- From: Jeremy Carroll <jjc@hplb.hpl.hp.com>
- Date: Tue, 26 Jun 2001 17:06:17 +0100
- To: <www-rdf-interest@w3.org>
I, foolishly, am looking at writing another RDF parser. (ARP?). In my design stages I stumbled across this little teaser: what statements does the following correspond: <rdf:RDF xmlns:rdf="..." xmlns:a="..."> <rdf:Description aboutEach="#self" bagId="self"> <a:prop>value</a:prop> </rdf:Description> </rdf:RDF> [The ... are uri's] or we can make it a little richer with: <rdf:RDF xmlns:rdf="..." xmlns:a="..."> <rdf:Description aboutEach="#self" bagId="self"> <a:prop>value</a:prop> </rdf:Description> <rdf:Description about="#self"> <rdf:_1 rdf:resource="http://bertrand-russell-lives.org/"> </rdf:Description> </rdf:RDF> One reading of M&S would say that this second description has an infinite model. "#self" is a bag, it includes at least one member, http://bertrand-russell-lives.org/ This resource is hence one of the distributive referent of a:prop="value" and so we add the statement model, and add its reification to the bag "#self". The reification is anonymous, and since it is in "#self" it is one of the distributive referent and so ... ad infinitum. A different reading is suggested by naming the reification of a statement with a distributive referent. e.g. <rdf:RDF xmlns:rdf="..." xmlns:a="..."> <rdf:Description aboutEach="#other" bagId="self"> <a:prop ID="ex">value</a:prop> </rdf:Description> <rdf:Bag ID="other> <rdf:li rdf:resource="http://bertrand-russell-lives.org/a1"/> <rdf:li rdf:resource="http://bertrand-russell-lives.org/a2"/> </rdf:Bag> </rdf:RDF> Then do we have two rdf:subject's for the reification "#ex". the bag "#self" can have only one member with uri "#ex" is there a second member corresponding to the second member of "#other" or is there just one member corresponding to the single statement "#ex" in the description. We could read reification within an aboutEach as generating one resource per property in the description, with one subject for each subject in the distributive referent. Enjoy ... Jeremy
Received on Tuesday, 26 June 2001 12:06:20 UTC