- From: Seth Russell <seth@robustai.net>
- Date: Thu, 27 Jun 2002 11:48:18 -0700
- To: "Danny Ayers" <danny666@virgilio.it>, "Jonathan Borden" <jonathan@openhealth.org>, "RDF Logic" <www-rdf-logic@w3.org>
From: "Danny Ayers" <danny666@virgilio.it> To: "Jonathan Borden" <jonathan@openhealth.org>; "RDF Logic" <www-rdf-logic@w3.org> > >Whether triples get asserted is not entirely a local issue, nor should it > >be. > > Hmm - I'm not convinced on either count. I believe the Model Theory does say that the triples are asserted in the document which contains them. If we assume that what is being called here "baseUri" is the uri of the context of the document; then reading that document into that context, should assert those triples in that context. Of course it's the agents prerogative whether it reads that document into that context. For example one could define a context called rdfs and assert that it's uri was <http://www.w3.org/2000/01/rdf-schema#> with the following triple: rdfs contextUri <http://www.w3.org/2000/01/rdf-schema#>. Then it would be natural to read in the definition triples of rdfs from some trusted source into that context. Now suppose some bozo comes along with a rdf file and puts a header on it that looks like: <rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#" xml:base="http://www.w3.org/1999/02/22-rdf-syntax-ns#"> > ... and asserts some really bizarre triples in it. Note the xml:base! Now if we read this bozo's file according to the way we are being led to believe is correct, we will totally krap on our ability to interpret rdfs files. But suppose we just read the bozo's file into it's own context. Now at our leisure we can examine Bozo's mutations to rdfs and decide whether we wish to accept or reject them for our own purposes. If we accept them, we copy his triples into our local context named rdfs. > but I'm not convinced of the need for formalization (although it probably > wouldn't hurt). I don't really see anything that might contradict my earlier > inclination to leave these matters to the agent. Allowing the agent to widen > or narrow the scope of the "current document", or in a similar way apply its > own join semantics would strike me as advantageous, if not essential for a > semantic web to get working. I totally agree :) Seth Russell A corollary to the axiom "Anybody can say anything about anything on the web." is "Anybody can read anything about anything on the web."
Received on Thursday, 27 June 2002 14:54:39 UTC