Re: PROV-ISSUE-29 (mutual-iVP-of): can two bobs be mutually "IVP of" each other [Conceptual Model]

On Mar 23, 2012 6:32 PM, "Stian Soiland-Reyes" <
soiland-reyes@cs.manchester.ac.uk> wrote:
>
> On Fri, Mar 23, 2012 at 16:44, Jim McCusker <mccusj@rpi.edu> wrote:
>
> > Proposed:
> > alternativeOf: An entity is alternate of another if they specialize a
common
> > Entity.
>
> ..  "specialize *some* common Entity" - we don't need to know which one
it is.

Exactly. Stating that it is an alternative means that there is a common
entity, whether or not we know what it is.

> From this definition alternativeOf is both reflective (entity(a) =>
> alternativeOf(a,a) ) and  symmetric (alternativeOf(a,b) =>
> alternativeOf(b,a)), but not neccessarily transitive
> (alternativeOf(a,b);  alternativeOf(b,c)  does not imply
> alternativeOf(a,c) - does A and C necessarily have a common
> specialization of parent?

No, transitivity should not be implied here, but reflectivity and symmetry
are both reasonable.

> > specializationOf: An entity is a specialization of another if they both
> > denote a common thing, but the former is a more constrained denotation
than
> > the former. Examples of denotational constraint may include:
abstraction,
> > context, and roles the entity has.
>
>
> What prevents everything from being a specializationOf say
> atomsInTheUniverse or thingsThatExists - and making everything
> alternativeOf each other? The "common thing" saves us? :)

Yes. They must have a common denotation at some level.

> We no longer require alternativesOf to have overlapping
> characterizations, right? What about attributes on a specializationOf
> parent and their characterization intervals? All gone?

We no longer talk about characterizations in the DM, so there is no need or
means by which we can talk about characterization intervals.

> .. not sure if we should include "roles" here as it would be confusing
> with prov:hadRole (the old EntityInRole discussion).

My intention is to use this to provide roles to entities within a
particular context. We could leave this out, if it's too confusing.

> The current example of Bob with Facebook account is not very good. Why
> would bobWithFacebook be an alternative of bobWithTwitter? Just
> because they share bob as a parent specialization? Why would you form
> such entities?

JimAtYale and JimAtRPI being specializations of JimMcCusker (in general)
are better examples, probably.

> The BBC News home page today is a specialization of the BBC home page.
> That could be a good one.
>
>
>
> The BBC news home page today is a specialization of the BBC news page
> in general. BBC does not provide a URI for a given day's news page, so
> we mint our own:
>
>
> specializationOf(bbcNews2012-03-23, <http://www.bbc.co.uk/news/>)
>
>
> The mobile news page is an alternative of the desktop news page. They
> are both specialization of (here unspecified) entity.
>
> alternativeOf(<http://www.bbc.co.uk/news/mobile/>, <
http://www.bbc.co.uk/news/>)

This is a perfect example.

> The mobile news page of today is a specialization of the mobile news page:
>
> specializationOf(bbcNewsMobile2012-03-23, <
http://www.bbc.co.uk/news/mobile/>)
>
>
> This implies (as /news/mobile and /news/ have a common specialization):
>
> alternativeOf(bbcNews2012-03-23, bbcNewsMobile2012-03-23)

Yes, this all correctly follows.

Jim