Re: [higgins-dev] idas model as entities

[Paul Trevithick <paul@xxxxxxxxxxxxxxxxx>]

Title: Re: [higgins-dev] idas model as entities

I’d prefer the data model to use rdfs:domain on each Attribute Definition, but have a IdAS offer your getAttributes() as a convenience method.

-Paul

[BTW, I do now see that you are not preventing reuse of Attribute Definitions across multiple classes. I misunderstood your proposal.]

On 9/3/08 7:35 PM, "Markus Sabadello" <msabadello@xxxxxxxxxxxxx> wrote:

> Well yes, in RDF it's the "other way round" than what we are doing with higgins:validAttributes. In RDF you can only connect attribute types to entity classes by saying that the attribute type is valid for entity classes X, Y and Z. That's what you do with rdfs:domain. And that's why I thought higgins:validAttributes could be defined in HOWL to be the owl:inverseOf of rdfs:domain.
>
> Don't know if both ways should be exposed in IdAS, i.e. if rdfs:domain should be exposed in addition to IEntity.getAttributes(). I agree it would be redundant.
>
> Markus
>
> On Thu, Sep 4, 2008 at 1:09 AM, Jim Sermersheim <jimse@xxxxxxxxxx> wrote:
> > I just re-read this and noticed I missed something:
> >
> > >>> Paul Trevithick <paul@xxxxxxxxxxxxxxxxx> 09/02/08 8:51 PM >>>
> > <snip>
> > >http://wiki.eclipse.org/Higgins/ModelAPIs#An_attribute_model:
> > >* I see that you've chosen not to return rdf:domain as an attribute. Is this because you're trying to avoid allowing your attribute models to be used on more than one of your entity models?
> >
> > In IdAS, a consumer finds the attributes of an entity by calling IEntity.getAttributes().  So it seems natural in the class definition for an entity type to call out which attributes the consumer might expect to find or might be able to add to entities of that type.  So, the relationship between an entity's class definition and the attributes which are valid for it are on the class definition (in the validAttributes attribute) rather than the relationship being at the attribute type definition, pointing back at all the entity class definitions that allow that attribute.
> >
> > Multiple entity classes may name the same attribute type in their validAttributes attr, so I'm not preventing that (nor do I want to).
> >
> > Is there a need to allow each attribute type definition to list all the entity classes which it's valid for?  This would just be redundant data, correct?
> >
> > _______________________________________________
> > higgins-dev mailing list
> > higgins-dev@xxxxxxxxxxx
> > https://dev.eclipse.org/mailman/listinfo/higgins-dev