them are listed as “expanded properties” (influenced and generated) and one as “qualified properties” (prov:wasInfluencedBy) in the Prov-O document.
> 
> I think that the two influence properties, prov:influenced and prov:wasInfluencedBy should be documented as SUPER PROPERTIES, in the Prov-O document maybe closer to the discussion on the associated axioms (the non OWL 2 RL ones with the unions).
> 
> Also, if prov:wasInfluencedBy is indeed a qualified property , then where are the prov:qualifiedForm members pointing to Influence and qualifiedInfluence?
> In the owl file I have used, I have:
>       <owl:ObjectProperty rdf:about="http://www.w3.org/ns/prov#wasInfluencedBy">
>                <sharesDefinitionWith rdf:resource="http://www.w3.org/ns/prov#Influence"/>
> 
> And in my view, the prov:generated property should be listed as a “Starting Point” property because it appears to be at the same level as its inverse (prov: wasGeneratedBy) .
> Or alternatively, its full declaration should be “toned down” in the OWL file if your intent is only to “recommend it”.
> 
> PS: also in the PROV-O document, the “Example 3” link for the caption for Figure 4 “Figure 4. An illustration of the PROV-O assertions in Example 3, where Derek publishes two versions of a blog for the National Newspaper, Inc” points to http://www.w3.org/TR/prov-o/#narrative-example-expanded-1 which is Example 2.
Received on Friday, 9 November 2012 14:25:48 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 9 November 2012 14:25:48 GMT