Re: Quick comment from new SKOS user

* Dickinson, Ian John (HP Labs, Bristol, UK) <ian.dickinson@hp.com> [2005-01-06 12:23-0000]
> 
> Hi Brian,
> I understand your choice, and I concur that it makes sense that way
> round. I actually wasn't advocating a change to the meaning, I was
> merely trying to point out that, as a new SKOS user, I tripped over the
> *potential* ambiguity inherent in the name.  I regard myself as
> moderately clueful in the arcane world of RDF, so I infer this may
> happen to other newcomers as well.  I accept that it's too late to
> change the name from skos:narrower to skos:hasNarrower or something else
> with directionality built-in, so my advocacy was to make the text in the
> guide (I was looking at version 2004-11-25) a bit more explicit.  Sorry
> if that wasn't clear from my message.

Some RDF vocabs do go down this route, with 'hasAge', 'hasName',
'hasHomepage' instead of 'age', 'name', 'homepage'. While
backwards-named properties like 'subClassOf' complicate the picture,
I prefer to have simple property names. I think an analogy with OO
class/property models is appropriate here (it isn't always, with RDF).
In Java, a Person class might have 'age', 'name', 'homepage' as fields 
or bean properties. Similarly in other objects. The fact that the 
property is attached to some object is enough to suggest the implicit 
'has'. Unfortunately, 'has-subClassOf' doesn't quite work, raising 
intuitions that the thing referenced is a sub-, rather than super- 
class of the subject resource. 

If we can minimise the deployment of 'isXYZof' properties, we can
hopefully get away without having to prefix everything else with
'has'. Perhaps... :)

Dan

Received on Thursday, 6 January 2005 12:35:23 UTC