Re: about EntityInRole and assumedBy

I think that from a point of view of a consumer of the data the current
approach is more confusing than the opmo modeling.

With the current approach if I ask for the "uses" of a process execution I
would have to check each
of the entities to see if they are entitiesInRole or not, and if they are,
I would have to see which of the
entities of the other usage relationships are they representing.

With the opmo modeling I would be sure that every usage is linked to a
separate entity. In case I wanted more information,
I would ask for the n-ary relationships.

The current approach is more simple, and I like it. But I don't see how we
can overcome this issue easily. We could solve it
by not making entityInRole subclass of Entity, but that would change
everything, so I can't consider it a solution.

Best,
Daniel

PS: Mutiple uses of the same role would lead to multiple instances of
EntityInRole. Each entityInRole instance is unique
per processExecution.

2011/10/30 Stian Soiland-Reyes <soiland-reyes@cs.manchester.ac.uk>

>
> On Oct 28, 2011 4:27 PM, "Daniel Garijo" <dgarijo@delicias.dia.fi.upm.es>
> wrote:
>
> > Something similar happens to the second approach: we can link directly
> the process to the entity with
> > the "used" property, and if we need more information, then we use the
> "EntityInRole" for adding the time,
> > location, role, etc. What imo can be confusing here is that we would
> have 2 entities for refering to the same
> > "usage" relationship.
>
> Which is the reason for prov:assumedBy being a more specific version of
> prov:wasComplementOf; instead of two good friends being used, it was the
> very same use, and so the use time etc stated about the EntityInRole also
> applies to the Entity. The EiR does complement the E in a sense, by adding
> any qualifier attributes and the prov:assumedRole attribute.
>
> This also easily allows multiple use of same entity with different
> roles/qualifiers. (not sure about multiple use of same role!)
>