Please approve this draft PROV-O response by Tuesday 5pm GMT

prov-wg,

Your approval is needed for the draft response to issue 552.

The response is at http://www.w3.org/2011/prov/wiki/ResponsesToPublicComments#ISSUE-552_.28Influence_subclasses.29

and is copied below.

Please raise objections before tomorrow 5pm UK time, so that we can close this out before the F2F.

Regards,
Tim

ISSUE-552 (Influence subclasses)
Original email: http://lists.w3.org/Archives/Public/public-prov-comments/2012Sep/0000.html
Tracker: https://www.w3.org/2011/prov/track/issues/552
Group Response:
On "subclassing Influence":
The WG agrees with the suggestion that the phrase "a particular case of derivation" should be expressed using rdfs:subClassOf.
Since the prov-dm's definitions for revision, quotation, and primary source mention that they are "particular case[s] of derivation", then it follows that each should be subclasses in the PROV-O encoding. We changed PROV-O to include these three classes as a subclass of Derivation.
The WG aggress with the reviewer that "a kind of" is a more natural phrasing than "a particular case", and so we have adopted it as suggested.
On the phrasing of definitions:
It was pointed out that the definitions for "{Entity,Agent,Activity}Influence" are inconsistent with that of their parent class "Influence".
The source of this inconsistency is that {Entity,Agent,Activity}Influence are not defined by prov-dm, but by prov-o as artifacts of encoding prov-dm's model into the paradigm of OWL (i.e., the use of the qualification pattern to describe binary relations).
The inconsistent definitions were "demoted" to rdfs:comments because they focus too heavily on the RDF and OWL paradigm and not enough on how they are expressing the abstract model of prov-dm.
New definitions were created to align with their parent class, with a focus on how the classes are expressing the abstract model of prov-dm.
On the inconsistency of subclasses according to "general understanding of the english terms":
The reviewer points out that the definitions of Influence, EntityInfluence, and Start illustrate an inconsistency: "influence is a capacity, an entity influence is a provider (of descriptions) and a start is a "when" (a time)".
The WG acknowledges that the definitions as shown support this concern.
The inconsistency between Influence and its immediate subclasses {Entity,Agent,Activity}Influence is addressed by the response to the earlier comment ("phrasing of definitions").
To address the inconsistency between {Influence, {Entity,Agent,Activity}Influence} and {Start,End}, PROV-DM updated the definitions for Start and End:
Start is when an activity is deemed to have been started by an an entity, known as trigger . The activity did not exist before its start. Any usage, generation, or invalidation involving an activity follows the activity's start. A start may refer to a trigger entity that set off the activity, or to an activity, known as starter , that generated the trigger. ref
End is when an activity is deemed to have been ended by an entity, known as trigger . The activity no longer exists after its end. Any usage, generation, or invalidation involving an activity precedes the activity's end. An end may refer to a trigger entity that terminated the activity, or to an activity, known as ender that generated the trigger. ref
References:
Changes to the document:
prov-dm updated the definitions for revision, quotation, and primary source to reinforce that each is a relation.
prov-o changed to add axioms:
prov:Revision rdfs:subClassOf prov:Derivation .
prov:PrimarySource rdfs:subClassOf prov:Derivation .
prov:Quotation rdfs:subClassOf prov:Derivation .
prov-o "demoted" the original definitions of {Entity,Agent,Activity}Influence to rdfs:comments.
prov-o created new definitions for {Entity,Agent,Activity}Influence to align with their parent class definition.
prov-o removed existing comments on {Entity,Agent,Activity}Influence that were very similar to the new "prov-dm centric" definitions. The removed comments had more of an OWL flavor to them instead of an abstract flavor. For example, the following comment was removed:
"ActivityInfluence is intended to be a general subclass of Influence of an Activity. It is a superclass for more specific kinds of Influences (e.g. Generation, Communication, and Invalidation)." in favor of the definition "ActivitiyInfluence is the capacity an activity to have an effect on the character, development, or behavior of another by means of generation, invalidation, communication, or other."
The latest draft of the PROV-O html document reflects the definitions changed in the PROV-O OWL file:
http://aquarius.tw.rpi.edu/prov-wg/prov-o#EntityInfluence,
http://aquarius.tw.rpi.edu/prov-wg/prov-o#ActivityInfluence,
http://aquarius.tw.rpi.edu/prov-wg/prov-o#AgentInfluence
PROV-DM's new definition for Start -> PROV-O's new definition for Start
PROV-DM's new definition for End -> PROV-O's new definition for End




On Sep 13, 2012, at 7:27 AM, Provenance Working Group Issue Tracker <sysbot+tracker@w3.org> wrote:

> PROV-ISSUE-552 (subclass-prov-o): Check subclass definitions in prov-o [Ontology]
> 
> http://www.w3.org/2011/prov/track/issues/552
> 
> Raised by: Paul Groth
> On product: Ontology
> 
> See email from Alan Ruttenberg
> 
> 
> 
> 

Received on Monday, 5 November 2012 13:56:19 UTC