[foaf-dev] Fwd: relationship vocab

Karen Coyle kcoyle at kcoyle.net
Mon Dec 13 15:07:45 CET 2010

Does it really make sense to try to shoehorn all of these concepts  
into foaf? It seems to me that in this and other projects we are  
trying to add general terms and relationships to more specific  
schemas. Why not elevate some concepts (like "next to", "name",  
"place") to be direct descendants of owl:Thing or owl:Class? They can  
still use the foaf namespace (it doesn't really matter what namespace  
is used) but people wouldn't struggle with trying to fit them into the  
original foaf realm of social networking. (Yes, it's true that RDF  
doesn't constrain you, but the human mind does. The fact is, using the  
term "Person" is not neutral; it means something to people, who in  
this case are less flexible than machines.) It's ok to outgrow the  
original concept, and not to try to keep everything inside its  

Besides, I see us re-defining the same general properties over and  
over, and most of them are buried in schemes where no one would think  
to look for them because it's just not the logical place to look.


Quoting Kevin Pickens <kbpickens at gmail.com>:

> Toby,
> I agree with your point regarding relationship properties.  Further, it
> would seem to me that relationship properties should be almost universally
> expanded to refer to and imply that the entity having the property is a
> foaf:Agent.  A simple example: the grocery store down the street is a
> neighbour of the pharmacy next to it.  Neither is likely to be considered a
> foaf:Person, but they are both reasonably foaf:Agents.  The only
> relationship properties that seem inconsistent with non-foaf:Person entities
> are some of the more formal ones (I just can't see the the grocery store
> settling down, getting rel:engagedTo and becoming the rel:spouseOf the
> pharmacy - what would they talk about?).
> To put this in a practical form, a person may want to say that they work for
> the grocery store next to the pharmacy.  With the current relationship
> properties, that doesn't work properly.  They could assert that they are a
> foaf:Person who is rel:employedBy (assuming the definition is expanded) the
> grocery and that they are the rel:neighborOf a foaf:Person who is
> rel:employedBy the pharmacy, but that would generally be interpreted as
> "we're neighbours and we work at the grocery and the pharmacy,
> respectively."
> Cheers,
> Kevin
> On Mon, Dec 13, 2010 at 7:54 AM, Toby Inkster <tai at g5n.co.uk> wrote:
>> On Mon, 2010-12-13 at 13:16 +0100, Dan Brickley wrote:
>> > If you want to say your cat is a person, you can do that. It's not
>> > wrong to do so, just a little unconventional or metaphorical. If you
>> > want to say something about your cat without asserting personhood,
>> > don't use foaf:Person.
>> But my broader point was that there are instances where one would want
>> to apply many of the relationship properties to non-human agents without
>> getting into a philosophical quandary.

Karen Coyle
kcoyle at kcoyle.net http://kcoyle.net
ph: 1-510-540-7596
m: 1-510-435-8234
skype: kcoylenet

More information about the foaf-dev mailing list