[rdfweb-dev] Relationship Schema Updated

Dan Brickley danbri at w3.org
Sun Mar 14 18:23:29 UTC 2004

* Richard Lennox <listserve at richardlennox.net> [2004-03-14 18:13-0000]
> From: "Dan Brickley" <danbri at w3.org>
> > See also http://xmlns.com/foaf/0.1/#term_gender
> Why can a group not have a gender? There are many groups that are gender
> specific.  ie. for only people of a certain gender.  The group, therefore,
> could be said to have a gender.  The use of foaf:gender would in that case
> specify that the group was solely for people of a certain gender.

The concept of 'gender' is larger, scarier and more contested than any 
single modest effort such as FOAF can attempt to encompass. You might as
well ask why foaf:gender doesn't apply also to natural language terms 
(such as those in French etc). We had to draw the line somewhere; in the
case of foaf:gender, the line includes non-Person Agents, but doesn't 
extend to every kind of thing that might be talked about as being, in
some way, gendered.

The Group usage you sketch could, however, be characterised using FOAF's 
OWL-based group rules mechanism. You can define a Group by listing its 
membership criteria. These could certainly include gender properties, 
alongside any other that are describable using some RDF vocabulary. The 
expressive limits here are those imposed by the OWL language, ie. no
notion of variables etc etc. So you could define, for eg., a sub-class
of Person which was FemalePerson, or SingleWhiteFemale for that matter, 
so long as singleness, white-ness, female-ness could each be  
picked out with an RDF property. Whether/why you'd want to is, of
course, a separate matter.


More information about the foaf-dev mailing list