[foaf-dev] No name for agents or groups?

Henry Story henry.story at bblfish.net
Sat Mar 5 18:09:35 CET 2011


On 5 Mar 2011, at 18:00, Bob Ferris wrote:

> Hi Henry,
> 
> Am 05.03.2011 17:30, schrieb Henry Story:
>> 
>> On 5 Mar 2011, at 17:22, Melvin Carvalho wrote:
>> 
>>> On 5 March 2011 17:15, Henry Story<henry.story at bblfish.net>  wrote:
>>>> I was looking at the definition of foaf:Agent again. I was surprised the there is no name relation for Agents or groups. It seems to me that all agents should be able to have human readable names.
>>> 
>>> foaf:name ?
>>> 
>>> http://xmlns.com/foaf/spec/#term_name
>> 
>> Thanks. The mistake I made was to look for it on the foaf:Agent class. Perhaps foaf:Agent should state that it is a subclass of rdf:Resource, so that one has a hint to look there.
> 
> I don't see a subclass relation of foaf:Agent to rdf:Resource. It's 
> simply the case that when one uses foaf:name, a reasoning engine can 
> infer that the domain of that property is owl:Thing. So, e.g.,
> 
> :me a foaf:Agent ;
>    foaf:name "me .
> 
> can be used to infer
> 
> :me a owl:Thing .
> 
> Please remember, the things in an ontology are often somehow loosely 
> coupled.

Yes, I know. But I make that mistake often, and I think so will most object
oriented programmers, which is a good constituency to watch out for, as they
work with OWL like tools all day. For example in javadoc you usually walk up 
the tree hierarchy. So automatically such developers will do the same here, 
and miss the TOP. 

In UI design you have to make hidden knowledge visible. I think this is a simple 
improvement one can do here, and in all other similar ontologies.

Henry


> 
> Cheers,
> 
> 
> Bob
> _______________________________________________
> foaf-dev mailing list
> foaf-dev at lists.foaf-project.org
> http://lists.foaf-project.org/mailman/listinfo/foaf-dev

Social Web Architect
http://bblfish.net/



More information about the foaf-dev mailing list