RE: Define/change the range of "supportedProperties" (ISSUE-37)

On Wednesday, March 05, 2014 10:13 PM, Ruben Verborgh wrote:
> > What do you think? Ruben, would that work for you?
> 
> No not really, sorry.

Too bad :-)

 
> >> However, I just re-evauated this and meanwhile think that Ruben had a
> >> rather technical view on this which only SemWeb experts might have.
> >> Ruben points out that one might get the impression that a
> >> `hydra:supportedProperty` is an `rdf:Property`
> >> but in fact it is only a pointer to such a thing (if I understood
> >> correctly).
> 
> But hat's not my main objection.
> 
> You're right that not everybody thinks of rdf:Property if they don't
> know about it.
> However.
> 
> > So you have
> >
> >  Some Class --- supportedProperty --> SupportedProperty
> >                                         |- property --> Real property
> 
> The problem remains even without thinking of rdf:Property.
> I rephrase, given the above.
> 
> A class can have a supportedProperty that is a SupportedProperty
> but not a Real property, but it has a property property that points to
> the Real property.

Hmm... yeah. Not optimal


> So same issue arises, for non-SemWeb'ers, without rdf:Property.
> 
> > Unfortunately, I can't think of a better term either.
> 
> You really need something different. I even think that
> 
>  Some Class - supportedProperty --> hydra:Property
>                                         |- hydra:attribute -> Real
> property
> 
> would be better.
> (i.e., keep calling a "supported property" a "property", and the "Real
> property" an attribute,
>  which is the opposite of what I said earlier.)

Really? Of all the options discussed so far, I find that to be the most
confusing. Sorry.


> We need distinct names for both things; they are different things.
> 
> > So perhaps a middle ground would be to keep *s*upportedProperty but
> rename *S*upportedProperty to something like "PropertyDescription"!?
> 
> Mmm I don't fully like PropertyDescription either, because
> "Description" can make it confusing on what level we are.
> For instance, the following is the description of a property:
> 
> foaf:knows a rdf:Property;
>     a owl:SymmetricProprerty.
>
> It is, however, not a PropertyDescription.

That's obviously nitpicking but isn't this the *definition* of the property
(which of course is also a description, but I think you understand what I'm
getting at)?


> > That would perhaps make it clearer that it is kind of a wrapper
> around a real property.
> 
> I really suggest find distinct terms for both concepts.

PropertyWrapper, PropertyProxy? :-) Sam recently experimented with
PropertyRequirement but that feels odd if "required" is set to false.


> Being able to name something is crucial to have clarity.
> I doesn't have to be "attribute"; anything that is distinct from
> "property"
> (i.e., does not have "property" in its name) is fine with me.

Fully agree.. the problem is we haven't been able to come up with something
better yet.. something everyone can live with.



--
Markus Lanthaler
@markuslanthaler

Received on Friday, 7 March 2014 19:45:46 UTC