Re: PROV-ISSUE-96 (entities and roles): Relating Roled Entities with non-Roled Entities [Conceptual Model]

I think we should use the second interpretation. The first one really isn't
valid, as you say, and the third requires more machinery than necessary.
Except that the role serves (in the OWL model, anyway) as a class of the
entity.

account(acc,
       asserter,
   entity(http://x.com/James,[**colour= <http://x.com/James,%5Bcolour=>
"blue"])
   entity(http://x.com/James2,[**role= <http://x.com/James2,%5Brole=>
"FirstAuthor"])
   wasComplementOf(http://x.com/**James,http://x.com/James2<http://x.com/James,http://x.com/James2>
))

would become in OWL (inside of the account):

Individual: http://x.com/James
   Types:
      prov:Entity
   Facts:
      ex:color "blue"

Individual: http://x.com/James2
   Types:
      prov:Entity
      ex:FirstAuthor
   Facts:
      prov:wasComplementOf James

I reversed the wasComplementOf because in this case, it's not symmetric.
Everything that's true of James is also true of James2, but not vice-versa.

Jim

On Fri, Sep 23, 2011 at 8:46 AM, Luc Moreau <l.moreau@ecs.soton.ac.uk>wrote:

>
> Hi James,
>
> I see three different ways we can tackle your issue in PROV-DM.
>
> Within one account, an identifier has a unique interpretation.
> Hence, if multiple entity expressions have the same identifier, one
> needs to take the union of them.
>
> This account
>
> account(acc,
>        asserter,
>    entity(http://x.com/James,[])
>    entity(http://x.com/James,[**role= <http://x.com/James,%5Brole=>
> "FirstAuthor"]))
>
> is equivalent to:
>
>
> account(acc,
>        asserter,
>    entity(http://x.com/James,[**role= <http://x.com/James,%5Brole=>
> "FirstAuthor"]))
>
>
> Hence, with a single account, and a single uri, you can't express two
> different entity expressions.
>
>
> Alternatively, within a single account, you can use two different
> identifiers.
> You can then express the wasComplementOf relation.
>
>
> account(acc,
>        asserter,
>    entity(http://x.com/James,[**colour= <http://x.com/James,%5Bcolour=>
> "blue"])
>    entity(http://x.com/James2,[**role= <http://x.com/James2,%5Brole=>
> "FirstAuthor"])
>    wasComplementOf(http://x.com/**James,http://x.com/James2<http://x.com/James,http://x.com/James2>
> ))
>
>
>
> In the third option, you use two accounts.  In each account, you are
> entitled to use the same uri to denote different entity expressions.
>
>
> account(http://p.com/acc1,
>        asserter1,
>    entity(http://x.com/James,[**colour= <http://x.com/James,%5Bcolour=>
> "blue"]))
>
>
> account(http://p.com/acc2,
>        asserter2,
>    entity(http://x.com/James,[**role= <http://x.com/James,%5Brole=>
> "FirstAuthor"]))
>
> The account construct therefore acts as a scope distinguishing the two
> entities.
>
> To express the wasComplementOf relation, you then need a form of
> qualified identifier that involves the accounts ursi.
>
> account(acc3,
>        asserter3,
>
>    wasComplementOf(   (http://p.com/acc1,  http://x.com/James)
>                       (http://p.com/acc2,  http://x.com/James) ))
>
> This last construct is not defined in the document.
> We are planning to introduce qualified identifiers in the next iteration.
>
> Cheers,
> Luc
>
>
> On 15/09/2011 18:46, Provenance Working Group Issue Tracker wrote:
>
>> PROV-ISSUE-96 (entities and roles): Relating Roled Entities with non-Roled
>> Entities [Conceptual Model]
>>
>> http://www.w3.org/2011/prov/**track/issues/96<http://www.w3.org/2011/prov/track/issues/96>
>>
>> Raised by: James McCusker
>> On product: Conceptual Model
>>
>> Let's say we have Entities "Jim McCusker" and "Jim McCusker as first
>> author of XYZ paper". Under the current proposal, "Jim McCusker as first
>> author of XYZ paper" is of type Entity and also of type FirstAuthor, which
>> is a subclass of Entity and is therefore a Role. Since "Jim McCusker as
>> first author of XYZ paper" is contextualized by that role, it's not correct
>> for that to share the same URI as "Jim McCusker". However, there is some
>> sort of relationship.
>>
>> I'd like to propose that complementOf, or whatever succeeds it, is that
>> relationship. It follows the same pattern of "same entity in different
>> contexts", as the role is a particular non-temporal, non-spatial
>> contextualization.
>>
>>
>>
>>
>>
>
>
>


-- 
Jim McCusker
Programmer Analyst
Krauthammer Lab, Pathology Informatics
Yale School of Medicine
james.mccusker@yale.edu | (203) 785-6330
http://krauthammerlab.med.yale.edu

PhD Student
Tetherless World Constellation
Rensselaer Polytechnic Institute
mccusj@cs.rpi.edu
http://tw.rpi.edu