RE: Definition of Entity

Perhaps we can just start from a common sense overview/example?

When we talk about things in the world in natural language and even when we assign identifiers, we are often imprecise in ways that make it difficult to clearly and unambiguously report provenance: a resource with a URL may be understood as referring to a report available at that URL, the version of the report available there today, the report independent of where it is hosted over time, etc. PIL defines the notion of entity as a specific characterization of a thing that has only one of these meanings. It does not assume any characterization is more important than any other and in fact, it is possible to describe the processing that occurred for the report to be commissioned, for individual versions to be created, for those versions to be published at the given URL, etc., each via a different entity that unambiguously characterizes the report appropriately. PIL also includes the notion of complementOf/viewOf to associate different characterizations of the same thing so that one can relate all of the perspectives on how the report came to be.

To me, everything else we're trying to say about entity falls out from that: a resource may be asserted to be an entity if it unambiguously defined to start with, entities have to have some mechanism by which you can understand how they are frozen - attributes, etc.

Perhaps this is more in the style of the primer, but maybe some part of the explanation - entities remove the ambiguity inherent in our usual definitions of 'things' in the world and this is critical for reporting provenance - is missing/too technical in the current model doc? 
 
Jim

> -----Original Message-----
> From: public-prov-wg-request@w3.org [mailto:public-prov-wg-
> request@w3.org] On Behalf Of Graham Klyne
> Sent: Friday, September 02, 2011 8:00 AM
> To: Luc Moreau
> Cc: W3C provenance WG
> Subject: Re: Definition of Entity
> 
> (Luc, This is not specifically about the issue raised, but was triggered by
> something you wrote there.  What I am discussing is that the definitions of
> the concepts are insufficiently clear, which may in turn be why I have not
> previously been able to clearly express agreement or disagreement with
> what you have proposed.)
> 
> On 02/09/2011 12:38, Luc Moreau wrote:
>  > So:
>  > Entity: is data model construct/assertion
> 
> This is new to me.  Where does it say this?  I've just looked again in sections
> 4 and 5, and am not seeing it.
> 
> AFAICT, "Entity" is not mentioned at all in section 4.
> 
>  From what you say above, it seems that "Entity" and "Entity assertion" are
> the same thing.  But I don't see that stated explicitly.  And if true, it
> completely overturns what I thought we meant by "Entity".
> 
>  > Thing: is the thing in world
> 
> I thought we were saying that "Entity" is a thing in the world about which
> assertions have been made.  But here you seem to be saying that the
> "Entity"
> *is* the assertions that have been said about a thing.
> 
> Specifically, section 5.1 needs tlo clearly state what concept is denoted by the
> term "Entity" - for me it does not do that.  Without such a clear and well-
> understood statement, it's hard to make concrete progress on the rest of the
> specification.
> 
> It seems to me that the section is insufficiently clear in distinguishing:
> language used to describe entities/things, descriptions of entities/things, and
> the actual things described.
> 
> #g
> --
> 
> On 02/09/2011 12:38, Luc Moreau wrote:
> > Hi Graham,
> >
> > I don't understand your comment.
> >
> > Are you discussing the difference between entity and entity assertion?
> >
> > Section 5 introduces different types of constructs.
> > Section 4 states that all representations constructed with PIDM are in
> > fact assertions by the asserter.
> >
> > So, when we write "An Entity represents an identifiable characterized
> > thing.", we refer to the construct entity, which allows us in PIDM to
> > build a representation of an identifiable characterized thing.
> > That construct contains an id and attribute-value pairs.
> >
> > So:
> > Entity: is data model construct/assertion
> > Thing: is the thing in world
> >
> > I don't see what is not correct in the issue I raised.
> >
> > Luc
> >
> > On 09/02/2011 12:04 PM, Graham Klyne wrote:
> >> Luc,
> >>
> >> I'm picking up a small matter here to illustrate things I've said previously.
> >>
> >> I notice in ISSUE 89 you say:
> >>
> >> "The conceptual model defines an entity in terms of an identifier and
> >> a list of attribute-value pairs. It is indeed crucial for the
> >> asserter to identify the attributes that have been frozen in a given entity."
> >>
> >> But when I look at
> >> http://dvcs.w3.org/hg/prov/raw-
> file/default/model/ProvenanceModel.htm
> >> l#concept-Entity, this is not what I see. What I do see is a
> >> description of an "Entity assertion" that contains a list of
> >> attribute pairs, which to my reading is not the same thing at all.
> >>
> >> This is a part of the problem I have when I say the model document is
> >> difficult to understand.
> >>
> >> (I'm not raising this as an issue, as I've already raised a different
> >> issue to say I think that an Entity doesn't need to be so
> >> complicated.)
> >>
> >> #g
> >

Received on Friday, 2 September 2011 14:06:10 UTC